2003-01-29 18:22:22 +01:00
|
|
|
/* Copyright (C) 2000-2003 MySQL AB
|
2000-08-21 13:35:27 +02:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
This program is free software; you can redistribute it and/or modify
|
|
|
|
it under the terms of the GNU General Public License as published by
|
|
|
|
the Free Software Foundation; either version 2 of the License, or
|
|
|
|
(at your option) any later version.
|
2000-08-21 13:35:27 +02:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
This program is distributed in the hope that it will be useful,
|
|
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
GNU General Public License for more details.
|
2000-08-21 13:35:27 +02:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
You should have received a copy of the GNU General Public License
|
|
|
|
along with this program; if not, write to the Free Software
|
|
|
|
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
|
|
|
|
|
2006-03-09 19:09:52 +01:00
|
|
|
#define MYSQL_LEX 1
|
2000-07-31 21:29:14 +02:00
|
|
|
#include "mysql_priv.h"
|
2000-09-30 01:20:26 +02:00
|
|
|
#include "sql_repl.h"
|
2006-05-02 19:56:43 +02:00
|
|
|
#include "rpl_filter.h"
|
2001-11-11 06:24:12 +01:00
|
|
|
#include "repl_failsafe.h"
|
2000-07-31 21:29:14 +02:00
|
|
|
#include <m_ctype.h>
|
|
|
|
#include <myisam.h>
|
|
|
|
#include <my_dir.h>
|
|
|
|
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
#include "sp_head.h"
|
2002-12-12 13:14:23 +01:00
|
|
|
#include "sp.h"
|
2005-07-30 10:19:57 +02:00
|
|
|
#include "sp_cache.h"
|
2006-05-02 19:56:43 +02:00
|
|
|
#include "event.h"
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
|
2001-12-05 12:03:00 +01:00
|
|
|
#ifdef HAVE_OPENSSL
|
|
|
|
/*
|
|
|
|
Without SSL the handshake consists of one packet. This packet
|
|
|
|
has both client capabilites and scrambled password.
|
|
|
|
With SSL the handshake might consist of two packets. If the first
|
|
|
|
packet (client capabilities) has CLIENT_SSL flag set, we have to
|
|
|
|
switch to SSL and read the second packet. The scrambled password
|
|
|
|
is in the second packet and client_capabilites field will be ignored.
|
|
|
|
Maybe it is better to accept flags other than CLIENT_SSL from the
|
|
|
|
second packet?
|
|
|
|
*/
|
2002-06-12 14:04:18 +02:00
|
|
|
#define SSL_HANDSHAKE_SIZE 2
|
|
|
|
#define NORMAL_HANDSHAKE_SIZE 6
|
|
|
|
#define MIN_HANDSHAKE_SIZE 2
|
2001-12-05 12:03:00 +01:00
|
|
|
#else
|
2002-06-12 14:04:18 +02:00
|
|
|
#define MIN_HANDSHAKE_SIZE 6
|
2001-12-05 12:03:00 +01:00
|
|
|
#endif /* HAVE_OPENSSL */
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2003-03-05 19:45:17 +01:00
|
|
|
/* Used in error handling only */
|
|
|
|
#define SP_TYPE_STRING(LP) \
|
|
|
|
((LP)->sphead->m_type == TYPE_ENUM_FUNCTION ? "FUNCTION" : "PROCEDURE")
|
|
|
|
#define SP_COM_STRING(LP) \
|
|
|
|
((LP)->sql_command == SQLCOM_CREATE_SPFUNCTION || \
|
|
|
|
(LP)->sql_command == SQLCOM_ALTER_FUNCTION || \
|
2005-01-03 19:53:54 +01:00
|
|
|
(LP)->sql_command == SQLCOM_SHOW_CREATE_FUNC || \
|
2003-03-05 19:45:17 +01:00
|
|
|
(LP)->sql_command == SQLCOM_DROP_FUNCTION ? \
|
|
|
|
"FUNCTION" : "PROCEDURE")
|
|
|
|
|
2000-08-15 19:09:37 +02:00
|
|
|
#ifdef SOLARIS
|
|
|
|
extern "C" int gethostname(char *name, int namelen);
|
|
|
|
#endif
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2005-02-09 03:43:27 +01:00
|
|
|
static void time_out_user_resource_limits(THD *thd, USER_CONN *uc);
|
2004-04-05 12:56:05 +02:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2002-10-02 12:33:08 +02:00
|
|
|
static int check_for_max_user_connections(THD *thd, USER_CONN *uc);
|
2004-04-05 12:56:05 +02:00
|
|
|
#endif
|
2002-05-15 12:50:38 +02:00
|
|
|
static void decrease_user_connections(USER_CONN *uc);
|
2000-08-21 02:07:54 +02:00
|
|
|
static bool check_db_used(THD *thd,TABLE_LIST *tables);
|
2005-01-31 16:57:32 +01:00
|
|
|
static bool check_multi_update_lock(THD *thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
static void remove_escape(char *name);
|
|
|
|
|
2000-09-30 01:20:26 +02:00
|
|
|
const char *any_db="*any*"; // Special symbol for check_access
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2006-05-02 19:56:43 +02:00
|
|
|
LEX_STRING command_name[]={
|
|
|
|
(char *)STRING_WITH_LEN("Sleep"),
|
|
|
|
(char *)STRING_WITH_LEN("Quit"),
|
|
|
|
(char *)STRING_WITH_LEN("Init DB"),
|
|
|
|
(char *)STRING_WITH_LEN("Query"),
|
|
|
|
(char *)STRING_WITH_LEN("Field List"),
|
|
|
|
(char *)STRING_WITH_LEN("Create DB"),
|
|
|
|
(char *)STRING_WITH_LEN("Drop DB"),
|
|
|
|
(char *)STRING_WITH_LEN("Refresh"),
|
|
|
|
(char *)STRING_WITH_LEN("Shutdown"),
|
|
|
|
(char *)STRING_WITH_LEN("Statistics"),
|
|
|
|
(char *)STRING_WITH_LEN("Processlist"),
|
|
|
|
(char *)STRING_WITH_LEN("Connect"),
|
|
|
|
(char *)STRING_WITH_LEN("Kill"),
|
|
|
|
(char *)STRING_WITH_LEN("Debug"),
|
|
|
|
(char *)STRING_WITH_LEN("Ping"),
|
|
|
|
(char *)STRING_WITH_LEN("Time"),
|
|
|
|
(char *)STRING_WITH_LEN("Delayed insert"),
|
|
|
|
(char *)STRING_WITH_LEN("Change user"),
|
|
|
|
(char *)STRING_WITH_LEN("Binlog Dump"),
|
|
|
|
(char *)STRING_WITH_LEN("Table Dump"),
|
|
|
|
(char *)STRING_WITH_LEN("Connect Out"),
|
|
|
|
(char *)STRING_WITH_LEN("Register Slave"),
|
|
|
|
(char *)STRING_WITH_LEN("Prepare"),
|
|
|
|
(char *)STRING_WITH_LEN("Execute"),
|
|
|
|
(char *)STRING_WITH_LEN("Long Data"),
|
|
|
|
(char *)STRING_WITH_LEN("Close stmt"),
|
|
|
|
(char *)STRING_WITH_LEN("Reset stmt"),
|
|
|
|
(char *)STRING_WITH_LEN("Set option"),
|
|
|
|
(char *)STRING_WITH_LEN("Fetch"),
|
|
|
|
(char *)STRING_WITH_LEN("Daemon"),
|
|
|
|
(char *)STRING_WITH_LEN("Error") // Last command number
|
2000-07-31 21:29:14 +02:00
|
|
|
};
|
|
|
|
|
2005-01-27 22:38:56 +01:00
|
|
|
const char *xa_state_names[]={
|
|
|
|
"NON-EXISTING", "ACTIVE", "IDLE", "PREPARED"
|
|
|
|
};
|
|
|
|
|
2002-10-02 12:33:08 +02:00
|
|
|
static char empty_c_string[1]= {0}; // Used for not defined 'db'
|
2000-07-31 21:29:14 +02:00
|
|
|
|
|
|
|
#ifdef __WIN__
|
|
|
|
static void test_signal(int sig_ptr)
|
|
|
|
{
|
2001-08-22 00:45:07 +02:00
|
|
|
#if !defined( DBUG_OFF)
|
2000-07-31 21:29:14 +02:00
|
|
|
MessageBox(NULL,"Test signal","DBUG",MB_OK);
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
static void init_signals(void)
|
|
|
|
{
|
|
|
|
int signals[7] = {SIGINT,SIGILL,SIGFPE,SIGSEGV,SIGTERM,SIGBREAK,SIGABRT } ;
|
2002-06-11 10:20:31 +02:00
|
|
|
for (int i=0 ; i < 7 ; i++)
|
2000-07-31 21:29:14 +02:00
|
|
|
signal( signals[i], test_signal) ;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2002-11-16 19:19:10 +01:00
|
|
|
static void unlock_locked_tables(THD *thd)
|
|
|
|
{
|
|
|
|
if (thd->locked_tables)
|
|
|
|
{
|
|
|
|
thd->lock=thd->locked_tables;
|
2004-10-07 00:45:06 +02:00
|
|
|
thd->locked_tables=0; // Will be automatically closed
|
2002-11-16 19:19:10 +01:00
|
|
|
close_thread_tables(thd); // Free tables
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2004-03-12 08:15:08 +01:00
|
|
|
|
2006-05-05 19:08:40 +02:00
|
|
|
bool end_active_trans(THD *thd)
|
2000-11-13 22:55:10 +01:00
|
|
|
{
|
2000-12-24 14:19:00 +01:00
|
|
|
int error=0;
|
2004-12-06 01:00:37 +01:00
|
|
|
DBUG_ENTER("end_active_trans");
|
2005-07-30 10:19:57 +02:00
|
|
|
if (unlikely(thd->in_sub_stmt))
|
2005-06-07 12:53:08 +02:00
|
|
|
{
|
|
|
|
my_error(ER_COMMIT_NOT_ALLOWED_IN_SF_OR_TRG, MYF(0));
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
2005-10-05 19:58:16 +02:00
|
|
|
if (thd->transaction.xid_state.xa_state != XA_NOTR)
|
|
|
|
{
|
|
|
|
my_error(ER_XAER_RMFAIL, MYF(0),
|
|
|
|
xa_state_names[thd->transaction.xid_state.xa_state]);
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
2002-07-23 17:31:22 +02:00
|
|
|
if (thd->options & (OPTION_NOT_AUTOCOMMIT | OPTION_BEGIN |
|
2001-09-08 00:02:41 +02:00
|
|
|
OPTION_TABLE_LOCK))
|
2000-11-13 22:55:10 +01:00
|
|
|
{
|
2004-12-06 01:00:37 +01:00
|
|
|
DBUG_PRINT("info",("options: 0x%lx", (ulong) thd->options));
|
|
|
|
/* Safety if one did "drop table" on locked tables */
|
|
|
|
if (!thd->locked_tables)
|
|
|
|
thd->options&= ~OPTION_TABLE_LOCK;
|
2000-11-24 00:51:18 +01:00
|
|
|
thd->server_status&= ~SERVER_STATUS_IN_TRANS;
|
2000-11-13 22:55:10 +01:00
|
|
|
if (ha_commit(thd))
|
2000-12-24 14:19:00 +01:00
|
|
|
error=1;
|
2000-11-13 22:55:10 +01:00
|
|
|
}
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
thd->options&= ~(OPTION_BEGIN | OPTION_STATUS_NO_TRANS_UPDATE |
|
|
|
|
OPTION_KEEP_LOG);
|
2004-12-06 01:00:37 +01:00
|
|
|
DBUG_RETURN(error);
|
2000-11-13 22:55:10 +01:00
|
|
|
}
|
|
|
|
|
2006-05-02 19:56:43 +02:00
|
|
|
bool begin_trans(THD *thd)
|
2005-02-01 20:48:05 +01:00
|
|
|
{
|
|
|
|
int error=0;
|
2005-07-30 10:19:57 +02:00
|
|
|
if (unlikely(thd->in_sub_stmt))
|
2005-06-07 12:53:08 +02:00
|
|
|
{
|
|
|
|
my_error(ER_COMMIT_NOT_ALLOWED_IN_SF_OR_TRG, MYF(0));
|
|
|
|
return 1;
|
|
|
|
}
|
2005-02-01 20:48:05 +01:00
|
|
|
if (thd->locked_tables)
|
|
|
|
{
|
|
|
|
thd->lock=thd->locked_tables;
|
|
|
|
thd->locked_tables=0; // Will be automatically closed
|
|
|
|
close_thread_tables(thd); // Free tables
|
|
|
|
}
|
|
|
|
if (end_active_trans(thd))
|
|
|
|
error= -1;
|
|
|
|
else
|
|
|
|
{
|
2005-02-02 00:43:09 +01:00
|
|
|
LEX *lex= thd->lex;
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
thd->options|= OPTION_BEGIN;
|
2005-02-01 20:48:05 +01:00
|
|
|
thd->server_status|= SERVER_STATUS_IN_TRANS;
|
|
|
|
if (lex->start_transaction_opt & MYSQL_START_TRANS_OPT_WITH_CONS_SNAPSHOT)
|
2005-02-14 21:50:09 +01:00
|
|
|
error= ha_start_consistent_snapshot(thd);
|
2005-02-01 20:48:05 +01:00
|
|
|
}
|
|
|
|
return error;
|
|
|
|
}
|
2000-11-13 22:55:10 +01:00
|
|
|
|
2004-03-17 23:09:13 +01:00
|
|
|
#ifdef HAVE_REPLICATION
|
2005-08-31 18:08:45 +02:00
|
|
|
/*
|
|
|
|
Returns true if all tables should be ignored
|
|
|
|
*/
|
2004-03-12 08:15:08 +01:00
|
|
|
inline bool all_tables_not_ok(THD *thd, TABLE_LIST *tables)
|
|
|
|
{
|
2006-05-02 19:56:43 +02:00
|
|
|
return rpl_filter->is_on() && tables && !thd->spcont &&
|
|
|
|
!rpl_filter->tables_ok(thd->db, tables);
|
2004-03-12 08:15:08 +01:00
|
|
|
}
|
2004-03-17 23:09:13 +01:00
|
|
|
#endif
|
2004-03-12 08:15:08 +01:00
|
|
|
|
|
|
|
|
2005-10-17 10:52:34 +02:00
|
|
|
static bool some_non_temp_table_to_be_updated(THD *thd, TABLE_LIST *tables)
|
|
|
|
{
|
|
|
|
for (TABLE_LIST *table= tables; table; table= table->next_global)
|
|
|
|
{
|
|
|
|
DBUG_ASSERT(table->db && table->table_name);
|
|
|
|
if (table->updating &&
|
|
|
|
!find_temporary_table(thd, table->db, table->table_name))
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2001-12-26 15:49:10 +01:00
|
|
|
static HASH hash_user_connections;
|
|
|
|
|
2002-02-13 21:37:19 +01:00
|
|
|
static int get_or_create_user_conn(THD *thd, const char *user,
|
|
|
|
const char *host,
|
2002-11-30 18:58:53 +01:00
|
|
|
USER_RESOURCES *mqh)
|
2002-02-01 19:53:24 +01:00
|
|
|
{
|
2005-06-28 03:03:14 +02:00
|
|
|
int return_val= 0;
|
2003-11-20 21:06:25 +01:00
|
|
|
uint temp_len, user_len;
|
2006-01-11 00:07:40 +01:00
|
|
|
char temp_user[USER_HOST_BUFF_SIZE];
|
2002-02-01 19:53:24 +01:00
|
|
|
struct user_conn *uc;
|
|
|
|
|
|
|
|
DBUG_ASSERT(user != 0);
|
|
|
|
DBUG_ASSERT(host != 0);
|
|
|
|
|
2005-06-28 03:03:14 +02:00
|
|
|
user_len= strlen(user);
|
2002-05-15 12:50:38 +02:00
|
|
|
temp_len= (strmov(strmov(temp_user, user)+1, host) - temp_user)+1;
|
2002-02-01 19:53:24 +01:00
|
|
|
(void) pthread_mutex_lock(&LOCK_user_conn);
|
2002-02-13 21:37:19 +01:00
|
|
|
if (!(uc = (struct user_conn *) hash_search(&hash_user_connections,
|
|
|
|
(byte*) temp_user, temp_len)))
|
2002-02-01 19:53:24 +01:00
|
|
|
{
|
2002-02-13 21:37:19 +01:00
|
|
|
/* First connection for user; Create a user connection object */
|
|
|
|
if (!(uc= ((struct user_conn*)
|
|
|
|
my_malloc(sizeof(struct user_conn) + temp_len+1,
|
2002-02-15 01:49:02 +01:00
|
|
|
MYF(MY_WME)))))
|
|
|
|
{
|
2004-10-20 15:06:54 +02:00
|
|
|
net_send_error(thd, 0, NullS); // Out of memory
|
2005-06-28 03:03:14 +02:00
|
|
|
return_val= 1;
|
2002-02-01 19:53:24 +01:00
|
|
|
goto end;
|
2002-02-15 01:49:02 +01:00
|
|
|
}
|
2002-02-01 19:53:24 +01:00
|
|
|
uc->user=(char*) (uc+1);
|
|
|
|
memcpy(uc->user,temp_user,temp_len+1);
|
2004-12-29 18:30:37 +01:00
|
|
|
uc->host= uc->user + user_len + 1;
|
2005-06-28 03:03:14 +02:00
|
|
|
uc->len= temp_len;
|
2004-12-29 18:30:37 +01:00
|
|
|
uc->connections= uc->questions= uc->updates= uc->conn_per_hour= 0;
|
2005-06-28 03:03:14 +02:00
|
|
|
uc->user_resources= *mqh;
|
|
|
|
uc->intime= thd->thr_create_time;
|
2003-09-19 11:44:31 +02:00
|
|
|
if (my_hash_insert(&hash_user_connections, (byte*) uc))
|
2002-02-01 19:53:24 +01:00
|
|
|
{
|
|
|
|
my_free((char*) uc,0);
|
2004-10-20 15:06:54 +02:00
|
|
|
net_send_error(thd, 0, NullS); // Out of memory
|
2005-06-28 03:03:14 +02:00
|
|
|
return_val= 1;
|
2002-02-01 19:53:24 +01:00
|
|
|
goto end;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
thd->user_connect=uc;
|
2003-10-27 15:14:03 +01:00
|
|
|
uc->connections++;
|
2002-02-01 19:53:24 +01:00
|
|
|
end:
|
|
|
|
(void) pthread_mutex_unlock(&LOCK_user_conn);
|
|
|
|
return return_val;
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2002-02-01 19:53:24 +01:00
|
|
|
}
|
2001-12-26 15:49:10 +01:00
|
|
|
|
|
|
|
|
|
|
|
/*
|
2005-08-20 10:00:00 +02:00
|
|
|
Check if user exist and password supplied is correct.
|
|
|
|
|
2003-02-14 10:47:41 +01:00
|
|
|
SYNOPSIS
|
|
|
|
check_user()
|
2005-09-28 16:43:46 +02:00
|
|
|
thd thread handle, thd->security_ctx->{host,user,ip} are used
|
2003-07-18 16:25:54 +02:00
|
|
|
command originator of the check: now check_user is called
|
|
|
|
during connect and change user procedures; used for
|
|
|
|
logging.
|
2004-10-07 00:45:06 +02:00
|
|
|
passwd scrambled password received from client
|
2003-07-18 16:25:54 +02:00
|
|
|
passwd_len length of scrambled password
|
|
|
|
db database name to connect to, may be NULL
|
|
|
|
check_count dont know exactly
|
|
|
|
|
2003-07-01 21:40:59 +02:00
|
|
|
Note, that host, user and passwd may point to communication buffer.
|
2004-10-07 00:45:06 +02:00
|
|
|
Current implementation does not depend on that, but future changes
|
2003-07-18 16:25:54 +02:00
|
|
|
should be done with this in mind; 'thd' is INOUT, all other params
|
|
|
|
are 'IN'.
|
|
|
|
|
2003-07-01 21:40:59 +02:00
|
|
|
RETURN VALUE
|
2005-09-28 16:43:46 +02:00
|
|
|
0 OK; thd->security_ctx->user/master_access/priv_user/db_access and
|
|
|
|
thd->db are updated; OK is sent to client;
|
2003-07-04 18:52:04 +02:00
|
|
|
-1 access denied or handshake error; error is sent to client;
|
|
|
|
>0 error, not sent to client
|
2000-07-31 21:29:14 +02:00
|
|
|
*/
|
|
|
|
|
2003-09-26 12:33:13 +02:00
|
|
|
int check_user(THD *thd, enum enum_server_command command,
|
|
|
|
const char *passwd, uint passwd_len, const char *db,
|
|
|
|
bool check_count)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2003-02-14 10:47:41 +01:00
|
|
|
DBUG_ENTER("check_user");
|
2003-05-19 15:35:49 +02:00
|
|
|
|
2003-12-19 15:25:50 +01:00
|
|
|
#ifdef NO_EMBEDDED_ACCESS_CHECKS
|
2005-09-21 07:29:47 +02:00
|
|
|
thd->main_security_ctx.master_access= GLOBAL_ACLS; // Full rights
|
2005-04-14 11:56:13 +02:00
|
|
|
/* Change database if necessary */
|
2004-02-14 17:26:21 +01:00
|
|
|
if (db && db[0])
|
|
|
|
{
|
2005-08-20 10:00:00 +02:00
|
|
|
/*
|
|
|
|
thd->db is saved in caller and needs to be freed by caller if this
|
|
|
|
function returns 0
|
|
|
|
*/
|
2004-02-14 17:26:21 +01:00
|
|
|
thd->db= 0;
|
|
|
|
thd->db_length= 0;
|
2005-08-12 02:04:16 +02:00
|
|
|
if (mysql_change_db(thd, db, FALSE))
|
2004-02-14 17:26:21 +01:00
|
|
|
{
|
2005-04-14 11:56:13 +02:00
|
|
|
/* Send the error to the client */
|
|
|
|
net_send_error(thd);
|
2004-02-14 17:26:21 +01:00
|
|
|
if (thd->user_connect)
|
|
|
|
decrease_user_connections(thd->user_connect);
|
|
|
|
DBUG_RETURN(-1);
|
|
|
|
}
|
|
|
|
}
|
2005-08-12 02:04:16 +02:00
|
|
|
send_ok(thd);
|
2004-02-14 17:26:21 +01:00
|
|
|
DBUG_RETURN(0);
|
2003-12-19 15:25:50 +01:00
|
|
|
#else
|
|
|
|
|
2003-07-08 00:36:14 +02:00
|
|
|
my_bool opt_secure_auth_local;
|
|
|
|
pthread_mutex_lock(&LOCK_global_system_variables);
|
|
|
|
opt_secure_auth_local= opt_secure_auth;
|
|
|
|
pthread_mutex_unlock(&LOCK_global_system_variables);
|
|
|
|
|
2003-05-13 17:58:26 +02:00
|
|
|
/*
|
2003-07-08 00:36:14 +02:00
|
|
|
If the server is running in secure auth mode, short scrambles are
|
|
|
|
forbidden.
|
2003-05-13 17:58:26 +02:00
|
|
|
*/
|
2003-07-08 00:36:14 +02:00
|
|
|
if (opt_secure_auth_local && passwd_len == SCRAMBLE_LENGTH_323)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-10-20 15:06:54 +02:00
|
|
|
net_printf_error(thd, ER_NOT_SUPPORTED_AUTH_MODE);
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, COM_CONNECT, ER(ER_NOT_SUPPORTED_AUTH_MODE));
|
2003-07-08 00:36:14 +02:00
|
|
|
DBUG_RETURN(-1);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2003-07-04 18:52:04 +02:00
|
|
|
if (passwd_len != 0 &&
|
|
|
|
passwd_len != SCRAMBLE_LENGTH &&
|
|
|
|
passwd_len != SCRAMBLE_LENGTH_323)
|
|
|
|
DBUG_RETURN(ER_HANDSHAKE_ERROR);
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2003-01-18 18:31:38 +01:00
|
|
|
/*
|
2003-07-18 16:25:54 +02:00
|
|
|
Clear thd->db as it points to something, that will be freed when
|
2004-10-07 00:45:06 +02:00
|
|
|
connection is closed. We don't want to accidentally free a wrong pointer
|
2003-07-18 16:25:54 +02:00
|
|
|
if connect failed. Also in case of 'CHANGE USER' failure, current
|
|
|
|
database will be switched to 'no database selected'.
|
2003-01-18 18:31:38 +01:00
|
|
|
*/
|
2003-07-01 21:40:59 +02:00
|
|
|
thd->db= 0;
|
|
|
|
thd->db_length= 0;
|
2003-07-04 18:52:04 +02:00
|
|
|
|
2003-07-01 21:40:59 +02:00
|
|
|
USER_RESOURCES ur;
|
2003-07-08 00:36:14 +02:00
|
|
|
int res= acl_getroot(thd, &ur, passwd, passwd_len);
|
2003-09-26 12:33:13 +02:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2003-07-04 18:52:04 +02:00
|
|
|
if (res == -1)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2003-07-04 18:52:04 +02:00
|
|
|
/*
|
|
|
|
This happens when client (new) sends password scrambled with
|
|
|
|
scramble(), but database holds old value (scrambled with
|
|
|
|
scramble_323()). Here we please client to send scrambled_password
|
|
|
|
in old format.
|
|
|
|
*/
|
2003-07-18 16:25:54 +02:00
|
|
|
NET *net= &thd->net;
|
2003-07-08 00:36:14 +02:00
|
|
|
if (opt_secure_auth_local)
|
2002-11-24 15:07:53 +01:00
|
|
|
{
|
2004-10-20 15:06:54 +02:00
|
|
|
net_printf_error(thd, ER_SERVER_IS_IN_SECURE_AUTH_MODE,
|
2005-09-20 20:20:38 +02:00
|
|
|
thd->main_security_ctx.user,
|
|
|
|
thd->main_security_ctx.host_or_ip);
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, COM_CONNECT, ER(ER_SERVER_IS_IN_SECURE_AUTH_MODE),
|
|
|
|
thd->main_security_ctx.user,
|
|
|
|
thd->main_security_ctx.host_or_ip);
|
2003-07-08 00:36:14 +02:00
|
|
|
DBUG_RETURN(-1);
|
|
|
|
}
|
2004-07-07 10:29:39 +02:00
|
|
|
/* We have to read very specific packet size */
|
2003-07-18 16:25:54 +02:00
|
|
|
if (send_old_password_request(thd) ||
|
2004-07-07 10:29:39 +02:00
|
|
|
my_net_read(net) != SCRAMBLE_LENGTH_323 + 1)
|
2005-09-20 20:20:38 +02:00
|
|
|
{
|
2003-07-04 18:52:04 +02:00
|
|
|
inc_host_errors(&thd->remote.sin_addr);
|
|
|
|
DBUG_RETURN(ER_HANDSHAKE_ERROR);
|
|
|
|
}
|
|
|
|
/* Final attempt to check the user based on reply */
|
|
|
|
/* So as passwd is short, errcode is always >= 0 */
|
2003-07-08 00:36:14 +02:00
|
|
|
res= acl_getroot(thd, &ur, (char *) net->read_pos, SCRAMBLE_LENGTH_323);
|
2003-07-04 18:52:04 +02:00
|
|
|
}
|
2003-09-26 12:33:13 +02:00
|
|
|
#endif /*EMBEDDED_LIBRARY*/
|
2003-07-04 18:52:04 +02:00
|
|
|
/* here res is always >= 0 */
|
|
|
|
if (res == 0)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2005-09-20 20:20:38 +02:00
|
|
|
if (!(thd->main_security_ctx.master_access &
|
|
|
|
NO_ACCESS)) // authentication is OK
|
2002-11-24 15:07:53 +01:00
|
|
|
{
|
2003-07-04 18:52:04 +02:00
|
|
|
DBUG_PRINT("info",
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
("Capabilities: %lx packet_length: %ld Host: '%s' "
|
2003-07-04 18:52:04 +02:00
|
|
|
"Login user: '%s' Priv_user: '%s' Using password: %s "
|
|
|
|
"Access: %u db: '%s'",
|
2005-09-20 20:20:38 +02:00
|
|
|
thd->client_capabilities,
|
|
|
|
thd->max_client_packet_length,
|
|
|
|
thd->main_security_ctx.host_or_ip,
|
|
|
|
thd->main_security_ctx.user,
|
|
|
|
thd->main_security_ctx.priv_user,
|
2003-07-04 18:52:04 +02:00
|
|
|
passwd_len ? "yes": "no",
|
2005-09-20 20:20:38 +02:00
|
|
|
thd->main_security_ctx.master_access,
|
|
|
|
(thd->db ? thd->db : "*none*")));
|
2003-07-04 18:52:04 +02:00
|
|
|
|
|
|
|
if (check_count)
|
2002-12-26 17:12:57 +01:00
|
|
|
{
|
2003-07-04 18:52:04 +02:00
|
|
|
VOID(pthread_mutex_lock(&LOCK_thread_count));
|
2005-03-10 00:10:10 +01:00
|
|
|
bool count_ok= thread_count <= max_connections + delayed_insert_threads
|
2005-09-20 20:20:38 +02:00
|
|
|
|| (thd->main_security_ctx.master_access & SUPER_ACL);
|
2003-07-04 18:52:04 +02:00
|
|
|
VOID(pthread_mutex_unlock(&LOCK_thread_count));
|
|
|
|
if (!count_ok)
|
2005-03-10 00:10:10 +01:00
|
|
|
{ // too many connections
|
2004-10-20 15:06:54 +02:00
|
|
|
net_send_error(thd, ER_CON_COUNT_ERROR);
|
2003-07-04 18:52:04 +02:00
|
|
|
DBUG_RETURN(-1);
|
|
|
|
}
|
2002-12-26 17:12:57 +01:00
|
|
|
}
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2003-07-04 18:52:04 +02:00
|
|
|
/* Why logging is performed before all checks've passed? */
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command,
|
|
|
|
(thd->main_security_ctx.priv_user ==
|
|
|
|
thd->main_security_ctx.user ?
|
|
|
|
(char*) "%s@%s on %s" :
|
|
|
|
(char*) "%s@%s as anonymous on %s"),
|
|
|
|
thd->main_security_ctx.user,
|
|
|
|
thd->main_security_ctx.host_or_ip,
|
|
|
|
db ? db : (char*) "");
|
2003-07-04 18:52:04 +02:00
|
|
|
|
2003-01-18 18:31:38 +01:00
|
|
|
/*
|
2003-07-18 16:25:54 +02:00
|
|
|
This is the default access rights for the current database. It's
|
|
|
|
set to 0 here because we don't have an active database yet (and we
|
|
|
|
may not have an active database to set.
|
2003-01-18 18:31:38 +01:00
|
|
|
*/
|
2005-09-20 20:20:38 +02:00
|
|
|
thd->main_security_ctx.db_access=0;
|
2003-07-04 18:52:04 +02:00
|
|
|
|
|
|
|
/* Don't allow user to connect if he has done too many queries */
|
2004-12-29 18:30:37 +01:00
|
|
|
if ((ur.questions || ur.updates || ur.conn_per_hour || ur.user_conn ||
|
2003-11-04 08:40:36 +01:00
|
|
|
max_user_connections) &&
|
2004-12-29 18:30:37 +01:00
|
|
|
get_or_create_user_conn(thd,
|
2005-09-20 20:20:38 +02:00
|
|
|
(opt_old_style_user_limits ? thd->main_security_ctx.user :
|
|
|
|
thd->main_security_ctx.priv_user),
|
|
|
|
(opt_old_style_user_limits ? thd->main_security_ctx.host_or_ip :
|
|
|
|
thd->main_security_ctx.priv_host),
|
2004-12-29 18:30:37 +01:00
|
|
|
&ur))
|
2003-11-04 08:40:36 +01:00
|
|
|
DBUG_RETURN(-1);
|
|
|
|
if (thd->user_connect &&
|
2004-12-29 18:30:37 +01:00
|
|
|
(thd->user_connect->user_resources.conn_per_hour ||
|
|
|
|
thd->user_connect->user_resources.user_conn ||
|
2003-11-04 08:40:36 +01:00
|
|
|
max_user_connections) &&
|
|
|
|
check_for_max_user_connections(thd, thd->user_connect))
|
|
|
|
DBUG_RETURN(-1);
|
2003-07-04 18:52:04 +02:00
|
|
|
|
2005-04-14 11:56:13 +02:00
|
|
|
/* Change database if necessary */
|
2003-07-04 18:52:04 +02:00
|
|
|
if (db && db[0])
|
2002-12-26 17:12:57 +01:00
|
|
|
{
|
2005-08-12 02:04:16 +02:00
|
|
|
if (mysql_change_db(thd, db, FALSE))
|
2003-07-04 18:52:04 +02:00
|
|
|
{
|
2005-04-14 11:56:13 +02:00
|
|
|
/* Send error to the client */
|
|
|
|
net_send_error(thd);
|
2003-07-04 18:52:04 +02:00
|
|
|
if (thd->user_connect)
|
|
|
|
decrease_user_connections(thd->user_connect);
|
|
|
|
DBUG_RETURN(-1);
|
|
|
|
}
|
2002-12-26 17:12:57 +01:00
|
|
|
}
|
2005-08-12 02:04:16 +02:00
|
|
|
send_ok(thd);
|
2003-07-04 18:52:04 +02:00
|
|
|
thd->password= test(passwd_len); // remember for error messages
|
|
|
|
/* Ready to handle queries */
|
|
|
|
DBUG_RETURN(0);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
}
|
2003-07-04 18:52:04 +02:00
|
|
|
else if (res == 2) // client gave short hash, server has long hash
|
2001-02-21 00:11:32 +01:00
|
|
|
{
|
2004-10-20 15:06:54 +02:00
|
|
|
net_printf_error(thd, ER_NOT_SUPPORTED_AUTH_MODE);
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, COM_CONNECT, ER(ER_NOT_SUPPORTED_AUTH_MODE));
|
2003-07-04 18:52:04 +02:00
|
|
|
DBUG_RETURN(-1);
|
2001-02-21 00:11:32 +01:00
|
|
|
}
|
2004-10-20 15:06:54 +02:00
|
|
|
net_printf_error(thd, ER_ACCESS_DENIED_ERROR,
|
2005-09-20 20:20:38 +02:00
|
|
|
thd->main_security_ctx.user,
|
|
|
|
thd->main_security_ctx.host_or_ip,
|
2004-10-20 15:06:54 +02:00
|
|
|
passwd_len ? ER(ER_YES) : ER(ER_NO));
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, COM_CONNECT, ER(ER_ACCESS_DENIED_ERROR),
|
|
|
|
thd->main_security_ctx.user,
|
|
|
|
thd->main_security_ctx.host_or_ip,
|
|
|
|
passwd_len ? ER(ER_YES) : ER(ER_NO));
|
2003-07-04 18:52:04 +02:00
|
|
|
DBUG_RETURN(-1);
|
2003-12-19 15:25:50 +01:00
|
|
|
#endif /* NO_EMBEDDED_ACCESS_CHECKS */
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
2001-02-21 00:11:32 +01:00
|
|
|
/*
|
2002-01-29 17:32:16 +01:00
|
|
|
Check for maximum allowable user connections, if the mysqld server is
|
|
|
|
started with corresponding variable that is greater then 0.
|
2001-02-21 00:11:32 +01:00
|
|
|
*/
|
|
|
|
|
2002-11-07 11:49:02 +01:00
|
|
|
extern "C" byte *get_key_conn(user_conn *buff, uint *length,
|
|
|
|
my_bool not_used __attribute__((unused)))
|
2002-01-29 17:32:16 +01:00
|
|
|
{
|
|
|
|
*length=buff->len;
|
|
|
|
return (byte*) buff->user;
|
|
|
|
}
|
|
|
|
|
2002-11-07 11:49:02 +01:00
|
|
|
extern "C" void free_user(struct user_conn *uc)
|
2002-01-29 17:32:16 +01:00
|
|
|
{
|
|
|
|
my_free((char*) uc,MYF(0));
|
|
|
|
}
|
|
|
|
|
2002-11-30 18:58:53 +01:00
|
|
|
void init_max_user_conn(void)
|
2001-02-21 00:11:32 +01:00
|
|
|
{
|
2002-03-14 18:44:42 +01:00
|
|
|
(void) hash_init(&hash_user_connections,system_charset_info,max_connections,
|
|
|
|
0,0,
|
2002-11-07 02:54:00 +01:00
|
|
|
(hash_get_key) get_key_conn, (hash_free_key) free_user,
|
2001-03-06 14:24:08 +01:00
|
|
|
0);
|
2001-02-21 00:11:32 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2003-11-04 08:40:36 +01:00
|
|
|
/*
|
|
|
|
check if user has already too many connections
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
check_for_max_user_connections()
|
|
|
|
thd Thread handle
|
|
|
|
uc User connect object
|
|
|
|
|
|
|
|
NOTES
|
|
|
|
If check fails, we decrease user connection count, which means one
|
|
|
|
shouldn't call decrease_user_connections() after this function.
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
0 ok
|
|
|
|
1 error
|
|
|
|
*/
|
|
|
|
|
2004-04-05 12:56:05 +02:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
|
|
|
|
2002-10-02 12:33:08 +02:00
|
|
|
static int check_for_max_user_connections(THD *thd, USER_CONN *uc)
|
2001-02-21 00:11:32 +01:00
|
|
|
{
|
2002-02-01 19:53:24 +01:00
|
|
|
int error=0;
|
2001-03-06 14:24:08 +01:00
|
|
|
DBUG_ENTER("check_for_max_user_connections");
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2003-10-27 15:14:03 +01:00
|
|
|
(void) pthread_mutex_lock(&LOCK_user_conn);
|
2004-12-29 18:30:37 +01:00
|
|
|
if (max_user_connections && !uc->user_resources.user_conn &&
|
2003-11-04 08:40:36 +01:00
|
|
|
max_user_connections < (uint) uc->connections)
|
2001-02-21 00:11:32 +01:00
|
|
|
{
|
2004-10-20 15:06:54 +02:00
|
|
|
net_printf_error(thd, ER_TOO_MANY_USER_CONNECTIONS, uc->user);
|
2002-02-01 19:53:24 +01:00
|
|
|
error=1;
|
|
|
|
goto end;
|
2001-02-21 00:11:32 +01:00
|
|
|
}
|
2005-02-09 03:43:27 +01:00
|
|
|
time_out_user_resource_limits(thd, uc);
|
2004-12-29 18:30:37 +01:00
|
|
|
if (uc->user_resources.user_conn &&
|
|
|
|
uc->user_resources.user_conn < uc->connections)
|
|
|
|
{
|
|
|
|
net_printf_error(thd, ER_USER_LIMIT_REACHED, uc->user,
|
|
|
|
"max_user_connections",
|
|
|
|
(long) uc->user_resources.user_conn);
|
|
|
|
error= 1;
|
|
|
|
goto end;
|
|
|
|
}
|
|
|
|
if (uc->user_resources.conn_per_hour &&
|
|
|
|
uc->user_resources.conn_per_hour <= uc->conn_per_hour)
|
2002-05-15 12:50:38 +02:00
|
|
|
{
|
2004-10-20 15:06:54 +02:00
|
|
|
net_printf_error(thd, ER_USER_LIMIT_REACHED, uc->user,
|
2005-05-02 19:19:37 +02:00
|
|
|
"max_connections_per_hour",
|
2004-12-29 18:30:37 +01:00
|
|
|
(long) uc->user_resources.conn_per_hour);
|
2002-05-15 12:50:38 +02:00
|
|
|
error=1;
|
|
|
|
goto end;
|
|
|
|
}
|
2003-10-27 15:14:03 +01:00
|
|
|
uc->conn_per_hour++;
|
2003-11-04 08:40:36 +01:00
|
|
|
|
|
|
|
end:
|
2003-10-29 08:33:31 +01:00
|
|
|
if (error)
|
|
|
|
uc->connections--; // no need for decrease_user_connections() here
|
2003-10-27 15:14:03 +01:00
|
|
|
(void) pthread_mutex_unlock(&LOCK_user_conn);
|
2001-03-06 14:24:08 +01:00
|
|
|
DBUG_RETURN(error);
|
2001-02-21 00:11:32 +01:00
|
|
|
}
|
2004-04-05 12:56:05 +02:00
|
|
|
#endif /* NO_EMBEDDED_ACCESS_CHECKS */
|
2001-02-21 00:11:32 +01:00
|
|
|
|
2003-11-04 08:40:36 +01:00
|
|
|
/*
|
|
|
|
Decrease user connection count
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
decrease_user_connections()
|
|
|
|
uc User connection object
|
|
|
|
|
|
|
|
NOTES
|
|
|
|
If there is a n user connection object for a connection
|
|
|
|
(which only happens if 'max_user_connections' is defined or
|
|
|
|
if someone has created a resource grant for a user), then
|
|
|
|
the connection count is always incremented on connect.
|
|
|
|
|
|
|
|
The user connect object is not freed if some users has
|
|
|
|
'max connections per hour' defined as we need to be able to hold
|
|
|
|
count over the lifetime of the connection.
|
|
|
|
*/
|
|
|
|
|
2002-05-15 12:50:38 +02:00
|
|
|
static void decrease_user_connections(USER_CONN *uc)
|
2001-02-21 00:11:32 +01:00
|
|
|
{
|
2002-02-01 19:53:24 +01:00
|
|
|
DBUG_ENTER("decrease_user_connections");
|
2003-10-27 15:14:03 +01:00
|
|
|
(void) pthread_mutex_lock(&LOCK_user_conn);
|
|
|
|
DBUG_ASSERT(uc->connections);
|
|
|
|
if (!--uc->connections && !mqh_used)
|
2001-02-21 00:11:32 +01:00
|
|
|
{
|
|
|
|
/* Last connection for user; Delete it */
|
2001-03-07 22:50:44 +01:00
|
|
|
(void) hash_delete(&hash_user_connections,(byte*) uc);
|
2001-02-21 00:11:32 +01:00
|
|
|
}
|
2003-10-27 15:14:03 +01:00
|
|
|
(void) pthread_mutex_unlock(&LOCK_user_conn);
|
2001-03-06 14:24:08 +01:00
|
|
|
DBUG_VOID_RETURN;
|
2001-02-21 00:11:32 +01:00
|
|
|
}
|
|
|
|
|
2001-03-06 14:24:08 +01:00
|
|
|
|
2001-02-21 00:11:32 +01:00
|
|
|
void free_max_user_conn(void)
|
|
|
|
{
|
|
|
|
hash_free(&hash_user_connections);
|
|
|
|
}
|
|
|
|
|
2002-01-29 17:32:16 +01:00
|
|
|
|
2002-09-16 14:55:19 +02:00
|
|
|
/*
|
|
|
|
Mark all commands that somehow changes a table
|
|
|
|
This is used to check number of updates / hour
|
2004-10-22 08:54:14 +02:00
|
|
|
|
|
|
|
sql_command is actually set to SQLCOM_END sometimes
|
|
|
|
so we need the +1 to include it in the array.
|
2005-01-27 22:38:56 +01:00
|
|
|
|
|
|
|
numbers are:
|
|
|
|
0 - read-only query
|
|
|
|
!= 0 - query that may change a table
|
|
|
|
2 - query that returns meaningful ROW_COUNT() -
|
|
|
|
a number of modified rows
|
2002-09-16 14:55:19 +02:00
|
|
|
*/
|
|
|
|
|
2004-10-22 08:54:14 +02:00
|
|
|
char uc_update_queries[SQLCOM_END+1];
|
2002-09-16 14:55:19 +02:00
|
|
|
|
|
|
|
void init_update_queries(void)
|
|
|
|
{
|
2004-10-22 08:54:14 +02:00
|
|
|
bzero((gptr) &uc_update_queries, sizeof(uc_update_queries));
|
|
|
|
|
2002-09-16 14:55:19 +02:00
|
|
|
uc_update_queries[SQLCOM_CREATE_TABLE]=1;
|
|
|
|
uc_update_queries[SQLCOM_CREATE_INDEX]=1;
|
|
|
|
uc_update_queries[SQLCOM_ALTER_TABLE]=1;
|
2005-01-16 13:16:23 +01:00
|
|
|
uc_update_queries[SQLCOM_UPDATE]=2;
|
|
|
|
uc_update_queries[SQLCOM_UPDATE_MULTI]=2;
|
|
|
|
uc_update_queries[SQLCOM_INSERT]=2;
|
|
|
|
uc_update_queries[SQLCOM_INSERT_SELECT]=2;
|
|
|
|
uc_update_queries[SQLCOM_DELETE]=2;
|
|
|
|
uc_update_queries[SQLCOM_DELETE_MULTI]=2;
|
2002-09-16 14:55:19 +02:00
|
|
|
uc_update_queries[SQLCOM_TRUNCATE]=1;
|
|
|
|
uc_update_queries[SQLCOM_DROP_TABLE]=1;
|
|
|
|
uc_update_queries[SQLCOM_LOAD]=1;
|
|
|
|
uc_update_queries[SQLCOM_CREATE_DB]=1;
|
|
|
|
uc_update_queries[SQLCOM_DROP_DB]=1;
|
2005-01-16 13:16:23 +01:00
|
|
|
uc_update_queries[SQLCOM_REPLACE]=2;
|
|
|
|
uc_update_queries[SQLCOM_REPLACE_SELECT]=2;
|
2002-09-16 14:55:19 +02:00
|
|
|
uc_update_queries[SQLCOM_RENAME_TABLE]=1;
|
|
|
|
uc_update_queries[SQLCOM_BACKUP_TABLE]=1;
|
|
|
|
uc_update_queries[SQLCOM_RESTORE_TABLE]=1;
|
|
|
|
uc_update_queries[SQLCOM_DROP_INDEX]=1;
|
2004-07-16 00:15:55 +02:00
|
|
|
uc_update_queries[SQLCOM_CREATE_VIEW]=1;
|
|
|
|
uc_update_queries[SQLCOM_DROP_VIEW]=1;
|
2006-05-02 19:56:43 +02:00
|
|
|
uc_update_queries[SQLCOM_CREATE_EVENT]=1;
|
|
|
|
uc_update_queries[SQLCOM_ALTER_EVENT]=1;
|
|
|
|
uc_update_queries[SQLCOM_DROP_EVENT]=1;
|
2002-09-16 14:55:19 +02:00
|
|
|
}
|
|
|
|
|
2003-01-30 18:39:54 +01:00
|
|
|
bool is_update_query(enum enum_sql_command command)
|
|
|
|
{
|
2004-10-22 08:54:14 +02:00
|
|
|
DBUG_ASSERT(command >= 0 && command <= SQLCOM_END);
|
2005-08-27 00:33:06 +02:00
|
|
|
return uc_update_queries[command] != 0;
|
2003-01-30 18:39:54 +01:00
|
|
|
}
|
2002-09-16 14:55:19 +02:00
|
|
|
|
2002-01-29 17:32:16 +01:00
|
|
|
/*
|
2005-02-09 03:43:27 +01:00
|
|
|
Reset per-hour user resource limits when it has been more than
|
|
|
|
an hour since they were last checked
|
2002-02-13 21:37:19 +01:00
|
|
|
|
2005-02-09 03:43:27 +01:00
|
|
|
SYNOPSIS:
|
|
|
|
time_out_user_resource_limits()
|
|
|
|
thd Thread handler
|
|
|
|
uc User connection details
|
2002-01-29 17:32:16 +01:00
|
|
|
|
2005-02-09 03:43:27 +01:00
|
|
|
NOTE:
|
|
|
|
This assumes that the LOCK_user_conn mutex has been acquired, so it is
|
|
|
|
safe to test and modify members of the USER_CONN structure.
|
|
|
|
*/
|
2002-05-15 12:50:38 +02:00
|
|
|
|
2005-02-19 00:23:17 +01:00
|
|
|
static void time_out_user_resource_limits(THD *thd, USER_CONN *uc)
|
2002-01-29 17:32:16 +01:00
|
|
|
{
|
2002-06-27 10:27:04 +02:00
|
|
|
time_t check_time = thd->start_time ? thd->start_time : time(NULL);
|
2005-02-09 03:43:27 +01:00
|
|
|
DBUG_ENTER("time_out_user_resource_limits");
|
2002-01-29 17:32:16 +01:00
|
|
|
|
2002-06-27 10:27:04 +02:00
|
|
|
/* If more than a hour since last check, reset resource checking */
|
2002-06-20 15:46:25 +02:00
|
|
|
if (check_time - uc->intime >= 3600)
|
|
|
|
{
|
|
|
|
uc->questions=1;
|
|
|
|
uc->updates=0;
|
|
|
|
uc->conn_per_hour=0;
|
|
|
|
uc->intime=check_time;
|
|
|
|
}
|
2005-02-09 03:43:27 +01:00
|
|
|
|
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Check if maximum queries per hour limit has been reached
|
|
|
|
returns 0 if OK.
|
|
|
|
*/
|
|
|
|
|
|
|
|
static bool check_mqh(THD *thd, uint check_command)
|
|
|
|
{
|
2005-02-19 00:23:17 +01:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
|
|
|
bool error= 0;
|
2005-02-09 03:43:27 +01:00
|
|
|
USER_CONN *uc=thd->user_connect;
|
|
|
|
DBUG_ENTER("check_mqh");
|
|
|
|
DBUG_ASSERT(uc != 0);
|
|
|
|
|
|
|
|
(void) pthread_mutex_lock(&LOCK_user_conn);
|
|
|
|
|
|
|
|
time_out_user_resource_limits(thd, uc);
|
|
|
|
|
2002-06-27 10:27:04 +02:00
|
|
|
/* Check that we have not done too many questions / hour */
|
2002-06-20 15:46:25 +02:00
|
|
|
if (uc->user_resources.questions &&
|
|
|
|
uc->questions++ >= uc->user_resources.questions)
|
|
|
|
{
|
2004-10-20 15:06:54 +02:00
|
|
|
net_printf_error(thd, ER_USER_LIMIT_REACHED, uc->user, "max_questions",
|
|
|
|
(long) uc->user_resources.questions);
|
2002-06-20 15:46:25 +02:00
|
|
|
error=1;
|
|
|
|
goto end;
|
|
|
|
}
|
2002-05-15 12:50:38 +02:00
|
|
|
if (check_command < (uint) SQLCOM_END)
|
2002-01-29 17:32:16 +01:00
|
|
|
{
|
2002-06-27 10:27:04 +02:00
|
|
|
/* Check that we have not done too many updates / hour */
|
|
|
|
if (uc->user_resources.updates && uc_update_queries[check_command] &&
|
|
|
|
uc->updates++ >= uc->user_resources.updates)
|
|
|
|
{
|
2004-10-20 15:06:54 +02:00
|
|
|
net_printf_error(thd, ER_USER_LIMIT_REACHED, uc->user, "max_updates",
|
|
|
|
(long) uc->user_resources.updates);
|
2002-06-27 10:27:04 +02:00
|
|
|
error=1;
|
|
|
|
goto end;
|
|
|
|
}
|
2002-01-29 17:32:16 +01:00
|
|
|
}
|
|
|
|
end:
|
2005-02-09 03:43:27 +01:00
|
|
|
(void) pthread_mutex_unlock(&LOCK_user_conn);
|
2002-02-01 19:53:24 +01:00
|
|
|
DBUG_RETURN(error);
|
2005-02-19 00:23:17 +01:00
|
|
|
#else
|
|
|
|
return (0);
|
2003-12-19 15:25:50 +01:00
|
|
|
#endif /* NO_EMBEDDED_ACCESS_CHECKS */
|
2002-01-29 17:32:16 +01:00
|
|
|
}
|
|
|
|
|
2002-02-13 21:37:19 +01:00
|
|
|
|
2004-12-06 16:15:54 +01:00
|
|
|
static void reset_mqh(LEX_USER *lu, bool get_them= 0)
|
2002-02-01 19:53:24 +01:00
|
|
|
{
|
2003-12-19 15:25:50 +01:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2002-02-13 21:37:19 +01:00
|
|
|
(void) pthread_mutex_lock(&LOCK_user_conn);
|
2002-11-30 18:58:53 +01:00
|
|
|
if (lu) // for GRANT
|
2002-02-01 19:53:24 +01:00
|
|
|
{
|
2002-05-15 12:50:38 +02:00
|
|
|
USER_CONN *uc;
|
2002-06-25 20:20:10 +02:00
|
|
|
uint temp_len=lu->user.length+lu->host.length+2;
|
2006-01-11 00:07:40 +01:00
|
|
|
char temp_user[USER_HOST_BUFF_SIZE];
|
2002-02-01 19:53:24 +01:00
|
|
|
|
2002-02-13 21:37:19 +01:00
|
|
|
memcpy(temp_user,lu->user.str,lu->user.length);
|
|
|
|
memcpy(temp_user+lu->user.length+1,lu->host.str,lu->host.length);
|
2002-05-15 12:50:38 +02:00
|
|
|
temp_user[lu->user.length]='\0'; temp_user[temp_len-1]=0;
|
2002-02-13 21:37:19 +01:00
|
|
|
if ((uc = (struct user_conn *) hash_search(&hash_user_connections,
|
2002-06-20 15:46:25 +02:00
|
|
|
(byte*) temp_user, temp_len)))
|
2002-02-01 19:53:24 +01:00
|
|
|
{
|
|
|
|
uc->questions=0;
|
2002-06-20 15:46:25 +02:00
|
|
|
get_mqh(temp_user,&temp_user[lu->user.length+1],uc);
|
2002-05-15 12:50:38 +02:00
|
|
|
uc->updates=0;
|
|
|
|
uc->conn_per_hour=0;
|
2002-02-01 19:53:24 +01:00
|
|
|
}
|
|
|
|
}
|
2004-07-07 10:29:39 +02:00
|
|
|
else
|
2002-02-01 19:53:24 +01:00
|
|
|
{
|
2004-07-07 10:29:39 +02:00
|
|
|
/* for FLUSH PRIVILEGES and FLUSH USER_RESOURCES */
|
2002-02-13 21:37:19 +01:00
|
|
|
for (uint idx=0;idx < hash_user_connections.records; idx++)
|
2002-02-01 19:53:24 +01:00
|
|
|
{
|
2003-12-19 15:25:50 +01:00
|
|
|
USER_CONN *uc=(struct user_conn *) hash_element(&hash_user_connections,
|
|
|
|
idx);
|
2002-05-15 12:50:38 +02:00
|
|
|
if (get_them)
|
|
|
|
get_mqh(uc->user,uc->host,uc);
|
|
|
|
uc->questions=0;
|
|
|
|
uc->updates=0;
|
|
|
|
uc->conn_per_hour=0;
|
2002-02-01 19:53:24 +01:00
|
|
|
}
|
|
|
|
}
|
2002-02-13 21:37:19 +01:00
|
|
|
(void) pthread_mutex_unlock(&LOCK_user_conn);
|
2003-12-19 15:25:50 +01:00
|
|
|
#endif /* NO_EMBEDDED_ACCESS_CHECKS */
|
2002-02-01 19:53:24 +01:00
|
|
|
}
|
2002-02-13 21:37:19 +01:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
/*
|
2003-07-01 21:40:59 +02:00
|
|
|
Perform handshake, authorize client and update thd ACL variables.
|
2003-02-14 10:47:41 +01:00
|
|
|
SYNOPSIS
|
2003-07-01 21:40:59 +02:00
|
|
|
check_connection()
|
2003-07-18 16:25:54 +02:00
|
|
|
thd thread handle
|
2003-02-14 10:47:41 +01:00
|
|
|
|
|
|
|
RETURN
|
2003-07-18 16:25:54 +02:00
|
|
|
0 success, OK is sent to user, thd is updated.
|
2003-07-01 21:40:59 +02:00
|
|
|
-1 error, which is sent to user
|
|
|
|
> 0 error code (not sent to user)
|
2000-07-31 21:29:14 +02:00
|
|
|
*/
|
|
|
|
|
2003-09-26 12:33:13 +02:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
|
|
|
static int check_connection(THD *thd)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2003-07-01 21:40:59 +02:00
|
|
|
uint connect_errors= 0;
|
2000-07-31 21:29:14 +02:00
|
|
|
NET *net= &thd->net;
|
2005-02-15 13:49:52 +01:00
|
|
|
ulong pkt_len= 0;
|
|
|
|
char *end;
|
2003-02-14 10:47:41 +01:00
|
|
|
|
2003-07-01 21:40:59 +02:00
|
|
|
DBUG_PRINT("info",
|
|
|
|
("New connection received on %s", vio_description(net->vio)));
|
2005-12-04 14:02:06 +01:00
|
|
|
#ifdef SIGNAL_WITH_VIO_CLOSE
|
2005-12-04 14:34:47 +01:00
|
|
|
thd->set_active_vio(net->vio);
|
2005-12-04 14:02:06 +01:00
|
|
|
#endif
|
2003-07-01 21:40:59 +02:00
|
|
|
|
2005-09-20 20:20:38 +02:00
|
|
|
if (!thd->main_security_ctx.host) // If TCP/IP connection
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2000-12-07 13:08:48 +01:00
|
|
|
char ip[30];
|
2000-08-21 13:35:27 +02:00
|
|
|
|
2003-02-17 21:07:26 +01:00
|
|
|
if (vio_peer_addr(net->vio, ip, &thd->peer_port))
|
2000-07-31 21:29:14 +02:00
|
|
|
return (ER_BAD_HOST_ERROR);
|
2005-09-20 20:20:38 +02:00
|
|
|
if (!(thd->main_security_ctx.ip= my_strdup(ip,MYF(0))))
|
2000-07-31 21:29:14 +02:00
|
|
|
return (ER_OUT_OF_RESOURCES);
|
2005-09-20 20:20:38 +02:00
|
|
|
thd->main_security_ctx.host_or_ip= thd->main_security_ctx.ip;
|
2005-02-15 13:49:52 +01:00
|
|
|
vio_in_addr(net->vio,&thd->remote.sin_addr);
|
2005-07-19 02:00:37 +02:00
|
|
|
if (!(specialflag & SPECIAL_NO_RESOLVE))
|
2003-07-04 22:06:19 +02:00
|
|
|
{
|
2005-07-19 02:00:37 +02:00
|
|
|
vio_in_addr(net->vio,&thd->remote.sin_addr);
|
2005-09-20 20:20:38 +02:00
|
|
|
thd->main_security_ctx.host=
|
|
|
|
ip_to_hostname(&thd->remote.sin_addr, &connect_errors);
|
2005-07-19 02:00:37 +02:00
|
|
|
/* Cut very long hostnames to avoid possible overflows */
|
2005-09-20 20:20:38 +02:00
|
|
|
if (thd->main_security_ctx.host)
|
2003-02-12 20:55:37 +01:00
|
|
|
{
|
2005-09-20 20:20:38 +02:00
|
|
|
if (thd->main_security_ctx.host != my_localhost)
|
|
|
|
thd->main_security_ctx.host[min(strlen(thd->main_security_ctx.host),
|
|
|
|
HOSTNAME_LENGTH)]= 0;
|
|
|
|
thd->main_security_ctx.host_or_ip= thd->main_security_ctx.host;
|
2003-02-12 20:55:37 +01:00
|
|
|
}
|
2005-07-19 02:00:37 +02:00
|
|
|
if (connect_errors > max_connect_errors)
|
|
|
|
return(ER_HOST_IS_BLOCKED);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2001-10-11 17:58:40 +02:00
|
|
|
DBUG_PRINT("info",("Host: %s ip: %s",
|
2005-09-20 20:20:38 +02:00
|
|
|
(thd->main_security_ctx.host ?
|
|
|
|
thd->main_security_ctx.host : "unknown host"),
|
|
|
|
(thd->main_security_ctx.ip ?
|
|
|
|
thd->main_security_ctx.ip : "unknown ip")));
|
|
|
|
if (acl_check_host(thd->main_security_ctx.host, thd->main_security_ctx.ip))
|
2000-07-31 21:29:14 +02:00
|
|
|
return(ER_HOST_NOT_PRIVILEGED);
|
|
|
|
}
|
2000-09-02 06:58:42 +02:00
|
|
|
else /* Hostname given means that the connection was on a socket */
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2005-09-20 20:20:38 +02:00
|
|
|
DBUG_PRINT("info",("Host: %s", thd->main_security_ctx.host));
|
|
|
|
thd->main_security_ctx.host_or_ip= thd->main_security_ctx.host;
|
|
|
|
thd->main_security_ctx.ip= 0;
|
2005-02-15 13:42:13 +01:00
|
|
|
/* Reset sin_addr */
|
|
|
|
bzero((char*) &thd->remote, sizeof(thd->remote));
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
vio_keepalive(net->vio, TRUE);
|
|
|
|
{
|
2001-02-15 17:00:30 +01:00
|
|
|
/* buff[] needs to big enough to hold the server_version variable */
|
2003-07-01 21:40:59 +02:00
|
|
|
char buff[SERVER_VERSION_LENGTH + SCRAMBLE_LENGTH + 64];
|
2003-05-26 18:01:20 +02:00
|
|
|
ulong client_flags = (CLIENT_LONG_FLAG | CLIENT_CONNECT_WITH_DB |
|
|
|
|
CLIENT_PROTOCOL_41 | CLIENT_SECURE_CONNECTION);
|
2001-12-05 12:03:00 +01:00
|
|
|
|
2001-03-06 14:24:08 +01:00
|
|
|
if (opt_using_transactions)
|
|
|
|
client_flags|=CLIENT_TRANSACTIONS;
|
|
|
|
#ifdef HAVE_COMPRESS
|
|
|
|
client_flags |= CLIENT_COMPRESS;
|
|
|
|
#endif /* HAVE_COMPRESS */
|
2001-12-05 12:03:00 +01:00
|
|
|
#ifdef HAVE_OPENSSL
|
|
|
|
if (ssl_acceptor_fd)
|
2004-10-07 00:45:06 +02:00
|
|
|
client_flags |= CLIENT_SSL; /* Wow, SSL is available! */
|
2001-12-05 12:03:00 +01:00
|
|
|
#endif /* HAVE_OPENSSL */
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2003-07-01 21:40:59 +02:00
|
|
|
end= strnmov(buff, server_version, SERVER_VERSION_LENGTH) + 1;
|
|
|
|
int4store((uchar*) end, thd->thread_id);
|
|
|
|
end+= 4;
|
|
|
|
/*
|
|
|
|
So as check_connection is the only entry point to authorization
|
|
|
|
procedure, scramble is set here. This gives us new scramble for
|
|
|
|
each handshake.
|
|
|
|
*/
|
|
|
|
create_random_string(thd->scramble, SCRAMBLE_LENGTH, &thd->rand);
|
|
|
|
/*
|
|
|
|
Old clients does not understand long scrambles, but can ignore packet
|
2003-07-04 18:52:04 +02:00
|
|
|
tail: that's why first part of the scramble is placed here, and second
|
2003-07-01 21:40:59 +02:00
|
|
|
part at the end of packet.
|
|
|
|
*/
|
2003-07-18 16:25:54 +02:00
|
|
|
end= strmake(end, thd->scramble, SCRAMBLE_LENGTH_323) + 1;
|
2003-07-01 21:40:59 +02:00
|
|
|
|
|
|
|
int2store(end, client_flags);
|
|
|
|
/* write server characteristics: up to 16 bytes allowed */
|
2003-06-03 11:59:17 +02:00
|
|
|
end[2]=(char) default_charset_info->number;
|
2003-07-01 21:40:59 +02:00
|
|
|
int2store(end+3, thd->server_status);
|
|
|
|
bzero(end+5, 13);
|
|
|
|
end+= 18;
|
|
|
|
/* write scramble tail */
|
|
|
|
end= strmake(end, thd->scramble + SCRAMBLE_LENGTH_323,
|
|
|
|
SCRAMBLE_LENGTH - SCRAMBLE_LENGTH_323) + 1;
|
|
|
|
|
|
|
|
/* At this point we write connection message and read reply */
|
|
|
|
if (net_write_command(net, (uchar) protocol_version, "", 0, buff,
|
2000-07-31 21:29:14 +02:00
|
|
|
(uint) (end-buff)) ||
|
2003-02-12 20:55:37 +01:00
|
|
|
(pkt_len= my_net_read(net)) == packet_error ||
|
2000-07-31 21:29:14 +02:00
|
|
|
pkt_len < MIN_HANDSHAKE_SIZE)
|
|
|
|
{
|
|
|
|
inc_host_errors(&thd->remote.sin_addr);
|
|
|
|
return(ER_HANDSHAKE_ERROR);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#ifdef _CUSTOMCONFIG_
|
|
|
|
#include "_cust_sql_parse.h"
|
|
|
|
#endif
|
|
|
|
if (connect_errors)
|
|
|
|
reset_host_errors(&thd->remote.sin_addr);
|
2002-07-23 17:31:22 +02:00
|
|
|
if (thd->packet.alloc(thd->variables.net_buffer_length))
|
2000-07-31 21:29:14 +02:00
|
|
|
return(ER_OUT_OF_RESOURCES);
|
|
|
|
|
|
|
|
thd->client_capabilities=uint2korr(net->read_pos);
|
2003-01-18 15:39:21 +01:00
|
|
|
if (thd->client_capabilities & CLIENT_PROTOCOL_41)
|
|
|
|
{
|
|
|
|
thd->client_capabilities|= ((ulong) uint2korr(net->read_pos+2)) << 16;
|
|
|
|
thd->max_client_packet_length= uint4korr(net->read_pos+4);
|
2003-08-18 23:08:08 +02:00
|
|
|
DBUG_PRINT("info", ("client_character_set: %d", (uint) net->read_pos[8]));
|
|
|
|
/*
|
|
|
|
Use server character set and collation if
|
2005-09-05 17:00:11 +02:00
|
|
|
- opt_character_set_client_handshake is not set
|
2003-08-18 23:08:08 +02:00
|
|
|
- client has not specified a character set
|
|
|
|
- client character set is the same as the servers
|
|
|
|
- client character set doesn't exists in server
|
|
|
|
*/
|
2005-09-05 17:00:11 +02:00
|
|
|
if (!opt_character_set_client_handshake ||
|
2005-08-31 10:03:24 +02:00
|
|
|
!(thd->variables.character_set_client=
|
2003-08-18 23:08:08 +02:00
|
|
|
get_charset((uint) net->read_pos[8], MYF(0))) ||
|
|
|
|
!my_strcasecmp(&my_charset_latin1,
|
|
|
|
global_system_variables.character_set_client->name,
|
|
|
|
thd->variables.character_set_client->name))
|
2003-05-28 10:54:18 +02:00
|
|
|
{
|
2003-05-26 18:01:20 +02:00
|
|
|
thd->variables.character_set_client=
|
|
|
|
global_system_variables.character_set_client;
|
2003-05-28 10:54:18 +02:00
|
|
|
thd->variables.collation_connection=
|
|
|
|
global_system_variables.collation_connection;
|
2003-05-30 20:09:35 +02:00
|
|
|
thd->variables.character_set_results=
|
|
|
|
global_system_variables.character_set_results;
|
2003-05-28 10:54:18 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2003-05-30 20:09:35 +02:00
|
|
|
thd->variables.character_set_results=
|
2003-05-28 10:54:18 +02:00
|
|
|
thd->variables.collation_connection=
|
|
|
|
thd->variables.character_set_client;
|
|
|
|
}
|
2003-08-18 23:08:08 +02:00
|
|
|
thd->update_charset();
|
2003-05-26 18:01:20 +02:00
|
|
|
end= (char*) net->read_pos+32;
|
2003-01-18 15:39:21 +01:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
thd->max_client_packet_length= uint3korr(net->read_pos+2);
|
|
|
|
end= (char*) net->read_pos+5;
|
|
|
|
}
|
|
|
|
|
2001-08-09 17:51:38 +02:00
|
|
|
if (thd->client_capabilities & CLIENT_IGNORE_SPACE)
|
2003-01-16 01:04:50 +01:00
|
|
|
thd->variables.sql_mode|= MODE_IGNORE_SPACE;
|
2000-07-31 21:29:14 +02:00
|
|
|
#ifdef HAVE_OPENSSL
|
2001-12-05 12:03:00 +01:00
|
|
|
DBUG_PRINT("info", ("client capabilities: %d", thd->client_capabilities));
|
2000-07-31 21:29:14 +02:00
|
|
|
if (thd->client_capabilities & CLIENT_SSL)
|
|
|
|
{
|
|
|
|
/* Do the SSL layering. */
|
2003-06-19 12:38:52 +02:00
|
|
|
if (!ssl_acceptor_fd)
|
|
|
|
{
|
|
|
|
inc_host_errors(&thd->remote.sin_addr);
|
|
|
|
return(ER_HANDSHAKE_ERROR);
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_PRINT("info", ("IO layer change in progress..."));
|
2002-11-05 09:05:58 +01:00
|
|
|
if (sslaccept(ssl_acceptor_fd, net->vio, thd->variables.net_wait_timeout))
|
|
|
|
{
|
2005-09-01 11:46:43 +02:00
|
|
|
DBUG_PRINT("error", ("Failed to accept new SSL connection"));
|
2002-11-05 09:05:58 +01:00
|
|
|
inc_host_errors(&thd->remote.sin_addr);
|
2002-11-30 18:58:53 +01:00
|
|
|
return(ER_HANDSHAKE_ERROR);
|
2002-11-05 09:05:58 +01:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_PRINT("info", ("Reading user information over SSL layer"));
|
2003-07-01 21:40:59 +02:00
|
|
|
if ((pkt_len= my_net_read(net)) == packet_error ||
|
2000-07-31 21:29:14 +02:00
|
|
|
pkt_len < NORMAL_HANDSHAKE_SIZE)
|
|
|
|
{
|
2001-12-05 12:03:00 +01:00
|
|
|
DBUG_PRINT("error", ("Failed to read user information (pkt_len= %lu)",
|
|
|
|
pkt_len));
|
2000-07-31 21:29:14 +02:00
|
|
|
inc_host_errors(&thd->remote.sin_addr);
|
|
|
|
return(ER_HANDSHAKE_ERROR);
|
|
|
|
}
|
|
|
|
}
|
2003-01-18 15:39:21 +01:00
|
|
|
#endif
|
|
|
|
|
|
|
|
if (end >= (char*) net->read_pos+ pkt_len +2)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2003-01-18 15:39:21 +01:00
|
|
|
inc_host_errors(&thd->remote.sin_addr);
|
|
|
|
return(ER_HANDSHAKE_ERROR);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
if (thd->client_capabilities & CLIENT_INTERACTIVE)
|
2003-02-12 20:55:37 +01:00
|
|
|
thd->variables.net_wait_timeout= thd->variables.net_interactive_timeout;
|
2001-03-21 19:13:46 +01:00
|
|
|
if ((thd->client_capabilities & CLIENT_TRANSACTIONS) &&
|
2003-02-12 20:55:37 +01:00
|
|
|
opt_using_transactions)
|
|
|
|
net->return_status= &thd->server_status;
|
2002-07-23 17:31:22 +02:00
|
|
|
net->read_timeout=(uint) thd->variables.net_read_timeout;
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2003-07-01 21:40:59 +02:00
|
|
|
char *user= end;
|
|
|
|
char *passwd= strend(user)+1;
|
2006-04-18 10:46:17 +02:00
|
|
|
uint user_len= passwd - user - 1;
|
2003-07-04 18:52:04 +02:00
|
|
|
char *db= passwd;
|
2005-01-24 15:48:25 +01:00
|
|
|
char db_buff[NAME_LEN+1]; // buffer to store db in utf8
|
2003-12-24 14:58:06 +01:00
|
|
|
char user_buff[USERNAME_LENGTH+1]; // buffer to store user in utf8
|
2005-01-24 15:48:25 +01:00
|
|
|
uint dummy_errors;
|
|
|
|
|
|
|
|
/*
|
2003-07-04 18:52:04 +02:00
|
|
|
Old clients send null-terminated string as password; new clients send
|
|
|
|
the size (1 byte) + string (not null-terminated). Hence in case of empty
|
|
|
|
password both send '\0'.
|
|
|
|
*/
|
2005-01-24 15:48:25 +01:00
|
|
|
uint passwd_len= thd->client_capabilities & CLIENT_SECURE_CONNECTION ?
|
2003-07-04 18:52:04 +02:00
|
|
|
*passwd++ : strlen(passwd);
|
|
|
|
db= thd->client_capabilities & CLIENT_CONNECT_WITH_DB ?
|
|
|
|
db + passwd_len + 1 : 0;
|
2006-04-25 23:06:04 +02:00
|
|
|
uint db_len= db ? strlen(db) : 0;
|
|
|
|
|
2006-04-26 02:12:06 +02:00
|
|
|
if (passwd + passwd_len + db_len > (char *)net->read_pos + pkt_len)
|
2006-04-25 23:06:04 +02:00
|
|
|
{
|
|
|
|
inc_host_errors(&thd->remote.sin_addr);
|
|
|
|
return ER_HANDSHAKE_ERROR;
|
|
|
|
}
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2003-07-31 15:11:52 +02:00
|
|
|
/* Since 4.1 all database names are stored in utf8 */
|
|
|
|
if (db)
|
2002-11-30 18:58:53 +01:00
|
|
|
{
|
2003-09-16 00:39:25 +02:00
|
|
|
db_buff[copy_and_convert(db_buff, sizeof(db_buff)-1,
|
|
|
|
system_charset_info,
|
2006-04-25 23:06:04 +02:00
|
|
|
db, db_len,
|
2004-10-29 14:00:39 +02:00
|
|
|
thd->charset(), &dummy_errors)]= 0;
|
2003-09-03 12:12:10 +02:00
|
|
|
db= db_buff;
|
2003-07-31 15:11:52 +02:00
|
|
|
}
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2006-04-18 10:46:17 +02:00
|
|
|
user_buff[user_len= copy_and_convert(user_buff, sizeof(user_buff)-1,
|
|
|
|
system_charset_info, user, user_len,
|
|
|
|
thd->charset(), &dummy_errors)]= '\0';
|
2005-01-24 15:48:25 +01:00
|
|
|
user= user_buff;
|
2003-12-24 14:58:06 +01:00
|
|
|
|
2006-04-18 10:46:17 +02:00
|
|
|
/* If username starts and ends in "'", chop them off */
|
|
|
|
if (user_len > 1 && user[0] == '\'' && user[user_len - 1] == '\'')
|
|
|
|
{
|
|
|
|
user[user_len-1]= 0;
|
|
|
|
user++;
|
|
|
|
user_len-= 2;
|
|
|
|
}
|
|
|
|
|
2005-09-20 20:20:38 +02:00
|
|
|
if (thd->main_security_ctx.user)
|
|
|
|
x_free(thd->main_security_ctx.user);
|
|
|
|
if (!(thd->main_security_ctx.user= my_strdup(user, MYF(0))))
|
2003-07-18 16:25:54 +02:00
|
|
|
return (ER_OUT_OF_RESOURCES);
|
2004-10-14 17:03:46 +02:00
|
|
|
return check_user(thd, COM_CONNECT, passwd, passwd_len, db, TRUE);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
2003-02-14 10:47:41 +01:00
|
|
|
|
2003-12-08 06:13:14 +01:00
|
|
|
void execute_init_command(THD *thd, sys_var_str *init_command_var,
|
|
|
|
rw_lock_t *var_mutex)
|
2003-07-18 11:11:01 +02:00
|
|
|
{
|
|
|
|
Vio* save_vio;
|
|
|
|
ulong save_client_capabilities;
|
|
|
|
|
2003-12-08 06:13:14 +01:00
|
|
|
thd->proc_info= "Execution of init_command";
|
|
|
|
/*
|
|
|
|
We need to lock init_command_var because
|
|
|
|
during execution of init_command_var query
|
|
|
|
values of init_command_var can't be changed
|
|
|
|
*/
|
|
|
|
rw_rdlock(var_mutex);
|
|
|
|
thd->query= init_command_var->value;
|
|
|
|
thd->query_length= init_command_var->value_length;
|
2003-07-18 11:11:01 +02:00
|
|
|
save_client_capabilities= thd->client_capabilities;
|
|
|
|
thd->client_capabilities|= CLIENT_MULTI_QUERIES;
|
2003-12-08 06:13:14 +01:00
|
|
|
/*
|
|
|
|
We don't need return result of execution to client side.
|
|
|
|
To forbid this we should set thd->net.vio to 0.
|
|
|
|
*/
|
2003-07-18 11:11:01 +02:00
|
|
|
save_vio= thd->net.vio;
|
|
|
|
thd->net.vio= 0;
|
2005-01-20 09:41:37 +01:00
|
|
|
thd->net.no_send_error= 0;
|
2003-07-18 11:11:01 +02:00
|
|
|
dispatch_command(COM_QUERY, thd, thd->query, thd->query_length+1);
|
2003-12-08 06:13:14 +01:00
|
|
|
rw_unlock(var_mutex);
|
2003-07-18 11:11:01 +02:00
|
|
|
thd->client_capabilities= save_client_capabilities;
|
|
|
|
thd->net.vio= save_vio;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2005-10-08 16:39:55 +02:00
|
|
|
pthread_handler_t handle_one_connection(void *arg)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
THD *thd=(THD*) arg;
|
|
|
|
uint launch_time =
|
2000-08-23 14:02:27 +02:00
|
|
|
(uint) ((thd->thr_create_time = time(NULL)) - thd->connect_time);
|
2000-07-31 21:29:14 +02:00
|
|
|
if (launch_time >= slow_launch_time)
|
|
|
|
statistic_increment(slow_launch_threads,&LOCK_status );
|
|
|
|
|
|
|
|
pthread_detach_this_thread();
|
|
|
|
|
2006-05-02 19:56:43 +02:00
|
|
|
#if !defined( __WIN__) // Win32 calls this in pthread_create
|
2004-07-07 10:29:39 +02:00
|
|
|
/* The following calls needs to be done before we call DBUG_ macros */
|
2002-06-03 23:40:27 +02:00
|
|
|
if (!(test_flags & TEST_NO_THREADS) & my_thread_init())
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2003-05-26 18:01:20 +02:00
|
|
|
close_connection(thd, ER_OUT_OF_RESOURCES, 1);
|
2002-08-22 15:50:58 +02:00
|
|
|
statistic_increment(aborted_connects,&LOCK_status);
|
2000-07-31 21:29:14 +02:00
|
|
|
end_thread(thd,0);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2002-01-15 23:42:52 +01:00
|
|
|
/*
|
|
|
|
handle_one_connection() is the only way a thread would start
|
|
|
|
and would always be on top of the stack, therefore, the thread
|
|
|
|
stack always starts at the address of the first local variable
|
|
|
|
of handle_one_connection, which is thd. We need to know the
|
|
|
|
start of the stack so that we could check for stack overruns.
|
|
|
|
*/
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_PRINT("info", ("handle_one_connection called by thread %d\n",
|
|
|
|
thd->thread_id));
|
2004-07-07 10:29:39 +02:00
|
|
|
/* now that we've called my_thread_init(), it is safe to call DBUG_* */
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2001-08-22 00:45:07 +02:00
|
|
|
#if defined(__WIN__)
|
2005-05-06 10:39:30 +02:00
|
|
|
init_signals();
|
2006-05-02 19:56:43 +02:00
|
|
|
#elif !defined(__NETWARE__)
|
2000-07-31 21:29:14 +02:00
|
|
|
sigset_t set;
|
|
|
|
VOID(sigemptyset(&set)); // Get mask in use
|
|
|
|
VOID(pthread_sigmask(SIG_UNBLOCK,&set,&thd->block_signals));
|
|
|
|
#endif
|
2005-11-23 19:18:10 +01:00
|
|
|
thd->thread_stack= (char*) &thd;
|
2000-07-31 21:29:14 +02:00
|
|
|
if (thd->store_globals())
|
|
|
|
{
|
2003-05-26 18:01:20 +02:00
|
|
|
close_connection(thd, ER_OUT_OF_RESOURCES, 1);
|
2002-08-22 15:50:58 +02:00
|
|
|
statistic_increment(aborted_connects,&LOCK_status);
|
2000-07-31 21:29:14 +02:00
|
|
|
end_thread(thd,0);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
do
|
|
|
|
{
|
|
|
|
int error;
|
|
|
|
NET *net= &thd->net;
|
2005-09-20 20:20:38 +02:00
|
|
|
Security_context *sctx= thd->security_ctx;
|
2005-01-20 09:41:37 +01:00
|
|
|
net->no_send_error= 0;
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2003-07-01 21:40:59 +02:00
|
|
|
if ((error=check_connection(thd)))
|
2000-07-31 21:29:14 +02:00
|
|
|
{ // Wrong permissions
|
|
|
|
if (error > 0)
|
2005-09-15 21:29:07 +02:00
|
|
|
net_printf_error(thd, error, sctx->host_or_ip);
|
2000-07-31 21:29:14 +02:00
|
|
|
#ifdef __NT__
|
|
|
|
if (vio_type(net->vio) == VIO_TYPE_NAMEDPIPE)
|
2003-02-12 20:55:37 +01:00
|
|
|
my_sleep(1000); /* must wait after eof() */
|
2000-07-31 21:29:14 +02:00
|
|
|
#endif
|
2002-08-22 15:50:58 +02:00
|
|
|
statistic_increment(aborted_connects,&LOCK_status);
|
2000-07-31 21:29:14 +02:00
|
|
|
goto end_thread;
|
|
|
|
}
|
2003-01-28 07:38:28 +01:00
|
|
|
#ifdef __NETWARE__
|
2005-09-15 21:29:07 +02:00
|
|
|
netware_reg_user(sctx->ip, sctx->user, "MySQL");
|
2003-01-28 07:38:28 +01:00
|
|
|
#endif
|
2002-12-20 13:58:27 +01:00
|
|
|
if (thd->variables.max_join_size == HA_POS_ERROR)
|
2000-07-31 21:29:14 +02:00
|
|
|
thd->options |= OPTION_BIG_SELECTS;
|
|
|
|
if (thd->client_capabilities & CLIENT_COMPRESS)
|
|
|
|
net->compress=1; // Use compression
|
|
|
|
|
2003-07-18 11:11:01 +02:00
|
|
|
thd->version= refresh_version;
|
2004-10-14 00:53:59 +02:00
|
|
|
thd->proc_info= 0;
|
2004-11-12 13:38:01 +01:00
|
|
|
thd->command= COM_SLEEP;
|
2004-10-14 00:53:59 +02:00
|
|
|
thd->set_time();
|
|
|
|
thd->init_for_queries();
|
2004-10-29 18:26:52 +02:00
|
|
|
|
2005-09-15 21:29:07 +02:00
|
|
|
if (sys_init_connect.value_length && !(sctx->master_access & SUPER_ACL))
|
2003-07-18 11:11:01 +02:00
|
|
|
{
|
2003-12-08 06:13:14 +01:00
|
|
|
execute_init_command(thd, &sys_init_connect, &LOCK_sys_init_connect);
|
|
|
|
if (thd->query_error)
|
2003-12-09 19:05:41 +01:00
|
|
|
thd->killed= THD::KILL_CONNECTION;
|
2005-08-15 17:15:12 +02:00
|
|
|
thd->proc_info=0;
|
|
|
|
thd->set_time();
|
|
|
|
thd->init_for_queries();
|
2003-07-18 11:11:01 +02:00
|
|
|
}
|
|
|
|
|
2005-04-06 16:22:21 +02:00
|
|
|
while (!net->error && net->vio != 0 &&
|
|
|
|
!(thd->killed == THD::KILL_CONNECTION))
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2005-01-20 09:41:37 +01:00
|
|
|
net->no_send_error= 0;
|
2000-07-31 21:29:14 +02:00
|
|
|
if (do_command(thd))
|
|
|
|
break;
|
|
|
|
}
|
2002-02-01 19:53:24 +01:00
|
|
|
if (thd->user_connect)
|
|
|
|
decrease_user_connections(thd->user_connect);
|
2002-09-03 08:50:36 +02:00
|
|
|
if (net->error && net->vio != 0 && net->report_error)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-06-01 16:29:24 +02:00
|
|
|
if (!thd->killed && thd->variables.log_warnings > 1)
|
2004-10-29 18:26:52 +02:00
|
|
|
sql_print_warning(ER(ER_NEW_ABORTING_CONNECTION),
|
2004-09-04 20:17:09 +02:00
|
|
|
thd->thread_id,(thd->db ? thd->db : "unconnected"),
|
2005-09-15 21:29:07 +02:00
|
|
|
sctx->user ? sctx->user : "unauthenticated",
|
|
|
|
sctx->host_or_ip,
|
2004-09-04 20:17:09 +02:00
|
|
|
(net->last_errno ? ER(net->last_errno) :
|
|
|
|
ER(ER_UNKNOWN_ERROR)));
|
2004-10-20 15:06:54 +02:00
|
|
|
net_send_error(thd, net->last_errno, NullS);
|
2002-11-16 19:19:10 +01:00
|
|
|
statistic_increment(aborted_threads,&LOCK_status);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2003-06-05 11:29:13 +02:00
|
|
|
else if (thd->killed)
|
|
|
|
{
|
|
|
|
statistic_increment(aborted_threads,&LOCK_status);
|
|
|
|
}
|
2002-02-01 19:53:24 +01:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
end_thread:
|
2003-05-26 18:01:20 +02:00
|
|
|
close_connection(thd, 0, 1);
|
2000-07-31 21:29:14 +02:00
|
|
|
end_thread(thd,1);
|
|
|
|
/*
|
|
|
|
If end_thread returns, we are either running with --one-thread
|
|
|
|
or this thread has been schedule to handle the next query
|
|
|
|
*/
|
|
|
|
thd= current_thd;
|
2005-12-03 17:13:06 +01:00
|
|
|
thd->thread_stack= (char*) &thd;
|
2000-07-31 21:29:14 +02:00
|
|
|
} while (!(test_flags & TEST_NO_THREADS));
|
|
|
|
/* The following is only executed if we are not using --one-thread */
|
|
|
|
return(0); /* purecov: deadcode */
|
|
|
|
}
|
|
|
|
|
2003-09-29 18:07:53 +02:00
|
|
|
#endif /* EMBEDDED_LIBRARY */
|
|
|
|
|
2000-11-28 03:47:47 +01:00
|
|
|
/*
|
|
|
|
Execute commands from bootstrap_file.
|
|
|
|
Used when creating the initial grant tables
|
|
|
|
*/
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2005-10-08 16:39:55 +02:00
|
|
|
pthread_handler_t handle_bootstrap(void *arg)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2000-11-28 03:47:47 +01:00
|
|
|
THD *thd=(THD*) arg;
|
|
|
|
FILE *file=bootstrap_file;
|
|
|
|
char *buff;
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2001-03-06 14:24:08 +01:00
|
|
|
/* The following must be called before DBUG_ENTER */
|
2005-11-24 01:36:28 +01:00
|
|
|
thd->thread_stack= (char*) &thd;
|
2000-11-28 03:47:47 +01:00
|
|
|
if (my_thread_init() || thd->store_globals())
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2003-09-29 18:07:53 +02:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2003-05-26 18:01:20 +02:00
|
|
|
close_connection(thd, ER_OUT_OF_RESOURCES, 1);
|
2003-09-29 18:07:53 +02:00
|
|
|
#endif
|
2003-01-30 21:15:44 +01:00
|
|
|
thd->fatal_error();
|
2000-11-28 03:47:47 +01:00
|
|
|
goto end;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2001-03-06 14:24:08 +01:00
|
|
|
DBUG_ENTER("handle_bootstrap");
|
|
|
|
|
2003-09-29 18:07:53 +02:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2001-03-06 14:24:08 +01:00
|
|
|
pthread_detach_this_thread();
|
|
|
|
thd->thread_stack= (char*) &thd;
|
2006-05-02 19:56:43 +02:00
|
|
|
#if !defined(__WIN__) && !defined(__NETWARE__)
|
2000-07-31 21:29:14 +02:00
|
|
|
sigset_t set;
|
2002-08-04 15:14:51 +02:00
|
|
|
VOID(sigemptyset(&set)); // Get mask in use
|
|
|
|
VOID(pthread_sigmask(SIG_UNBLOCK,&set,&thd->block_signals));
|
2000-07-31 21:29:14 +02:00
|
|
|
#endif
|
2003-09-29 18:07:53 +02:00
|
|
|
#endif /* EMBEDDED_LIBRARY */
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2002-12-20 13:58:27 +01:00
|
|
|
if (thd->variables.max_join_size == HA_POS_ERROR)
|
2000-07-31 21:29:14 +02:00
|
|
|
thd->options |= OPTION_BIG_SELECTS;
|
|
|
|
|
|
|
|
thd->proc_info=0;
|
|
|
|
thd->version=refresh_version;
|
2005-09-15 21:29:07 +02:00
|
|
|
thd->security_ctx->priv_user=
|
|
|
|
thd->security_ctx->user= (char*) my_strdup("boot", MYF(MY_WME));
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->security_ctx->priv_host[0]=0;
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2000-11-28 03:47:47 +01:00
|
|
|
buff= (char*) thd->net.buff;
|
2003-12-21 20:26:45 +01:00
|
|
|
thd->init_for_queries();
|
2000-07-31 21:29:14 +02:00
|
|
|
while (fgets(buff, thd->net.max_packet, file))
|
|
|
|
{
|
2004-12-22 12:54:39 +01:00
|
|
|
ulong length= (ulong) strlen(buff);
|
|
|
|
while (buff[length-1] != '\n' && !feof(file))
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
We got only a part of the current string. Will try to increase
|
|
|
|
net buffer then read the rest of the current string.
|
|
|
|
*/
|
|
|
|
if (net_realloc(&(thd->net), 2 * thd->net.max_packet))
|
|
|
|
{
|
|
|
|
net_send_error(thd, ER_NET_PACKET_TOO_LARGE, NullS);
|
|
|
|
thd->fatal_error();
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
buff= (char*) thd->net.buff;
|
|
|
|
fgets(buff + length, thd->net.max_packet - length, file);
|
|
|
|
length+= (ulong) strlen(buff + length);
|
|
|
|
}
|
|
|
|
if (thd->is_fatal_error)
|
|
|
|
break;
|
|
|
|
|
2003-03-15 14:24:21 +01:00
|
|
|
while (length && (my_isspace(thd->charset(), buff[length-1]) ||
|
2002-03-12 18:37:58 +01:00
|
|
|
buff[length-1] == ';'))
|
2000-07-31 21:29:14 +02:00
|
|
|
length--;
|
|
|
|
buff[length]=0;
|
2001-10-08 22:20:19 +02:00
|
|
|
thd->query_length=length;
|
2005-08-24 01:45:15 +02:00
|
|
|
thd->query= thd->memdup_w_gap(buff, length+1,
|
|
|
|
thd->db_length+1+QUERY_CACHE_FLAGS_SIZE);
|
2001-12-21 06:00:58 +01:00
|
|
|
thd->query[length] = '\0';
|
2004-03-01 07:15:33 +01:00
|
|
|
/*
|
|
|
|
We don't need to obtain LOCK_thread_count here because in bootstrap
|
|
|
|
mode we have only one thread.
|
|
|
|
*/
|
2005-01-16 13:16:23 +01:00
|
|
|
thd->query_id=next_query_id();
|
2000-07-31 21:29:14 +02:00
|
|
|
mysql_parse(thd,thd->query,length);
|
|
|
|
close_thread_tables(thd); // Free tables
|
2003-01-30 21:15:44 +01:00
|
|
|
if (thd->is_fatal_error)
|
2000-11-28 03:47:47 +01:00
|
|
|
break;
|
2004-11-08 00:13:54 +01:00
|
|
|
free_root(thd->mem_root,MYF(MY_KEEP_PREALLOC));
|
2005-01-16 13:16:23 +01:00
|
|
|
#ifdef USING_TRANSACTIONS
|
2002-03-15 22:57:31 +01:00
|
|
|
free_root(&thd->transaction.mem_root,MYF(MY_KEEP_PREALLOC));
|
2005-01-16 13:16:23 +01:00
|
|
|
#endif
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2000-11-28 03:47:47 +01:00
|
|
|
|
|
|
|
/* thd->fatal_error should be set in case something went wrong */
|
|
|
|
end:
|
2005-01-16 13:16:23 +01:00
|
|
|
bootstrap_error= thd->is_fatal_error;
|
|
|
|
|
|
|
|
net_end(&thd->net);
|
|
|
|
thd->cleanup();
|
|
|
|
delete thd;
|
|
|
|
|
2003-09-29 18:07:53 +02:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2000-11-28 03:47:47 +01:00
|
|
|
(void) pthread_mutex_lock(&LOCK_thread_count);
|
|
|
|
thread_count--;
|
|
|
|
(void) pthread_mutex_unlock(&LOCK_thread_count);
|
2002-08-22 15:50:58 +02:00
|
|
|
(void) pthread_cond_broadcast(&COND_thread_count);
|
2000-11-28 03:47:47 +01:00
|
|
|
my_thread_end();
|
|
|
|
pthread_exit(0);
|
2003-09-29 18:07:53 +02:00
|
|
|
#endif
|
2005-01-16 13:16:23 +01:00
|
|
|
DBUG_RETURN(0);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
2002-10-02 12:33:08 +02:00
|
|
|
/* This works because items are allocated with sql_alloc() */
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2002-10-02 12:33:08 +02:00
|
|
|
void free_items(Item *item)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-11-28 15:53:17 +01:00
|
|
|
Item *next;
|
2004-07-15 03:19:07 +02:00
|
|
|
DBUG_ENTER("free_items");
|
2004-11-28 15:53:17 +01:00
|
|
|
for (; item ; item=next)
|
|
|
|
{
|
|
|
|
next=item->next;
|
2004-01-20 17:55:47 +01:00
|
|
|
item->delete_self();
|
2004-11-28 15:53:17 +01:00
|
|
|
}
|
2004-07-15 03:19:07 +02:00
|
|
|
DBUG_VOID_RETURN;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
2003-11-18 12:47:27 +01:00
|
|
|
/* This works because items are allocated with sql_alloc() */
|
|
|
|
|
|
|
|
void cleanup_items(Item *item)
|
|
|
|
{
|
2005-05-20 15:14:35 +02:00
|
|
|
DBUG_ENTER("cleanup_items");
|
2003-11-18 12:47:27 +01:00
|
|
|
for (; item ; item=item->next)
|
|
|
|
item->cleanup();
|
2005-05-20 15:14:35 +02:00
|
|
|
DBUG_VOID_RETURN;
|
2003-11-18 12:47:27 +01:00
|
|
|
}
|
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
int mysql_table_dump(THD* thd, char* db, char* tbl_name, int fd)
|
|
|
|
{
|
|
|
|
TABLE* table;
|
|
|
|
TABLE_LIST* table_list;
|
|
|
|
int error = 0;
|
|
|
|
DBUG_ENTER("mysql_table_dump");
|
|
|
|
db = (db && db[0]) ? db : thd->db;
|
2001-07-11 09:36:22 +02:00
|
|
|
if (!(table_list = (TABLE_LIST*) thd->calloc(sizeof(TABLE_LIST))))
|
2001-02-21 00:11:32 +01:00
|
|
|
DBUG_RETURN(1); // out of memory
|
2004-07-16 00:15:55 +02:00
|
|
|
table_list->db= db;
|
2005-01-06 12:00:13 +01:00
|
|
|
table_list->table_name= table_list->alias= tbl_name;
|
2004-07-16 00:15:55 +02:00
|
|
|
table_list->lock_type= TL_READ_NO_INSERT;
|
|
|
|
table_list->prev_global= &table_list; // can be removed after merge with 4.1
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2001-01-31 03:47:25 +01:00
|
|
|
if (!db || check_db_name(db))
|
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_WRONG_DB_NAME ,MYF(0), db ? db : "NULL");
|
2001-01-31 03:47:25 +01:00
|
|
|
goto err;
|
|
|
|
}
|
2003-01-29 17:56:34 +01:00
|
|
|
if (lower_case_table_names)
|
2003-02-07 14:47:24 +01:00
|
|
|
my_casedn_str(files_charset_info, tbl_name);
|
2005-01-06 12:00:13 +01:00
|
|
|
remove_escape(table_list->table_name);
|
2003-01-29 17:56:34 +01:00
|
|
|
|
|
|
|
if (!(table=open_ltable(thd, table_list, TL_READ_NO_INSERT)))
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
|
2004-04-10 00:14:32 +02:00
|
|
|
if (check_one_table_access(thd, SELECT_ACL, table_list))
|
2000-07-31 21:29:14 +02:00
|
|
|
goto err;
|
|
|
|
thd->free_list = 0;
|
2001-12-05 12:03:00 +01:00
|
|
|
thd->query_length=(uint) strlen(tbl_name);
|
2000-07-31 21:29:14 +02:00
|
|
|
thd->query = tbl_name;
|
2004-11-18 10:16:06 +01:00
|
|
|
if ((error = mysqld_dump_create_info(thd, table_list, -1)))
|
2001-10-08 22:20:19 +02:00
|
|
|
{
|
2003-02-07 14:47:24 +01:00
|
|
|
my_error(ER_GET_ERRNO, MYF(0), my_errno);
|
2001-10-08 22:20:19 +02:00
|
|
|
goto err;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
net_flush(&thd->net);
|
2003-01-19 00:31:43 +01:00
|
|
|
if ((error= table->file->dump(thd,fd)))
|
2004-06-10 21:18:57 +02:00
|
|
|
my_error(ER_GET_ERRNO, MYF(0), error);
|
2000-08-21 23:39:08 +02:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
err:
|
2000-08-21 23:39:08 +02:00
|
|
|
DBUG_RETURN(error);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
2005-02-01 20:48:05 +01:00
|
|
|
/*
|
|
|
|
Ends the current transaction and (maybe) begin the next
|
|
|
|
|
|
|
|
SYNOPSIS
|
2005-02-17 13:52:16 +01:00
|
|
|
end_trans()
|
2005-02-01 20:48:05 +01:00
|
|
|
thd Current thread
|
|
|
|
completion Completion type
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
0 - OK
|
|
|
|
*/
|
|
|
|
|
2005-02-17 13:52:16 +01:00
|
|
|
int end_trans(THD *thd, enum enum_mysql_completiontype completion)
|
2005-02-01 20:48:05 +01:00
|
|
|
{
|
|
|
|
bool do_release= 0;
|
|
|
|
int res= 0;
|
2005-02-17 13:52:16 +01:00
|
|
|
DBUG_ENTER("end_trans");
|
2005-02-01 20:48:05 +01:00
|
|
|
|
2005-07-30 10:19:57 +02:00
|
|
|
if (unlikely(thd->in_sub_stmt))
|
2005-06-07 12:53:08 +02:00
|
|
|
{
|
|
|
|
my_error(ER_COMMIT_NOT_ALLOWED_IN_SF_OR_TRG, MYF(0));
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
2005-10-05 19:58:16 +02:00
|
|
|
if (thd->transaction.xid_state.xa_state != XA_NOTR)
|
|
|
|
{
|
|
|
|
my_error(ER_XAER_RMFAIL, MYF(0),
|
|
|
|
xa_state_names[thd->transaction.xid_state.xa_state]);
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
2005-02-01 20:48:05 +01:00
|
|
|
switch (completion) {
|
|
|
|
case COMMIT:
|
|
|
|
/*
|
|
|
|
We don't use end_active_trans() here to ensure that this works
|
|
|
|
even if there is a problem with the OPTION_AUTO_COMMIT flag
|
|
|
|
(Which of course should never happen...)
|
|
|
|
*/
|
|
|
|
thd->server_status&= ~SERVER_STATUS_IN_TRANS;
|
2005-02-17 13:52:16 +01:00
|
|
|
res= ha_commit(thd);
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
thd->options&= ~(ulong) (OPTION_BEGIN | OPTION_STATUS_NO_TRANS_UPDATE |
|
|
|
|
OPTION_KEEP_LOG);
|
2005-02-01 20:48:05 +01:00
|
|
|
break;
|
|
|
|
case COMMIT_RELEASE:
|
2005-02-14 21:50:09 +01:00
|
|
|
do_release= 1; /* fall through */
|
2005-02-01 20:48:05 +01:00
|
|
|
case COMMIT_AND_CHAIN:
|
|
|
|
res= end_active_trans(thd);
|
|
|
|
if (!res && completion == COMMIT_AND_CHAIN)
|
|
|
|
res= begin_trans(thd);
|
|
|
|
break;
|
|
|
|
case ROLLBACK_RELEASE:
|
2005-02-14 21:50:09 +01:00
|
|
|
do_release= 1; /* fall through */
|
2005-02-01 20:48:05 +01:00
|
|
|
case ROLLBACK:
|
|
|
|
case ROLLBACK_AND_CHAIN:
|
|
|
|
{
|
|
|
|
thd->server_status&= ~SERVER_STATUS_IN_TRANS;
|
2005-02-16 17:34:02 +01:00
|
|
|
if (ha_rollback(thd))
|
2005-02-01 20:48:05 +01:00
|
|
|
res= -1;
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
thd->options&= ~(ulong) (OPTION_BEGIN | OPTION_STATUS_NO_TRANS_UPDATE |
|
|
|
|
OPTION_KEEP_LOG);
|
2005-02-01 20:48:05 +01:00
|
|
|
if (!res && (completion == ROLLBACK_AND_CHAIN))
|
|
|
|
res= begin_trans(thd);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
default:
|
|
|
|
res= -1;
|
|
|
|
my_error(ER_UNKNOWN_COM_ERROR, MYF(0));
|
|
|
|
DBUG_RETURN(-1);
|
|
|
|
}
|
2005-02-14 21:50:09 +01:00
|
|
|
|
2005-02-01 20:48:05 +01:00
|
|
|
if (res < 0)
|
|
|
|
my_error(thd->killed_errno(), MYF(0));
|
|
|
|
else if ((res == 0) && do_release)
|
2005-02-14 21:50:09 +01:00
|
|
|
thd->killed= THD::KILL_CONNECTION;
|
|
|
|
|
2005-02-01 20:48:05 +01:00
|
|
|
DBUG_RETURN(res);
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2003-02-07 14:47:24 +01:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2004-03-08 15:47:42 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
Read one command from socket and execute it (query or simple command).
|
|
|
|
This function is called in loop from thread function.
|
|
|
|
SYNOPSIS
|
|
|
|
do_command()
|
|
|
|
RETURN VALUE
|
|
|
|
0 success
|
|
|
|
1 request of thread shutdown (see dispatch_command() description)
|
|
|
|
*/
|
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
bool do_command(THD *thd)
|
|
|
|
{
|
|
|
|
char *packet;
|
2001-12-05 12:03:00 +01:00
|
|
|
uint old_timeout;
|
|
|
|
ulong packet_length;
|
2000-07-31 21:29:14 +02:00
|
|
|
NET *net;
|
|
|
|
enum enum_server_command command;
|
|
|
|
DBUG_ENTER("do_command");
|
|
|
|
|
|
|
|
net= &thd->net;
|
2003-07-01 18:05:08 +02:00
|
|
|
/*
|
|
|
|
indicator of uninitialized lex => normal flow of errors handling
|
|
|
|
(see my_message_sql)
|
|
|
|
*/
|
2003-07-09 17:07:12 +02:00
|
|
|
thd->lex->current_select= 0;
|
2000-07-31 21:29:14 +02:00
|
|
|
|
|
|
|
packet=0;
|
2002-07-23 17:31:22 +02:00
|
|
|
old_timeout=net->read_timeout;
|
2004-07-07 10:29:39 +02:00
|
|
|
/* Wait max for 8 hours */
|
2002-07-23 17:31:22 +02:00
|
|
|
net->read_timeout=(uint) thd->variables.net_wait_timeout;
|
2002-11-03 23:56:25 +01:00
|
|
|
thd->clear_error(); // Clear error message
|
2000-07-31 21:29:14 +02:00
|
|
|
|
|
|
|
net_new_transaction(net);
|
|
|
|
if ((packet_length=my_net_read(net)) == packet_error)
|
|
|
|
{
|
2003-01-04 14:17:16 +01:00
|
|
|
DBUG_PRINT("info",("Got error %d reading command from socket %s",
|
|
|
|
net->error,
|
|
|
|
vio_description(net->vio)));
|
|
|
|
/* Check if we can continue without closing the connection */
|
|
|
|
if (net->error != 3)
|
2003-06-05 11:29:13 +02:00
|
|
|
{
|
|
|
|
statistic_increment(aborted_threads,&LOCK_status);
|
2003-01-04 14:17:16 +01:00
|
|
|
DBUG_RETURN(TRUE); // We have to close it.
|
2003-06-05 11:29:13 +02:00
|
|
|
}
|
2004-10-20 15:06:54 +02:00
|
|
|
net_send_error(thd, net->last_errno, NullS);
|
2003-01-04 14:17:16 +01:00
|
|
|
net->error= 0;
|
2003-01-02 15:21:22 +01:00
|
|
|
DBUG_RETURN(FALSE);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
packet=(char*) net->read_pos;
|
|
|
|
command = (enum enum_server_command) (uchar) packet[0];
|
2003-01-04 14:17:16 +01:00
|
|
|
if (command >= COM_END)
|
|
|
|
command= COM_END; // Wrong command
|
2001-10-11 17:58:40 +02:00
|
|
|
DBUG_PRINT("info",("Command on %s = %d (%s)",
|
|
|
|
vio_description(net->vio), command,
|
|
|
|
command_name[command]));
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2002-07-23 17:31:22 +02:00
|
|
|
net->read_timeout=old_timeout; // restore it
|
2004-03-02 20:39:50 +01:00
|
|
|
/*
|
|
|
|
packet_length contains length of data, as it was stored in packet
|
|
|
|
header. In case of malformed header, packet_length can be zero.
|
|
|
|
If packet_length is not zero, my_net_read ensures that this number
|
|
|
|
of bytes was actually read from network. Additionally my_net_read
|
|
|
|
sets packet[packet_length]= 0 (thus if packet_length == 0,
|
|
|
|
command == packet[0] == COM_SLEEP).
|
|
|
|
In dispatch_command packet[packet_length] points beyond the end of packet.
|
|
|
|
*/
|
2001-12-05 12:03:00 +01:00
|
|
|
DBUG_RETURN(dispatch_command(command,thd, packet+1, (uint) packet_length));
|
2001-04-11 13:04:03 +02:00
|
|
|
}
|
2002-06-17 13:24:51 +02:00
|
|
|
#endif /* EMBEDDED_LIBRARY */
|
2001-04-11 13:04:03 +02:00
|
|
|
|
2005-02-13 20:47:00 +01:00
|
|
|
|
2004-03-02 20:39:50 +01:00
|
|
|
/*
|
|
|
|
Perform one connection-level (COM_XXXX) command.
|
2005-08-15 17:15:12 +02:00
|
|
|
|
2004-03-02 20:39:50 +01:00
|
|
|
SYNOPSIS
|
|
|
|
dispatch_command()
|
|
|
|
thd connection handle
|
|
|
|
command type of command to perform
|
|
|
|
packet data for the command, packet is always null-terminated
|
|
|
|
packet_length length of packet + 1 (to show that data is
|
|
|
|
null-terminated) except for COM_SLEEP, where it
|
|
|
|
can be zero.
|
|
|
|
RETURN VALUE
|
|
|
|
0 ok
|
|
|
|
1 request of thread shutdown, i. e. if command is
|
|
|
|
COM_QUIT/COM_SHUTDOWN
|
|
|
|
*/
|
2003-02-14 10:47:41 +01:00
|
|
|
|
2001-04-11 13:04:03 +02:00
|
|
|
bool dispatch_command(enum enum_server_command command, THD *thd,
|
|
|
|
char* packet, uint packet_length)
|
|
|
|
{
|
|
|
|
NET *net= &thd->net;
|
2003-02-14 10:47:41 +01:00
|
|
|
bool error= 0;
|
2003-11-25 13:28:43 +01:00
|
|
|
DBUG_ENTER("dispatch_command");
|
|
|
|
|
2005-05-09 13:26:06 +02:00
|
|
|
if (thd->killed == THD::KILL_QUERY || thd->killed == THD::KILL_BAD_DATA)
|
|
|
|
thd->killed= THD::NOT_KILLED;
|
|
|
|
|
2003-11-25 13:28:43 +01:00
|
|
|
thd->command=command;
|
2002-02-13 21:37:19 +01:00
|
|
|
/*
|
2005-06-16 21:05:38 +02:00
|
|
|
Commands which always take a long time are logged into
|
|
|
|
the slow log only if opt_log_slow_admin_statements is set.
|
2002-02-13 21:37:19 +01:00
|
|
|
*/
|
2005-06-16 21:05:38 +02:00
|
|
|
thd->enable_slow_log= TRUE;
|
2004-09-07 14:50:56 +02:00
|
|
|
thd->lex->sql_command= SQLCOM_END; /* to avoid confusing VIEW detectors */
|
2002-07-23 17:31:22 +02:00
|
|
|
thd->set_time();
|
2000-07-31 21:29:14 +02:00
|
|
|
VOID(pthread_mutex_lock(&LOCK_thread_count));
|
|
|
|
thd->query_id=query_id;
|
|
|
|
if (command != COM_STATISTICS && command != COM_PING)
|
2005-01-16 13:16:23 +01:00
|
|
|
next_query_id();
|
2000-07-31 21:29:14 +02:00
|
|
|
thread_running++;
|
2004-06-18 23:50:04 +02:00
|
|
|
/* TODO: set thd->lex->sql_command to SQLCOM_END here */
|
2000-07-31 21:29:14 +02:00
|
|
|
VOID(pthread_mutex_unlock(&LOCK_thread_count));
|
2002-07-23 17:31:22 +02:00
|
|
|
|
2003-12-06 23:21:09 +01:00
|
|
|
thd->server_status&=
|
|
|
|
~(SERVER_QUERY_NO_INDEX_USED | SERVER_QUERY_NO_GOOD_INDEX_USED);
|
2001-04-11 13:04:03 +02:00
|
|
|
switch (command) {
|
2000-07-31 21:29:14 +02:00
|
|
|
case COM_INIT_DB:
|
2003-08-18 23:08:08 +02:00
|
|
|
{
|
|
|
|
LEX_STRING tmp;
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_stat[SQLCOM_CHANGE_DB],
|
|
|
|
&LOCK_status);
|
2003-08-18 23:08:08 +02:00
|
|
|
thd->convert_string(&tmp, system_charset_info,
|
|
|
|
packet, strlen(packet), thd->charset());
|
2005-08-12 02:04:16 +02:00
|
|
|
if (!mysql_change_db(thd, tmp.str, FALSE))
|
|
|
|
{
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, "%s",thd->db);
|
2005-08-12 02:04:16 +02:00
|
|
|
send_ok(thd);
|
|
|
|
}
|
2003-08-18 23:08:08 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-09-08 12:08:53 +02:00
|
|
|
#ifdef HAVE_REPLICATION
|
2001-05-31 02:50:56 +02:00
|
|
|
case COM_REGISTER_SLAVE:
|
|
|
|
{
|
2002-08-21 21:04:22 +02:00
|
|
|
if (!register_slave(thd, (uchar*)packet, packet_length))
|
2002-10-02 12:33:08 +02:00
|
|
|
send_ok(thd);
|
2001-05-31 02:50:56 +02:00
|
|
|
break;
|
|
|
|
}
|
2002-12-16 14:33:29 +01:00
|
|
|
#endif
|
2000-07-31 21:29:14 +02:00
|
|
|
case COM_TABLE_DUMP:
|
2003-08-19 15:00:12 +02:00
|
|
|
{
|
|
|
|
char *db, *tbl_name;
|
|
|
|
uint db_len= *(uchar*) packet;
|
2006-04-25 23:06:04 +02:00
|
|
|
if (db_len >= packet_length || db_len > NAME_LEN)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2006-04-26 07:39:59 +02:00
|
|
|
my_message(ER_UNKNOWN_COM_ERROR, ER(ER_UNKNOWN_COM_ERROR), MYF(0));
|
2006-04-25 23:06:04 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-08-19 15:00:12 +02:00
|
|
|
uint tbl_len= *(uchar*) (packet + db_len + 1);
|
2006-04-25 23:06:04 +02:00
|
|
|
if (db_len+tbl_len+2 > packet_length || tbl_len > NAME_LEN)
|
|
|
|
{
|
2006-04-26 07:39:59 +02:00
|
|
|
my_message(ER_UNKNOWN_COM_ERROR, ER(ER_UNKNOWN_COM_ERROR), MYF(0));
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-08-19 15:00:12 +02:00
|
|
|
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_other, &LOCK_status);
|
2005-06-16 21:05:38 +02:00
|
|
|
thd->enable_slow_log= opt_log_slow_admin_statements;
|
2003-08-19 15:00:12 +02:00
|
|
|
db= thd->alloc(db_len + tbl_len + 2);
|
2006-01-03 17:54:54 +01:00
|
|
|
if (!db)
|
|
|
|
{
|
|
|
|
my_message(ER_OUT_OF_RESOURCES, ER(ER_OUT_OF_RESOURCES), MYF(0));
|
|
|
|
break;
|
|
|
|
}
|
2003-08-19 15:00:12 +02:00
|
|
|
tbl_name= strmake(db, packet + 1, db_len)+1;
|
|
|
|
strmake(tbl_name, packet + db_len + 2, tbl_len);
|
2004-10-20 03:04:37 +02:00
|
|
|
mysql_table_dump(thd, db, tbl_name, -1);
|
2003-08-19 15:00:12 +02:00
|
|
|
break;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case COM_CHANGE_USER:
|
|
|
|
{
|
2002-11-16 19:19:10 +01:00
|
|
|
thd->change_user();
|
2003-07-01 21:40:59 +02:00
|
|
|
thd->clear_error(); // if errors from rollback
|
2002-11-16 19:19:10 +01:00
|
|
|
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_other, &LOCK_status);
|
2003-07-01 21:40:59 +02:00
|
|
|
char *user= (char*) packet;
|
2000-07-31 21:29:14 +02:00
|
|
|
char *passwd= strend(user)+1;
|
2006-01-03 17:54:54 +01:00
|
|
|
/*
|
2003-07-04 18:52:04 +02:00
|
|
|
Old clients send null-terminated string ('\0' for empty string) for
|
|
|
|
password. New clients send the size (1 byte) + string (not null
|
|
|
|
terminated, so also '\0' for empty string).
|
|
|
|
*/
|
2006-01-03 17:54:54 +01:00
|
|
|
char db_buff[NAME_LEN+1]; // buffer to store db in utf8
|
2003-07-04 18:52:04 +02:00
|
|
|
char *db= passwd;
|
2006-01-03 17:54:54 +01:00
|
|
|
uint passwd_len= thd->client_capabilities & CLIENT_SECURE_CONNECTION ?
|
2003-07-04 18:52:04 +02:00
|
|
|
*passwd++ : strlen(passwd);
|
|
|
|
db+= passwd_len + 1;
|
2004-02-14 17:26:21 +01:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2004-10-07 00:45:06 +02:00
|
|
|
/* Small check for incoming packet */
|
2000-07-31 21:29:14 +02:00
|
|
|
if ((uint) ((uchar*) db - net->read_pos) > packet_length)
|
2003-07-01 21:40:59 +02:00
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_UNKNOWN_COM_ERROR, ER(ER_UNKNOWN_COM_ERROR), MYF(0));
|
2003-07-01 21:40:59 +02:00
|
|
|
break;
|
|
|
|
}
|
2004-02-14 17:26:21 +01:00
|
|
|
#endif
|
2003-09-03 15:29:51 +02:00
|
|
|
/* Convert database name to utf8 */
|
2004-10-29 14:00:39 +02:00
|
|
|
uint dummy_errors;
|
2003-09-16 00:39:25 +02:00
|
|
|
db_buff[copy_and_convert(db_buff, sizeof(db_buff)-1,
|
|
|
|
system_charset_info, db, strlen(db),
|
2004-10-29 14:00:39 +02:00
|
|
|
thd->charset(), &dummy_errors)]= 0;
|
2003-09-16 00:39:25 +02:00
|
|
|
db= db_buff;
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2003-07-01 21:40:59 +02:00
|
|
|
/* Save user and privileges */
|
|
|
|
uint save_db_length= thd->db_length;
|
|
|
|
char *save_db= thd->db;
|
2005-09-20 20:20:38 +02:00
|
|
|
Security_context save_security_ctx= *thd->security_ctx;
|
2003-10-30 00:01:53 +01:00
|
|
|
USER_CONN *save_user_connect= thd->user_connect;
|
2005-09-15 21:29:07 +02:00
|
|
|
|
|
|
|
if (!(thd->security_ctx->user= my_strdup(user, MYF(0))))
|
2003-07-01 21:40:59 +02:00
|
|
|
{
|
2005-09-15 21:29:07 +02:00
|
|
|
thd->security_ctx->user= save_security_ctx.user;
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_OUT_OF_RESOURCES, ER(ER_OUT_OF_RESOURCES), MYF(0));
|
2003-07-01 21:40:59 +02:00
|
|
|
break;
|
|
|
|
}
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2003-10-30 00:01:53 +01:00
|
|
|
/* Clear variables that are allocated */
|
|
|
|
thd->user_connect= 0;
|
2004-03-30 01:32:41 +02:00
|
|
|
int res= check_user(thd, COM_CHANGE_USER, passwd, passwd_len, db, FALSE);
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2003-07-01 21:40:59 +02:00
|
|
|
if (res)
|
|
|
|
{
|
2004-10-07 00:45:06 +02:00
|
|
|
/* authentication failure, we shall restore old user */
|
2003-07-01 21:40:59 +02:00
|
|
|
if (res > 0)
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_UNKNOWN_COM_ERROR, ER(ER_UNKNOWN_COM_ERROR), MYF(0));
|
2005-09-15 21:29:07 +02:00
|
|
|
x_free(thd->security_ctx->user);
|
|
|
|
*thd->security_ctx= save_security_ctx;
|
2003-10-30 00:01:53 +01:00
|
|
|
thd->user_connect= save_user_connect;
|
2003-07-01 21:40:59 +02:00
|
|
|
thd->db= save_db;
|
|
|
|
thd->db_length= save_db_length;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
/* we've authenticated new user */
|
2003-11-04 08:40:36 +01:00
|
|
|
if (save_user_connect)
|
|
|
|
decrease_user_connections(save_user_connect);
|
2003-07-01 21:40:59 +02:00
|
|
|
x_free((gptr) save_db);
|
2005-09-15 21:29:07 +02:00
|
|
|
x_free((gptr) save_security_ctx.user);
|
2003-07-01 21:40:59 +02:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2005-06-17 21:26:25 +02:00
|
|
|
case COM_STMT_EXECUTE:
|
2002-06-12 23:13:12 +02:00
|
|
|
{
|
2004-03-15 18:20:47 +01:00
|
|
|
mysql_stmt_execute(thd, packet, packet_length);
|
2002-06-12 23:13:12 +02:00
|
|
|
break;
|
|
|
|
}
|
2005-06-17 21:26:25 +02:00
|
|
|
case COM_STMT_FETCH:
|
2004-08-03 12:32:21 +02:00
|
|
|
{
|
|
|
|
mysql_stmt_fetch(thd, packet, packet_length);
|
|
|
|
break;
|
|
|
|
}
|
2005-06-17 21:26:25 +02:00
|
|
|
case COM_STMT_SEND_LONG_DATA:
|
2002-06-12 23:13:12 +02:00
|
|
|
{
|
2002-10-02 12:33:08 +02:00
|
|
|
mysql_stmt_get_longdata(thd, packet, packet_length);
|
2002-06-12 23:13:12 +02:00
|
|
|
break;
|
|
|
|
}
|
2005-06-17 21:26:25 +02:00
|
|
|
case COM_STMT_PREPARE:
|
2002-06-12 23:13:12 +02:00
|
|
|
{
|
Implement WL#2661 "Prepared Statements: Dynamic SQL in Stored Procedures".
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
2005-09-03 01:13:18 +02:00
|
|
|
mysql_stmt_prepare(thd, packet, packet_length);
|
2002-06-12 23:13:12 +02:00
|
|
|
break;
|
|
|
|
}
|
2005-06-17 21:26:25 +02:00
|
|
|
case COM_STMT_CLOSE:
|
2002-11-27 03:51:38 +01:00
|
|
|
{
|
2005-06-17 21:26:25 +02:00
|
|
|
mysql_stmt_close(thd, packet);
|
2002-11-27 03:51:38 +01:00
|
|
|
break;
|
|
|
|
}
|
2005-06-17 21:26:25 +02:00
|
|
|
case COM_STMT_RESET:
|
2003-07-08 11:27:21 +02:00
|
|
|
{
|
|
|
|
mysql_stmt_reset(thd, packet);
|
|
|
|
break;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case COM_QUERY:
|
|
|
|
{
|
2002-10-02 12:33:08 +02:00
|
|
|
if (alloc_query(thd, packet, packet_length))
|
|
|
|
break; // fatal error is set
|
2005-02-14 23:47:17 +01:00
|
|
|
char *packet_end= thd->query + thd->query_length;
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, "%.*b", thd->query_length, thd->query);
|
2003-01-04 14:17:16 +01:00
|
|
|
DBUG_PRINT("query",("%-.4096s",thd->query));
|
Implement WL#2661 "Prepared Statements: Dynamic SQL in Stored Procedures".
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
2005-09-03 01:13:18 +02:00
|
|
|
|
|
|
|
if (!(specialflag & SPECIAL_NO_PRIOR))
|
|
|
|
my_pthread_setprio(pthread_self(),QUERY_PRIOR);
|
|
|
|
|
2002-10-02 12:33:08 +02:00
|
|
|
mysql_parse(thd,thd->query, thd->query_length);
|
2003-01-20 23:00:50 +01:00
|
|
|
|
2005-01-16 13:16:23 +01:00
|
|
|
while (!thd->killed && thd->lex->found_semicolon && !thd->net.report_error)
|
2003-01-20 23:00:50 +01:00
|
|
|
{
|
2005-01-16 13:16:23 +01:00
|
|
|
char *packet= thd->lex->found_semicolon;
|
2005-05-27 15:01:09 +02:00
|
|
|
net->no_send_error= 0;
|
2003-11-14 13:50:19 +01:00
|
|
|
/*
|
2003-01-20 23:00:50 +01:00
|
|
|
Multiple queries exits, execute them individually
|
|
|
|
*/
|
2005-03-04 14:35:28 +01:00
|
|
|
if (thd->lock || thd->open_tables || thd->derived_tables ||
|
|
|
|
thd->prelocked_mode)
|
2003-11-14 13:50:19 +01:00
|
|
|
close_thread_tables(thd);
|
2005-02-14 23:47:17 +01:00
|
|
|
ulong length= (ulong)(packet_end-packet);
|
2003-11-14 13:50:19 +01:00
|
|
|
|
2005-06-16 22:11:48 +02:00
|
|
|
log_slow_statement(thd);
|
2005-02-17 20:48:44 +01:00
|
|
|
|
2003-01-20 23:00:50 +01:00
|
|
|
/* Remove garbage at start of query */
|
2003-03-15 14:24:21 +01:00
|
|
|
while (my_isspace(thd->charset(), *packet) && length > 0)
|
2003-01-20 23:00:50 +01:00
|
|
|
{
|
|
|
|
packet++;
|
|
|
|
length--;
|
|
|
|
}
|
2004-05-17 00:52:13 +02:00
|
|
|
VOID(pthread_mutex_lock(&LOCK_thread_count));
|
2003-01-20 23:00:50 +01:00
|
|
|
thd->query_length= length;
|
2003-01-23 20:49:28 +01:00
|
|
|
thd->query= packet;
|
2005-01-16 13:16:23 +01:00
|
|
|
thd->query_id= next_query_id();
|
2005-02-17 20:48:44 +01:00
|
|
|
thd->set_time(); /* Reset the query start time. */
|
2004-06-18 23:50:04 +02:00
|
|
|
/* TODO: set thd->lex->sql_command to SQLCOM_END here */
|
2003-01-23 20:49:28 +01:00
|
|
|
VOID(pthread_mutex_unlock(&LOCK_thread_count));
|
2003-01-20 23:00:50 +01:00
|
|
|
mysql_parse(thd, packet, length);
|
|
|
|
}
|
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
if (!(specialflag & SPECIAL_NO_PRIOR))
|
|
|
|
my_pthread_setprio(pthread_self(),WAIT_PRIOR);
|
|
|
|
DBUG_PRINT("info",("query ready"));
|
|
|
|
break;
|
|
|
|
}
|
2001-09-27 20:45:48 +02:00
|
|
|
case COM_FIELD_LIST: // This isn't actually needed
|
2000-07-31 21:29:14 +02:00
|
|
|
#ifdef DONT_ALLOW_SHOW_COMMANDS
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_NOT_ALLOWED_COMMAND, ER(ER_NOT_ALLOWED_COMMAND),
|
|
|
|
MYF(0)); /* purecov: inspected */
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
#else
|
|
|
|
{
|
2003-05-30 20:09:35 +02:00
|
|
|
char *fields, *pend;
|
2005-02-13 20:47:00 +01:00
|
|
|
/* Locked closure of all tables */
|
|
|
|
TABLE_LIST *locked_tables= NULL;
|
2000-07-31 21:29:14 +02:00
|
|
|
TABLE_LIST table_list;
|
2003-08-18 23:08:08 +02:00
|
|
|
LEX_STRING conv_name;
|
2005-02-13 20:47:00 +01:00
|
|
|
/* Saved variable value */
|
2006-05-02 19:56:43 +02:00
|
|
|
my_bool old_innodb_table_locks= thd->variables.innodb_table_locks;
|
|
|
|
|
|
|
|
|
2005-11-02 14:17:57 +01:00
|
|
|
/* used as fields initializator */
|
|
|
|
lex_start(thd, 0, 0);
|
2005-02-13 20:47:00 +01:00
|
|
|
|
2003-08-18 23:08:08 +02:00
|
|
|
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_stat[SQLCOM_SHOW_FIELDS],
|
|
|
|
&LOCK_status);
|
2000-07-31 21:29:14 +02:00
|
|
|
bzero((char*) &table_list,sizeof(table_list));
|
|
|
|
if (!(table_list.db=thd->db))
|
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_NO_DB_ERROR, ER(ER_NO_DB_ERROR), MYF(0));
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-05-30 20:09:35 +02:00
|
|
|
pend= strend(packet);
|
2003-08-18 23:08:08 +02:00
|
|
|
thd->convert_string(&conv_name, system_charset_info,
|
|
|
|
packet, (uint) (pend-packet), thd->charset());
|
2005-01-06 12:00:13 +01:00
|
|
|
table_list.alias= table_list.table_name= conv_name.str;
|
2003-05-30 20:09:35 +02:00
|
|
|
packet= pend+1;
|
2004-12-18 10:48:01 +01:00
|
|
|
|
|
|
|
if (!my_strcasecmp(system_charset_info, table_list.db,
|
|
|
|
information_schema_name.str))
|
|
|
|
{
|
|
|
|
ST_SCHEMA_TABLE *schema_table= find_schema_table(thd, table_list.alias);
|
|
|
|
if (schema_table)
|
|
|
|
table_list.schema_table= schema_table;
|
|
|
|
}
|
|
|
|
|
2005-01-12 15:04:45 +01:00
|
|
|
thd->query_length= strlen(packet); // for simplicity: don't optimize
|
2001-12-17 02:02:58 +01:00
|
|
|
if (!(thd->query=fields=thd->memdup(packet,thd->query_length+1)))
|
|
|
|
break;
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, "%s %s", table_list.table_name, fields);
|
2003-01-29 17:56:34 +01:00
|
|
|
if (lower_case_table_names)
|
2005-01-06 12:00:13 +01:00
|
|
|
my_casedn_str(files_charset_info, table_list.table_name);
|
|
|
|
remove_escape(table_list.table_name); // This can't have wildcards
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2004-02-16 09:03:25 +01:00
|
|
|
if (check_access(thd,SELECT_ACL,table_list.db,&table_list.grant.privilege,
|
2005-09-13 13:07:38 +02:00
|
|
|
0, 0, test(table_list.schema_table)))
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2004-04-10 00:14:32 +02:00
|
|
|
if (grant_option &&
|
|
|
|
check_grant(thd, SELECT_ACL, &table_list, 2, UINT_MAX, 0))
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2004-09-07 14:50:56 +02:00
|
|
|
/* init structures for VIEW processing */
|
|
|
|
table_list.select_lex= &(thd->lex->select_lex);
|
|
|
|
mysql_init_query(thd, (uchar*)"", 0);
|
|
|
|
thd->lex->
|
|
|
|
select_lex.table_list.link_in_list((byte*) &table_list,
|
|
|
|
(byte**) &table_list.next_local);
|
2005-05-11 01:31:13 +02:00
|
|
|
thd->lex->add_to_query_tables(&table_list);
|
2004-09-07 14:50:56 +02:00
|
|
|
|
2004-10-10 10:01:05 +02:00
|
|
|
/* switch on VIEW optimisation: do not fill temporary tables */
|
|
|
|
thd->lex->sql_command= SQLCOM_SHOW_FIELDS;
|
2000-07-31 21:29:14 +02:00
|
|
|
mysqld_list_fields(thd,&table_list,fields);
|
2004-09-07 14:50:56 +02:00
|
|
|
thd->lex->unit.cleanup();
|
2004-09-15 21:10:31 +02:00
|
|
|
thd->cleanup_after_query();
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
case COM_QUIT:
|
2001-12-13 19:06:44 +01:00
|
|
|
/* We don't calculate statistics for this command */
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, NullS);
|
2000-07-31 21:29:14 +02:00
|
|
|
net->error=0; // Don't give 'abort' message
|
|
|
|
error=TRUE; // End server
|
|
|
|
break;
|
|
|
|
|
2001-09-03 04:16:15 +02:00
|
|
|
case COM_CREATE_DB: // QQ: To be removed
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2003-12-30 12:14:21 +01:00
|
|
|
char *db=thd->strdup(packet), *alias;
|
2004-10-17 11:59:46 +02:00
|
|
|
HA_CREATE_INFO create_info;
|
2003-12-30 12:14:21 +01:00
|
|
|
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_stat[SQLCOM_CREATE_DB],
|
|
|
|
&LOCK_status);
|
2001-01-31 03:47:25 +01:00
|
|
|
// null test to handle EOM
|
2004-03-05 19:13:33 +01:00
|
|
|
if (!db || !(alias= thd->strdup(db)) || check_db_name(db))
|
2001-01-31 03:47:25 +01:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_WRONG_DB_NAME, MYF(0), db ? db : "NULL");
|
2001-01-31 03:47:25 +01:00
|
|
|
break;
|
|
|
|
}
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd,CREATE_ACL,db,0,1,0,is_schema_db(db)))
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, packet);
|
2004-10-17 11:59:46 +02:00
|
|
|
bzero(&create_info, sizeof(create_info));
|
2004-11-16 12:39:10 +01:00
|
|
|
mysql_create_db(thd, (lower_case_table_names == 2 ? alias : db),
|
|
|
|
&create_info, 0);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2001-09-03 04:16:15 +02:00
|
|
|
case COM_DROP_DB: // QQ: To be removed
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_stat[SQLCOM_DROP_DB],
|
|
|
|
&LOCK_status);
|
2005-02-25 15:53:22 +01:00
|
|
|
char *db=thd->strdup(packet);
|
2004-07-07 10:29:39 +02:00
|
|
|
/* null test to handle EOM */
|
2005-02-25 15:53:22 +01:00
|
|
|
if (!db || check_db_name(db))
|
2001-01-31 03:47:25 +01:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_WRONG_DB_NAME, MYF(0), db ? db : "NULL");
|
2001-01-31 03:47:25 +01:00
|
|
|
break;
|
|
|
|
}
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd,DROP_ACL,db,0,1,0,is_schema_db(db)))
|
2002-12-20 13:58:27 +01:00
|
|
|
break;
|
2001-09-03 04:16:15 +02:00
|
|
|
if (thd->locked_tables || thd->active_transaction())
|
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_LOCK_OR_ACTIVE_TRANSACTION,
|
|
|
|
ER(ER_LOCK_OR_ACTIVE_TRANSACTION), MYF(0));
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2001-09-03 04:16:15 +02:00
|
|
|
}
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, db);
|
2005-02-21 13:47:57 +01:00
|
|
|
mysql_rm_db(thd, db, 0, 0);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2002-12-16 14:33:29 +01:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2000-07-31 21:29:14 +02:00
|
|
|
case COM_BINLOG_DUMP:
|
|
|
|
{
|
2004-11-10 17:07:39 +01:00
|
|
|
ulong pos;
|
|
|
|
ushort flags;
|
|
|
|
uint32 slave_server_id;
|
|
|
|
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_other,&LOCK_status);
|
2005-06-16 21:05:38 +02:00
|
|
|
thd->enable_slow_log= opt_log_slow_admin_statements;
|
2002-06-12 14:04:18 +02:00
|
|
|
if (check_global_access(thd, REPL_SLAVE_ACL))
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2000-08-21 23:39:08 +02:00
|
|
|
|
2002-08-21 21:04:22 +02:00
|
|
|
/* TODO: The following has to be changed to an 8 byte integer */
|
2001-04-11 13:04:03 +02:00
|
|
|
pos = uint4korr(packet);
|
|
|
|
flags = uint2korr(packet + 4);
|
2001-10-12 17:37:25 +02:00
|
|
|
thd->server_id=0; /* avoid suicide */
|
2003-07-03 01:08:31 +02:00
|
|
|
if ((slave_server_id= uint4korr(packet+6))) // mysqlbinlog.server_id==0
|
2003-07-02 22:56:27 +02:00
|
|
|
kill_zombie_dump_threads(slave_server_id);
|
2000-09-30 01:20:26 +02:00
|
|
|
thd->server_id = slave_server_id;
|
2004-11-10 17:07:39 +01:00
|
|
|
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, "Log: '%s' Pos: %ld", packet+10,
|
2004-11-10 17:07:39 +01:00
|
|
|
(long) pos);
|
2002-08-21 21:04:22 +02:00
|
|
|
mysql_binlog_send(thd, thd->strdup(packet + 10), (my_off_t) pos, flags);
|
2001-10-12 17:37:25 +02:00
|
|
|
unregister_slave(thd,1,1);
|
2004-07-07 10:29:39 +02:00
|
|
|
/* fake COM_QUIT -- if we get here, the thread needs to terminate */
|
2000-10-03 01:59:12 +02:00
|
|
|
error = TRUE;
|
|
|
|
net->error = 0;
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2002-12-16 14:33:29 +01:00
|
|
|
#endif
|
2000-07-31 21:29:14 +02:00
|
|
|
case COM_REFRESH:
|
2005-08-11 14:58:15 +02:00
|
|
|
{
|
|
|
|
bool not_used;
|
|
|
|
statistic_increment(thd->status_var.com_stat[SQLCOM_FLUSH],
|
|
|
|
&LOCK_status);
|
|
|
|
ulong options= (ulong) (uchar) packet[0];
|
|
|
|
if (check_global_access(thd,RELOAD_ACL))
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, NullS);
|
2005-08-11 14:58:15 +02:00
|
|
|
if (!reload_acl_and_cache(thd, options, (TABLE_LIST*) 0, ¬_used))
|
|
|
|
send_ok(thd);
|
|
|
|
break;
|
|
|
|
}
|
2002-12-16 14:33:29 +01:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2000-07-31 21:29:14 +02:00
|
|
|
case COM_SHUTDOWN:
|
2004-06-15 11:35:23 +02:00
|
|
|
{
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_other, &LOCK_status);
|
2002-06-12 14:04:18 +02:00
|
|
|
if (check_global_access(thd,SHUTDOWN_ACL))
|
2000-07-31 21:29:14 +02:00
|
|
|
break; /* purecov: inspected */
|
2004-06-15 11:35:23 +02:00
|
|
|
/*
|
2004-06-18 23:50:04 +02:00
|
|
|
If the client is < 4.1.3, it is going to send us no argument; then
|
|
|
|
packet_length is 1, packet[0] is the end 0 of the packet. Note that
|
|
|
|
SHUTDOWN_DEFAULT is 0. If client is >= 4.1.3, the shutdown level is in
|
|
|
|
packet[0].
|
2004-06-15 11:35:23 +02:00
|
|
|
*/
|
2004-08-19 20:48:00 +02:00
|
|
|
enum mysql_enum_shutdown_level level=
|
|
|
|
(enum mysql_enum_shutdown_level) (uchar) packet[0];
|
2004-06-18 23:50:04 +02:00
|
|
|
DBUG_PRINT("quit",("Got shutdown command for level %u", level));
|
2004-06-15 11:35:23 +02:00
|
|
|
if (level == SHUTDOWN_DEFAULT)
|
|
|
|
level= SHUTDOWN_WAIT_ALL_BUFFERS; // soon default will be configurable
|
|
|
|
else if (level != SHUTDOWN_WAIT_ALL_BUFFERS)
|
|
|
|
{
|
|
|
|
my_error(ER_NOT_SUPPORTED_YET, MYF(0), "this shutdown level");
|
|
|
|
break;
|
|
|
|
}
|
2004-06-18 23:50:04 +02:00
|
|
|
DBUG_PRINT("quit",("Got shutdown command for level %u", level));
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, NullS);
|
2002-10-02 12:33:08 +02:00
|
|
|
send_eof(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
#ifdef __WIN__
|
|
|
|
sleep(1); // must wait after eof()
|
|
|
|
#endif
|
2002-10-02 12:33:08 +02:00
|
|
|
send_eof(thd); // This is for 'quit request'
|
2003-05-26 18:01:20 +02:00
|
|
|
close_connection(thd, 0, 1);
|
2000-07-31 21:29:14 +02:00
|
|
|
close_thread_tables(thd); // Free before kill
|
|
|
|
kill_mysql();
|
|
|
|
error=TRUE;
|
|
|
|
break;
|
2004-06-15 11:35:23 +02:00
|
|
|
}
|
2002-12-16 14:33:29 +01:00
|
|
|
#endif
|
2000-07-31 21:29:14 +02:00
|
|
|
case COM_STATISTICS:
|
|
|
|
{
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, NullS);
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_stat[SQLCOM_SHOW_STATUS],
|
|
|
|
&LOCK_status);
|
2003-10-04 16:28:08 +02:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2000-07-31 21:29:14 +02:00
|
|
|
char buff[200];
|
2003-10-04 16:28:08 +02:00
|
|
|
#else
|
|
|
|
char *buff= thd->net.last_error;
|
|
|
|
#endif
|
2006-06-06 13:21:07 +02:00
|
|
|
|
|
|
|
STATUS_VAR current_global_status_var;
|
|
|
|
calc_sum_of_all_status(¤t_global_status_var);
|
|
|
|
|
2000-09-20 03:54:10 +02:00
|
|
|
ulong uptime = (ulong) (thd->start_time - start_time);
|
2000-07-31 21:29:14 +02:00
|
|
|
sprintf((char*) buff,
|
2005-04-05 22:14:56 +02:00
|
|
|
"Uptime: %lu Threads: %d Questions: %lu Slow queries: %lu Opens: %lu Flush tables: %lu Open tables: %u Queries per second avg: %.3f",
|
2000-07-31 21:29:14 +02:00
|
|
|
uptime,
|
2005-04-01 14:04:50 +02:00
|
|
|
(int) thread_count, (ulong) thd->query_id,
|
2006-06-06 13:21:07 +02:00
|
|
|
current_global_status_var.long_query_count,
|
2006-06-09 21:27:28 +02:00
|
|
|
current_global_status_var.opened_tables, refresh_version,
|
|
|
|
cached_open_tables(),
|
2005-04-05 22:14:56 +02:00
|
|
|
(uptime ? (ulonglong2double(thd->query_id) / (double) uptime) :
|
|
|
|
(double) 0));
|
2000-07-31 21:29:14 +02:00
|
|
|
#ifdef SAFEMALLOC
|
2003-06-12 21:39:45 +02:00
|
|
|
if (sf_malloc_cur_memory) // Using SAFEMALLOC
|
2000-07-31 21:29:14 +02:00
|
|
|
sprintf(strend(buff), " Memory in use: %ldK Max memory used: %ldK",
|
2003-06-12 21:39:45 +02:00
|
|
|
(sf_malloc_cur_memory+1023L)/1024L,
|
|
|
|
(sf_malloc_max_memory+1023L)/1024L);
|
2003-10-04 16:28:08 +02:00
|
|
|
#endif
|
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2000-08-21 23:18:32 +02:00
|
|
|
VOID(my_net_write(net, buff,(uint) strlen(buff)));
|
2000-07-31 21:29:14 +02:00
|
|
|
VOID(net_flush(net));
|
2003-10-04 16:28:08 +02:00
|
|
|
#endif
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case COM_PING:
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_other, &LOCK_status);
|
2002-10-02 12:33:08 +02:00
|
|
|
send_ok(thd); // Tell client we are alive
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
case COM_PROCESS_INFO:
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_stat[SQLCOM_SHOW_PROCESSLIST],
|
|
|
|
&LOCK_status);
|
2005-09-15 21:29:07 +02:00
|
|
|
if (!thd->security_ctx->priv_user[0] &&
|
|
|
|
check_global_access(thd, PROCESS_ACL))
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, NullS);
|
2003-09-26 12:33:13 +02:00
|
|
|
mysqld_list_processes(thd,
|
2005-09-15 21:29:07 +02:00
|
|
|
thd->security_ctx->master_access & PROCESS_ACL ?
|
|
|
|
NullS : thd->security_ctx->priv_user, 0);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
case COM_PROCESS_KILL:
|
|
|
|
{
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_stat[SQLCOM_KILL], &LOCK_status);
|
2001-04-11 13:04:03 +02:00
|
|
|
ulong id=(ulong) uint4korr(packet);
|
2006-05-22 20:46:13 +02:00
|
|
|
sql_kill(thd,id,false);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-11-18 12:47:27 +01:00
|
|
|
case COM_SET_OPTION:
|
|
|
|
{
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_stat[SQLCOM_SET_OPTION],
|
|
|
|
&LOCK_status);
|
2003-11-18 12:47:27 +01:00
|
|
|
enum_mysql_set_option command= (enum_mysql_set_option) uint2korr(packet);
|
|
|
|
switch (command) {
|
|
|
|
case MYSQL_OPTION_MULTI_STATEMENTS_ON:
|
|
|
|
thd->client_capabilities|= CLIENT_MULTI_STATEMENTS;
|
2004-01-09 21:28:29 +01:00
|
|
|
send_eof(thd);
|
2003-11-18 12:47:27 +01:00
|
|
|
break;
|
|
|
|
case MYSQL_OPTION_MULTI_STATEMENTS_OFF:
|
|
|
|
thd->client_capabilities&= ~CLIENT_MULTI_STATEMENTS;
|
2004-01-09 21:28:29 +01:00
|
|
|
send_eof(thd);
|
2003-11-18 12:47:27 +01:00
|
|
|
break;
|
|
|
|
default:
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_UNKNOWN_COM_ERROR, ER(ER_UNKNOWN_COM_ERROR), MYF(0));
|
2003-11-18 12:47:27 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case COM_DEBUG:
|
2004-09-13 15:48:01 +02:00
|
|
|
statistic_increment(thd->status_var.com_other, &LOCK_status);
|
2002-06-12 14:04:18 +02:00
|
|
|
if (check_global_access(thd, SUPER_ACL))
|
2000-07-31 21:29:14 +02:00
|
|
|
break; /* purecov: inspected */
|
2005-02-17 13:07:28 +01:00
|
|
|
mysql_print_status();
|
2006-05-02 19:56:43 +02:00
|
|
|
general_log_print(thd, command, NullS);
|
2002-10-02 12:33:08 +02:00
|
|
|
send_eof(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
case COM_SLEEP:
|
|
|
|
case COM_CONNECT: // Impossible here
|
|
|
|
case COM_TIME: // Impossible from client
|
|
|
|
case COM_DELAYED_INSERT:
|
2003-01-04 14:17:16 +01:00
|
|
|
case COM_END:
|
2000-07-31 21:29:14 +02:00
|
|
|
default:
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_UNKNOWN_COM_ERROR, ER(ER_UNKNOWN_COM_ERROR), MYF(0));
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2005-03-04 14:35:28 +01:00
|
|
|
if (thd->lock || thd->open_tables || thd->derived_tables ||
|
|
|
|
thd->prelocked_mode)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
thd->proc_info="closing tables";
|
|
|
|
close_thread_tables(thd); /* Free tables */
|
|
|
|
}
|
2005-01-16 13:16:23 +01:00
|
|
|
/*
|
|
|
|
assume handlers auto-commit (if some doesn't - transaction handling
|
|
|
|
in MySQL should be redesigned to support it; it's a big change,
|
|
|
|
and it's not worth it - better to commit explicitly only writing
|
|
|
|
transactions, read-only ones should better take care of themselves.
|
|
|
|
saves some work in 2pc too)
|
|
|
|
see also sql_base.cc - close_thread_tables()
|
|
|
|
*/
|
|
|
|
bzero(&thd->transaction.stmt, sizeof(thd->transaction.stmt));
|
|
|
|
if (!thd->active_transaction())
|
2005-08-12 21:15:01 +02:00
|
|
|
thd->transaction.xid_state.xid.null();
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2004-10-20 03:04:37 +02:00
|
|
|
/* report error issued during command execution */
|
|
|
|
if (thd->killed_errno() && !thd->net.report_error)
|
|
|
|
thd->send_kill_message();
|
2004-11-12 13:34:00 +01:00
|
|
|
if (thd->net.report_error)
|
2004-10-20 15:06:54 +02:00
|
|
|
net_send_error(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2005-06-16 22:11:48 +02:00
|
|
|
log_slow_statement(thd);
|
2005-02-17 20:48:44 +01:00
|
|
|
|
|
|
|
thd->proc_info="cleaning up";
|
|
|
|
VOID(pthread_mutex_lock(&LOCK_thread_count)); // For process list
|
|
|
|
thd->proc_info=0;
|
|
|
|
thd->command=COM_SLEEP;
|
|
|
|
thd->query=0;
|
|
|
|
thd->query_length=0;
|
|
|
|
thread_running--;
|
|
|
|
VOID(pthread_mutex_unlock(&LOCK_thread_count));
|
|
|
|
thd->packet.shrink(thd->variables.net_buffer_length); // Reclaim some memory
|
|
|
|
free_root(thd->mem_root,MYF(MY_KEEP_PREALLOC));
|
|
|
|
DBUG_RETURN(error);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2005-06-16 22:11:48 +02:00
|
|
|
void log_slow_statement(THD *thd)
|
2005-02-17 20:48:44 +01:00
|
|
|
{
|
2005-08-15 17:15:12 +02:00
|
|
|
time_t start_of_query;
|
2006-05-02 19:56:43 +02:00
|
|
|
DBUG_ENTER("log_slow_statement");
|
2005-08-15 17:15:12 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
The following should never be true with our current code base,
|
|
|
|
but better to keep this here so we don't accidently try to log a
|
|
|
|
statement in a trigger or stored function
|
|
|
|
*/
|
|
|
|
if (unlikely(thd->in_sub_stmt))
|
2006-05-02 19:56:43 +02:00
|
|
|
DBUG_VOID_RETURN; // Don't set time for sub stmt
|
2005-08-15 17:15:12 +02:00
|
|
|
|
|
|
|
start_of_query= thd->start_time;
|
2000-10-14 02:16:35 +02:00
|
|
|
thd->end_time(); // Set start time
|
2001-02-16 22:46:52 +01:00
|
|
|
|
2005-06-16 21:05:38 +02:00
|
|
|
/*
|
|
|
|
Do not log administrative statements unless the appropriate option is
|
|
|
|
set; do not log into slow log if reading from backup.
|
|
|
|
*/
|
|
|
|
if (thd->enable_slow_log && !thd->user_time)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2001-02-16 22:46:52 +01:00
|
|
|
thd->proc_info="logging slow query";
|
|
|
|
|
2002-06-28 18:30:09 +02:00
|
|
|
if ((ulong) (thd->start_time - thd->time_after_lock) >
|
|
|
|
thd->variables.long_query_time ||
|
2003-12-06 23:21:09 +01:00
|
|
|
((thd->server_status &
|
|
|
|
(SERVER_QUERY_NO_INDEX_USED | SERVER_QUERY_NO_GOOD_INDEX_USED)) &&
|
2003-10-07 09:05:35 +02:00
|
|
|
(specialflag & SPECIAL_LOG_QUERIES_NOT_USING_INDEXES)))
|
2000-10-14 02:16:35 +02:00
|
|
|
{
|
2004-09-13 15:48:01 +02:00
|
|
|
thd->status_var.long_query_count++;
|
2006-05-02 19:56:43 +02:00
|
|
|
slow_log_print(thd, thd->query, thd->query_length, start_of_query);
|
2000-10-14 02:16:35 +02:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2006-05-02 19:56:43 +02:00
|
|
|
DBUG_VOID_RETURN;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
2002-10-02 12:33:08 +02:00
|
|
|
|
2004-11-13 11:56:39 +01:00
|
|
|
int prepare_schema_table(THD *thd, LEX *lex, Table_ident *table_ident,
|
|
|
|
enum enum_schema_tables schema_table_idx)
|
|
|
|
{
|
|
|
|
DBUG_ENTER("prepare_schema_table");
|
|
|
|
SELECT_LEX *sel= 0;
|
2005-01-12 02:38:53 +01:00
|
|
|
switch (schema_table_idx) {
|
2004-11-13 11:56:39 +01:00
|
|
|
case SCH_SCHEMATA:
|
|
|
|
#if defined(DONT_ALLOW_SHOW_COMMANDS)
|
2004-11-13 22:26:15 +01:00
|
|
|
my_message(ER_NOT_ALLOWED_COMMAND,
|
|
|
|
ER(ER_NOT_ALLOWED_COMMAND), MYF(0)); /* purecov: inspected */
|
2004-11-13 11:56:39 +01:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
#else
|
|
|
|
if ((specialflag & SPECIAL_SKIP_SHOW_DB) &&
|
|
|
|
check_global_access(thd, SHOW_DB_ACL))
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
break;
|
|
|
|
#endif
|
|
|
|
case SCH_TABLE_NAMES:
|
|
|
|
case SCH_TABLES:
|
|
|
|
case SCH_VIEWS:
|
2005-07-19 18:06:49 +02:00
|
|
|
case SCH_TRIGGERS:
|
2006-05-02 19:56:43 +02:00
|
|
|
case SCH_EVENTS:
|
2004-11-13 11:56:39 +01:00
|
|
|
#ifdef DONT_ALLOW_SHOW_COMMANDS
|
2004-11-13 22:26:15 +01:00
|
|
|
my_message(ER_NOT_ALLOWED_COMMAND,
|
|
|
|
ER(ER_NOT_ALLOWED_COMMAND), MYF(0)); /* purecov: inspected */
|
2004-11-13 11:56:39 +01:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
#else
|
|
|
|
{
|
|
|
|
char *db= lex->select_lex.db ? lex->select_lex.db : thd->db;
|
|
|
|
if (!db)
|
|
|
|
{
|
2004-11-13 22:26:15 +01:00
|
|
|
my_message(ER_NO_DB_ERROR,
|
|
|
|
ER(ER_NO_DB_ERROR), MYF(0)); /* purecov: inspected */
|
2004-11-13 11:56:39 +01:00
|
|
|
DBUG_RETURN(1); /* purecov: inspected */
|
|
|
|
}
|
|
|
|
remove_escape(db); // Fix escaped '_'
|
|
|
|
if (check_db_name(db))
|
|
|
|
{
|
2004-11-13 22:26:15 +01:00
|
|
|
my_error(ER_WRONG_DB_NAME, MYF(0), db);
|
2004-11-13 11:56:39 +01:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd, SELECT_ACL, db, &thd->col_access, 0, 0,
|
|
|
|
is_schema_db(db)))
|
2004-11-13 11:56:39 +01:00
|
|
|
DBUG_RETURN(1); /* purecov: inspected */
|
|
|
|
if (!thd->col_access && check_grant_db(thd,db))
|
|
|
|
{
|
2004-11-13 22:26:15 +01:00
|
|
|
my_error(ER_DBACCESS_DENIED_ERROR, MYF(0),
|
2005-09-15 21:29:07 +02:00
|
|
|
thd->security_ctx->priv_user, thd->security_ctx->priv_host,
|
|
|
|
db);
|
2004-11-13 11:56:39 +01:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
2005-01-12 02:38:53 +01:00
|
|
|
/*
|
|
|
|
We need to do a copy to make this prepared statement safe if this
|
|
|
|
was thd->db
|
|
|
|
*/
|
|
|
|
lex->select_lex.db= thd->strdup(db);
|
2004-11-13 11:56:39 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
case SCH_COLUMNS:
|
|
|
|
case SCH_STATISTICS:
|
|
|
|
#ifdef DONT_ALLOW_SHOW_COMMANDS
|
2004-11-13 22:26:15 +01:00
|
|
|
my_message(ER_NOT_ALLOWED_COMMAND,
|
|
|
|
ER(ER_NOT_ALLOWED_COMMAND), MYF(0)); /* purecov: inspected */
|
2004-11-13 11:56:39 +01:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
#else
|
|
|
|
if (table_ident)
|
|
|
|
{
|
|
|
|
TABLE_LIST **query_tables_last= lex->query_tables_last;
|
|
|
|
sel= new SELECT_LEX();
|
2005-08-12 16:57:19 +02:00
|
|
|
/* 'parent_lex' is used in init_query() so it must be before it. */
|
|
|
|
sel->parent_lex= lex;
|
2004-11-13 11:56:39 +01:00
|
|
|
sel->init_query();
|
2005-07-18 14:33:18 +02:00
|
|
|
if (!sel->add_table_to_list(thd, table_ident, 0, 0, TL_READ,
|
2004-11-13 11:56:39 +01:00
|
|
|
(List<String> *) 0, (List<String> *) 0))
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
lex->query_tables_last= query_tables_last;
|
|
|
|
TABLE_LIST *table_list= (TABLE_LIST*) sel->table_list.first;
|
|
|
|
char *db= table_list->db;
|
|
|
|
remove_escape(db); // Fix escaped '_'
|
2005-01-06 12:00:13 +01:00
|
|
|
remove_escape(table_list->table_name);
|
2004-11-13 11:56:39 +01:00
|
|
|
if (check_access(thd,SELECT_ACL | EXTRA_ACL,db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&table_list->grant.privilege, 0, 0,
|
|
|
|
test(table_list->schema_table)))
|
2004-11-13 11:56:39 +01:00
|
|
|
DBUG_RETURN(1); /* purecov: inspected */
|
|
|
|
if (grant_option && check_grant(thd, SELECT_ACL, table_list, 2,
|
|
|
|
UINT_MAX, 0))
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
2004-12-30 13:20:40 +01:00
|
|
|
case SCH_OPEN_TABLES:
|
|
|
|
case SCH_VARIABLES:
|
|
|
|
case SCH_STATUS:
|
2004-11-13 11:56:39 +01:00
|
|
|
case SCH_PROCEDURES:
|
|
|
|
case SCH_CHARSETS:
|
2006-05-02 19:56:43 +02:00
|
|
|
case SCH_ENGINES:
|
2004-11-13 11:56:39 +01:00
|
|
|
case SCH_COLLATIONS:
|
|
|
|
case SCH_COLLATION_CHARACTER_SET_APPLICABILITY:
|
|
|
|
case SCH_USER_PRIVILEGES:
|
|
|
|
case SCH_SCHEMA_PRIVILEGES:
|
|
|
|
case SCH_TABLE_PRIVILEGES:
|
|
|
|
case SCH_COLUMN_PRIVILEGES:
|
|
|
|
case SCH_TABLE_CONSTRAINTS:
|
|
|
|
case SCH_KEY_COLUMN_USAGE:
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
SELECT_LEX *select_lex= lex->current_select;
|
|
|
|
if (make_schema_select(thd, select_lex, schema_table_idx))
|
|
|
|
{
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
TABLE_LIST *table_list= (TABLE_LIST*) select_lex->table_list.first;
|
|
|
|
table_list->schema_select_lex= sel;
|
2005-01-24 16:44:54 +01:00
|
|
|
table_list->schema_table_reformed= 1;
|
2005-07-13 10:30:39 +02:00
|
|
|
statistic_increment(thd->status_var.com_stat[lex->orig_sql_command],
|
|
|
|
&LOCK_status);
|
2004-11-13 11:56:39 +01:00
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2002-10-02 12:33:08 +02:00
|
|
|
/*
|
|
|
|
Read query from packet and store in thd->query
|
2005-06-17 21:26:25 +02:00
|
|
|
Used in COM_QUERY and COM_STMT_PREPARE
|
2002-10-02 12:33:08 +02:00
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
Sets the following THD variables:
|
|
|
|
query
|
|
|
|
query_length
|
|
|
|
|
|
|
|
RETURN VALUES
|
2004-10-20 03:04:37 +02:00
|
|
|
FALSE ok
|
|
|
|
TRUE error; In this case thd->fatal_error is set
|
2002-10-02 12:33:08 +02:00
|
|
|
*/
|
|
|
|
|
Implement WL#2661 "Prepared Statements: Dynamic SQL in Stored Procedures".
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
2005-09-03 01:13:18 +02:00
|
|
|
bool alloc_query(THD *thd, const char *packet, uint packet_length)
|
2002-10-02 12:33:08 +02:00
|
|
|
{
|
|
|
|
packet_length--; // Remove end null
|
2003-01-20 23:00:50 +01:00
|
|
|
/* Remove garbage at start and end of query */
|
2003-03-15 14:24:21 +01:00
|
|
|
while (my_isspace(thd->charset(),packet[0]) && packet_length > 0)
|
2002-10-02 12:33:08 +02:00
|
|
|
{
|
|
|
|
packet++;
|
|
|
|
packet_length--;
|
|
|
|
}
|
Implement WL#2661 "Prepared Statements: Dynamic SQL in Stored Procedures".
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
2005-09-03 01:13:18 +02:00
|
|
|
const char *pos= packet + packet_length; // Point at end null
|
2002-11-30 18:58:53 +01:00
|
|
|
while (packet_length > 0 &&
|
2003-03-15 14:24:21 +01:00
|
|
|
(pos[-1] == ';' || my_isspace(thd->charset() ,pos[-1])))
|
2002-10-02 12:33:08 +02:00
|
|
|
{
|
|
|
|
pos--;
|
|
|
|
packet_length--;
|
|
|
|
}
|
|
|
|
/* We must allocate some extra memory for query cache */
|
2004-05-17 00:52:13 +02:00
|
|
|
thd->query_length= 0; // Extra safety: Avoid races
|
2002-10-02 12:33:08 +02:00
|
|
|
if (!(thd->query= (char*) thd->memdup_w_gap((gptr) (packet),
|
|
|
|
packet_length,
|
2003-10-02 11:02:05 +02:00
|
|
|
thd->db_length+ 1 +
|
|
|
|
QUERY_CACHE_FLAGS_SIZE)))
|
2004-10-20 03:04:37 +02:00
|
|
|
return TRUE;
|
2002-10-02 12:33:08 +02:00
|
|
|
thd->query[packet_length]=0;
|
|
|
|
thd->query_length= packet_length;
|
2004-05-25 00:03:49 +02:00
|
|
|
|
|
|
|
/* Reclaim some memory */
|
|
|
|
thd->packet.shrink(thd->variables.net_buffer_length);
|
|
|
|
thd->convert_buffer.shrink(thd->variables.net_buffer_length);
|
2002-10-02 12:33:08 +02:00
|
|
|
|
2004-10-20 03:04:37 +02:00
|
|
|
return FALSE;
|
2002-10-02 12:33:08 +02:00
|
|
|
}
|
|
|
|
|
2005-08-31 18:08:45 +02:00
|
|
|
static void reset_one_shot_variables(THD *thd)
|
|
|
|
{
|
|
|
|
thd->variables.character_set_client=
|
|
|
|
global_system_variables.character_set_client;
|
|
|
|
thd->variables.collation_connection=
|
|
|
|
global_system_variables.collation_connection;
|
|
|
|
thd->variables.collation_database=
|
|
|
|
global_system_variables.collation_database;
|
|
|
|
thd->variables.collation_server=
|
|
|
|
global_system_variables.collation_server;
|
|
|
|
thd->update_charset();
|
|
|
|
thd->variables.time_zone=
|
|
|
|
global_system_variables.time_zone;
|
|
|
|
thd->one_shot_set= 0;
|
|
|
|
}
|
|
|
|
|
2005-08-25 15:34:34 +02:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
/****************************************************************************
|
|
|
|
** mysql_execute_command
|
|
|
|
** Execute command saved in thd and current_lex->sql_command
|
|
|
|
****************************************************************************/
|
|
|
|
|
2004-10-20 03:04:37 +02:00
|
|
|
bool
|
2002-10-02 12:33:08 +02:00
|
|
|
mysql_execute_command(THD *thd)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-10-20 03:04:37 +02:00
|
|
|
bool res= FALSE;
|
2004-11-25 01:23:13 +01:00
|
|
|
int result= 0;
|
2003-05-05 20:54:37 +02:00
|
|
|
LEX *lex= thd->lex;
|
2004-10-29 18:26:52 +02:00
|
|
|
/* first SELECT_LEX (have special meaning for many of non-SELECTcommands) */
|
2002-10-03 15:35:08 +02:00
|
|
|
SELECT_LEX *select_lex= &lex->select_lex;
|
2004-07-16 00:15:55 +02:00
|
|
|
/* first table of first SELECT_LEX */
|
2004-10-29 18:26:52 +02:00
|
|
|
TABLE_LIST *first_table= (TABLE_LIST*) select_lex->table_list.first;
|
2004-07-16 00:15:55 +02:00
|
|
|
/* list of all tables in query */
|
|
|
|
TABLE_LIST *all_tables;
|
|
|
|
/* most outer SELECT_LEX_UNIT of query */
|
2002-05-08 22:14:40 +02:00
|
|
|
SELECT_LEX_UNIT *unit= &lex->unit;
|
2005-02-08 20:52:50 +01:00
|
|
|
/* Saved variable value */
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_ENTER("mysql_execute_command");
|
2005-01-20 09:41:37 +01:00
|
|
|
thd->net.no_send_error= 0;
|
2006-05-02 19:56:43 +02:00
|
|
|
#ifdef WITH_PARTITION_STORAGE_ENGINE
|
|
|
|
thd->work_part_info= 0;
|
|
|
|
#endif
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2004-07-16 00:15:55 +02:00
|
|
|
/*
|
|
|
|
In many cases first table of main SELECT_LEX have special meaning =>
|
|
|
|
check that it is first table in global list and relink it first in
|
|
|
|
queries_tables list if it is necessary (we need such relinking only
|
|
|
|
for queries with subqueries in select list, in this case tables of
|
|
|
|
subqueries will go to global list first)
|
|
|
|
|
|
|
|
all_tables will differ from first_table only if most upper SELECT_LEX
|
|
|
|
do not contain tables.
|
|
|
|
|
|
|
|
Because of above in place where should be at least one table in most
|
|
|
|
outer SELECT_LEX we have following check:
|
|
|
|
DBUG_ASSERT(first_table == all_tables);
|
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
*/
|
|
|
|
lex->first_lists_tables_same();
|
2004-09-03 20:43:04 +02:00
|
|
|
/* should be assigned after making first tables same */
|
2004-07-16 00:15:55 +02:00
|
|
|
all_tables= lex->query_tables;
|
2005-07-01 06:05:42 +02:00
|
|
|
/* set context for commands which do not use setup_tables */
|
|
|
|
select_lex->
|
|
|
|
context.resolve_in_table_list_only((TABLE_LIST*)select_lex->
|
|
|
|
table_list.first);
|
2004-07-16 00:15:55 +02:00
|
|
|
|
2002-10-02 12:33:08 +02:00
|
|
|
/*
|
|
|
|
Reset warning count for each query that uses tables
|
|
|
|
A better approach would be to reset this for any commands
|
|
|
|
that is not a SHOW command or a select that only access local
|
|
|
|
variables, but for now this is probably good enough.
|
2005-04-22 12:53:48 +02:00
|
|
|
Don't reset warnings when executing a stored routine.
|
2002-10-02 12:33:08 +02:00
|
|
|
*/
|
2005-04-22 12:53:48 +02:00
|
|
|
if ((all_tables || &lex->select_lex != lex->all_selects_list ||
|
2005-07-18 22:55:37 +02:00
|
|
|
lex->sroutines.records) && !thd->spcont ||
|
2005-06-17 17:56:04 +02:00
|
|
|
lex->time_zone_tables_used)
|
2005-03-16 15:11:01 +01:00
|
|
|
mysql_reset_errors(thd, 0);
|
2002-10-02 12:33:08 +02:00
|
|
|
|
2003-09-29 08:47:37 +02:00
|
|
|
#ifdef HAVE_REPLICATION
|
2005-10-17 10:52:34 +02:00
|
|
|
if (unlikely(thd->slave_thread))
|
2001-12-19 05:56:36 +01:00
|
|
|
{
|
2002-11-30 18:58:53 +01:00
|
|
|
/*
|
2005-10-10 15:10:14 +02:00
|
|
|
Check if statment should be skipped because of slave filtering
|
|
|
|
rules
|
2005-08-31 18:08:45 +02:00
|
|
|
|
|
|
|
Exceptions are:
|
2005-10-10 15:10:14 +02:00
|
|
|
- UPDATE MULTI: For this statement, we want to check the filtering
|
|
|
|
rules later in the code
|
2005-08-31 18:08:45 +02:00
|
|
|
- SET: we always execute it (Not that many SET commands exists in
|
2005-09-01 17:58:00 +02:00
|
|
|
the binary log anyway -- only 4.1 masters write SET statements,
|
|
|
|
in 5.0 there are no SET statements in the binary log)
|
2005-08-31 18:08:45 +02:00
|
|
|
- DROP TEMPORARY TABLE IF EXISTS: we always execute it (otherwise we
|
|
|
|
have stale files on slave caused by exclusion of one tmp table).
|
2001-12-20 12:52:04 +01:00
|
|
|
*/
|
2005-10-10 15:10:14 +02:00
|
|
|
if (!(lex->sql_command == SQLCOM_UPDATE_MULTI) &&
|
|
|
|
!(lex->sql_command == SQLCOM_SET_OPTION) &&
|
2005-08-31 18:08:45 +02:00
|
|
|
!(lex->sql_command == SQLCOM_DROP_TABLE &&
|
2005-02-14 23:47:17 +01:00
|
|
|
lex->drop_temporary && lex->drop_if_exists) &&
|
2005-02-16 21:03:58 +01:00
|
|
|
all_tables_not_ok(thd, all_tables))
|
2003-07-08 15:50:57 +02:00
|
|
|
{
|
|
|
|
/* we warn the slave SQL thread */
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_SLAVE_IGNORED_TABLE, ER(ER_SLAVE_IGNORED_TABLE), MYF(0));
|
2005-08-31 18:08:45 +02:00
|
|
|
reset_one_shot_variables(thd);
|
2003-02-12 16:17:03 +01:00
|
|
|
DBUG_RETURN(0);
|
2003-07-08 15:50:57 +02:00
|
|
|
}
|
2001-12-19 05:56:36 +01:00
|
|
|
}
|
2005-10-17 10:52:34 +02:00
|
|
|
else
|
2006-01-23 17:03:09 +01:00
|
|
|
{
|
2005-10-10 19:23:13 +02:00
|
|
|
#endif /* HAVE_REPLICATION */
|
2006-01-23 17:03:09 +01:00
|
|
|
/*
|
|
|
|
When option readonly is set deny operations which change non-temporary
|
|
|
|
tables. Except for the replication thread and the 'super' users.
|
|
|
|
*/
|
|
|
|
if (opt_readonly &&
|
|
|
|
!(thd->security_ctx->master_access & SUPER_ACL) &&
|
|
|
|
uc_update_queries[lex->sql_command] &&
|
|
|
|
!((lex->sql_command == SQLCOM_CREATE_TABLE) &&
|
|
|
|
(lex->create_info.options & HA_LEX_CREATE_TMP_TABLE)) &&
|
|
|
|
((lex->sql_command != SQLCOM_UPDATE_MULTI) &&
|
|
|
|
some_non_temp_table_to_be_updated(thd, all_tables)))
|
|
|
|
{
|
|
|
|
my_error(ER_OPTION_PREVENTS_STATEMENT, MYF(0), "--read-only");
|
|
|
|
DBUG_RETURN(-1);
|
|
|
|
}
|
|
|
|
#ifdef HAVE_REPLICATION
|
|
|
|
} /* endif unlikely slave */
|
|
|
|
#endif
|
2005-07-13 10:30:39 +02:00
|
|
|
if(lex->orig_sql_command == SQLCOM_END)
|
|
|
|
statistic_increment(thd->status_var.com_stat[lex->sql_command],
|
|
|
|
&LOCK_status);
|
2001-08-18 14:29:21 +02:00
|
|
|
|
2006-04-24 16:39:33 +02:00
|
|
|
#ifdef HAVE_ROW_BASED_REPLICATION
|
2006-05-02 19:56:43 +02:00
|
|
|
if (lex->binlog_row_based_if_mixed)
|
|
|
|
thd->set_current_stmt_binlog_row_based_if_mixed();
|
2006-04-24 16:39:33 +02:00
|
|
|
#endif /*HAVE_ROW_BASED_REPLICATION*/
|
2006-05-02 19:56:43 +02:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
switch (lex->sql_command) {
|
|
|
|
case SQLCOM_SELECT:
|
|
|
|
{
|
2004-05-05 20:21:41 +02:00
|
|
|
/* assign global limit variable if limit is not given */
|
|
|
|
{
|
|
|
|
SELECT_LEX *param= lex->unit.global_parameters;
|
|
|
|
if (!param->explicit_limit)
|
2005-06-07 12:11:36 +02:00
|
|
|
param->select_limit=
|
|
|
|
new Item_int((ulonglong)thd->variables.select_limit);
|
2004-05-05 20:21:41 +02:00
|
|
|
}
|
|
|
|
|
2002-10-16 15:55:08 +02:00
|
|
|
select_result *result=lex->result;
|
2004-07-16 00:15:55 +02:00
|
|
|
if (all_tables)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2005-07-19 10:49:43 +02:00
|
|
|
if (lex->orig_sql_command != SQLCOM_SHOW_STATUS_PROC &&
|
2006-05-02 19:56:43 +02:00
|
|
|
lex->orig_sql_command != SQLCOM_SHOW_STATUS_FUNC &&
|
|
|
|
lex->orig_sql_command != SQLCOM_SHOW_EVENTS)
|
2005-07-19 10:49:43 +02:00
|
|
|
res= check_table_access(thd,
|
|
|
|
lex->exchange ? SELECT_ACL | FILE_ACL :
|
|
|
|
SELECT_ACL,
|
|
|
|
all_tables, 0);
|
2006-05-02 19:56:43 +02:00
|
|
|
else if (lex->orig_sql_command == SQLCOM_SHOW_EVENTS)
|
|
|
|
res= check_access(thd, EVENT_ACL, thd->lex->select_lex.db, 0, 0, 0,
|
|
|
|
is_schema_db(thd->lex->select_lex.db));
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
else
|
2004-07-16 00:15:55 +02:00
|
|
|
res= check_access(thd,
|
|
|
|
lex->exchange ? SELECT_ACL | FILE_ACL : SELECT_ACL,
|
2005-09-13 13:07:38 +02:00
|
|
|
any_db, 0, 0, 0, 0);
|
2000-07-31 21:29:14 +02:00
|
|
|
if (res)
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2004-07-16 00:15:55 +02:00
|
|
|
if (!(res= open_and_lock_tables(thd, all_tables)))
|
2001-12-02 13:34:01 +01:00
|
|
|
{
|
2002-10-16 20:17:57 +02:00
|
|
|
if (lex->describe)
|
2002-09-26 22:08:22 +02:00
|
|
|
{
|
2005-09-09 14:23:12 +02:00
|
|
|
/*
|
|
|
|
We always use select_send for EXPLAIN, even if it's an EXPLAIN
|
|
|
|
for SELECT ... INTO OUTFILE: a user application should be able
|
|
|
|
to prepend EXPLAIN to any query and receive output for it,
|
|
|
|
even if the query itself redirects the output.
|
|
|
|
*/
|
2002-11-09 14:40:46 +01:00
|
|
|
if (!(result= new select_send()))
|
2003-01-15 15:39:36 +01:00
|
|
|
goto error;
|
2002-11-09 14:40:46 +01:00
|
|
|
else
|
|
|
|
thd->send_explain_fields(result);
|
2003-05-05 20:54:37 +02:00
|
|
|
res= mysql_explain_union(thd, &thd->lex->unit, result);
|
2003-10-16 14:54:47 +02:00
|
|
|
if (lex->describe & DESCRIBE_EXTENDED)
|
|
|
|
{
|
|
|
|
char buff[1024];
|
|
|
|
String str(buff,(uint32) sizeof(buff), system_charset_info);
|
|
|
|
str.length(0);
|
2003-11-19 16:59:35 +01:00
|
|
|
thd->lex->unit.print(&str);
|
2003-10-16 14:54:47 +02:00
|
|
|
str.append('\0');
|
|
|
|
push_warning(thd, MYSQL_ERROR::WARN_LEVEL_NOTE,
|
|
|
|
ER_YES, str.ptr());
|
|
|
|
}
|
2002-11-09 14:40:46 +01:00
|
|
|
result->send_eof();
|
2004-08-24 18:17:11 +02:00
|
|
|
delete result;
|
2002-09-26 22:08:22 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2004-08-24 18:17:11 +02:00
|
|
|
if (!result && !(result= new select_send()))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2004-07-16 00:15:55 +02:00
|
|
|
query_cache_store_query(thd, all_tables);
|
2005-01-03 20:04:33 +01:00
|
|
|
res= handle_select(thd, lex, result, 0);
|
2004-08-24 18:17:11 +02:00
|
|
|
if (result != lex->result)
|
|
|
|
delete result;
|
2002-09-26 22:08:22 +02:00
|
|
|
}
|
2001-12-02 13:34:01 +01:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2004-04-05 17:43:37 +02:00
|
|
|
case SQLCOM_PREPARE:
|
2004-06-07 10:09:10 +02:00
|
|
|
{
|
Implement WL#2661 "Prepared Statements: Dynamic SQL in Stored Procedures".
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
2005-09-03 01:13:18 +02:00
|
|
|
mysql_sql_stmt_prepare(thd);
|
2004-04-05 17:43:37 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_EXECUTE:
|
|
|
|
{
|
Implement WL#2661 "Prepared Statements: Dynamic SQL in Stored Procedures".
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
2005-09-03 01:13:18 +02:00
|
|
|
mysql_sql_stmt_execute(thd);
|
2004-04-05 17:43:37 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_DEALLOCATE_PREPARE:
|
|
|
|
{
|
Implement WL#2661 "Prepared Statements: Dynamic SQL in Stored Procedures".
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
2005-09-03 01:13:18 +02:00
|
|
|
mysql_sql_stmt_close(thd);
|
2004-04-05 17:43:37 +02:00
|
|
|
break;
|
|
|
|
}
|
2001-12-17 18:59:20 +01:00
|
|
|
case SQLCOM_DO:
|
2005-03-04 14:35:28 +01:00
|
|
|
if (check_table_access(thd, SELECT_ACL, all_tables, 0) ||
|
|
|
|
open_and_lock_tables(thd, all_tables))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2003-01-03 16:08:00 +01:00
|
|
|
|
|
|
|
res= mysql_do(thd, *lex->insert_list);
|
2001-12-17 18:59:20 +01:00
|
|
|
break;
|
|
|
|
|
2002-02-14 14:04:14 +01:00
|
|
|
case SQLCOM_EMPTY_QUERY:
|
2002-10-02 12:33:08 +02:00
|
|
|
send_ok(thd);
|
2002-02-14 14:04:14 +01:00
|
|
|
break;
|
|
|
|
|
2002-10-28 14:44:19 +01:00
|
|
|
case SQLCOM_HELP:
|
|
|
|
res= mysqld_help(thd,lex->help_arg);
|
|
|
|
break;
|
|
|
|
|
2002-12-16 14:33:29 +01:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2000-10-27 06:11:55 +02:00
|
|
|
case SQLCOM_PURGE:
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2002-06-12 14:04:18 +02:00
|
|
|
if (check_global_access(thd, SUPER_ACL))
|
2001-08-14 19:33:49 +02:00
|
|
|
goto error;
|
2004-07-07 10:29:39 +02:00
|
|
|
/* PURGE MASTER LOGS TO 'file' */
|
2001-08-14 19:33:49 +02:00
|
|
|
res = purge_master_logs(thd, lex->to_log);
|
|
|
|
break;
|
|
|
|
}
|
2003-02-16 17:39:12 +01:00
|
|
|
case SQLCOM_PURGE_BEFORE:
|
|
|
|
{
|
2004-11-25 16:13:06 +01:00
|
|
|
Item *it;
|
|
|
|
|
2003-02-16 17:39:12 +01:00
|
|
|
if (check_global_access(thd, SUPER_ACL))
|
|
|
|
goto error;
|
2004-07-07 10:29:39 +02:00
|
|
|
/* PURGE MASTER LOGS BEFORE 'data' */
|
2004-11-25 16:13:06 +01:00
|
|
|
it= (Item *)lex->value_list.head();
|
2005-07-01 06:05:42 +02:00
|
|
|
if ((!it->fixed && it->fix_fields(lex->thd, &it)) ||
|
2005-02-08 23:50:45 +01:00
|
|
|
it->check_cols(1))
|
2004-11-25 16:13:06 +01:00
|
|
|
{
|
|
|
|
my_error(ER_WRONG_ARGUMENTS, MYF(0), "PURGE LOGS BEFORE");
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
it= new Item_func_unix_timestamp(it);
|
|
|
|
/*
|
|
|
|
it is OK only emulate fix_fieds, because we need only
|
|
|
|
value of constant
|
|
|
|
*/
|
|
|
|
it->quick_fix_field();
|
|
|
|
res = purge_master_logs_before_date(thd, (ulong)it->val_int());
|
2003-02-16 17:39:12 +01:00
|
|
|
break;
|
|
|
|
}
|
2002-12-16 14:33:29 +01:00
|
|
|
#endif
|
2002-06-12 23:13:12 +02:00
|
|
|
case SQLCOM_SHOW_WARNS:
|
|
|
|
{
|
2002-10-02 12:33:08 +02:00
|
|
|
res= mysqld_show_warnings(thd, (ulong)
|
|
|
|
((1L << (uint) MYSQL_ERROR::WARN_LEVEL_NOTE) |
|
2003-01-04 14:37:20 +01:00
|
|
|
(1L << (uint) MYSQL_ERROR::WARN_LEVEL_WARN) |
|
|
|
|
(1L << (uint) MYSQL_ERROR::WARN_LEVEL_ERROR)
|
|
|
|
));
|
2002-06-12 23:13:12 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_SHOW_ERRORS:
|
|
|
|
{
|
2002-10-02 12:33:08 +02:00
|
|
|
res= mysqld_show_warnings(thd, (ulong)
|
|
|
|
(1L << (uint) MYSQL_ERROR::WARN_LEVEL_ERROR));
|
2002-06-12 23:13:12 +02:00
|
|
|
break;
|
|
|
|
}
|
2001-07-05 01:14:31 +02:00
|
|
|
case SQLCOM_SHOW_NEW_MASTER:
|
|
|
|
{
|
2002-06-12 14:04:18 +02:00
|
|
|
if (check_global_access(thd, REPL_SLAVE_ACL))
|
2001-07-05 01:14:31 +02:00
|
|
|
goto error;
|
2003-08-20 01:38:31 +02:00
|
|
|
/* This query don't work now. See comment in repl_failsafe.cc */
|
2002-09-11 05:40:08 +02:00
|
|
|
#ifndef WORKING_NEW_MASTER
|
2004-10-20 03:04:37 +02:00
|
|
|
my_error(ER_NOT_SUPPORTED_YET, MYF(0), "SHOW NEW MASTER");
|
|
|
|
goto error;
|
2002-09-11 05:40:08 +02:00
|
|
|
#else
|
2001-07-05 01:14:31 +02:00
|
|
|
res = show_new_master(thd);
|
|
|
|
break;
|
2004-10-20 03:04:37 +02:00
|
|
|
#endif
|
2001-07-05 01:14:31 +02:00
|
|
|
}
|
2002-12-16 14:33:29 +01:00
|
|
|
|
2003-09-08 12:08:53 +02:00
|
|
|
#ifdef HAVE_REPLICATION
|
2001-05-31 02:50:56 +02:00
|
|
|
case SQLCOM_SHOW_SLAVE_HOSTS:
|
|
|
|
{
|
2002-06-12 14:04:18 +02:00
|
|
|
if (check_global_access(thd, REPL_SLAVE_ACL))
|
2001-05-31 02:50:56 +02:00
|
|
|
goto error;
|
|
|
|
res = show_slave_hosts(thd);
|
|
|
|
break;
|
|
|
|
}
|
2001-06-21 21:19:24 +02:00
|
|
|
case SQLCOM_SHOW_BINLOG_EVENTS:
|
|
|
|
{
|
2002-06-12 14:04:18 +02:00
|
|
|
if (check_global_access(thd, REPL_SLAVE_ACL))
|
2001-06-21 21:19:24 +02:00
|
|
|
goto error;
|
2005-01-16 13:16:23 +01:00
|
|
|
res = mysql_show_binlog_events(thd);
|
2001-06-21 21:19:24 +02:00
|
|
|
break;
|
|
|
|
}
|
2002-12-16 14:33:29 +01:00
|
|
|
#endif
|
|
|
|
|
2000-09-15 00:34:50 +02:00
|
|
|
case SQLCOM_BACKUP_TABLE:
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_db_used(thd, all_tables) ||
|
|
|
|
check_table_access(thd, SELECT_ACL, all_tables, 0) ||
|
2002-06-12 14:04:18 +02:00
|
|
|
check_global_access(thd, FILE_ACL))
|
2001-08-14 19:33:49 +02:00
|
|
|
goto error; /* purecov: inspected */
|
2005-06-16 21:05:38 +02:00
|
|
|
thd->enable_slow_log= opt_log_slow_admin_statements;
|
2004-07-16 00:15:55 +02:00
|
|
|
res = mysql_backup_table(thd, first_table);
|
2005-12-07 16:15:02 +01:00
|
|
|
select_lex->table_list.first= (byte*) first_table;
|
2005-12-03 15:02:09 +01:00
|
|
|
lex->query_tables=all_tables;
|
2001-08-14 19:33:49 +02:00
|
|
|
break;
|
|
|
|
}
|
2000-09-15 00:34:50 +02:00
|
|
|
case SQLCOM_RESTORE_TABLE:
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_db_used(thd, all_tables) ||
|
|
|
|
check_table_access(thd, INSERT_ACL, all_tables, 0) ||
|
2002-06-12 14:04:18 +02:00
|
|
|
check_global_access(thd, FILE_ACL))
|
2001-08-14 19:33:49 +02:00
|
|
|
goto error; /* purecov: inspected */
|
2005-06-16 21:05:38 +02:00
|
|
|
thd->enable_slow_log= opt_log_slow_admin_statements;
|
2004-07-16 00:15:55 +02:00
|
|
|
res = mysql_restore_table(thd, first_table);
|
2005-12-07 16:15:02 +01:00
|
|
|
select_lex->table_list.first= (byte*) first_table;
|
2005-12-03 15:02:09 +01:00
|
|
|
lex->query_tables=all_tables;
|
2001-08-14 19:33:49 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-07-16 21:30:49 +02:00
|
|
|
case SQLCOM_ASSIGN_TO_KEYCACHE:
|
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_db_used(thd, all_tables) ||
|
|
|
|
check_access(thd, INDEX_ACL, first_table->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&first_table->grant.privilege, 0, 0,
|
|
|
|
test(first_table->schema_table)))
|
2003-08-02 12:26:02 +02:00
|
|
|
goto error;
|
2005-01-16 13:16:23 +01:00
|
|
|
res= mysql_assign_to_keycache(thd, first_table, &lex->ident);
|
2003-07-16 21:30:49 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-06-12 13:29:02 +02:00
|
|
|
case SQLCOM_PRELOAD_KEYS:
|
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_db_used(thd, all_tables) ||
|
|
|
|
check_access(thd, INDEX_ACL, first_table->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&first_table->grant.privilege, 0, 0,
|
|
|
|
test(first_table->schema_table)))
|
2003-11-14 13:50:19 +01:00
|
|
|
goto error;
|
2004-07-16 00:15:55 +02:00
|
|
|
res = mysql_preload_keys(thd, first_table);
|
2003-06-12 13:29:02 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-09-08 12:08:53 +02:00
|
|
|
#ifdef HAVE_REPLICATION
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_CHANGE_MASTER:
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2002-06-12 14:04:18 +02:00
|
|
|
if (check_global_access(thd, SUPER_ACL))
|
2001-08-14 19:33:49 +02:00
|
|
|
goto error;
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_lock(&LOCK_active_mi);
|
2002-01-20 03:16:52 +01:00
|
|
|
res = change_master(thd,active_mi);
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_unlock(&LOCK_active_mi);
|
2001-08-14 19:33:49 +02:00
|
|
|
break;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_SHOW_SLAVE_STAT:
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2003-05-14 00:55:23 +02:00
|
|
|
/* Accept one of two privileges */
|
|
|
|
if (check_global_access(thd, SUPER_ACL | REPL_CLIENT_ACL))
|
2001-08-14 19:33:49 +02:00
|
|
|
goto error;
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_lock(&LOCK_active_mi);
|
2002-01-20 03:16:52 +01:00
|
|
|
res = show_master_info(thd,active_mi);
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_unlock(&LOCK_active_mi);
|
2001-08-14 19:33:49 +02:00
|
|
|
break;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_SHOW_MASTER_STAT:
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2003-05-14 00:55:23 +02:00
|
|
|
/* Accept one of two privileges */
|
|
|
|
if (check_global_access(thd, SUPER_ACL | REPL_CLIENT_ACL))
|
2001-08-14 19:33:49 +02:00
|
|
|
goto error;
|
|
|
|
res = show_binlog_info(thd);
|
|
|
|
break;
|
|
|
|
}
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2001-05-29 03:18:23 +02:00
|
|
|
case SQLCOM_LOAD_MASTER_DATA: // sync with master
|
2002-06-12 14:04:18 +02:00
|
|
|
if (check_global_access(thd, SUPER_ACL))
|
2001-05-29 03:18:23 +02:00
|
|
|
goto error;
|
2002-08-08 02:12:02 +02:00
|
|
|
if (end_active_trans(thd))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
res = load_master_data(thd);
|
2001-05-29 03:18:23 +02:00
|
|
|
break;
|
2003-09-08 12:08:53 +02:00
|
|
|
#endif /* HAVE_REPLICATION */
|
2006-05-02 19:56:43 +02:00
|
|
|
case SQLCOM_SHOW_ENGINE_STATUS:
|
2002-07-08 18:34:49 +02:00
|
|
|
{
|
2002-08-08 02:12:02 +02:00
|
|
|
if (check_global_access(thd, SUPER_ACL))
|
2006-05-02 19:56:43 +02:00
|
|
|
goto error;
|
|
|
|
res = ha_show_status(thd, lex->create_info.db_type, HA_ENGINE_STATUS);
|
2002-07-08 18:34:49 +02:00
|
|
|
break;
|
|
|
|
}
|
2006-05-02 19:56:43 +02:00
|
|
|
case SQLCOM_SHOW_ENGINE_MUTEX:
|
2004-12-24 12:13:32 +01:00
|
|
|
{
|
|
|
|
if (check_global_access(thd, SUPER_ACL))
|
2004-12-24 13:31:21 +01:00
|
|
|
goto error;
|
2006-05-02 19:56:43 +02:00
|
|
|
res = ha_show_status(thd, lex->create_info.db_type, HA_ENGINE_MUTEX);
|
2004-12-24 12:13:32 +01:00
|
|
|
break;
|
|
|
|
}
|
2003-09-08 12:08:53 +02:00
|
|
|
#ifdef HAVE_REPLICATION
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_LOAD_MASTER_TABLE:
|
2002-01-20 03:16:52 +01:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (!first_table->db)
|
|
|
|
first_table->db= thd->db;
|
|
|
|
if (check_access(thd, CREATE_ACL, first_table->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&first_table->grant.privilege, 0, 0,
|
|
|
|
test(first_table->schema_table)))
|
2000-07-31 21:29:14 +02:00
|
|
|
goto error; /* purecov: inspected */
|
|
|
|
if (grant_option)
|
|
|
|
{
|
|
|
|
/* Check that the first table has CREATE privilege */
|
2004-07-16 00:15:55 +02:00
|
|
|
if (check_grant(thd, CREATE_ACL, all_tables, 0, 1, 0))
|
2001-08-14 19:33:49 +02:00
|
|
|
goto error;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2005-01-06 12:00:13 +01:00
|
|
|
if (strlen(first_table->table_name) > NAME_LEN)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2005-01-06 12:00:13 +01:00
|
|
|
my_error(ER_WRONG_TABLE_NAME, MYF(0), first_table->table_name);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_lock(&LOCK_active_mi);
|
2003-09-11 23:17:28 +02:00
|
|
|
/*
|
|
|
|
fetch_master_table will send the error to the client on failure.
|
|
|
|
Give error if the table already exists.
|
|
|
|
*/
|
2005-01-06 12:00:13 +01:00
|
|
|
if (!fetch_master_table(thd, first_table->db, first_table->table_name,
|
2003-09-11 23:17:28 +02:00
|
|
|
active_mi, 0, 0))
|
2002-01-20 03:16:52 +01:00
|
|
|
{
|
2002-10-02 12:33:08 +02:00
|
|
|
send_ok(thd);
|
2002-01-20 03:16:52 +01:00
|
|
|
}
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_unlock(&LOCK_active_mi);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2002-01-20 03:16:52 +01:00
|
|
|
}
|
2003-09-08 12:08:53 +02:00
|
|
|
#endif /* HAVE_REPLICATION */
|
2002-12-16 14:33:29 +01:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_CREATE_TABLE:
|
2002-06-12 14:04:18 +02:00
|
|
|
{
|
2005-05-27 04:17:33 +02:00
|
|
|
/* If CREATE TABLE of non-temporary table, do implicit commit */
|
2005-06-01 15:52:32 +02:00
|
|
|
if (!(lex->create_info.options & HA_LEX_CREATE_TMP_TABLE))
|
|
|
|
{
|
|
|
|
if (end_active_trans(thd))
|
|
|
|
{
|
|
|
|
res= -1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
bool link_to_local;
|
|
|
|
// Skip first table, which is the table we are creating
|
|
|
|
TABLE_LIST *create_table= lex->unlink_first_table(&link_to_local);
|
|
|
|
TABLE_LIST *select_tables= lex->query_tables;
|
2003-12-19 15:25:50 +01:00
|
|
|
|
2004-07-16 00:15:55 +02:00
|
|
|
if ((res= create_table_precheck(thd, select_tables, create_table)))
|
2005-06-06 23:31:53 +02:00
|
|
|
goto end_with_restore_list;
|
2004-04-10 00:14:32 +02:00
|
|
|
|
2001-11-03 20:33:11 +01:00
|
|
|
#ifndef HAVE_READLINK
|
|
|
|
lex->create_info.data_file_name=lex->create_info.index_file_name=0;
|
|
|
|
#else
|
2001-06-01 03:27:59 +02:00
|
|
|
/* Fix names if symlinked tables */
|
2001-12-05 12:03:00 +01:00
|
|
|
if (append_file_to_dir(thd, &lex->create_info.data_file_name,
|
2005-01-06 12:00:13 +01:00
|
|
|
create_table->table_name) ||
|
2004-07-16 00:15:55 +02:00
|
|
|
append_file_to_dir(thd, &lex->create_info.index_file_name,
|
2005-01-06 12:00:13 +01:00
|
|
|
create_table->table_name))
|
2005-06-06 23:31:53 +02:00
|
|
|
goto end_with_restore_list;
|
2001-11-03 20:33:11 +01:00
|
|
|
#endif
|
2003-11-18 12:47:27 +01:00
|
|
|
/*
|
2004-10-07 00:45:06 +02:00
|
|
|
If we are using SET CHARSET without DEFAULT, add an implicit
|
2003-11-18 12:47:27 +01:00
|
|
|
DEFAULT to not confuse old users. (This may change).
|
|
|
|
*/
|
|
|
|
if ((lex->create_info.used_fields &
|
|
|
|
(HA_CREATE_USED_DEFAULT_CHARSET | HA_CREATE_USED_CHARSET)) ==
|
|
|
|
HA_CREATE_USED_CHARSET)
|
|
|
|
{
|
|
|
|
lex->create_info.used_fields&= ~HA_CREATE_USED_CHARSET;
|
|
|
|
lex->create_info.used_fields|= HA_CREATE_USED_DEFAULT_CHARSET;
|
|
|
|
lex->create_info.default_table_charset= lex->create_info.table_charset;
|
|
|
|
lex->create_info.table_charset= 0;
|
|
|
|
}
|
2005-05-31 11:08:14 +02:00
|
|
|
/*
|
|
|
|
The create-select command will open and read-lock the select table
|
|
|
|
and then create, open and write-lock the new table. If a global
|
|
|
|
read lock steps in, we get a deadlock. The write lock waits for
|
|
|
|
the global read lock, while the global read lock waits for the
|
|
|
|
select table to be closed. So we wait until the global readlock is
|
|
|
|
gone before starting both steps. Note that
|
|
|
|
wait_if_global_read_lock() sets a protection against a new global
|
|
|
|
read lock when it succeeds. This needs to be released by
|
|
|
|
start_waiting_global_read_lock(). We protect the normal CREATE
|
|
|
|
TABLE in the same way. That way we avoid that a new table is
|
|
|
|
created during a gobal read lock.
|
|
|
|
*/
|
|
|
|
if (wait_if_global_read_lock(thd, 0, 1))
|
|
|
|
{
|
2005-06-06 23:31:53 +02:00
|
|
|
res= 1;
|
|
|
|
goto end_with_restore_list;
|
2005-05-31 11:08:14 +02:00
|
|
|
}
|
2006-05-04 19:18:55 +02:00
|
|
|
#ifdef WITH_PARTITION_STORAGE_ENGINE
|
|
|
|
{
|
|
|
|
partition_info *part_info= thd->lex->part_info;
|
|
|
|
if (part_info && !(part_info= thd->lex->part_info->get_clone()))
|
|
|
|
{
|
|
|
|
res= -1;
|
|
|
|
goto end_with_restore_list;
|
|
|
|
}
|
|
|
|
thd->work_part_info= part_info;
|
|
|
|
}
|
|
|
|
#endif
|
2001-06-07 13:10:58 +02:00
|
|
|
if (select_lex->item_list.elements) // With select
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
select_result *result;
|
2004-09-04 14:59:49 +02:00
|
|
|
|
2003-03-12 06:51:08 +01:00
|
|
|
select_lex->options|= SELECT_NO_UNLOCK;
|
2005-05-30 18:54:37 +02:00
|
|
|
unit->set_limit(select_lex);
|
2001-08-14 19:33:49 +02:00
|
|
|
|
2004-07-16 00:15:55 +02:00
|
|
|
if (!(res= open_and_lock_tables(thd, select_tables)))
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2004-09-08 09:18:04 +02:00
|
|
|
/*
|
|
|
|
Is table which we are changing used somewhere in other parts
|
|
|
|
of query
|
|
|
|
*/
|
2005-08-02 21:54:49 +02:00
|
|
|
if (!(lex->create_info.options & HA_LEX_CREATE_TMP_TABLE))
|
2004-09-08 09:18:04 +02:00
|
|
|
{
|
2005-08-02 21:54:49 +02:00
|
|
|
TABLE_LIST *duplicate;
|
2005-12-20 16:35:05 +01:00
|
|
|
if ((duplicate= unique_table(thd, create_table, select_tables)))
|
2005-08-02 21:54:49 +02:00
|
|
|
{
|
|
|
|
update_non_unique_table_error(create_table, "CREATE", duplicate);
|
|
|
|
res= 1;
|
|
|
|
goto end_with_restart_wait;
|
|
|
|
}
|
2004-09-08 09:18:04 +02:00
|
|
|
}
|
2004-09-10 01:22:44 +02:00
|
|
|
/* If we create merge table, we have to test tables in merge, too */
|
|
|
|
if (lex->create_info.used_fields & HA_CREATE_USED_UNION)
|
|
|
|
{
|
|
|
|
TABLE_LIST *tab;
|
|
|
|
for (tab= (TABLE_LIST*) lex->create_info.merge_list.first;
|
|
|
|
tab;
|
|
|
|
tab= tab->next_local)
|
|
|
|
{
|
2005-08-02 21:54:49 +02:00
|
|
|
TABLE_LIST *duplicate;
|
2005-12-20 16:35:05 +01:00
|
|
|
if ((duplicate= unique_table(thd, tab, select_tables)))
|
2004-09-10 01:22:44 +02:00
|
|
|
{
|
2005-08-02 21:54:49 +02:00
|
|
|
update_non_unique_table_error(tab, "CREATE", duplicate);
|
2005-06-06 23:31:53 +02:00
|
|
|
res= 1;
|
|
|
|
goto end_with_restart_wait;
|
2004-09-10 01:22:44 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2004-09-08 09:18:04 +02:00
|
|
|
|
2004-07-16 00:15:55 +02:00
|
|
|
if ((result= new select_create(create_table,
|
|
|
|
&lex->create_info,
|
|
|
|
lex->create_list,
|
|
|
|
lex->key_list,
|
|
|
|
select_lex->item_list,
|
2005-01-03 22:04:52 +01:00
|
|
|
lex->duplicates,
|
|
|
|
lex->ignore)))
|
2004-06-13 21:39:09 +02:00
|
|
|
{
|
|
|
|
/*
|
|
|
|
CREATE from SELECT give its SELECT_LEX for SELECT,
|
|
|
|
and item_list belong to SELECT
|
|
|
|
*/
|
2005-01-03 20:04:33 +01:00
|
|
|
res= handle_select(thd, lex, result, 0);
|
2004-09-17 02:08:23 +02:00
|
|
|
delete result;
|
2004-06-13 21:39:09 +02:00
|
|
|
}
|
2004-07-07 10:29:39 +02:00
|
|
|
/* reset for PS */
|
2004-04-07 23:16:17 +02:00
|
|
|
lex->create_list.empty();
|
|
|
|
lex->key_list.empty();
|
2001-08-14 19:33:49 +02:00
|
|
|
}
|
|
|
|
}
|
2004-07-07 10:29:39 +02:00
|
|
|
else
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
/* So that CREATE TEMPORARY TABLE gets to binlog at commit/rollback */
|
|
|
|
if (lex->create_info.options & HA_LEX_CREATE_TMP_TABLE)
|
|
|
|
thd->options|= OPTION_KEEP_LOG;
|
2004-07-07 10:29:39 +02:00
|
|
|
/* regular create */
|
2006-05-02 19:56:43 +02:00
|
|
|
if (lex->like_name)
|
2003-12-19 15:25:50 +01:00
|
|
|
res= mysql_create_like_table(thd, create_table, &lex->create_info,
|
2006-05-02 19:56:43 +02:00
|
|
|
lex->like_name);
|
2002-12-28 10:38:29 +01:00
|
|
|
else
|
2003-08-11 15:18:34 +02:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
res= mysql_create_table(thd, create_table->db,
|
2005-01-06 12:00:13 +01:00
|
|
|
create_table->table_name, &lex->create_info,
|
2004-07-16 00:15:55 +02:00
|
|
|
lex->create_list,
|
2004-09-06 14:14:10 +02:00
|
|
|
lex->key_list, 0, 0);
|
2003-08-11 15:18:34 +02:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
if (!res)
|
2002-10-02 12:33:08 +02:00
|
|
|
send_ok(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2004-04-05 23:10:43 +02:00
|
|
|
|
2005-06-06 23:31:53 +02:00
|
|
|
end_with_restart_wait:
|
2005-06-01 13:22:17 +02:00
|
|
|
/*
|
|
|
|
Release the protection against the global read lock and wake
|
|
|
|
everyone, who might want to set a global read lock.
|
|
|
|
*/
|
|
|
|
start_waiting_global_read_lock(thd);
|
|
|
|
|
2004-07-07 10:29:39 +02:00
|
|
|
/* put tables back for PS rexecuting */
|
2005-06-06 23:31:53 +02:00
|
|
|
end_with_restore_list:
|
2004-07-16 00:15:55 +02:00
|
|
|
lex->link_first_table_back(create_table, link_to_local);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2002-06-12 14:04:18 +02:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_CREATE_INDEX:
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_one_table_access(thd, INDEX_ACL, all_tables))
|
2000-07-31 21:29:14 +02:00
|
|
|
goto error; /* purecov: inspected */
|
2005-06-16 21:05:38 +02:00
|
|
|
thd->enable_slow_log= opt_log_slow_admin_statements;
|
2000-12-07 13:08:48 +01:00
|
|
|
if (end_active_trans(thd))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2006-05-05 19:08:40 +02:00
|
|
|
res= mysql_create_index(thd, first_table, lex->key_list);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
|
2003-09-08 12:08:53 +02:00
|
|
|
#ifdef HAVE_REPLICATION
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_SLAVE_START:
|
2002-01-20 03:16:52 +01:00
|
|
|
{
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_lock(&LOCK_active_mi);
|
2002-01-20 03:16:52 +01:00
|
|
|
start_slave(thd,active_mi,1 /* net report*/);
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_unlock(&LOCK_active_mi);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2002-01-20 03:16:52 +01:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_SLAVE_STOP:
|
2002-12-13 15:36:08 +01:00
|
|
|
/*
|
|
|
|
If the client thread has locked tables, a deadlock is possible.
|
|
|
|
Assume that
|
|
|
|
- the client thread does LOCK TABLE t READ.
|
|
|
|
- then the master updates t.
|
|
|
|
- then the SQL slave thread wants to update t,
|
2002-12-13 15:01:51 +01:00
|
|
|
so it waits for the client thread because t is locked by it.
|
2002-12-13 15:36:08 +01:00
|
|
|
- then the client thread does SLAVE STOP.
|
2002-12-13 15:01:51 +01:00
|
|
|
SLAVE STOP waits for the SQL slave thread to terminate its
|
|
|
|
update t, which waits for the client thread because t is locked by it.
|
2002-12-13 15:36:08 +01:00
|
|
|
To prevent that, refuse SLAVE STOP if the
|
|
|
|
client thread has locked tables
|
|
|
|
*/
|
2005-10-08 14:46:04 +02:00
|
|
|
if (thd->locked_tables || thd->active_transaction() || thd->global_read_lock)
|
2002-12-13 15:36:08 +01:00
|
|
|
{
|
2005-08-12 12:54:42 +02:00
|
|
|
my_message(ER_LOCK_OR_ACTIVE_TRANSACTION,
|
|
|
|
ER(ER_LOCK_OR_ACTIVE_TRANSACTION), MYF(0));
|
2003-01-15 15:39:36 +01:00
|
|
|
goto error;
|
2002-12-13 15:36:08 +01:00
|
|
|
}
|
2002-01-20 03:16:52 +01:00
|
|
|
{
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_lock(&LOCK_active_mi);
|
2002-01-20 03:16:52 +01:00
|
|
|
stop_slave(thd,active_mi,1/* net report*/);
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_unlock(&LOCK_active_mi);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2002-01-20 03:16:52 +01:00
|
|
|
}
|
2003-09-08 12:08:53 +02:00
|
|
|
#endif /* HAVE_REPLICATION */
|
2002-12-16 14:33:29 +01:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_ALTER_TABLE:
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
2000-07-31 21:29:14 +02:00
|
|
|
#if defined(DONT_ALLOW_SHOW_COMMANDS)
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_NOT_ALLOWED_COMMAND, ER(ER_NOT_ALLOWED_COMMAND),
|
|
|
|
MYF(0)); /* purecov: inspected */
|
2003-01-15 15:39:36 +01:00
|
|
|
goto error;
|
2000-07-31 21:29:14 +02:00
|
|
|
#else
|
|
|
|
{
|
2002-06-12 14:04:18 +02:00
|
|
|
ulong priv=0;
|
2006-05-02 19:56:43 +02:00
|
|
|
ulong priv_needed= ALTER_ACL;
|
|
|
|
/* We also require DROP priv for ALTER TABLE ... DROP PARTITION */
|
|
|
|
if (lex->alter_info.flags & ALTER_DROP_PARTITION)
|
|
|
|
priv_needed|= DROP_ACL;
|
|
|
|
|
2002-11-12 16:41:31 +01:00
|
|
|
if (lex->name && (!lex->name[0] || strlen(lex->name) > NAME_LEN))
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_WRONG_TABLE_NAME, MYF(0), lex->name);
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2001-06-07 13:10:58 +02:00
|
|
|
if (!select_lex->db)
|
2005-08-29 16:54:33 +02:00
|
|
|
{
|
|
|
|
/*
|
|
|
|
In the case of ALTER TABLE ... RENAME we should supply the
|
|
|
|
default database if the new name is not explicitly qualified
|
|
|
|
by a database. (Bug #11493)
|
|
|
|
*/
|
|
|
|
if (lex->alter_info.flags & ALTER_RENAME)
|
|
|
|
{
|
|
|
|
if (! thd->db)
|
|
|
|
{
|
2005-08-29 18:41:39 +02:00
|
|
|
my_message(ER_NO_DB_ERROR, ER(ER_NO_DB_ERROR), MYF(0));
|
2005-08-29 16:54:33 +02:00
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
select_lex->db= thd->db;
|
|
|
|
}
|
|
|
|
else
|
2005-08-29 18:18:30 +02:00
|
|
|
select_lex->db= first_table->db;
|
2005-08-29 16:54:33 +02:00
|
|
|
}
|
2006-05-02 19:56:43 +02:00
|
|
|
if (check_access(thd, priv_needed, first_table->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&first_table->grant.privilege, 0, 0,
|
|
|
|
test(first_table->schema_table)) ||
|
|
|
|
check_access(thd,INSERT_ACL | CREATE_ACL,select_lex->db,&priv,0,0,
|
|
|
|
is_schema_db(select_lex->db))||
|
2004-07-16 00:15:55 +02:00
|
|
|
check_merge_table_access(thd, first_table->db,
|
2000-09-14 01:39:07 +02:00
|
|
|
(TABLE_LIST *)
|
|
|
|
lex->create_info.merge_list.first))
|
|
|
|
goto error; /* purecov: inspected */
|
2000-07-31 21:29:14 +02:00
|
|
|
if (grant_option)
|
|
|
|
{
|
2006-05-02 19:56:43 +02:00
|
|
|
if (check_grant(thd, priv_needed, all_tables, 0, UINT_MAX, 0))
|
2000-07-31 21:29:14 +02:00
|
|
|
goto error;
|
|
|
|
if (lex->name && !test_all_bits(priv,INSERT_ACL | CREATE_ACL))
|
|
|
|
{ // Rename of table
|
|
|
|
TABLE_LIST tmp_table;
|
|
|
|
bzero((char*) &tmp_table,sizeof(tmp_table));
|
2005-01-06 12:00:13 +01:00
|
|
|
tmp_table.table_name=lex->name;
|
2001-06-07 13:10:58 +02:00
|
|
|
tmp_table.db=select_lex->db;
|
2000-07-31 21:29:14 +02:00
|
|
|
tmp_table.grant.privilege=priv;
|
2004-04-10 00:14:32 +02:00
|
|
|
if (check_grant(thd, INSERT_ACL | CREATE_ACL, &tmp_table, 0,
|
|
|
|
UINT_MAX, 0))
|
2000-07-31 21:29:14 +02:00
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
}
|
2001-06-01 03:27:59 +02:00
|
|
|
/* Don't yet allow changing of symlinks with ALTER TABLE */
|
|
|
|
lex->create_info.data_file_name=lex->create_info.index_file_name=0;
|
2000-07-31 21:29:14 +02:00
|
|
|
/* ALTER TABLE ends previous transaction */
|
2000-11-13 22:55:10 +01:00
|
|
|
if (end_active_trans(thd))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2006-05-05 19:08:40 +02:00
|
|
|
|
|
|
|
thd->enable_slow_log= opt_log_slow_admin_statements;
|
|
|
|
res= mysql_alter_table(thd, select_lex->db, lex->name,
|
|
|
|
&lex->create_info,
|
|
|
|
first_table, lex->create_list,
|
|
|
|
lex->key_list,
|
|
|
|
select_lex->order_list.elements,
|
|
|
|
(ORDER *) select_lex->order_list.first,
|
|
|
|
lex->duplicates, lex->ignore, &lex->alter_info,
|
|
|
|
1);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-09-26 12:33:13 +02:00
|
|
|
#endif /*DONT_ALLOW_SHOW_COMMANDS*/
|
2000-08-21 02:07:54 +02:00
|
|
|
case SQLCOM_RENAME_TABLE:
|
2000-08-21 23:18:32 +02:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
2000-08-21 23:18:32 +02:00
|
|
|
TABLE_LIST *table;
|
2004-07-16 00:15:55 +02:00
|
|
|
if (check_db_used(thd, all_tables))
|
2000-08-21 02:07:54 +02:00
|
|
|
goto error;
|
2004-07-16 00:15:55 +02:00
|
|
|
for (table= first_table; table; table= table->next_local->next_local)
|
2000-08-21 23:18:32 +02:00
|
|
|
{
|
2000-08-29 18:38:32 +02:00
|
|
|
if (check_access(thd, ALTER_ACL | DROP_ACL, table->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&table->grant.privilege,0,0, test(table->schema_table)) ||
|
2004-07-16 00:15:55 +02:00
|
|
|
check_access(thd, INSERT_ACL | CREATE_ACL, table->next_local->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&table->next_local->grant.privilege, 0, 0,
|
|
|
|
test(table->next_local->schema_table)))
|
2000-08-21 23:18:32 +02:00
|
|
|
goto error;
|
|
|
|
if (grant_option)
|
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
TABLE_LIST old_list, new_list;
|
2004-03-23 18:24:35 +01:00
|
|
|
/*
|
|
|
|
we do not need initialize old_list and new_list because we will
|
|
|
|
come table[0] and table->next[0] there
|
|
|
|
*/
|
2004-07-16 00:15:55 +02:00
|
|
|
old_list= table[0];
|
|
|
|
new_list= table->next_local[0];
|
|
|
|
if (check_grant(thd, ALTER_ACL, &old_list, 0, 1, 0) ||
|
|
|
|
(!test_all_bits(table->next_local->grant.privilege,
|
2001-08-14 19:33:49 +02:00
|
|
|
INSERT_ACL | CREATE_ACL) &&
|
2004-07-16 00:15:55 +02:00
|
|
|
check_grant(thd, INSERT_ACL | CREATE_ACL, &new_list, 0, 1, 0)))
|
2000-08-21 23:18:32 +02:00
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
}
|
2004-07-16 00:15:55 +02:00
|
|
|
query_cache_invalidate3(thd, first_table, 0);
|
2006-05-02 19:56:43 +02:00
|
|
|
if (end_active_trans(thd) || mysql_rename_tables(thd, first_table, 0))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2000-08-21 02:07:54 +02:00
|
|
|
break;
|
2000-08-21 23:18:32 +02:00
|
|
|
}
|
2002-12-16 14:33:29 +01:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2000-10-27 06:11:55 +02:00
|
|
|
case SQLCOM_SHOW_BINLOGS:
|
|
|
|
#ifdef DONT_ALLOW_SHOW_COMMANDS
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_NOT_ALLOWED_COMMAND, ER(ER_NOT_ALLOWED_COMMAND),
|
|
|
|
MYF(0)); /* purecov: inspected */
|
2003-01-15 15:39:36 +01:00
|
|
|
goto error;
|
2000-10-27 06:11:55 +02:00
|
|
|
#else
|
|
|
|
{
|
2002-06-12 14:04:18 +02:00
|
|
|
if (check_global_access(thd, SUPER_ACL))
|
2000-10-27 06:11:55 +02:00
|
|
|
goto error;
|
|
|
|
res = show_binlogs(thd);
|
|
|
|
break;
|
|
|
|
}
|
2002-11-30 18:58:53 +01:00
|
|
|
#endif
|
2002-12-16 14:33:29 +01:00
|
|
|
#endif /* EMBEDDED_LIBRARY */
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_SHOW_CREATE:
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
2000-08-21 02:07:54 +02:00
|
|
|
#ifdef DONT_ALLOW_SHOW_COMMANDS
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_NOT_ALLOWED_COMMAND, ER(ER_NOT_ALLOWED_COMMAND),
|
|
|
|
MYF(0)); /* purecov: inspected */
|
2003-01-15 15:39:36 +01:00
|
|
|
goto error;
|
2000-08-21 02:07:54 +02:00
|
|
|
#else
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2005-04-25 02:00:35 +02:00
|
|
|
/* Ignore temporary tables if this is "SHOW CREATE VIEW" */
|
|
|
|
if (lex->only_view)
|
|
|
|
first_table->skip_temporary= 1;
|
|
|
|
|
2004-07-16 00:15:55 +02:00
|
|
|
if (check_db_used(thd, all_tables) ||
|
|
|
|
check_access(thd, SELECT_ACL | EXTRA_ACL, first_table->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&first_table->grant.privilege, 0, 0,
|
|
|
|
test(first_table->schema_table)))
|
2000-08-21 02:07:54 +02:00
|
|
|
goto error;
|
2004-12-22 12:54:39 +01:00
|
|
|
if (grant_option && check_grant(thd, SELECT_ACL, all_tables, 2, UINT_MAX, 0))
|
2004-12-07 20:57:54 +01:00
|
|
|
goto error;
|
2004-12-22 12:54:39 +01:00
|
|
|
res= mysqld_show_create(thd, first_table);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2000-08-21 02:07:54 +02:00
|
|
|
#endif
|
2003-08-21 16:15:06 +02:00
|
|
|
case SQLCOM_CHECKSUM:
|
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_db_used(thd, all_tables) ||
|
|
|
|
check_table_access(thd, SELECT_ACL | EXTRA_ACL, all_tables, 0))
|
2003-08-21 16:15:06 +02:00
|
|
|
goto error; /* purecov: inspected */
|
2004-07-16 00:15:55 +02:00
|
|
|
res = mysql_checksum_table(thd, first_table, &lex->check_opt);
|
2003-08-21 16:15:06 +02:00
|
|
|
break;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_REPAIR:
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_db_used(thd, all_tables) ||
|
|
|
|
check_table_access(thd, SELECT_ACL | INSERT_ACL, all_tables, 0))
|
2001-08-14 19:33:49 +02:00
|
|
|
goto error; /* purecov: inspected */
|
2005-06-16 21:05:38 +02:00
|
|
|
thd->enable_slow_log= opt_log_slow_admin_statements;
|
2004-07-16 00:15:55 +02:00
|
|
|
res= mysql_repair_table(thd, first_table, &lex->check_opt);
|
2003-05-15 18:35:39 +02:00
|
|
|
/* ! we write after unlocking the table */
|
|
|
|
if (!res && !lex->no_write_to_binlog)
|
|
|
|
{
|
|
|
|
if (mysql_bin_log.is_open())
|
|
|
|
{
|
2004-11-03 18:23:58 +01:00
|
|
|
thd->clear_error(); // No binlog error generated
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::STMT_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, 0, FALSE);
|
2003-05-15 18:35:39 +02:00
|
|
|
}
|
|
|
|
}
|
2005-12-07 16:15:02 +01:00
|
|
|
select_lex->table_list.first= (byte*) first_table;
|
2005-12-03 15:02:09 +01:00
|
|
|
lex->query_tables=all_tables;
|
2001-08-14 19:33:49 +02:00
|
|
|
break;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_CHECK:
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_db_used(thd, all_tables) ||
|
|
|
|
check_table_access(thd, SELECT_ACL | EXTRA_ACL , all_tables, 0))
|
2001-08-14 19:33:49 +02:00
|
|
|
goto error; /* purecov: inspected */
|
2005-06-16 21:05:38 +02:00
|
|
|
thd->enable_slow_log= opt_log_slow_admin_statements;
|
2004-07-16 00:15:55 +02:00
|
|
|
res = mysql_check_table(thd, first_table, &lex->check_opt);
|
2005-12-07 16:15:02 +01:00
|
|
|
select_lex->table_list.first= (byte*) first_table;
|
2005-12-03 15:02:09 +01:00
|
|
|
lex->query_tables=all_tables;
|
2001-08-14 19:33:49 +02:00
|
|
|
break;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_ANALYZE:
|
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_db_used(thd, all_tables) ||
|
|
|
|
check_table_access(thd, SELECT_ACL | INSERT_ACL, all_tables, 0))
|
2000-07-31 21:29:14 +02:00
|
|
|
goto error; /* purecov: inspected */
|
2005-06-16 21:05:38 +02:00
|
|
|
thd->enable_slow_log= opt_log_slow_admin_statements;
|
2006-05-05 19:08:40 +02:00
|
|
|
res= mysql_analyze_table(thd, first_table, &lex->check_opt);
|
2003-05-15 18:35:39 +02:00
|
|
|
/* ! we write after unlocking the table */
|
|
|
|
if (!res && !lex->no_write_to_binlog)
|
|
|
|
{
|
|
|
|
if (mysql_bin_log.is_open())
|
|
|
|
{
|
2004-11-03 18:23:58 +01:00
|
|
|
thd->clear_error(); // No binlog error generated
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::STMT_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, 0, FALSE);
|
2003-05-15 18:35:39 +02:00
|
|
|
}
|
|
|
|
}
|
2005-12-07 16:15:02 +01:00
|
|
|
select_lex->table_list.first= (byte*) first_table;
|
2005-12-03 15:02:09 +01:00
|
|
|
lex->query_tables=all_tables;
|
2000-08-21 02:07:54 +02:00
|
|
|
break;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2000-09-12 02:02:33 +02:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_OPTIMIZE:
|
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_db_used(thd, all_tables) ||
|
|
|
|
check_table_access(thd, SELECT_ACL | INSERT_ACL, all_tables, 0))
|
2000-07-31 21:29:14 +02:00
|
|
|
goto error; /* purecov: inspected */
|
2005-06-16 21:05:38 +02:00
|
|
|
thd->enable_slow_log= opt_log_slow_admin_statements;
|
2004-06-10 16:41:24 +02:00
|
|
|
res= (specialflag & (SPECIAL_SAFE_MODE | SPECIAL_NO_NEW_FUNC)) ?
|
2004-07-16 00:15:55 +02:00
|
|
|
mysql_recreate_table(thd, first_table, 1) :
|
|
|
|
mysql_optimize_table(thd, first_table, &lex->check_opt);
|
2003-05-15 18:35:39 +02:00
|
|
|
/* ! we write after unlocking the table */
|
|
|
|
if (!res && !lex->no_write_to_binlog)
|
|
|
|
{
|
|
|
|
if (mysql_bin_log.is_open())
|
|
|
|
{
|
2004-11-03 18:23:58 +01:00
|
|
|
thd->clear_error(); // No binlog error generated
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::STMT_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, 0, FALSE);
|
2003-05-15 18:35:39 +02:00
|
|
|
}
|
|
|
|
}
|
2005-12-07 16:15:02 +01:00
|
|
|
select_lex->table_list.first= (byte*) first_table;
|
2005-12-03 15:02:09 +01:00
|
|
|
lex->query_tables=all_tables;
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_UPDATE:
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (update_precheck(thd, all_tables))
|
2004-04-10 00:14:32 +02:00
|
|
|
break;
|
2005-06-07 12:11:36 +02:00
|
|
|
DBUG_ASSERT(select_lex->offset_limit == 0);
|
|
|
|
unit->set_limit(select_lex);
|
2004-11-25 01:23:13 +01:00
|
|
|
res= (result= mysql_update(thd, all_tables,
|
|
|
|
select_lex->item_list,
|
|
|
|
lex->value_list,
|
|
|
|
select_lex->where,
|
|
|
|
select_lex->order_list.elements,
|
|
|
|
(ORDER *) select_lex->order_list.first,
|
2005-06-07 12:11:36 +02:00
|
|
|
unit->select_limit_cnt,
|
2005-01-03 22:04:52 +01:00
|
|
|
lex->duplicates, lex->ignore));
|
2004-11-25 08:28:32 +01:00
|
|
|
/* mysql_update return 2 if we need to switch to multi-update */
|
2004-11-25 01:23:13 +01:00
|
|
|
if (result != 2)
|
2004-09-15 22:42:56 +02:00
|
|
|
break;
|
2002-11-29 13:17:54 +01:00
|
|
|
case SQLCOM_UPDATE_MULTI:
|
2005-10-12 10:18:46 +02:00
|
|
|
{
|
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
/* if we switched from normal update, rights are checked */
|
|
|
|
if (result != 2)
|
2004-09-15 22:42:56 +02:00
|
|
|
{
|
2005-10-12 10:18:46 +02:00
|
|
|
if ((res= multi_update_precheck(thd, all_tables)))
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
res= 0;
|
2004-11-25 01:23:13 +01:00
|
|
|
|
2006-01-23 17:03:09 +01:00
|
|
|
res= mysql_multi_update_prepare(thd);
|
2005-10-10 15:10:14 +02:00
|
|
|
|
2005-10-10 21:42:14 +02:00
|
|
|
#ifdef HAVE_REPLICATION
|
2005-10-12 10:18:46 +02:00
|
|
|
/* Check slave filtering rules */
|
2005-10-17 10:52:34 +02:00
|
|
|
if (unlikely(thd->slave_thread))
|
2005-10-12 10:18:46 +02:00
|
|
|
{
|
2005-10-17 10:52:34 +02:00
|
|
|
if (all_tables_not_ok(thd, all_tables))
|
|
|
|
{
|
2006-01-23 17:03:09 +01:00
|
|
|
if (res!= 0)
|
|
|
|
{
|
|
|
|
res= 0; /* don't care of prev failure */
|
|
|
|
thd->clear_error(); /* filters are of highest prior */
|
|
|
|
}
|
2005-10-17 10:52:34 +02:00
|
|
|
/* we warn the slave SQL thread */
|
|
|
|
my_error(ER_SLAVE_IGNORED_TABLE, MYF(0));
|
|
|
|
break;
|
|
|
|
}
|
2006-01-23 17:03:09 +01:00
|
|
|
if (res)
|
|
|
|
break;
|
2005-10-12 10:18:46 +02:00
|
|
|
}
|
2005-10-17 10:52:34 +02:00
|
|
|
else
|
|
|
|
{
|
2006-01-23 17:03:09 +01:00
|
|
|
#endif /* HAVE_REPLICATION */
|
|
|
|
if (res)
|
|
|
|
break;
|
|
|
|
if (opt_readonly &&
|
|
|
|
!(thd->security_ctx->master_access & SUPER_ACL) &&
|
|
|
|
some_non_temp_table_to_be_updated(thd, all_tables))
|
|
|
|
{
|
|
|
|
my_error(ER_OPTION_PREVENTS_STATEMENT, MYF(0), "--read-only");
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#ifdef HAVE_REPLICATION
|
|
|
|
} /* unlikely */
|
|
|
|
#endif
|
2005-10-10 15:10:14 +02:00
|
|
|
|
2005-10-12 10:18:46 +02:00
|
|
|
res= mysql_multi_update(thd, all_tables,
|
|
|
|
&select_lex->item_list,
|
|
|
|
&lex->value_list,
|
|
|
|
select_lex->where,
|
|
|
|
select_lex->options,
|
|
|
|
lex->duplicates, lex->ignore, unit, select_lex);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2004-02-11 00:06:46 +01:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_REPLACE:
|
2002-11-28 17:25:41 +01:00
|
|
|
case SQLCOM_INSERT:
|
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
2004-10-29 18:26:52 +02:00
|
|
|
if ((res= insert_precheck(thd, all_tables)))
|
2004-04-10 00:14:32 +02:00
|
|
|
break;
|
2005-07-01 06:05:42 +02:00
|
|
|
/* Skip first table, which is the table we are inserting in */
|
|
|
|
select_lex->context.table_list= first_table->next_local;
|
2004-07-16 00:15:55 +02:00
|
|
|
res= mysql_insert(thd, all_tables, lex->field_list, lex->many_values,
|
2004-12-22 12:54:39 +01:00
|
|
|
lex->update_list, lex->value_list,
|
2004-12-31 11:04:35 +01:00
|
|
|
lex->duplicates, lex->ignore);
|
2004-07-16 00:15:55 +02:00
|
|
|
if (first_table->view && !first_table->contain_auto_increment)
|
|
|
|
thd->last_insert_id= 0; // do not show last insert ID if VIEW have not it
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2002-11-28 17:25:41 +01:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_REPLACE_SELECT:
|
|
|
|
case SQLCOM_INSERT_SELECT:
|
|
|
|
{
|
2005-07-03 13:17:52 +02:00
|
|
|
select_result *result;
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
2004-12-22 12:54:39 +01:00
|
|
|
if ((res= insert_precheck(thd, all_tables)))
|
2004-04-07 23:16:17 +02:00
|
|
|
break;
|
2003-12-17 16:35:34 +01:00
|
|
|
|
2003-12-14 05:39:52 +01:00
|
|
|
/* Fix lock for first table */
|
2004-07-16 00:15:55 +02:00
|
|
|
if (first_table->lock_type == TL_WRITE_DELAYED)
|
|
|
|
first_table->lock_type= TL_WRITE;
|
2003-12-14 05:39:52 +01:00
|
|
|
|
2003-03-12 06:51:08 +01:00
|
|
|
/* Don't unlock tables until command is written to binary log */
|
|
|
|
select_lex->options|= SELECT_NO_UNLOCK;
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2005-05-30 18:54:37 +02:00
|
|
|
unit->set_limit(select_lex);
|
2004-07-16 00:15:55 +02:00
|
|
|
if (!(res= open_and_lock_tables(thd, all_tables)))
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2004-09-15 22:42:56 +02:00
|
|
|
/* Skip first table, which is the table we are inserting in */
|
2005-08-12 16:57:19 +02:00
|
|
|
TABLE_LIST *second_table= first_table->next_local;
|
|
|
|
select_lex->table_list.first= (byte*) second_table;
|
2005-08-18 02:12:42 +02:00
|
|
|
select_lex->context.table_list=
|
|
|
|
select_lex->context.first_name_resolution_table= second_table;
|
2004-09-15 22:42:56 +02:00
|
|
|
res= mysql_insert_select_prepare(thd);
|
|
|
|
if (!res && (result= new select_insert(first_table, first_table->table,
|
|
|
|
&lex->field_list,
|
2005-07-01 06:05:42 +02:00
|
|
|
&lex->update_list,
|
|
|
|
&lex->value_list,
|
2005-01-03 22:04:52 +01:00
|
|
|
lex->duplicates, lex->ignore)))
|
2004-09-17 02:08:23 +02:00
|
|
|
{
|
2005-08-12 16:57:19 +02:00
|
|
|
/*
|
|
|
|
Skip first table, which is the table we are inserting in.
|
|
|
|
Below we set context.table_list again because the call above to
|
|
|
|
mysql_insert_select_prepare() calls resolve_in_table_list_only(),
|
|
|
|
which in turn resets context.table_list and
|
|
|
|
context.first_name_resolution_table.
|
|
|
|
*/
|
2005-08-18 02:12:42 +02:00
|
|
|
select_lex->context.table_list=
|
|
|
|
select_lex->context.first_name_resolution_table= second_table;
|
2005-01-03 20:04:33 +01:00
|
|
|
res= handle_select(thd, lex, result, OPTION_SETUP_TABLES_DONE);
|
2006-03-15 18:15:52 +01:00
|
|
|
/*
|
|
|
|
Invalidate the table in the query cache if something changed
|
|
|
|
after unlocking when changes become visible.
|
|
|
|
TODO: this is workaround. right way will be move invalidating in
|
|
|
|
the unlock procedure.
|
|
|
|
*/
|
|
|
|
if (first_table->lock_type == TL_WRITE_CONCURRENT_INSERT &&
|
|
|
|
thd->lock)
|
|
|
|
{
|
|
|
|
mysql_unlock_tables(thd, thd->lock);
|
|
|
|
query_cache_invalidate3(thd, first_table, 1);
|
|
|
|
thd->lock=0;
|
|
|
|
}
|
2004-09-17 02:08:23 +02:00
|
|
|
delete result;
|
|
|
|
}
|
2004-09-15 22:42:56 +02:00
|
|
|
/* revert changes for SP */
|
2005-05-26 21:01:55 +02:00
|
|
|
select_lex->table_list.first= (byte*) first_table;
|
2001-08-14 19:33:49 +02:00
|
|
|
}
|
2004-07-16 00:15:55 +02:00
|
|
|
|
|
|
|
if (first_table->view && !first_table->contain_auto_increment)
|
|
|
|
thd->last_insert_id= 0; // do not show last insert ID if VIEW have not it
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2000-11-13 22:55:10 +01:00
|
|
|
case SQLCOM_TRUNCATE:
|
2005-05-27 04:17:33 +02:00
|
|
|
if (end_active_trans(thd))
|
|
|
|
{
|
|
|
|
res= -1;
|
|
|
|
break;
|
|
|
|
}
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_one_table_access(thd, DELETE_ACL, all_tables))
|
2003-03-07 11:55:07 +01:00
|
|
|
goto error;
|
2001-09-02 12:47:00 +02:00
|
|
|
/*
|
|
|
|
Don't allow this within a transaction because we want to use
|
|
|
|
re-generate table
|
|
|
|
*/
|
2006-04-06 12:19:01 +02:00
|
|
|
if (thd->locked_tables || thd->active_transaction())
|
2001-09-02 12:47:00 +02:00
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_LOCK_OR_ACTIVE_TRANSACTION,
|
|
|
|
ER(ER_LOCK_OR_ACTIVE_TRANSACTION), MYF(0));
|
2001-09-02 12:47:00 +02:00
|
|
|
goto error;
|
|
|
|
}
|
2004-07-16 00:15:55 +02:00
|
|
|
|
2004-09-06 14:14:10 +02:00
|
|
|
res= mysql_truncate(thd, first_table, 0);
|
2001-09-02 12:47:00 +02:00
|
|
|
break;
|
2000-11-29 04:09:28 +01:00
|
|
|
case SQLCOM_DELETE:
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if ((res= delete_precheck(thd, all_tables)))
|
2004-04-10 00:14:32 +02:00
|
|
|
break;
|
2005-06-07 12:11:36 +02:00
|
|
|
DBUG_ASSERT(select_lex->offset_limit == 0);
|
|
|
|
unit->set_limit(select_lex);
|
2004-07-16 00:15:55 +02:00
|
|
|
res = mysql_delete(thd, all_tables, select_lex->where,
|
2003-12-19 19:16:26 +01:00
|
|
|
&select_lex->order_list,
|
2005-08-30 11:39:20 +02:00
|
|
|
unit->select_limit_cnt, select_lex->options,
|
|
|
|
FALSE);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2001-12-13 21:03:09 +01:00
|
|
|
case SQLCOM_DELETE_MULTI:
|
2001-06-15 04:03:15 +02:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
2004-04-07 23:16:17 +02:00
|
|
|
TABLE_LIST *aux_tables=
|
|
|
|
(TABLE_LIST *)thd->lex->auxilliary_table_list.first;
|
2001-06-15 04:03:15 +02:00
|
|
|
multi_delete *result;
|
2001-12-06 13:10:51 +01:00
|
|
|
|
2005-06-08 23:07:52 +02:00
|
|
|
if ((res= multi_delete_precheck(thd, all_tables)))
|
2004-04-07 23:16:17 +02:00
|
|
|
break;
|
2001-12-06 13:10:51 +01:00
|
|
|
|
2004-04-12 02:26:32 +02:00
|
|
|
/* condition will be TRUE on SP re-excuting */
|
2004-04-07 23:16:17 +02:00
|
|
|
if (select_lex->item_list.elements != 0)
|
|
|
|
select_lex->item_list.empty();
|
2002-12-06 20:11:27 +01:00
|
|
|
if (add_item_to_list(thd, new Item_null()))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2004-04-07 23:16:17 +02:00
|
|
|
|
2001-06-15 04:03:15 +02:00
|
|
|
thd->proc_info="init";
|
2004-07-16 00:15:55 +02:00
|
|
|
if ((res= open_and_lock_tables(thd, all_tables)))
|
|
|
|
break;
|
|
|
|
|
|
|
|
if ((res= mysql_multi_delete_prepare(thd)))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2004-02-01 14:30:32 +01:00
|
|
|
|
2005-08-09 22:23:56 +02:00
|
|
|
if (!thd->is_fatal_error && (result= new multi_delete(aux_tables,
|
2005-06-08 23:07:52 +02:00
|
|
|
lex->table_count)))
|
2001-06-15 04:03:15 +02:00
|
|
|
{
|
2003-01-25 01:25:52 +01:00
|
|
|
res= mysql_select(thd, &select_lex->ref_pointer_array,
|
|
|
|
select_lex->get_table_list(),
|
|
|
|
select_lex->with_wild,
|
2002-11-26 21:33:33 +01:00
|
|
|
select_lex->item_list,
|
2002-09-03 08:50:36 +02:00
|
|
|
select_lex->where,
|
2003-01-25 01:25:52 +01:00
|
|
|
0, (ORDER *)NULL, (ORDER *)NULL, (Item *)NULL,
|
2002-09-03 08:50:36 +02:00
|
|
|
(ORDER *)NULL,
|
|
|
|
select_lex->options | thd->options |
|
2005-01-03 20:04:33 +01:00
|
|
|
SELECT_NO_JOIN_CACHE | SELECT_NO_UNLOCK |
|
|
|
|
OPTION_SETUP_TABLES_DONE,
|
2003-11-23 01:01:15 +01:00
|
|
|
result, unit, select_lex);
|
2002-01-12 18:51:10 +01:00
|
|
|
delete result;
|
2001-06-15 04:03:15 +02:00
|
|
|
}
|
|
|
|
else
|
2005-02-21 13:47:57 +01:00
|
|
|
res= TRUE; // Error
|
2001-06-15 04:03:15 +02:00
|
|
|
break;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_DROP_TABLE:
|
2001-08-02 05:29:50 +02:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
2003-01-04 14:37:20 +01:00
|
|
|
if (!lex->drop_temporary)
|
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
if (check_table_access(thd, DROP_ACL, all_tables, 0))
|
2003-01-04 14:37:20 +01:00
|
|
|
goto error; /* purecov: inspected */
|
|
|
|
if (end_active_trans(thd))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2003-01-04 14:37:20 +01:00
|
|
|
}
|
2001-08-02 05:29:50 +02:00
|
|
|
else
|
2003-04-03 20:19:12 +02:00
|
|
|
{
|
|
|
|
/*
|
|
|
|
If this is a slave thread, we may sometimes execute some
|
|
|
|
DROP / * 40005 TEMPORARY * / TABLE
|
|
|
|
that come from parts of binlogs (likely if we use RESET SLAVE or CHANGE
|
|
|
|
MASTER TO), while the temporary table has already been dropped.
|
2004-02-11 00:06:46 +01:00
|
|
|
To not generate such irrelevant "table does not exist errors",
|
|
|
|
we silently add IF EXISTS if TEMPORARY was used.
|
2003-04-03 20:19:12 +02:00
|
|
|
*/
|
|
|
|
if (thd->slave_thread)
|
|
|
|
lex->drop_if_exists= 1;
|
2005-06-01 15:52:32 +02:00
|
|
|
|
2005-06-07 11:47:59 +02:00
|
|
|
/* So that DROP TEMPORARY TABLE gets to binlog at commit/rollback */
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
thd->options|= OPTION_KEEP_LOG;
|
2003-04-03 20:19:12 +02:00
|
|
|
}
|
2004-07-16 00:15:55 +02:00
|
|
|
res= mysql_rm_table(thd, first_table, lex->drop_if_exists,
|
|
|
|
lex->drop_temporary);
|
2001-08-02 05:29:50 +02:00
|
|
|
}
|
|
|
|
break;
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_DROP_INDEX:
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_one_table_access(thd, INDEX_ACL, all_tables))
|
2000-07-31 21:29:14 +02:00
|
|
|
goto error; /* purecov: inspected */
|
2000-12-07 13:08:48 +01:00
|
|
|
if (end_active_trans(thd))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2006-05-05 19:08:40 +02:00
|
|
|
res= mysql_drop_index(thd, first_table, &lex->alter_info);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
case SQLCOM_SHOW_PROCESSLIST:
|
2005-09-15 21:29:07 +02:00
|
|
|
if (!thd->security_ctx->priv_user[0] &&
|
|
|
|
check_global_access(thd,PROCESS_ACL))
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2003-09-26 12:33:13 +02:00
|
|
|
mysqld_list_processes(thd,
|
2005-09-15 21:29:07 +02:00
|
|
|
(thd->security_ctx->master_access & PROCESS_ACL ?
|
|
|
|
NullS :
|
|
|
|
thd->security_ctx->priv_user),
|
|
|
|
lex->verbose);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2003-12-17 23:52:03 +01:00
|
|
|
case SQLCOM_SHOW_STORAGE_ENGINES:
|
|
|
|
res= mysqld_show_storage_engines(thd);
|
2002-06-12 23:13:12 +02:00
|
|
|
break;
|
2006-05-02 19:56:43 +02:00
|
|
|
case SQLCOM_SHOW_AUTHORS:
|
|
|
|
res= mysqld_show_authors(thd);
|
|
|
|
break;
|
2006-05-03 02:53:26 +02:00
|
|
|
case SQLCOM_SHOW_CONTRIBUTORS:
|
|
|
|
res= mysqld_show_contributors(thd);
|
|
|
|
break;
|
2002-06-12 23:13:12 +02:00
|
|
|
case SQLCOM_SHOW_PRIVILEGES:
|
|
|
|
res= mysqld_show_privileges(thd);
|
|
|
|
break;
|
|
|
|
case SQLCOM_SHOW_COLUMN_TYPES:
|
|
|
|
res= mysqld_show_column_types(thd);
|
|
|
|
break;
|
2006-05-02 19:56:43 +02:00
|
|
|
case SQLCOM_SHOW_ENGINE_LOGS:
|
2000-12-15 12:18:52 +01:00
|
|
|
#ifdef DONT_ALLOW_SHOW_COMMANDS
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_NOT_ALLOWED_COMMAND, ER(ER_NOT_ALLOWED_COMMAND),
|
|
|
|
MYF(0)); /* purecov: inspected */
|
2003-01-15 15:39:36 +01:00
|
|
|
goto error;
|
2000-12-15 12:18:52 +01:00
|
|
|
#else
|
|
|
|
{
|
2005-09-13 13:07:38 +02:00
|
|
|
if (grant_option && check_access(thd, FILE_ACL, any_db,0,0,0,0))
|
2000-12-15 12:18:52 +01:00
|
|
|
goto error;
|
2006-05-02 19:56:43 +02:00
|
|
|
res= ha_show_status(thd, lex->create_info.db_type, HA_ENGINE_LOGS);
|
2000-12-15 12:18:52 +01:00
|
|
|
break;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
#endif
|
|
|
|
case SQLCOM_CHANGE_DB:
|
2005-08-12 02:04:16 +02:00
|
|
|
if (!mysql_change_db(thd,select_lex->db,FALSE))
|
|
|
|
send_ok(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2003-09-06 15:50:30 +02:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_LOAD:
|
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
2000-07-31 21:29:14 +02:00
|
|
|
uint privilege= (lex->duplicates == DUP_REPLACE ?
|
2005-03-16 02:32:47 +01:00
|
|
|
INSERT_ACL | DELETE_ACL : INSERT_ACL) |
|
|
|
|
(lex->local_file ? 0 : FILE_ACL);
|
2002-02-13 20:53:26 +01:00
|
|
|
|
2005-03-16 02:32:47 +01:00
|
|
|
if (lex->local_file)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2002-02-13 20:53:26 +01:00
|
|
|
if (!(thd->client_capabilities & CLIENT_LOCAL_FILES) ||
|
2005-09-01 11:46:43 +02:00
|
|
|
!opt_local_infile)
|
2002-02-13 20:53:26 +01:00
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_NOT_ALLOWED_COMMAND, ER(ER_NOT_ALLOWED_COMMAND), MYF(0));
|
2002-02-13 20:53:26 +01:00
|
|
|
goto error;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2005-03-16 02:32:47 +01:00
|
|
|
|
|
|
|
if (check_one_table_access(thd, privilege, all_tables))
|
|
|
|
goto error;
|
|
|
|
|
2004-07-16 00:15:55 +02:00
|
|
|
res= mysql_load(thd, lex->exchange, first_table, lex->field_list,
|
2005-03-16 02:32:47 +01:00
|
|
|
lex->update_list, lex->value_list, lex->duplicates,
|
2005-03-16 10:13:35 +01:00
|
|
|
lex->ignore, (bool) lex->local_file);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-09-06 15:50:30 +02:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_SET_OPTION:
|
2004-06-03 23:17:18 +02:00
|
|
|
{
|
|
|
|
List<set_var_base> *lex_var_list= &lex->var_list;
|
2005-03-04 14:35:28 +01:00
|
|
|
if ((check_table_access(thd, SELECT_ACL, all_tables, 0) ||
|
2004-10-20 03:04:37 +02:00
|
|
|
open_and_lock_tables(thd, all_tables)))
|
|
|
|
goto error;
|
2004-06-03 23:17:18 +02:00
|
|
|
if (lex->one_shot_set && not_all_support_one_shot(lex_var_list))
|
|
|
|
{
|
2004-10-20 03:04:37 +02:00
|
|
|
my_error(ER_RESERVED_SYNTAX, MYF(0), "SET ONE_SHOT");
|
|
|
|
goto error;
|
2004-06-03 23:17:18 +02:00
|
|
|
}
|
|
|
|
if (!(res= sql_set_variables(thd, lex_var_list)))
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
If the previous command was a SET ONE_SHOT, we don't want to forget
|
|
|
|
about the ONE_SHOT property of that SET. So we use a |= instead of = .
|
|
|
|
*/
|
|
|
|
thd->one_shot_set|= lex->one_shot_set;
|
2002-10-02 12:33:08 +02:00
|
|
|
send_ok(thd);
|
2004-06-03 23:17:18 +02:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2004-06-03 23:17:18 +02:00
|
|
|
}
|
2003-01-04 15:33:42 +01:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_UNLOCK_TABLES:
|
2004-11-10 17:56:45 +01:00
|
|
|
/*
|
|
|
|
It is critical for mysqldump --single-transaction --master-data that
|
|
|
|
UNLOCK TABLES does not implicitely commit a connection which has only
|
|
|
|
done FLUSH TABLES WITH READ LOCK + BEGIN. If this assumption becomes
|
|
|
|
false, mysqldump will not work.
|
|
|
|
*/
|
2002-11-16 19:19:10 +01:00
|
|
|
unlock_locked_tables(thd);
|
2001-09-08 00:02:41 +02:00
|
|
|
if (thd->options & OPTION_TABLE_LOCK)
|
|
|
|
{
|
2001-04-19 19:41:19 +02:00
|
|
|
end_active_trans(thd);
|
2001-09-08 00:02:41 +02:00
|
|
|
thd->options&= ~(ulong) (OPTION_TABLE_LOCK);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
if (thd->global_read_lock)
|
2001-08-14 19:33:49 +02:00
|
|
|
unlock_global_read_lock(thd);
|
2002-10-02 12:33:08 +02:00
|
|
|
send_ok(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
case SQLCOM_LOCK_TABLES:
|
2002-11-16 19:19:10 +01:00
|
|
|
unlock_locked_tables(thd);
|
2004-07-16 00:15:55 +02:00
|
|
|
if (check_db_used(thd, all_tables) || end_active_trans(thd))
|
2000-07-31 21:29:14 +02:00
|
|
|
goto error;
|
2004-07-16 00:15:55 +02:00
|
|
|
if (check_table_access(thd, LOCK_TABLES_ACL | SELECT_ACL, all_tables, 0))
|
2001-06-03 16:07:26 +02:00
|
|
|
goto error;
|
2000-07-31 21:29:14 +02:00
|
|
|
thd->in_lock_tables=1;
|
2001-09-08 00:02:41 +02:00
|
|
|
thd->options|= OPTION_TABLE_LOCK;
|
2004-07-16 00:15:55 +02:00
|
|
|
|
2005-03-04 14:35:28 +01:00
|
|
|
if (!(res= simple_open_n_lock_tables(thd, all_tables)))
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-03-04 17:32:55 +01:00
|
|
|
#ifdef HAVE_QUERY_CACHE
|
|
|
|
if (thd->variables.query_cache_wlock_invalidate)
|
2004-07-16 00:15:55 +02:00
|
|
|
query_cache.invalidate_locked_for_write(first_table);
|
2004-03-04 17:32:55 +01:00
|
|
|
#endif /*HAVE_QUERY_CACHE*/
|
2000-07-31 21:29:14 +02:00
|
|
|
thd->locked_tables=thd->lock;
|
|
|
|
thd->lock=0;
|
2002-10-02 12:33:08 +02:00
|
|
|
send_ok(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2001-09-08 00:02:41 +02:00
|
|
|
else
|
|
|
|
thd->options&= ~(ulong) (OPTION_TABLE_LOCK);
|
2000-07-31 21:29:14 +02:00
|
|
|
thd->in_lock_tables=0;
|
|
|
|
break;
|
|
|
|
case SQLCOM_CREATE_DB:
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2005-05-27 04:17:33 +02:00
|
|
|
if (end_active_trans(thd))
|
|
|
|
{
|
|
|
|
res= -1;
|
|
|
|
break;
|
|
|
|
}
|
2003-12-30 12:14:21 +01:00
|
|
|
char *alias;
|
2004-03-05 19:13:33 +01:00
|
|
|
if (!(alias=thd->strdup(lex->name)) || check_db_name(lex->name))
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_WRONG_DB_NAME, MYF(0), lex->name);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-01-25 14:07:51 +01:00
|
|
|
/*
|
|
|
|
If in a slave thread :
|
|
|
|
CREATE DATABASE DB was certainly not preceded by USE DB.
|
2004-12-27 12:08:22 +01:00
|
|
|
For that reason, db_ok() in sql/slave.cc did not check the
|
2003-01-25 14:07:51 +01:00
|
|
|
do_db/ignore_db. And as this query involves no tables, tables_ok()
|
|
|
|
above was not called. So we have to check rules again here.
|
|
|
|
*/
|
2003-02-07 14:47:24 +01:00
|
|
|
#ifdef HAVE_REPLICATION
|
2006-05-02 19:56:43 +02:00
|
|
|
if (thd->slave_thread &&
|
|
|
|
(!rpl_filter->db_ok(lex->name) ||
|
|
|
|
!rpl_filter->db_ok_with_wild_table(lex->name)))
|
2003-07-08 15:50:57 +02:00
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_SLAVE_IGNORED_TABLE, ER(ER_SLAVE_IGNORED_TABLE), MYF(0));
|
2003-01-25 14:07:51 +01:00
|
|
|
break;
|
2003-07-08 15:50:57 +02:00
|
|
|
}
|
2003-02-07 14:47:24 +01:00
|
|
|
#endif
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd,CREATE_ACL,lex->name,0,1,0,is_schema_db(lex->name)))
|
2001-08-14 19:33:49 +02:00
|
|
|
break;
|
2004-02-11 00:06:46 +01:00
|
|
|
res= mysql_create_db(thd,(lower_case_table_names == 2 ? alias : lex->name),
|
2004-02-16 09:03:25 +01:00
|
|
|
&lex->create_info, 0);
|
2001-08-14 19:33:49 +02:00
|
|
|
break;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_DROP_DB:
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2005-05-27 04:17:33 +02:00
|
|
|
if (end_active_trans(thd))
|
|
|
|
{
|
|
|
|
res= -1;
|
|
|
|
break;
|
|
|
|
}
|
2005-02-25 15:53:22 +01:00
|
|
|
if (check_db_name(lex->name))
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_WRONG_DB_NAME, MYF(0), lex->name);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2003-01-25 14:07:51 +01:00
|
|
|
/*
|
|
|
|
If in a slave thread :
|
|
|
|
DROP DATABASE DB may not be preceded by USE DB.
|
|
|
|
For that reason, maybe db_ok() in sql/slave.cc did not check the
|
|
|
|
do_db/ignore_db. And as this query involves no tables, tables_ok()
|
|
|
|
above was not called. So we have to check rules again here.
|
|
|
|
*/
|
2003-02-07 14:47:24 +01:00
|
|
|
#ifdef HAVE_REPLICATION
|
2003-01-25 14:07:51 +01:00
|
|
|
if (thd->slave_thread &&
|
2006-05-02 19:56:43 +02:00
|
|
|
(!rpl_filter->db_ok(lex->name) ||
|
|
|
|
!rpl_filter->db_ok_with_wild_table(lex->name)))
|
2003-07-08 15:50:57 +02:00
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_SLAVE_IGNORED_TABLE, ER(ER_SLAVE_IGNORED_TABLE), MYF(0));
|
2003-01-25 14:07:51 +01:00
|
|
|
break;
|
2003-07-08 15:50:57 +02:00
|
|
|
}
|
2003-02-07 14:47:24 +01:00
|
|
|
#endif
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd,DROP_ACL,lex->name,0,1,0,is_schema_db(lex->name)))
|
2001-08-14 19:33:49 +02:00
|
|
|
break;
|
2001-09-02 12:47:00 +02:00
|
|
|
if (thd->locked_tables || thd->active_transaction())
|
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_LOCK_OR_ACTIVE_TRANSACTION,
|
|
|
|
ER(ER_LOCK_OR_ACTIVE_TRANSACTION), MYF(0));
|
2001-09-02 12:47:00 +02:00
|
|
|
goto error;
|
|
|
|
}
|
2005-02-21 13:47:57 +01:00
|
|
|
res= mysql_rm_db(thd, lex->name, lex->drop_if_exists, 0);
|
2001-08-14 19:33:49 +02:00
|
|
|
break;
|
|
|
|
}
|
2006-05-02 19:56:43 +02:00
|
|
|
case SQLCOM_RENAME_DB:
|
|
|
|
{
|
|
|
|
LEX_STRING *olddb, *newdb;
|
|
|
|
List_iterator <LEX_STRING> db_list(lex->db_list);
|
|
|
|
olddb= db_list++;
|
|
|
|
newdb= db_list++;
|
|
|
|
if (end_active_trans(thd))
|
|
|
|
{
|
|
|
|
res= 1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#ifdef HAVE_REPLICATION
|
|
|
|
if (thd->slave_thread &&
|
|
|
|
(!rpl_filter->db_ok(olddb->str) ||
|
|
|
|
!rpl_filter->db_ok(newdb->str) ||
|
|
|
|
!rpl_filter->db_ok_with_wild_table(olddb->str) ||
|
|
|
|
!rpl_filter->db_ok_with_wild_table(newdb->str)))
|
|
|
|
{
|
|
|
|
res= 1;
|
|
|
|
my_message(ER_SLAVE_IGNORED_TABLE, ER(ER_SLAVE_IGNORED_TABLE), MYF(0));
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
if (check_access(thd,ALTER_ACL,olddb->str,0,1,0,is_schema_db(olddb->str)) ||
|
|
|
|
check_access(thd,DROP_ACL,olddb->str,0,1,0,is_schema_db(olddb->str)) ||
|
|
|
|
check_access(thd,CREATE_ACL,newdb->str,0,1,0,is_schema_db(newdb->str)))
|
|
|
|
{
|
|
|
|
res= 1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (thd->locked_tables || thd->active_transaction())
|
|
|
|
{
|
|
|
|
res= 1;
|
|
|
|
my_message(ER_LOCK_OR_ACTIVE_TRANSACTION,
|
|
|
|
ER(ER_LOCK_OR_ACTIVE_TRANSACTION), MYF(0));
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
res= mysql_rename_db(thd, olddb, newdb);
|
|
|
|
if (!res)
|
|
|
|
send_ok(thd);
|
|
|
|
break;
|
|
|
|
}
|
2002-06-27 11:41:02 +02:00
|
|
|
case SQLCOM_ALTER_DB:
|
|
|
|
{
|
2004-12-06 17:01:51 +01:00
|
|
|
char *db= lex->name ? lex->name : thd->db;
|
|
|
|
if (!db)
|
2002-06-27 11:41:02 +02:00
|
|
|
{
|
2004-12-22 12:54:39 +01:00
|
|
|
my_message(ER_NO_DB_ERROR, ER(ER_NO_DB_ERROR), MYF(0));
|
|
|
|
break;
|
2004-12-06 17:01:51 +01:00
|
|
|
}
|
|
|
|
if (!strip_sp(db) || check_db_name(db))
|
2002-06-27 11:41:02 +02:00
|
|
|
{
|
2006-05-02 19:56:43 +02:00
|
|
|
my_error(ER_WRONG_DB_NAME, MYF(0), db);
|
2002-06-27 11:41:02 +02:00
|
|
|
break;
|
|
|
|
}
|
2004-02-28 22:22:29 +01:00
|
|
|
/*
|
|
|
|
If in a slave thread :
|
|
|
|
ALTER DATABASE DB may not be preceded by USE DB.
|
2005-01-24 15:48:25 +01:00
|
|
|
For that reason, maybe db_ok() in sql/slave.cc did not check the
|
2004-02-28 22:22:29 +01:00
|
|
|
do_db/ignore_db. And as this query involves no tables, tables_ok()
|
|
|
|
above was not called. So we have to check rules again here.
|
|
|
|
*/
|
|
|
|
#ifdef HAVE_REPLICATION
|
2005-01-24 15:48:25 +01:00
|
|
|
if (thd->slave_thread &&
|
2006-05-02 19:56:43 +02:00
|
|
|
(!rpl_filter->db_ok(db) ||
|
|
|
|
!rpl_filter->db_ok_with_wild_table(db)))
|
2004-02-28 22:22:29 +01:00
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_SLAVE_IGNORED_TABLE, ER(ER_SLAVE_IGNORED_TABLE), MYF(0));
|
2004-02-28 22:22:29 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd, ALTER_ACL, db, 0, 1, 0, is_schema_db(db)))
|
2002-06-27 11:41:02 +02:00
|
|
|
break;
|
|
|
|
if (thd->locked_tables || thd->active_transaction())
|
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_LOCK_OR_ACTIVE_TRANSACTION,
|
|
|
|
ER(ER_LOCK_OR_ACTIVE_TRANSACTION), MYF(0));
|
2002-06-27 11:41:02 +02:00
|
|
|
goto error;
|
|
|
|
}
|
2004-12-06 17:01:51 +01:00
|
|
|
res= mysql_alter_db(thd, db, &lex->create_info);
|
2002-06-27 11:41:02 +02:00
|
|
|
break;
|
|
|
|
}
|
2002-07-02 11:31:54 +02:00
|
|
|
case SQLCOM_SHOW_CREATE_DB:
|
|
|
|
{
|
|
|
|
if (!strip_sp(lex->name) || check_db_name(lex->name))
|
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_WRONG_DB_NAME, MYF(0), lex->name);
|
2002-07-02 11:31:54 +02:00
|
|
|
break;
|
|
|
|
}
|
2002-11-06 09:01:38 +01:00
|
|
|
res=mysqld_show_create_db(thd,lex->name,&lex->create_info);
|
2002-07-02 11:31:54 +02:00
|
|
|
break;
|
|
|
|
}
|
2006-05-02 19:56:43 +02:00
|
|
|
case SQLCOM_CREATE_EVENT:
|
|
|
|
case SQLCOM_ALTER_EVENT:
|
|
|
|
case SQLCOM_DROP_EVENT:
|
|
|
|
{
|
|
|
|
uint rows_affected= 1;
|
|
|
|
DBUG_ASSERT(lex->et);
|
|
|
|
do {
|
|
|
|
if (! lex->et->dbname.str)
|
|
|
|
{
|
|
|
|
my_message(ER_NO_DB_ERROR, ER(ER_NO_DB_ERROR), MYF(0));
|
|
|
|
res= true;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (check_access(thd, EVENT_ACL, lex->et->dbname.str, 0, 0, 0,
|
|
|
|
is_schema_db(lex->et->dbname.str)))
|
|
|
|
break;
|
|
|
|
|
|
|
|
if (end_active_trans(thd))
|
|
|
|
{
|
|
|
|
res= -1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
switch (lex->sql_command) {
|
|
|
|
case SQLCOM_CREATE_EVENT:
|
2006-05-22 20:46:13 +02:00
|
|
|
res= Events::create_event(thd, lex->et,
|
2006-05-29 10:39:45 +02:00
|
|
|
(uint) lex->create_info.options,
|
|
|
|
&rows_affected);
|
2006-05-02 19:56:43 +02:00
|
|
|
break;
|
|
|
|
case SQLCOM_ALTER_EVENT:
|
2006-05-22 20:46:13 +02:00
|
|
|
res= Events::update_event(thd, lex->et, lex->spname,
|
2006-05-29 10:39:45 +02:00
|
|
|
&rows_affected);
|
2006-05-02 19:56:43 +02:00
|
|
|
break;
|
|
|
|
case SQLCOM_DROP_EVENT:
|
2006-05-22 20:46:13 +02:00
|
|
|
res= Events::drop_event(thd, lex->et, lex->drop_if_exists,
|
2006-05-29 10:39:45 +02:00
|
|
|
&rows_affected);
|
2006-05-02 19:56:43 +02:00
|
|
|
default:;
|
|
|
|
}
|
|
|
|
DBUG_PRINT("info", ("CREATE/ALTER/DROP returned error code=%d af_rows=%d",
|
|
|
|
res, rows_affected));
|
|
|
|
if (!res)
|
|
|
|
send_ok(thd, rows_affected);
|
|
|
|
|
|
|
|
/* lex->unit.cleanup() is called outside, no need to call it here */
|
|
|
|
} while (0);
|
|
|
|
if (!thd->spcont)
|
|
|
|
{
|
|
|
|
lex->et->free_sphead_on_delete= true;
|
|
|
|
lex->et->free_sp();
|
|
|
|
lex->et->deinit_mutexes();
|
|
|
|
}
|
|
|
|
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_SHOW_CREATE_EVENT:
|
|
|
|
{
|
|
|
|
DBUG_ASSERT(lex->spname);
|
|
|
|
DBUG_ASSERT(lex->et);
|
|
|
|
if (! lex->spname->m_db.str)
|
|
|
|
{
|
|
|
|
my_message(ER_NO_DB_ERROR, ER(ER_NO_DB_ERROR), MYF(0));
|
|
|
|
res= true;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (check_access(thd, EVENT_ACL, lex->spname->m_db.str, 0, 0, 0,
|
|
|
|
is_schema_db(lex->spname->m_db.str)))
|
|
|
|
break;
|
|
|
|
|
|
|
|
if (lex->spname->m_name.length > NAME_LEN)
|
|
|
|
{
|
|
|
|
my_error(ER_TOO_LONG_IDENT, MYF(0), lex->spname->m_name.str);
|
|
|
|
goto error;
|
|
|
|
}
|
2006-05-29 10:39:45 +02:00
|
|
|
res= Events::show_create_event(thd, lex->spname);
|
2006-05-02 19:56:43 +02:00
|
|
|
break;
|
|
|
|
}
|
2006-05-22 20:46:13 +02:00
|
|
|
#ifndef DBUG_OFF
|
|
|
|
case SQLCOM_SHOW_SCHEDULER_STATUS:
|
|
|
|
{
|
|
|
|
res= Events::dump_internal_status(thd);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif
|
2004-07-07 10:29:39 +02:00
|
|
|
case SQLCOM_CREATE_FUNCTION: // UDF function
|
2003-12-21 11:48:03 +01:00
|
|
|
{
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd,INSERT_ACL,"mysql",0,1,0,0))
|
2003-12-21 11:48:03 +01:00
|
|
|
break;
|
2000-07-31 21:29:14 +02:00
|
|
|
#ifdef HAVE_DLOPEN
|
2005-11-23 00:11:19 +01:00
|
|
|
if (sp_find_routine(thd, TYPE_ENUM_FUNCTION, lex->spname,
|
|
|
|
&thd->sp_func_cache, FALSE))
|
2003-12-21 11:48:03 +01:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_UDF_EXISTS, MYF(0), lex->spname->m_name.str);
|
2003-12-21 11:48:03 +01:00
|
|
|
goto error;
|
|
|
|
}
|
2004-11-23 19:19:09 +01:00
|
|
|
if (!(res = mysql_create_function(thd, &lex->udf)))
|
2003-12-21 11:48:03 +01:00
|
|
|
send_ok(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
#else
|
2005-08-10 13:02:36 +02:00
|
|
|
my_error(ER_CANT_OPEN_LIBRARY, MYF(0), lex->udf.dl, 0, "feature disabled");
|
2004-10-20 03:04:37 +02:00
|
|
|
res= TRUE;
|
2000-07-31 21:29:14 +02:00
|
|
|
#endif
|
|
|
|
break;
|
2003-12-21 11:48:03 +01:00
|
|
|
}
|
2003-09-26 12:33:13 +02:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2004-11-25 21:55:49 +01:00
|
|
|
case SQLCOM_CREATE_USER:
|
|
|
|
{
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd, INSERT_ACL, "mysql", 0, 1, 1, 0) &&
|
2005-03-22 15:54:18 +01:00
|
|
|
check_global_access(thd,CREATE_USER_ACL))
|
2004-11-25 21:55:49 +01:00
|
|
|
break;
|
2005-10-13 11:12:17 +02:00
|
|
|
if (end_active_trans(thd))
|
|
|
|
goto error;
|
2004-11-25 21:55:49 +01:00
|
|
|
if (!(res= mysql_create_user(thd, lex->users_list)))
|
|
|
|
{
|
|
|
|
if (mysql_bin_log.is_open())
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::MYSQL_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, FALSE, FALSE);
|
2004-11-25 21:55:49 +01:00
|
|
|
send_ok(thd);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
2003-06-06 14:43:23 +02:00
|
|
|
case SQLCOM_DROP_USER:
|
|
|
|
{
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd, DELETE_ACL, "mysql", 0, 1, 1, 0) &&
|
2005-03-22 15:54:18 +01:00
|
|
|
check_global_access(thd,CREATE_USER_ACL))
|
2003-06-06 14:43:23 +02:00
|
|
|
break;
|
2005-10-13 11:12:17 +02:00
|
|
|
if (end_active_trans(thd))
|
|
|
|
goto error;
|
2003-06-06 14:43:23 +02:00
|
|
|
if (!(res= mysql_drop_user(thd, lex->users_list)))
|
|
|
|
{
|
|
|
|
if (mysql_bin_log.is_open())
|
|
|
|
{
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::MYSQL_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, FALSE, FALSE);
|
2004-11-25 21:55:49 +01:00
|
|
|
}
|
|
|
|
send_ok(thd);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_RENAME_USER:
|
|
|
|
{
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd, UPDATE_ACL, "mysql", 0, 1, 1, 0) &&
|
2005-03-22 15:54:18 +01:00
|
|
|
check_global_access(thd,CREATE_USER_ACL))
|
2004-11-25 21:55:49 +01:00
|
|
|
break;
|
2005-10-13 11:12:17 +02:00
|
|
|
if (end_active_trans(thd))
|
|
|
|
goto error;
|
2004-11-25 21:55:49 +01:00
|
|
|
if (!(res= mysql_rename_user(thd, lex->users_list)))
|
|
|
|
{
|
|
|
|
if (mysql_bin_log.is_open())
|
|
|
|
{
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::MYSQL_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, FALSE, FALSE);
|
2003-06-06 14:43:23 +02:00
|
|
|
}
|
|
|
|
send_ok(thd);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_REVOKE_ALL:
|
|
|
|
{
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd, UPDATE_ACL, "mysql", 0, 1, 1, 0) &&
|
2005-03-22 15:54:18 +01:00
|
|
|
check_global_access(thd,CREATE_USER_ACL))
|
2003-06-06 14:43:23 +02:00
|
|
|
break;
|
|
|
|
if (!(res = mysql_revoke_all(thd, lex->users_list)))
|
|
|
|
{
|
|
|
|
if (mysql_bin_log.is_open())
|
|
|
|
{
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::MYSQL_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, FALSE, FALSE);
|
2003-06-06 14:43:23 +02:00
|
|
|
}
|
|
|
|
send_ok(thd);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
2001-08-14 19:33:49 +02:00
|
|
|
case SQLCOM_REVOKE:
|
|
|
|
case SQLCOM_GRANT:
|
|
|
|
{
|
|
|
|
if (check_access(thd, lex->grant | lex->grant_tot_col | GRANT_ACL,
|
2005-01-24 19:41:42 +01:00
|
|
|
first_table ? first_table->db : select_lex->db,
|
2004-07-16 00:15:55 +02:00
|
|
|
first_table ? &first_table->grant.privilege : 0,
|
2005-09-13 13:07:38 +02:00
|
|
|
first_table ? 0 : 1, 0,
|
|
|
|
first_table ? (bool) first_table->schema_table :
|
|
|
|
select_lex->db ? is_schema_db(select_lex->db) : 0))
|
2001-08-14 19:33:49 +02:00
|
|
|
goto error;
|
|
|
|
|
2005-09-15 21:29:07 +02:00
|
|
|
if (thd->security_ctx->user) // If not replication
|
2005-03-17 07:16:56 +01:00
|
|
|
{
|
|
|
|
LEX_USER *user;
|
2005-03-18 12:32:28 +01:00
|
|
|
|
2005-03-17 07:16:56 +01:00
|
|
|
List_iterator <LEX_USER> user_list(lex->users_list);
|
2005-03-22 15:54:18 +01:00
|
|
|
while ((user= user_list++))
|
2005-03-17 07:16:56 +01:00
|
|
|
{
|
2005-03-22 15:54:18 +01:00
|
|
|
if (specialflag & SPECIAL_NO_RESOLVE &&
|
|
|
|
hostname_requires_resolving(user->host.str))
|
|
|
|
push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
|
|
|
|
ER_WARN_HOSTNAME_WONT_WORK,
|
|
|
|
ER(ER_WARN_HOSTNAME_WONT_WORK),
|
|
|
|
user->host.str);
|
|
|
|
// Are we trying to change a password of another user
|
|
|
|
DBUG_ASSERT(user->host.str != 0);
|
2005-09-15 21:29:07 +02:00
|
|
|
if (strcmp(thd->security_ctx->user, user->user.str) ||
|
2005-03-22 15:54:18 +01:00
|
|
|
my_strcasecmp(system_charset_info,
|
2005-09-15 21:29:07 +02:00
|
|
|
user->host.str, thd->security_ctx->host_or_ip))
|
2005-03-22 15:54:18 +01:00
|
|
|
{
|
|
|
|
// TODO: use check_change_password()
|
2006-05-06 09:25:59 +02:00
|
|
|
if (is_acl_user(user->host.str, user->user.str) &&
|
|
|
|
user->password.str &&
|
2005-09-13 13:07:38 +02:00
|
|
|
check_access(thd, UPDATE_ACL,"mysql",0,1,1,0))
|
2005-03-22 15:54:18 +01:00
|
|
|
{
|
|
|
|
my_message(ER_PASSWORD_NOT_ALLOWED,
|
|
|
|
ER(ER_PASSWORD_NOT_ALLOWED), MYF(0));
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
}
|
2003-11-20 09:55:48 +01:00
|
|
|
}
|
|
|
|
}
|
2004-07-16 00:15:55 +02:00
|
|
|
if (first_table)
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2005-05-17 20:54:20 +02:00
|
|
|
if (lex->type == TYPE_ENUM_PROCEDURE ||
|
|
|
|
lex->type == TYPE_ENUM_FUNCTION)
|
2004-12-23 11:46:24 +01:00
|
|
|
{
|
|
|
|
uint grants= lex->all_privileges
|
|
|
|
? (PROC_ACLS & ~GRANT_ACL) | (lex->grant & GRANT_ACL)
|
|
|
|
: lex->grant;
|
|
|
|
if (grant_option &&
|
2005-05-17 20:54:20 +02:00
|
|
|
check_grant_routine(thd, grants | GRANT_ACL, all_tables,
|
|
|
|
lex->type == TYPE_ENUM_PROCEDURE, 0))
|
2004-12-23 11:46:24 +01:00
|
|
|
goto error;
|
2005-05-17 20:54:20 +02:00
|
|
|
res= mysql_routine_grant(thd, all_tables,
|
|
|
|
lex->type == TYPE_ENUM_PROCEDURE,
|
|
|
|
lex->users_list, grants,
|
|
|
|
lex->sql_command == SQLCOM_REVOKE, 0);
|
2004-12-23 11:46:24 +01:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (grant_option && check_grant(thd,
|
|
|
|
(lex->grant | lex->grant_tot_col |
|
|
|
|
GRANT_ACL),
|
|
|
|
all_tables, 0, UINT_MAX, 0))
|
|
|
|
goto error;
|
|
|
|
res= mysql_table_grant(thd, all_tables, lex->users_list,
|
|
|
|
lex->columns, lex->grant,
|
|
|
|
lex->sql_command == SQLCOM_REVOKE);
|
|
|
|
}
|
|
|
|
if (!res && mysql_bin_log.is_open())
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2003-12-19 18:03:27 +01:00
|
|
|
thd->clear_error();
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::MYSQL_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, FALSE, FALSE);
|
2001-08-14 19:33:49 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2005-05-17 20:54:20 +02:00
|
|
|
if (lex->columns.elements || lex->type)
|
2001-08-14 19:33:49 +02:00
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_ILLEGAL_GRANT_FOR_TABLE, ER(ER_ILLEGAL_GRANT_FOR_TABLE),
|
|
|
|
MYF(0));
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2001-08-14 19:33:49 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
res = mysql_grant(thd, select_lex->db, lex->users_list, lex->grant,
|
|
|
|
lex->sql_command == SQLCOM_REVOKE);
|
|
|
|
if (!res)
|
|
|
|
{
|
|
|
|
if (mysql_bin_log.is_open())
|
|
|
|
{
|
2003-12-16 11:10:50 +01:00
|
|
|
thd->clear_error();
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::MYSQL_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, FALSE, FALSE);
|
2001-08-14 19:33:49 +02:00
|
|
|
}
|
2004-12-29 18:30:37 +01:00
|
|
|
if (lex->sql_command == SQLCOM_GRANT)
|
2002-02-01 19:53:24 +01:00
|
|
|
{
|
2002-02-13 21:37:19 +01:00
|
|
|
List_iterator <LEX_USER> str_list(lex->users_list);
|
|
|
|
LEX_USER *user;
|
|
|
|
while ((user=str_list++))
|
2004-12-06 16:15:54 +01:00
|
|
|
reset_mqh(user);
|
2002-02-01 19:53:24 +01:00
|
|
|
}
|
2001-08-14 19:33:49 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
2003-09-29 08:47:37 +02:00
|
|
|
#endif /*!NO_EMBEDDED_ACCESS_CHECKS*/
|
2000-10-14 10:16:17 +02:00
|
|
|
case SQLCOM_RESET:
|
2005-02-26 11:19:02 +01:00
|
|
|
/*
|
|
|
|
RESET commands are never written to the binary log, so we have to
|
|
|
|
initialize this variable because RESET shares the same code as FLUSH
|
2003-05-15 18:35:39 +02:00
|
|
|
*/
|
|
|
|
lex->no_write_to_binlog= 1;
|
|
|
|
case SQLCOM_FLUSH:
|
|
|
|
{
|
2005-08-11 14:58:15 +02:00
|
|
|
bool write_to_binlog;
|
2004-07-16 00:15:55 +02:00
|
|
|
if (check_global_access(thd,RELOAD_ACL) || check_db_used(thd, all_tables))
|
2000-07-31 21:29:14 +02:00
|
|
|
goto error;
|
2003-05-15 18:35:39 +02:00
|
|
|
/*
|
|
|
|
reload_acl_and_cache() will tell us if we are allowed to write to the
|
|
|
|
binlog or not.
|
|
|
|
*/
|
2004-10-20 03:04:37 +02:00
|
|
|
if (!reload_acl_and_cache(thd, lex->type, first_table, &write_to_binlog))
|
2003-05-15 18:35:39 +02:00
|
|
|
{
|
|
|
|
/*
|
|
|
|
We WANT to write and we CAN write.
|
|
|
|
! we write after unlocking the table.
|
|
|
|
*/
|
|
|
|
if (!lex->no_write_to_binlog && write_to_binlog)
|
|
|
|
{
|
|
|
|
if (mysql_bin_log.is_open())
|
|
|
|
{
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::STMT_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, 0, FALSE);
|
2003-05-15 18:35:39 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
send_ok(thd);
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2003-05-15 18:35:39 +02:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_KILL:
|
2004-11-25 16:13:06 +01:00
|
|
|
{
|
|
|
|
Item *it= (Item *)lex->value_list.head();
|
|
|
|
|
2005-07-01 06:05:42 +02:00
|
|
|
if ((!it->fixed && it->fix_fields(lex->thd, &it)) || it->check_cols(1))
|
2004-11-25 16:13:06 +01:00
|
|
|
{
|
|
|
|
my_message(ER_SET_CONSTANTS_ONLY, ER(ER_SET_CONSTANTS_ONLY),
|
|
|
|
MYF(0));
|
|
|
|
goto error;
|
|
|
|
}
|
2006-05-22 20:46:13 +02:00
|
|
|
sql_kill(thd, (ulong)it->val_int(), lex->type & ONLY_KILL_QUERY);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2004-11-25 16:13:06 +01:00
|
|
|
}
|
2003-09-26 12:33:13 +02:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_SHOW_GRANTS:
|
2005-09-15 21:29:07 +02:00
|
|
|
if ((thd->security_ctx->priv_user &&
|
|
|
|
!strcmp(thd->security_ctx->priv_user, lex->grant_user->user.str)) ||
|
2005-09-13 13:07:38 +02:00
|
|
|
!check_access(thd, SELECT_ACL, "mysql",0,1,0,0))
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
res = mysql_show_grants(thd,lex->grant_user);
|
|
|
|
}
|
|
|
|
break;
|
2003-09-26 12:33:13 +02:00
|
|
|
#endif
|
2001-04-07 00:18:33 +02:00
|
|
|
case SQLCOM_HA_OPEN:
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_db_used(thd, all_tables) ||
|
|
|
|
check_table_access(thd, SELECT_ACL, all_tables, 0))
|
2001-04-07 00:18:33 +02:00
|
|
|
goto error;
|
2005-03-16 15:11:01 +01:00
|
|
|
res= mysql_ha_open(thd, first_table, 0);
|
2001-04-07 00:18:33 +02:00
|
|
|
break;
|
|
|
|
case SQLCOM_HA_CLOSE:
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
|
|
|
if (check_db_used(thd, all_tables))
|
2001-04-07 00:18:33 +02:00
|
|
|
goto error;
|
2004-07-16 00:15:55 +02:00
|
|
|
res= mysql_ha_close(thd, first_table);
|
2001-04-07 00:18:33 +02:00
|
|
|
break;
|
|
|
|
case SQLCOM_HA_READ:
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ASSERT(first_table == all_tables && first_table != 0);
|
2003-07-03 10:55:36 +02:00
|
|
|
/*
|
|
|
|
There is no need to check for table permissions here, because
|
|
|
|
if a user has no permissions to read a table, he won't be
|
|
|
|
able to open it (with SQLCOM_HA_OPEN) in the first place.
|
|
|
|
*/
|
2004-07-16 00:15:55 +02:00
|
|
|
if (check_db_used(thd, all_tables))
|
2001-04-07 00:18:33 +02:00
|
|
|
goto error;
|
2005-06-07 12:11:36 +02:00
|
|
|
unit->set_limit(select_lex);
|
2005-01-16 13:16:23 +01:00
|
|
|
res= mysql_ha_read(thd, first_table, lex->ha_read_mode, lex->ident.str,
|
2004-07-16 00:15:55 +02:00
|
|
|
lex->insert_list, lex->ha_rkey_mode, select_lex->where,
|
2005-06-07 12:11:36 +02:00
|
|
|
unit->select_limit_cnt, unit->offset_limit_cnt);
|
2001-04-07 00:18:33 +02:00
|
|
|
break;
|
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
case SQLCOM_BEGIN:
|
2005-10-05 19:58:16 +02:00
|
|
|
if (thd->transaction.xid_state.xa_state != XA_NOTR)
|
|
|
|
{
|
|
|
|
my_error(ER_XAER_RMFAIL, MYF(0),
|
|
|
|
xa_state_names[thd->transaction.xid_state.xa_state]);
|
|
|
|
break;
|
|
|
|
}
|
2005-02-01 20:48:05 +01:00
|
|
|
if (begin_trans(thd))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2005-02-14 21:50:09 +01:00
|
|
|
send_ok(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
case SQLCOM_COMMIT:
|
2005-02-17 13:52:16 +01:00
|
|
|
if (end_trans(thd, lex->tx_release ? COMMIT_RELEASE :
|
2005-02-16 17:34:02 +01:00
|
|
|
lex->tx_chain ? COMMIT_AND_CHAIN : COMMIT))
|
2004-10-20 03:04:37 +02:00
|
|
|
goto error;
|
2005-02-17 13:52:16 +01:00
|
|
|
send_ok(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
case SQLCOM_ROLLBACK:
|
2005-02-17 13:52:16 +01:00
|
|
|
if (end_trans(thd, lex->tx_release ? ROLLBACK_RELEASE :
|
2005-02-16 17:34:02 +01:00
|
|
|
lex->tx_chain ? ROLLBACK_AND_CHAIN : ROLLBACK))
|
2005-02-01 20:48:05 +01:00
|
|
|
goto error;
|
2005-02-17 13:52:16 +01:00
|
|
|
send_ok(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2005-02-14 21:50:09 +01:00
|
|
|
case SQLCOM_RELEASE_SAVEPOINT:
|
2005-02-16 17:34:02 +01:00
|
|
|
{
|
2005-03-11 21:10:41 +01:00
|
|
|
SAVEPOINT *sv;
|
|
|
|
for (sv=thd->transaction.savepoints; sv; sv=sv->prev)
|
2005-02-14 21:50:09 +01:00
|
|
|
{
|
|
|
|
if (my_strnncoll(system_charset_info,
|
|
|
|
(uchar *)lex->ident.str, lex->ident.length,
|
2005-03-11 21:10:41 +01:00
|
|
|
(uchar *)sv->name, sv->length) == 0)
|
2005-02-14 21:50:09 +01:00
|
|
|
break;
|
|
|
|
}
|
2005-03-11 21:10:41 +01:00
|
|
|
if (sv)
|
2005-02-14 21:50:09 +01:00
|
|
|
{
|
2005-03-11 21:10:41 +01:00
|
|
|
if (ha_release_savepoint(thd, sv))
|
2005-02-14 21:50:09 +01:00
|
|
|
res= TRUE; // cannot happen
|
2005-02-16 17:34:02 +01:00
|
|
|
else
|
|
|
|
send_ok(thd);
|
2005-03-11 21:10:41 +01:00
|
|
|
thd->transaction.savepoints=sv->prev;
|
2005-02-14 21:50:09 +01:00
|
|
|
}
|
2005-01-16 13:16:23 +01:00
|
|
|
else
|
2005-02-14 21:50:09 +01:00
|
|
|
my_error(ER_SP_DOES_NOT_EXIST, MYF(0), "SAVEPOINT", lex->ident.str);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
2005-02-14 21:50:09 +01:00
|
|
|
}
|
2003-06-15 00:04:28 +02:00
|
|
|
case SQLCOM_ROLLBACK_TO_SAVEPOINT:
|
2005-02-16 17:34:02 +01:00
|
|
|
{
|
2005-03-11 21:10:41 +01:00
|
|
|
SAVEPOINT *sv;
|
|
|
|
for (sv=thd->transaction.savepoints; sv; sv=sv->prev)
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
|
|
|
if (my_strnncoll(system_charset_info,
|
|
|
|
(uchar *)lex->ident.str, lex->ident.length,
|
2005-03-11 21:10:41 +01:00
|
|
|
(uchar *)sv->name, sv->length) == 0)
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
|
|
|
}
|
2005-03-11 21:10:41 +01:00
|
|
|
if (sv)
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
2005-03-11 21:10:41 +01:00
|
|
|
if (ha_rollback_to_savepoint(thd, sv))
|
2005-01-16 13:16:23 +01:00
|
|
|
res= TRUE; // cannot happen
|
|
|
|
else
|
2005-02-14 21:50:09 +01:00
|
|
|
{
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
if ((thd->options &
|
|
|
|
(OPTION_STATUS_NO_TRANS_UPDATE | OPTION_KEEP_LOG)) &&
|
2005-02-14 21:50:09 +01:00
|
|
|
!thd->slave_thread)
|
|
|
|
push_warning(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
|
|
|
|
ER_WARNING_NOT_COMPLETE_ROLLBACK,
|
|
|
|
ER(ER_WARNING_NOT_COMPLETE_ROLLBACK));
|
|
|
|
send_ok(thd);
|
|
|
|
}
|
2005-03-11 21:10:41 +01:00
|
|
|
thd->transaction.savepoints=sv;
|
2003-06-15 00:04:28 +02:00
|
|
|
}
|
|
|
|
else
|
2005-01-16 13:16:23 +01:00
|
|
|
my_error(ER_SP_DOES_NOT_EXIST, MYF(0), "SAVEPOINT", lex->ident.str);
|
2003-06-15 00:04:28 +02:00
|
|
|
break;
|
2005-01-16 13:16:23 +01:00
|
|
|
}
|
2003-06-06 03:18:58 +02:00
|
|
|
case SQLCOM_SAVEPOINT:
|
2005-11-19 13:09:23 +01:00
|
|
|
if (!(thd->options & (OPTION_NOT_AUTOCOMMIT | OPTION_BEGIN) ||
|
|
|
|
thd->in_sub_stmt) || !opt_using_transactions)
|
2003-08-18 23:08:08 +02:00
|
|
|
send_ok(thd);
|
2003-06-15 00:04:28 +02:00
|
|
|
else
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
|
|
|
SAVEPOINT **sv, *newsv;
|
|
|
|
for (sv=&thd->transaction.savepoints; *sv; sv=&(*sv)->prev)
|
|
|
|
{
|
|
|
|
if (my_strnncoll(system_charset_info,
|
|
|
|
(uchar *)lex->ident.str, lex->ident.length,
|
|
|
|
(uchar *)(*sv)->name, (*sv)->length) == 0)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (*sv) /* old savepoint of the same name exists */
|
|
|
|
{
|
|
|
|
newsv=*sv;
|
|
|
|
ha_release_savepoint(thd, *sv); // it cannot fail
|
|
|
|
*sv=(*sv)->prev;
|
|
|
|
}
|
|
|
|
else if ((newsv=(SAVEPOINT *) alloc_root(&thd->transaction.mem_root,
|
|
|
|
savepoint_alloc_size)) == 0)
|
|
|
|
{
|
|
|
|
my_error(ER_OUT_OF_RESOURCES, MYF(0));
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
newsv->name=strmake_root(&thd->transaction.mem_root,
|
|
|
|
lex->ident.str, lex->ident.length);
|
|
|
|
newsv->length=lex->ident.length;
|
|
|
|
/*
|
|
|
|
if we'll get an error here, don't add new savepoint to the list.
|
|
|
|
we'll lose a little bit of memory in transaction mem_root, but it'll
|
|
|
|
be free'd when transaction ends anyway
|
|
|
|
*/
|
|
|
|
if (ha_savepoint(thd, newsv))
|
|
|
|
res= TRUE;
|
|
|
|
else
|
|
|
|
{
|
|
|
|
newsv->prev=thd->transaction.savepoints;
|
|
|
|
thd->transaction.savepoints=newsv;
|
|
|
|
send_ok(thd);
|
|
|
|
}
|
|
|
|
}
|
2005-02-01 20:48:05 +01:00
|
|
|
break;
|
2003-03-05 19:45:17 +01:00
|
|
|
case SQLCOM_CREATE_PROCEDURE:
|
|
|
|
case SQLCOM_CREATE_SPFUNCTION:
|
2003-12-21 11:48:03 +01:00
|
|
|
{
|
2004-10-07 15:32:36 +02:00
|
|
|
uint namelen;
|
2004-12-23 11:46:24 +01:00
|
|
|
char *name, *db;
|
2004-10-20 03:04:37 +02:00
|
|
|
int result;
|
2004-10-07 15:32:36 +02:00
|
|
|
|
2005-02-25 15:53:22 +01:00
|
|
|
DBUG_ASSERT(lex->sphead != 0);
|
2004-10-07 15:32:36 +02:00
|
|
|
|
2005-10-19 12:46:32 +02:00
|
|
|
if (!lex->sphead->m_db.str || !lex->sphead->m_db.str[0])
|
|
|
|
{
|
2005-11-11 19:01:48 +01:00
|
|
|
if (!thd->db)
|
2005-10-19 12:46:32 +02:00
|
|
|
{
|
|
|
|
my_message(ER_NO_DB_ERROR, ER(ER_NO_DB_ERROR), MYF(0));
|
|
|
|
delete lex->sphead;
|
|
|
|
lex->sphead= 0;
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
lex->sphead->m_db.length= strlen(thd->db);
|
|
|
|
lex->sphead->m_db.str= thd->db;
|
|
|
|
}
|
|
|
|
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd, CREATE_PROC_ACL, lex->sphead->m_db.str, 0, 0, 0,
|
|
|
|
is_schema_db(lex->sphead->m_db.str)))
|
2004-10-07 15:32:36 +02:00
|
|
|
{
|
|
|
|
delete lex->sphead;
|
|
|
|
lex->sphead= 0;
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
|
2005-09-14 20:42:39 +02:00
|
|
|
if (end_active_trans(thd))
|
2004-12-23 11:46:24 +01:00
|
|
|
{
|
2005-10-19 12:46:32 +02:00
|
|
|
delete lex->sphead;
|
|
|
|
lex->sphead= 0;
|
|
|
|
goto error;
|
2004-12-23 11:46:24 +01:00
|
|
|
}
|
2004-10-07 15:32:36 +02:00
|
|
|
|
|
|
|
name= lex->sphead->name(&namelen);
|
2003-02-26 19:22:29 +01:00
|
|
|
#ifdef HAVE_DLOPEN
|
2003-12-21 11:48:03 +01:00
|
|
|
if (lex->sphead->m_type == TYPE_ENUM_FUNCTION)
|
|
|
|
{
|
|
|
|
udf_func *udf = find_udf(name, namelen);
|
2003-04-03 20:00:52 +02:00
|
|
|
|
2003-12-21 11:48:03 +01:00
|
|
|
if (udf)
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_UDF_EXISTS, MYF(0), name);
|
2003-12-15 13:24:16 +01:00
|
|
|
delete lex->sphead;
|
2004-10-07 15:32:36 +02:00
|
|
|
lex->sphead= 0;
|
2003-02-21 17:37:05 +01:00
|
|
|
goto error;
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
}
|
2003-12-21 11:48:03 +01:00
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2006-03-02 13:18:49 +01:00
|
|
|
/*
|
|
|
|
If the definer is not specified, this means that CREATE-statement missed
|
|
|
|
DEFINER-clause. DEFINER-clause can be missed in two cases:
|
|
|
|
|
|
|
|
- The user submitted a statement w/o the clause. This is a normal
|
|
|
|
case, we should assign CURRENT_USER as definer.
|
|
|
|
|
|
|
|
- Our slave received an updated from the master, that does not
|
|
|
|
replicate definer for stored rountines. We should also assign
|
|
|
|
CURRENT_USER as definer here, but also we should mark this routine
|
|
|
|
as NON-SUID. This is essential for the sake of backward
|
|
|
|
compatibility.
|
|
|
|
|
|
|
|
The problem is the slave thread is running under "special" user (@),
|
|
|
|
that actually does not exist. In the older versions we do not fail
|
|
|
|
execution of a stored routine if its definer does not exist and
|
|
|
|
continue the execution under the authorization of the invoker
|
|
|
|
(BUG#13198). And now if we try to switch to slave-current-user (@),
|
|
|
|
we will fail.
|
|
|
|
|
|
|
|
Actually, this leads to the inconsistent state of master and
|
|
|
|
slave (different definers, different SUID behaviour), but it seems,
|
|
|
|
this is the best we can do.
|
|
|
|
*/
|
|
|
|
|
|
|
|
if (!lex->definer)
|
|
|
|
{
|
|
|
|
bool res= FALSE;
|
|
|
|
Query_arena original_arena;
|
|
|
|
Query_arena *ps_arena = thd->activate_stmt_arena_if_needed(&original_arena);
|
|
|
|
|
|
|
|
if (!(lex->definer= create_default_definer(thd)))
|
|
|
|
res= TRUE;
|
|
|
|
|
|
|
|
if (ps_arena)
|
|
|
|
thd->restore_active_arena(ps_arena, &original_arena);
|
|
|
|
|
|
|
|
if (res)
|
|
|
|
{
|
|
|
|
/* Error has been already reported. */
|
|
|
|
delete lex->sphead;
|
|
|
|
lex->sphead= 0;
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (thd->slave_thread)
|
|
|
|
lex->sphead->m_chistics->suid= SP_IS_NOT_SUID;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
If the specified definer differs from the current user, we should check
|
|
|
|
that the current user has SUPER privilege (in order to create a stored
|
|
|
|
routine under another user one must have SUPER privilege).
|
|
|
|
*/
|
|
|
|
|
|
|
|
else if (strcmp(lex->definer->user.str, thd->security_ctx->priv_user) ||
|
|
|
|
my_strcasecmp(system_charset_info,
|
|
|
|
lex->definer->host.str,
|
|
|
|
thd->security_ctx->priv_host))
|
|
|
|
{
|
|
|
|
if (check_global_access(thd, SUPER_ACL))
|
|
|
|
{
|
|
|
|
my_error(ER_SPECIFIC_ACCESS_DENIED_ERROR, MYF(0), "SUPER");
|
|
|
|
delete lex->sphead;
|
|
|
|
lex->sphead= 0;
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Check that the specified definer exists. Emit a warning if not. */
|
|
|
|
|
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
|
|
|
if (!is_acl_user(lex->definer->host.str,
|
|
|
|
lex->definer->user.str))
|
|
|
|
{
|
|
|
|
push_warning_printf(thd,
|
|
|
|
MYSQL_ERROR::WARN_LEVEL_NOTE,
|
|
|
|
ER_NO_SUCH_USER,
|
|
|
|
ER(ER_NO_SUCH_USER),
|
|
|
|
lex->definer->user.str,
|
|
|
|
lex->definer->host.str);
|
|
|
|
}
|
|
|
|
#endif /* NO_EMBEDDED_ACCESS_CHECKS */
|
|
|
|
|
2005-10-19 12:46:32 +02:00
|
|
|
/*
|
|
|
|
We need to copy name and db in order to use them for
|
|
|
|
check_routine_access which is called after lex->sphead has
|
|
|
|
been deleted.
|
|
|
|
*/
|
2004-12-23 11:46:24 +01:00
|
|
|
name= thd->strdup(name);
|
2005-10-19 12:46:32 +02:00
|
|
|
lex->sphead->m_db.str= db= thd->strmake(lex->sphead->m_db.str,
|
|
|
|
lex->sphead->m_db.length);
|
2004-10-20 03:04:37 +02:00
|
|
|
res= (result= lex->sphead->create(thd));
|
2005-08-09 14:29:33 +02:00
|
|
|
if (result == SP_OK)
|
|
|
|
{
|
2005-10-19 12:46:32 +02:00
|
|
|
/*
|
|
|
|
We must cleanup the unit and the lex here because
|
|
|
|
sp_grant_privileges calls (indirectly) db_find_routine,
|
2006-03-10 01:44:08 +01:00
|
|
|
which in turn may call MYSQLparse with THD::lex.
|
2005-10-19 12:46:32 +02:00
|
|
|
TODO: fix db_find_routine to use a temporary lex.
|
|
|
|
*/
|
2004-08-24 12:59:06 +02:00
|
|
|
lex->unit.cleanup();
|
2003-12-21 11:48:03 +01:00
|
|
|
delete lex->sphead;
|
|
|
|
lex->sphead= 0;
|
2004-12-23 12:42:57 +01:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2004-12-23 11:46:24 +01:00
|
|
|
/* only add privileges if really neccessary */
|
2005-05-26 16:44:46 +02:00
|
|
|
if (sp_automatic_privileges && !opt_noacl &&
|
2005-05-17 20:54:20 +02:00
|
|
|
check_routine_access(thd, DEFAULT_CREATE_PROC_ACLS,
|
|
|
|
db, name,
|
|
|
|
lex->sql_command == SQLCOM_CREATE_PROCEDURE, 1))
|
2004-12-23 11:46:24 +01:00
|
|
|
{
|
2005-05-17 20:54:20 +02:00
|
|
|
if (sp_grant_privileges(thd, db, name,
|
|
|
|
lex->sql_command == SQLCOM_CREATE_PROCEDURE))
|
2004-12-23 11:46:24 +01:00
|
|
|
push_warning(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
|
|
|
|
ER_PROC_AUTO_GRANT_FAIL,
|
|
|
|
ER(ER_PROC_AUTO_GRANT_FAIL));
|
2006-05-02 19:56:43 +02:00
|
|
|
close_thread_tables(thd);
|
2004-12-23 11:46:24 +01:00
|
|
|
}
|
2004-12-23 12:42:57 +01:00
|
|
|
#endif
|
2004-12-23 11:46:24 +01:00
|
|
|
send_ok(thd);
|
2005-08-09 14:29:33 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
switch (result) {
|
|
|
|
case SP_WRITE_ROW_FAILED:
|
|
|
|
my_error(ER_SP_ALREADY_EXISTS, MYF(0), SP_TYPE_STRING(lex), name);
|
|
|
|
break;
|
|
|
|
case SP_NO_DB_ERROR:
|
|
|
|
my_error(ER_BAD_DB_ERROR, MYF(0), lex->sphead->m_db.str);
|
|
|
|
break;
|
|
|
|
case SP_BAD_IDENTIFIER:
|
|
|
|
my_error(ER_TOO_LONG_IDENT, MYF(0), name);
|
|
|
|
break;
|
|
|
|
case SP_BODY_TOO_LONG:
|
|
|
|
my_error(ER_TOO_LONG_BODY, MYF(0), name);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
my_error(ER_SP_STORE_FAILED, MYF(0), SP_TYPE_STRING(lex), name);
|
|
|
|
break;
|
|
|
|
}
|
2004-08-24 12:59:06 +02:00
|
|
|
lex->unit.cleanup();
|
2003-12-21 11:48:03 +01:00
|
|
|
delete lex->sphead;
|
|
|
|
lex->sphead= 0;
|
|
|
|
goto error;
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
}
|
2003-12-21 11:48:03 +01:00
|
|
|
break;
|
|
|
|
}
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
case SQLCOM_CALL:
|
|
|
|
{
|
|
|
|
sp_head *sp;
|
|
|
|
|
2005-03-04 14:35:28 +01:00
|
|
|
/*
|
|
|
|
This will cache all SP and SF and open and lock all tables
|
|
|
|
required for execution.
|
|
|
|
*/
|
|
|
|
if (check_table_access(thd, SELECT_ACL, all_tables, 0) ||
|
|
|
|
open_and_lock_tables(thd, all_tables))
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
/*
|
2005-07-30 10:19:57 +02:00
|
|
|
By this moment all needed SPs should be in cache so no need to look
|
|
|
|
into DB.
|
2005-03-04 14:35:28 +01:00
|
|
|
*/
|
2005-11-23 00:11:19 +01:00
|
|
|
if (!(sp= sp_find_routine(thd, TYPE_ENUM_PROCEDURE, lex->spname,
|
|
|
|
&thd->sp_proc_cache, TRUE)))
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_SP_DOES_NOT_EXIST, MYF(0), "PROCEDURE",
|
2004-10-20 03:04:37 +02:00
|
|
|
lex->spname->m_qname.str);
|
2003-01-15 15:39:36 +01:00
|
|
|
goto error;
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2003-12-16 14:15:27 +01:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2005-09-20 20:20:38 +02:00
|
|
|
Security_context *save_ctx;
|
2003-12-16 14:15:27 +01:00
|
|
|
#endif
|
2004-05-14 16:00:57 +02:00
|
|
|
ha_rows select_limit;
|
2004-12-30 23:44:00 +01:00
|
|
|
/* bits that should be cleared in thd->server_status */
|
|
|
|
uint bits_to_be_cleared= 0;
|
Implement WL#2661 "Prepared Statements: Dynamic SQL in Stored Procedures".
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
2005-09-03 01:13:18 +02:00
|
|
|
/*
|
|
|
|
Check that the stored procedure doesn't contain Dynamic SQL
|
|
|
|
and doesn't return result sets: such stored procedures can't
|
|
|
|
be called from a function or trigger.
|
|
|
|
*/
|
|
|
|
if (thd->in_sub_stmt)
|
|
|
|
{
|
|
|
|
const char *where= (thd->in_sub_stmt & SUB_STMT_TRIGGER ?
|
|
|
|
"trigger" : "function");
|
|
|
|
if (sp->is_not_allowed_in_function(where))
|
|
|
|
goto error;
|
|
|
|
}
|
2003-04-23 09:22:54 +02:00
|
|
|
|
2003-01-15 15:39:36 +01:00
|
|
|
my_bool nsok= thd->net.no_send_ok;
|
|
|
|
thd->net.no_send_ok= TRUE;
|
Implement WL#2661 "Prepared Statements: Dynamic SQL in Stored Procedures".
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
2005-09-03 01:13:18 +02:00
|
|
|
if (sp->m_flags & sp_head::MULTI_RESULTS)
|
2003-04-23 09:22:54 +02:00
|
|
|
{
|
2003-08-27 17:04:33 +02:00
|
|
|
if (! (thd->client_capabilities & CLIENT_MULTI_RESULTS))
|
2003-04-23 09:22:54 +02:00
|
|
|
{
|
Implement WL#2661 "Prepared Statements: Dynamic SQL in Stored Procedures".
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
2005-09-03 01:13:18 +02:00
|
|
|
/*
|
|
|
|
The client does not support multiple result sets being sent
|
|
|
|
back
|
|
|
|
*/
|
2005-04-27 16:35:49 +02:00
|
|
|
my_error(ER_SP_BADSELECT, MYF(0), sp->m_qname.str);
|
2003-04-23 09:22:54 +02:00
|
|
|
thd->net.no_send_ok= nsok;
|
|
|
|
goto error;
|
|
|
|
}
|
2004-12-30 23:44:00 +01:00
|
|
|
/*
|
|
|
|
If SERVER_MORE_RESULTS_EXISTS is not set,
|
|
|
|
then remember that it should be cleared
|
|
|
|
*/
|
|
|
|
bits_to_be_cleared= (~thd->server_status &
|
|
|
|
SERVER_MORE_RESULTS_EXISTS);
|
|
|
|
thd->server_status|= SERVER_MORE_RESULTS_EXISTS;
|
2003-04-23 09:22:54 +02:00
|
|
|
}
|
|
|
|
|
2003-12-16 14:15:27 +01:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2005-09-15 21:29:07 +02:00
|
|
|
if (check_routine_access(thd, EXECUTE_ACL,
|
|
|
|
sp->m_db.str, sp->m_name.str, TRUE, 0) ||
|
|
|
|
sp_change_security_context(thd, sp, &save_ctx))
|
2004-12-23 11:46:24 +01:00
|
|
|
{
|
|
|
|
thd->net.no_send_ok= nsok;
|
|
|
|
goto error;
|
|
|
|
}
|
2005-09-15 21:29:07 +02:00
|
|
|
if (save_ctx &&
|
|
|
|
check_routine_access(thd, EXECUTE_ACL,
|
|
|
|
sp->m_db.str, sp->m_name.str, TRUE, 0))
|
2004-12-23 11:46:24 +01:00
|
|
|
{
|
|
|
|
thd->net.no_send_ok= nsok;
|
2005-09-15 21:29:07 +02:00
|
|
|
sp_restore_security_context(thd, save_ctx);
|
2004-12-23 11:46:24 +01:00
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
|
2003-12-16 14:15:27 +01:00
|
|
|
#endif
|
2004-05-14 16:00:57 +02:00
|
|
|
select_limit= thd->variables.select_limit;
|
|
|
|
thd->variables.select_limit= HA_POS_ERROR;
|
2003-12-13 16:40:52 +01:00
|
|
|
|
2005-07-30 10:19:57 +02:00
|
|
|
thd->row_count_func= 0;
|
2005-08-25 15:34:34 +02:00
|
|
|
|
|
|
|
/*
|
Implement WL#2661 "Prepared Statements: Dynamic SQL in Stored Procedures".
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
2005-09-03 01:13:18 +02:00
|
|
|
We never write CALL statements into binlog:
|
2005-08-25 15:34:34 +02:00
|
|
|
- If the mode is non-prelocked, each statement will be logged
|
|
|
|
separately.
|
|
|
|
- If the mode is prelocked, the invoking statement will care
|
|
|
|
about writing into binlog.
|
|
|
|
So just execute the statement.
|
2005-05-05 14:20:53 +02:00
|
|
|
*/
|
2005-08-25 15:34:34 +02:00
|
|
|
res= sp->execute_procedure(thd, &lex->value_list);
|
2005-02-26 11:19:02 +01:00
|
|
|
/*
|
|
|
|
If warnings have been cleared, we have to clear total_warn_count
|
|
|
|
too, otherwise the clients get confused.
|
2005-02-24 13:56:09 +01:00
|
|
|
*/
|
|
|
|
if (thd->warn_list.is_empty())
|
|
|
|
thd->total_warn_count= 0;
|
|
|
|
|
2004-05-14 16:00:57 +02:00
|
|
|
thd->variables.select_limit= select_limit;
|
2003-12-16 14:15:27 +01:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2005-09-15 21:29:07 +02:00
|
|
|
sp_restore_security_context(thd, save_ctx);
|
2003-12-16 14:15:27 +01:00
|
|
|
#endif
|
2003-12-13 16:40:52 +01:00
|
|
|
|
2003-01-15 15:39:36 +01:00
|
|
|
thd->net.no_send_ok= nsok;
|
2004-12-30 23:44:00 +01:00
|
|
|
thd->server_status&= ~bits_to_be_cleared;
|
2003-01-15 15:39:36 +01:00
|
|
|
|
2004-10-20 03:04:37 +02:00
|
|
|
if (!res)
|
2004-12-30 23:44:00 +01:00
|
|
|
send_ok(thd, (ulong) (thd->row_count_func < 0 ? 0 :
|
|
|
|
thd->row_count_func));
|
2003-01-15 15:39:36 +01:00
|
|
|
else
|
|
|
|
goto error; // Substatement should already have sent error
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
}
|
2003-04-23 09:22:54 +02:00
|
|
|
break;
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
}
|
|
|
|
case SQLCOM_ALTER_PROCEDURE:
|
2003-02-21 17:37:05 +01:00
|
|
|
case SQLCOM_ALTER_FUNCTION:
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
{
|
2004-10-20 03:04:37 +02:00
|
|
|
int result;
|
2004-10-22 20:29:06 +02:00
|
|
|
sp_head *sp;
|
|
|
|
st_sp_chistics chistics;
|
|
|
|
|
|
|
|
memcpy(&chistics, &lex->sp_chistics, sizeof(chistics));
|
2003-11-17 18:21:36 +01:00
|
|
|
if (lex->sql_command == SQLCOM_ALTER_PROCEDURE)
|
2005-11-23 00:11:19 +01:00
|
|
|
sp= sp_find_routine(thd, TYPE_ENUM_PROCEDURE, lex->spname,
|
|
|
|
&thd->sp_proc_cache, FALSE);
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
else
|
2005-11-23 00:11:19 +01:00
|
|
|
sp= sp_find_routine(thd, TYPE_ENUM_FUNCTION, lex->spname,
|
|
|
|
&thd->sp_func_cache, FALSE);
|
2005-03-16 15:11:01 +01:00
|
|
|
mysql_reset_errors(thd, 0);
|
2004-10-22 20:29:06 +02:00
|
|
|
if (! sp)
|
2005-04-12 14:52:54 +02:00
|
|
|
{
|
|
|
|
if (lex->spname->m_db.str)
|
|
|
|
result= SP_KEY_NOT_FOUND;
|
|
|
|
else
|
|
|
|
{
|
|
|
|
my_message(ER_NO_DB_ERROR, ER(ER_NO_DB_ERROR), MYF(0));
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
}
|
2004-10-22 20:29:06 +02:00
|
|
|
else
|
|
|
|
{
|
2005-05-17 20:54:20 +02:00
|
|
|
if (check_routine_access(thd, ALTER_PROC_ACL, sp->m_db.str,
|
|
|
|
sp->m_name.str,
|
|
|
|
lex->sql_command == SQLCOM_ALTER_PROCEDURE, 0))
|
2004-12-23 11:46:24 +01:00
|
|
|
goto error;
|
2005-09-14 20:42:39 +02:00
|
|
|
|
|
|
|
if (end_active_trans(thd))
|
|
|
|
goto error;
|
2004-10-22 20:29:06 +02:00
|
|
|
memcpy(&lex->sp_chistics, &chistics, sizeof(lex->sp_chistics));
|
2005-11-10 17:50:51 +01:00
|
|
|
if ((sp->m_type == TYPE_ENUM_FUNCTION) &&
|
|
|
|
!trust_function_creators && mysql_bin_log.is_open() &&
|
2005-05-05 14:20:53 +02:00
|
|
|
!sp->m_chistics->detistic &&
|
|
|
|
(chistics.daccess == SP_CONTAINS_SQL ||
|
|
|
|
chistics.daccess == SP_MODIFIES_SQL_DATA))
|
|
|
|
{
|
|
|
|
my_message(ER_BINLOG_UNSAFE_ROUTINE,
|
|
|
|
ER(ER_BINLOG_UNSAFE_ROUTINE), MYF(0));
|
|
|
|
result= SP_INTERNAL_ERROR;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2005-11-10 17:50:51 +01:00
|
|
|
/*
|
|
|
|
Note that if you implement the capability of ALTER FUNCTION to
|
|
|
|
alter the body of the function, this command should be made to
|
|
|
|
follow the restrictions that log-bin-trust-function-creators=0
|
|
|
|
already puts on CREATE FUNCTION.
|
|
|
|
*/
|
2005-05-05 14:20:53 +02:00
|
|
|
if (lex->sql_command == SQLCOM_ALTER_PROCEDURE)
|
|
|
|
result= sp_update_procedure(thd, lex->spname, &lex->sp_chistics);
|
|
|
|
else
|
|
|
|
result= sp_update_function(thd, lex->spname, &lex->sp_chistics);
|
|
|
|
}
|
2004-10-22 20:29:06 +02:00
|
|
|
}
|
2004-10-20 03:04:37 +02:00
|
|
|
switch (result)
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
{
|
2003-11-17 18:21:36 +01:00
|
|
|
case SP_OK:
|
2005-05-05 14:20:53 +02:00
|
|
|
if (mysql_bin_log.is_open())
|
|
|
|
{
|
|
|
|
thd->clear_error();
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::MYSQL_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, FALSE, FALSE);
|
2005-05-05 14:20:53 +02:00
|
|
|
}
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
send_ok(thd);
|
2003-11-17 18:21:36 +01:00
|
|
|
break;
|
|
|
|
case SP_KEY_NOT_FOUND:
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_SP_DOES_NOT_EXIST, MYF(0),
|
|
|
|
SP_COM_STRING(lex), lex->spname->m_qname.str);
|
2003-11-17 18:21:36 +01:00
|
|
|
goto error;
|
|
|
|
default:
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_SP_CANT_ALTER, MYF(0),
|
|
|
|
SP_COM_STRING(lex), lex->spname->m_qname.str);
|
2003-11-17 18:21:36 +01:00
|
|
|
goto error;
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
}
|
2003-04-23 09:22:54 +02:00
|
|
|
break;
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
}
|
|
|
|
case SQLCOM_DROP_PROCEDURE:
|
2003-02-21 17:37:05 +01:00
|
|
|
case SQLCOM_DROP_FUNCTION:
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
{
|
2004-10-20 03:04:37 +02:00
|
|
|
int result;
|
2006-01-26 13:29:46 +01:00
|
|
|
int type= (lex->sql_command == SQLCOM_DROP_PROCEDURE ?
|
|
|
|
TYPE_ENUM_PROCEDURE : TYPE_ENUM_FUNCTION);
|
2004-10-22 20:29:06 +02:00
|
|
|
|
2006-01-26 13:29:46 +01:00
|
|
|
result= sp_routine_exists_in_table(thd, type, lex->spname);
|
2005-03-16 15:11:01 +01:00
|
|
|
mysql_reset_errors(thd, 0);
|
2006-01-26 13:29:46 +01:00
|
|
|
if (result == SP_OK)
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
{
|
2006-01-26 13:29:46 +01:00
|
|
|
char *db= lex->spname->m_db.str;
|
|
|
|
char *name= lex->spname->m_name.str;
|
|
|
|
|
2005-05-17 20:54:20 +02:00
|
|
|
if (check_routine_access(thd, ALTER_PROC_ACL, db, name,
|
|
|
|
lex->sql_command == SQLCOM_DROP_PROCEDURE, 0))
|
2004-10-28 10:02:48 +02:00
|
|
|
goto error;
|
2005-09-14 20:42:39 +02:00
|
|
|
|
|
|
|
if (end_active_trans(thd))
|
|
|
|
goto error;
|
2004-12-23 12:42:57 +01:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2005-05-26 16:44:46 +02:00
|
|
|
if (sp_automatic_privileges && !opt_noacl &&
|
2005-05-17 20:54:20 +02:00
|
|
|
sp_revoke_privileges(thd, db, name,
|
|
|
|
lex->sql_command == SQLCOM_DROP_PROCEDURE))
|
2004-12-23 11:46:24 +01:00
|
|
|
{
|
|
|
|
push_warning(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
|
|
|
|
ER_PROC_AUTO_REVOKE_FAIL,
|
|
|
|
ER(ER_PROC_AUTO_REVOKE_FAIL));
|
|
|
|
}
|
2004-12-23 12:42:57 +01:00
|
|
|
#endif
|
2004-10-22 20:29:06 +02:00
|
|
|
if (lex->sql_command == SQLCOM_DROP_PROCEDURE)
|
2004-10-28 10:02:48 +02:00
|
|
|
result= sp_drop_procedure(thd, lex->spname);
|
2004-10-22 20:29:06 +02:00
|
|
|
else
|
2004-10-28 10:02:48 +02:00
|
|
|
result= sp_drop_function(thd, lex->spname);
|
2004-11-23 19:19:09 +01:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2004-10-22 20:29:06 +02:00
|
|
|
#ifdef HAVE_DLOPEN
|
2004-11-23 19:19:09 +01:00
|
|
|
if (lex->sql_command == SQLCOM_DROP_FUNCTION)
|
|
|
|
{
|
|
|
|
udf_func *udf = find_udf(lex->spname->m_name.str,
|
|
|
|
lex->spname->m_name.length);
|
|
|
|
if (udf)
|
|
|
|
{
|
2005-09-13 13:07:38 +02:00
|
|
|
if (check_access(thd, DELETE_ACL, "mysql", 0, 1, 0, 0))
|
2004-11-23 19:19:09 +01:00
|
|
|
goto error;
|
|
|
|
if (!(res = mysql_drop_function(thd, &lex->spname->m_name)))
|
2003-02-21 17:37:05 +01:00
|
|
|
{
|
2004-11-23 19:19:09 +01:00
|
|
|
send_ok(thd);
|
|
|
|
break;
|
2003-02-21 17:37:05 +01:00
|
|
|
}
|
|
|
|
}
|
2004-10-22 20:29:06 +02:00
|
|
|
}
|
2004-11-23 19:19:09 +01:00
|
|
|
#endif
|
2005-04-12 14:52:54 +02:00
|
|
|
if (lex->spname->m_db.str)
|
|
|
|
result= SP_KEY_NOT_FOUND;
|
|
|
|
else
|
|
|
|
{
|
|
|
|
my_message(ER_NO_DB_ERROR, ER(ER_NO_DB_ERROR), MYF(0));
|
|
|
|
goto error;
|
|
|
|
}
|
2003-02-21 17:37:05 +01:00
|
|
|
}
|
2004-10-20 03:04:37 +02:00
|
|
|
res= result;
|
|
|
|
switch (result)
|
2003-02-21 17:37:05 +01:00
|
|
|
{
|
|
|
|
case SP_OK:
|
2005-05-05 14:20:53 +02:00
|
|
|
if (mysql_bin_log.is_open())
|
|
|
|
{
|
|
|
|
thd->clear_error();
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::MYSQL_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, FALSE, FALSE);
|
2005-05-05 14:20:53 +02:00
|
|
|
}
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
send_ok(thd);
|
2003-02-21 17:37:05 +01:00
|
|
|
break;
|
|
|
|
case SP_KEY_NOT_FOUND:
|
2003-03-26 12:29:58 +01:00
|
|
|
if (lex->drop_if_exists)
|
|
|
|
{
|
2004-09-09 05:59:26 +02:00
|
|
|
push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_NOTE,
|
2003-03-26 12:29:58 +01:00
|
|
|
ER_SP_DOES_NOT_EXIST, ER(ER_SP_DOES_NOT_EXIST),
|
2004-02-17 17:36:53 +01:00
|
|
|
SP_COM_STRING(lex), lex->spname->m_name.str);
|
2004-10-20 03:04:37 +02:00
|
|
|
res= FALSE;
|
2003-03-26 12:29:58 +01:00
|
|
|
send_ok(thd);
|
|
|
|
break;
|
|
|
|
}
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_SP_DOES_NOT_EXIST, MYF(0),
|
|
|
|
SP_COM_STRING(lex), lex->spname->m_qname.str);
|
2003-02-21 17:37:05 +01:00
|
|
|
goto error;
|
|
|
|
default:
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_SP_DROP_FAILED, MYF(0),
|
|
|
|
SP_COM_STRING(lex), lex->spname->m_qname.str);
|
2003-02-21 17:37:05 +01:00
|
|
|
goto error;
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
}
|
2003-04-23 09:22:54 +02:00
|
|
|
break;
|
Simplistic, experimental framework for Stored Procedures (SPs).
Implements creation and dropping of PROCEDUREs, IN, OUT, and INOUT parameters,
single-statement procedures, rudimentary multi-statement (begin-end) prodedures
(when the client can handle it), and local variables.
Missing most of the embedded SQL language, all attributes, FUNCTIONs, error handling,
reparses procedures at each call (no caching), etc, etc.
Certainly buggy too, but procedures can actually be created and called....
2002-12-08 19:59:22 +01:00
|
|
|
}
|
2003-11-17 18:21:36 +01:00
|
|
|
case SQLCOM_SHOW_CREATE_PROC:
|
|
|
|
{
|
2004-02-17 17:36:53 +01:00
|
|
|
if (lex->spname->m_name.length > NAME_LEN)
|
2003-11-17 18:21:36 +01:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_TOO_LONG_IDENT, MYF(0), lex->spname->m_name.str);
|
2003-11-17 18:21:36 +01:00
|
|
|
goto error;
|
|
|
|
}
|
2004-10-20 03:04:37 +02:00
|
|
|
if (sp_show_create_procedure(thd, lex->spname) != SP_OK)
|
2003-11-20 15:07:22 +01:00
|
|
|
{ /* We don't distinguish between errors for now */
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_SP_DOES_NOT_EXIST, MYF(0),
|
|
|
|
SP_COM_STRING(lex), lex->spname->m_name.str);
|
2003-11-17 18:21:36 +01:00
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_SHOW_CREATE_FUNC:
|
|
|
|
{
|
2004-02-17 17:36:53 +01:00
|
|
|
if (lex->spname->m_name.length > NAME_LEN)
|
2003-11-17 18:21:36 +01:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_TOO_LONG_IDENT, MYF(0), lex->spname->m_name.str);
|
2003-11-17 18:21:36 +01:00
|
|
|
goto error;
|
|
|
|
}
|
2004-10-20 03:04:37 +02:00
|
|
|
if (sp_show_create_function(thd, lex->spname) != SP_OK)
|
2003-11-20 15:07:22 +01:00
|
|
|
{ /* We don't distinguish between errors for now */
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_SP_DOES_NOT_EXIST, MYF(0),
|
|
|
|
SP_COM_STRING(lex), lex->spname->m_name.str);
|
2003-11-17 18:21:36 +01:00
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_SHOW_STATUS_PROC:
|
|
|
|
{
|
2003-11-20 15:07:22 +01:00
|
|
|
res= sp_show_status_procedure(thd, (lex->wild ?
|
2003-11-17 18:21:36 +01:00
|
|
|
lex->wild->ptr() : NullS));
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_SHOW_STATUS_FUNC:
|
|
|
|
{
|
2003-11-20 15:07:22 +01:00
|
|
|
res= sp_show_status_function(thd, (lex->wild ?
|
2003-11-17 18:21:36 +01:00
|
|
|
lex->wild->ptr() : NullS));
|
|
|
|
break;
|
|
|
|
}
|
2005-11-17 11:11:48 +01:00
|
|
|
#ifndef DBUG_OFF
|
|
|
|
case SQLCOM_SHOW_PROC_CODE:
|
|
|
|
case SQLCOM_SHOW_FUNC_CODE:
|
|
|
|
{
|
|
|
|
sp_head *sp;
|
|
|
|
|
|
|
|
if (lex->spname->m_name.length > NAME_LEN)
|
|
|
|
{
|
|
|
|
my_error(ER_TOO_LONG_IDENT, MYF(0), lex->spname->m_name.str);
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
if (lex->sql_command == SQLCOM_SHOW_PROC_CODE)
|
2005-11-23 01:49:44 +01:00
|
|
|
sp= sp_find_routine(thd, TYPE_ENUM_PROCEDURE, lex->spname,
|
|
|
|
&thd->sp_proc_cache, FALSE);
|
2005-11-17 11:11:48 +01:00
|
|
|
else
|
2005-11-23 01:49:44 +01:00
|
|
|
sp= sp_find_routine(thd, TYPE_ENUM_FUNCTION, lex->spname,
|
|
|
|
&thd->sp_func_cache, FALSE);
|
2006-01-26 17:26:25 +01:00
|
|
|
if (!sp || sp->show_routine_code(thd))
|
2005-11-18 16:30:27 +01:00
|
|
|
{
|
|
|
|
/* We don't distinguish between errors for now */
|
2005-11-17 11:11:48 +01:00
|
|
|
my_error(ER_SP_DOES_NOT_EXIST, MYF(0),
|
|
|
|
SP_COM_STRING(lex), lex->spname->m_name.str);
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
#endif // ifndef DBUG_OFF
|
2004-07-16 00:15:55 +02:00
|
|
|
case SQLCOM_CREATE_VIEW:
|
|
|
|
{
|
2005-10-13 11:12:17 +02:00
|
|
|
if (end_active_trans(thd))
|
|
|
|
goto error;
|
|
|
|
|
2005-01-17 00:18:08 +01:00
|
|
|
if (!(res= mysql_create_view(thd, thd->lex->create_view_mode)) &&
|
|
|
|
mysql_bin_log.is_open())
|
|
|
|
{
|
2005-09-14 09:53:09 +02:00
|
|
|
String buff;
|
2005-09-21 07:29:47 +02:00
|
|
|
const LEX_STRING command[3]=
|
|
|
|
{{(char *)STRING_WITH_LEN("CREATE ")},
|
|
|
|
{(char *)STRING_WITH_LEN("ALTER ")},
|
|
|
|
{(char *)STRING_WITH_LEN("CREATE OR REPLACE ")}};
|
2005-02-22 11:43:49 +01:00
|
|
|
thd->clear_error();
|
2005-09-14 09:53:09 +02:00
|
|
|
|
|
|
|
buff.append(command[thd->lex->create_view_mode].str,
|
|
|
|
command[thd->lex->create_view_mode].length);
|
|
|
|
view_store_options(thd, first_table, &buff);
|
2005-11-20 19:47:07 +01:00
|
|
|
buff.append(STRING_WITH_LEN("VIEW "));
|
2005-10-06 16:54:43 +02:00
|
|
|
/* Test if user supplied a db (ie: we did not use thd->db) */
|
|
|
|
if (first_table->db != thd->db && first_table->db[0])
|
2005-09-14 09:53:09 +02:00
|
|
|
{
|
|
|
|
append_identifier(thd, &buff, first_table->db,
|
|
|
|
first_table->db_length);
|
|
|
|
buff.append('.');
|
|
|
|
}
|
|
|
|
append_identifier(thd, &buff, first_table->table_name,
|
|
|
|
first_table->table_name_length);
|
2005-11-20 19:47:07 +01:00
|
|
|
buff.append(STRING_WITH_LEN(" AS "));
|
2005-09-14 09:53:09 +02:00
|
|
|
buff.append(first_table->source.str, first_table->source.length);
|
|
|
|
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::STMT_QUERY_TYPE,
|
|
|
|
buff.ptr(), buff.length(), FALSE, FALSE);
|
2005-01-17 00:18:08 +01:00
|
|
|
}
|
2004-07-16 00:15:55 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_DROP_VIEW:
|
|
|
|
{
|
2004-10-20 03:04:37 +02:00
|
|
|
if (check_table_access(thd, DROP_ACL, all_tables, 0) ||
|
|
|
|
end_active_trans(thd))
|
|
|
|
goto error;
|
2005-01-17 00:18:08 +01:00
|
|
|
if (!(res= mysql_drop_view(thd, first_table, thd->lex->drop_mode)) &&
|
|
|
|
mysql_bin_log.is_open())
|
|
|
|
{
|
2005-02-22 11:43:49 +01:00
|
|
|
thd->clear_error();
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->binlog_query(THD::STMT_QUERY_TYPE,
|
|
|
|
thd->query, thd->query_length, FALSE, FALSE);
|
2005-01-17 00:18:08 +01:00
|
|
|
}
|
2004-07-16 00:15:55 +02:00
|
|
|
break;
|
|
|
|
}
|
2004-09-07 14:29:46 +02:00
|
|
|
case SQLCOM_CREATE_TRIGGER:
|
|
|
|
{
|
2005-10-13 11:12:17 +02:00
|
|
|
if (end_active_trans(thd))
|
|
|
|
goto error;
|
|
|
|
|
2004-11-24 10:24:02 +01:00
|
|
|
res= mysql_create_or_drop_trigger(thd, all_tables, 1);
|
|
|
|
|
|
|
|
/* We don't care about trigger body after this point */
|
2004-09-07 14:29:46 +02:00
|
|
|
delete lex->sphead;
|
|
|
|
lex->sphead= 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case SQLCOM_DROP_TRIGGER:
|
|
|
|
{
|
2005-10-13 11:12:17 +02:00
|
|
|
if (end_active_trans(thd))
|
|
|
|
goto error;
|
|
|
|
|
2004-09-07 14:29:46 +02:00
|
|
|
res= mysql_create_or_drop_trigger(thd, all_tables, 0);
|
|
|
|
break;
|
|
|
|
}
|
2005-01-16 13:16:23 +01:00
|
|
|
case SQLCOM_XA_START:
|
2005-08-12 21:15:01 +02:00
|
|
|
if (thd->transaction.xid_state.xa_state == XA_IDLE &&
|
|
|
|
thd->lex->xa_opt == XA_RESUME)
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
2005-08-12 21:15:01 +02:00
|
|
|
if (! thd->transaction.xid_state.xid.eq(thd->lex->xid))
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
|
|
|
my_error(ER_XAER_NOTA, MYF(0));
|
|
|
|
break;
|
|
|
|
}
|
2005-08-12 21:15:01 +02:00
|
|
|
thd->transaction.xid_state.xa_state=XA_ACTIVE;
|
2005-01-16 13:16:23 +01:00
|
|
|
send_ok(thd);
|
|
|
|
break;
|
|
|
|
}
|
2005-04-04 00:50:05 +02:00
|
|
|
if (thd->lex->xa_opt != XA_NONE)
|
2005-01-16 13:16:23 +01:00
|
|
|
{ // JOIN is not supported yet. TODO
|
|
|
|
my_error(ER_XAER_INVAL, MYF(0));
|
|
|
|
break;
|
|
|
|
}
|
2005-08-12 21:15:01 +02:00
|
|
|
if (thd->transaction.xid_state.xa_state != XA_NOTR)
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
2005-01-27 22:38:56 +01:00
|
|
|
my_error(ER_XAER_RMFAIL, MYF(0),
|
2005-08-12 21:15:01 +02:00
|
|
|
xa_state_names[thd->transaction.xid_state.xa_state]);
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (thd->active_transaction() || thd->locked_tables)
|
|
|
|
{
|
|
|
|
my_error(ER_XAER_OUTSIDE, MYF(0));
|
|
|
|
break;
|
|
|
|
}
|
2005-08-12 21:15:01 +02:00
|
|
|
if (xid_cache_search(thd->lex->xid))
|
|
|
|
{
|
|
|
|
my_error(ER_XAER_DUPID, MYF(0));
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
DBUG_ASSERT(thd->transaction.xid_state.xid.is_null());
|
|
|
|
thd->transaction.xid_state.xa_state=XA_ACTIVE;
|
|
|
|
thd->transaction.xid_state.xid.set(thd->lex->xid);
|
|
|
|
xid_cache_insert(&thd->transaction.xid_state);
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
thd->options= ((thd->options & ~(OPTION_STATUS_NO_TRANS_UPDATE |
|
|
|
|
OPTION_KEEP_LOG)) |
|
2005-01-16 13:16:23 +01:00
|
|
|
OPTION_BEGIN);
|
|
|
|
thd->server_status|= SERVER_STATUS_IN_TRANS;
|
|
|
|
send_ok(thd);
|
|
|
|
break;
|
|
|
|
case SQLCOM_XA_END:
|
|
|
|
/* fake it */
|
|
|
|
if (thd->lex->xa_opt != XA_NONE)
|
|
|
|
{ // SUSPEND and FOR MIGRATE are not supported yet. TODO
|
|
|
|
my_error(ER_XAER_INVAL, MYF(0));
|
|
|
|
break;
|
|
|
|
}
|
2005-08-12 21:15:01 +02:00
|
|
|
if (thd->transaction.xid_state.xa_state != XA_ACTIVE)
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
2005-01-27 22:38:56 +01:00
|
|
|
my_error(ER_XAER_RMFAIL, MYF(0),
|
2005-08-12 21:15:01 +02:00
|
|
|
xa_state_names[thd->transaction.xid_state.xa_state]);
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
|
|
|
}
|
2005-08-12 21:15:01 +02:00
|
|
|
if (!thd->transaction.xid_state.xid.eq(thd->lex->xid))
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
|
|
|
my_error(ER_XAER_NOTA, MYF(0));
|
|
|
|
break;
|
|
|
|
}
|
2005-08-12 21:15:01 +02:00
|
|
|
thd->transaction.xid_state.xa_state=XA_IDLE;
|
2005-01-16 13:16:23 +01:00
|
|
|
send_ok(thd);
|
|
|
|
break;
|
|
|
|
case SQLCOM_XA_PREPARE:
|
2005-08-12 21:15:01 +02:00
|
|
|
if (thd->transaction.xid_state.xa_state != XA_IDLE)
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
2005-01-27 22:38:56 +01:00
|
|
|
my_error(ER_XAER_RMFAIL, MYF(0),
|
2005-08-12 21:15:01 +02:00
|
|
|
xa_state_names[thd->transaction.xid_state.xa_state]);
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
|
|
|
}
|
2005-08-12 21:15:01 +02:00
|
|
|
if (!thd->transaction.xid_state.xid.eq(thd->lex->xid))
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
|
|
|
my_error(ER_XAER_NOTA, MYF(0));
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (ha_prepare(thd))
|
|
|
|
{
|
|
|
|
my_error(ER_XA_RBROLLBACK, MYF(0));
|
2005-08-12 21:15:01 +02:00
|
|
|
xid_cache_delete(&thd->transaction.xid_state);
|
|
|
|
thd->transaction.xid_state.xa_state=XA_NOTR;
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
|
|
|
}
|
2005-08-12 21:15:01 +02:00
|
|
|
thd->transaction.xid_state.xa_state=XA_PREPARED;
|
2005-01-16 13:16:23 +01:00
|
|
|
send_ok(thd);
|
|
|
|
break;
|
|
|
|
case SQLCOM_XA_COMMIT:
|
2005-08-12 21:15:01 +02:00
|
|
|
if (!thd->transaction.xid_state.xid.eq(thd->lex->xid))
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
2005-08-12 21:15:01 +02:00
|
|
|
XID_STATE *xs=xid_cache_search(thd->lex->xid);
|
|
|
|
if (!xs || xs->in_thd)
|
2005-01-16 13:16:23 +01:00
|
|
|
my_error(ER_XAER_NOTA, MYF(0));
|
2005-03-16 08:42:06 +01:00
|
|
|
else
|
2005-08-12 21:15:01 +02:00
|
|
|
{
|
|
|
|
ha_commit_or_rollback_by_xid(thd->lex->xid, 1);
|
|
|
|
xid_cache_delete(xs);
|
2005-03-16 08:42:06 +01:00
|
|
|
send_ok(thd);
|
2005-08-12 21:15:01 +02:00
|
|
|
}
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
|
|
|
}
|
2005-08-12 21:15:01 +02:00
|
|
|
if (thd->transaction.xid_state.xa_state == XA_IDLE &&
|
2005-06-05 16:01:20 +02:00
|
|
|
thd->lex->xa_opt == XA_ONE_PHASE)
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
2005-01-27 22:38:56 +01:00
|
|
|
int r;
|
|
|
|
if ((r= ha_commit(thd)))
|
|
|
|
my_error(r == 1 ? ER_XA_RBROLLBACK : ER_XAER_RMERR, MYF(0));
|
2005-01-16 13:16:23 +01:00
|
|
|
else
|
|
|
|
send_ok(thd);
|
|
|
|
}
|
2005-08-12 21:15:01 +02:00
|
|
|
else if (thd->transaction.xid_state.xa_state == XA_PREPARED &&
|
2005-06-05 16:01:20 +02:00
|
|
|
thd->lex->xa_opt == XA_NONE)
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
2005-04-12 17:15:54 +02:00
|
|
|
if (wait_if_global_read_lock(thd, 0, 0))
|
|
|
|
{
|
|
|
|
ha_rollback(thd);
|
2005-01-16 13:16:23 +01:00
|
|
|
my_error(ER_XAER_RMERR, MYF(0));
|
2005-04-12 17:15:54 +02:00
|
|
|
}
|
2005-01-16 13:16:23 +01:00
|
|
|
else
|
2005-04-12 17:15:54 +02:00
|
|
|
{
|
|
|
|
if (ha_commit_one_phase(thd, 1))
|
|
|
|
my_error(ER_XAER_RMERR, MYF(0));
|
|
|
|
else
|
|
|
|
send_ok(thd);
|
|
|
|
start_waiting_global_read_lock(thd);
|
|
|
|
}
|
2005-01-16 13:16:23 +01:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2005-01-27 22:38:56 +01:00
|
|
|
my_error(ER_XAER_RMFAIL, MYF(0),
|
2005-08-12 21:15:01 +02:00
|
|
|
xa_state_names[thd->transaction.xid_state.xa_state]);
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
|
|
|
}
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
thd->options&= ~(OPTION_BEGIN | OPTION_STATUS_NO_TRANS_UPDATE |
|
|
|
|
OPTION_KEEP_LOG);
|
2005-01-16 13:16:23 +01:00
|
|
|
thd->server_status&= ~SERVER_STATUS_IN_TRANS;
|
2005-08-12 21:15:01 +02:00
|
|
|
xid_cache_delete(&thd->transaction.xid_state);
|
|
|
|
thd->transaction.xid_state.xa_state=XA_NOTR;
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
|
|
|
case SQLCOM_XA_ROLLBACK:
|
2005-08-12 21:15:01 +02:00
|
|
|
if (!thd->transaction.xid_state.xid.eq(thd->lex->xid))
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
2005-08-12 21:15:01 +02:00
|
|
|
XID_STATE *xs=xid_cache_search(thd->lex->xid);
|
|
|
|
if (!xs || xs->in_thd)
|
2005-01-16 13:16:23 +01:00
|
|
|
my_error(ER_XAER_NOTA, MYF(0));
|
2005-03-16 08:42:06 +01:00
|
|
|
else
|
2005-08-12 21:15:01 +02:00
|
|
|
{
|
|
|
|
ha_commit_or_rollback_by_xid(thd->lex->xid, 0);
|
|
|
|
xid_cache_delete(xs);
|
2005-03-16 08:42:06 +01:00
|
|
|
send_ok(thd);
|
2005-08-12 21:15:01 +02:00
|
|
|
}
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
|
|
|
}
|
2005-08-12 21:15:01 +02:00
|
|
|
if (thd->transaction.xid_state.xa_state != XA_IDLE &&
|
|
|
|
thd->transaction.xid_state.xa_state != XA_PREPARED)
|
2005-01-16 13:16:23 +01:00
|
|
|
{
|
2005-01-27 22:38:56 +01:00
|
|
|
my_error(ER_XAER_RMFAIL, MYF(0),
|
2005-08-12 21:15:01 +02:00
|
|
|
xa_state_names[thd->transaction.xid_state.xa_state]);
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (ha_rollback(thd))
|
|
|
|
my_error(ER_XAER_RMERR, MYF(0));
|
|
|
|
else
|
|
|
|
send_ok(thd);
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
thd->options&= ~(OPTION_BEGIN | OPTION_STATUS_NO_TRANS_UPDATE |
|
|
|
|
OPTION_KEEP_LOG);
|
2005-01-16 13:16:23 +01:00
|
|
|
thd->server_status&= ~SERVER_STATUS_IN_TRANS;
|
2005-08-12 21:15:01 +02:00
|
|
|
xid_cache_delete(&thd->transaction.xid_state);
|
|
|
|
thd->transaction.xid_state.xa_state=XA_NOTR;
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
|
|
|
case SQLCOM_XA_RECOVER:
|
2005-02-21 13:47:57 +01:00
|
|
|
res= mysql_xa_recover(thd);
|
2005-01-16 13:16:23 +01:00
|
|
|
break;
|
2006-05-02 19:56:43 +02:00
|
|
|
case SQLCOM_ALTER_TABLESPACE:
|
|
|
|
if (check_access(thd, ALTER_ACL, thd->db, 0, 1, 0, thd->db ? is_schema_db(thd->db) : 0))
|
|
|
|
break;
|
|
|
|
if (!(res= mysql_alter_tablespace(thd, lex->alter_tablespace_info)))
|
|
|
|
send_ok(thd);
|
|
|
|
break;
|
|
|
|
case SQLCOM_INSTALL_PLUGIN:
|
|
|
|
if (! (res= mysql_install_plugin(thd, &thd->lex->comment,
|
|
|
|
&thd->lex->ident)))
|
|
|
|
send_ok(thd);
|
|
|
|
break;
|
|
|
|
case SQLCOM_UNINSTALL_PLUGIN:
|
|
|
|
if (! (res= mysql_uninstall_plugin(thd, &thd->lex->comment)))
|
|
|
|
send_ok(thd);
|
|
|
|
break;
|
|
|
|
case SQLCOM_BINLOG_BASE64_EVENT:
|
|
|
|
{
|
|
|
|
#ifndef EMBEDDED_LIBRARY
|
|
|
|
mysql_client_binlog_statement(thd);
|
|
|
|
#else /* EMBEDDED_LIBRARY */
|
|
|
|
my_error(ER_OPTION_PREVENTS_STATEMENT, MYF(0), "embedded");
|
|
|
|
#endif /* EMBEDDED_LIBRARY */
|
|
|
|
break;
|
|
|
|
}
|
2005-01-16 13:16:23 +01:00
|
|
|
default:
|
2006-02-24 17:34:15 +01:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2005-01-16 13:16:23 +01:00
|
|
|
DBUG_ASSERT(0); /* Impossible */
|
2006-02-24 17:34:15 +01:00
|
|
|
#endif
|
2002-10-02 12:33:08 +02:00
|
|
|
send_ok(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
break;
|
|
|
|
}
|
2006-05-02 19:56:43 +02:00
|
|
|
|
|
|
|
end:
|
2004-07-07 10:29:39 +02:00
|
|
|
thd->proc_info="query end";
|
2005-08-31 18:08:45 +02:00
|
|
|
|
|
|
|
/*
|
2006-05-02 19:56:43 +02:00
|
|
|
Binlog-related cleanup:
|
2005-08-31 18:08:45 +02:00
|
|
|
Reset system variables temporarily modified by SET ONE SHOT.
|
|
|
|
|
|
|
|
Exception: If this is a SET, do nothing. This is to allow
|
|
|
|
mysqlbinlog to print many SET commands (in this case we want the
|
|
|
|
charset temp setting to live until the real query). This is also
|
|
|
|
needed so that SET CHARACTER_SET_CLIENT... does not cancel itself
|
|
|
|
immediately.
|
|
|
|
*/
|
|
|
|
if (thd->one_shot_set && lex->sql_command != SQLCOM_SET_OPTION)
|
|
|
|
reset_one_shot_variables(thd);
|
2006-04-24 16:39:33 +02:00
|
|
|
#ifdef HAVE_ROW_BASED_REPLICATION
|
2006-05-02 19:56:43 +02:00
|
|
|
thd->reset_current_stmt_binlog_row_based();
|
2006-04-24 16:39:33 +02:00
|
|
|
#endif /*HAVE_ROW_BASED_REPLICATION*/
|
2005-08-31 18:08:45 +02:00
|
|
|
|
2004-05-04 17:18:47 +02:00
|
|
|
/*
|
2005-11-23 23:57:26 +01:00
|
|
|
The return value for ROW_COUNT() is "implementation dependent" if the
|
|
|
|
statement is not DELETE, INSERT or UPDATE, but -1 is what JDBC and ODBC
|
|
|
|
wants.
|
|
|
|
|
|
|
|
We do not change the value for a CALL or EXECUTE statement, so the value
|
|
|
|
generated by the last called (or executed) statement is preserved.
|
2004-05-04 17:18:47 +02:00
|
|
|
*/
|
2005-11-23 23:57:26 +01:00
|
|
|
if (lex->sql_command != SQLCOM_CALL && lex->sql_command != SQLCOM_EXECUTE &&
|
|
|
|
uc_update_queries[lex->sql_command]<2)
|
2004-05-04 17:18:47 +02:00
|
|
|
thd->row_count_func= -1;
|
2006-01-05 23:47:49 +01:00
|
|
|
DBUG_RETURN(res || thd->net.report_error);
|
2000-07-31 21:29:14 +02:00
|
|
|
|
|
|
|
error:
|
2006-05-02 19:56:43 +02:00
|
|
|
res= 1; // would be better to set res=1 before "goto error"
|
|
|
|
goto end;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2003-05-28 01:06:56 +02:00
|
|
|
/*
|
|
|
|
Check grants for commands which work only with one table and all other
|
2004-12-09 11:31:46 +01:00
|
|
|
tables belonging to subselects or implicitly opened tables.
|
2003-05-28 01:06:56 +02:00
|
|
|
|
2003-07-01 21:40:59 +02:00
|
|
|
SYNOPSIS
|
2004-02-11 00:06:46 +01:00
|
|
|
check_one_table_access()
|
|
|
|
thd Thread handler
|
2004-10-07 00:45:06 +02:00
|
|
|
privilege requested privilege
|
2004-07-16 00:15:55 +02:00
|
|
|
all_tables global table list of query
|
2003-05-28 01:06:56 +02:00
|
|
|
|
|
|
|
RETURN
|
|
|
|
0 - OK
|
2004-02-11 00:06:46 +01:00
|
|
|
1 - access denied, error is sent to client
|
2003-05-28 01:06:56 +02:00
|
|
|
*/
|
|
|
|
|
2004-09-03 20:43:04 +02:00
|
|
|
bool check_one_table_access(THD *thd, ulong privilege, TABLE_LIST *all_tables)
|
2003-05-28 01:06:56 +02:00
|
|
|
{
|
2006-05-26 10:47:53 +02:00
|
|
|
Security_context * backup_ctx= thd->security_ctx;
|
|
|
|
|
|
|
|
/* we need to switch to the saved context (if any) */
|
|
|
|
if (all_tables->security_ctx)
|
|
|
|
thd->security_ctx= all_tables->security_ctx;
|
|
|
|
|
2004-07-16 00:15:55 +02:00
|
|
|
if (check_access(thd, privilege, all_tables->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&all_tables->grant.privilege, 0, 0,
|
|
|
|
test(all_tables->schema_table)))
|
2006-05-26 10:47:53 +02:00
|
|
|
goto deny;
|
2003-05-28 01:06:56 +02:00
|
|
|
|
2004-04-12 02:26:32 +02:00
|
|
|
/* Show only 1 table for check_grant */
|
2004-07-16 00:15:55 +02:00
|
|
|
if (grant_option && check_grant(thd, privilege, all_tables, 0, 1, 0))
|
2006-05-26 10:47:53 +02:00
|
|
|
goto deny;
|
|
|
|
|
|
|
|
thd->security_ctx= backup_ctx;
|
2003-05-28 01:06:56 +02:00
|
|
|
|
2004-12-09 11:31:46 +01:00
|
|
|
/* Check rights on tables of subselects and implictly opened tables */
|
2004-04-10 00:14:32 +02:00
|
|
|
TABLE_LIST *subselects_tables;
|
2004-07-16 00:15:55 +02:00
|
|
|
if ((subselects_tables= all_tables->next_global))
|
2003-12-19 15:25:50 +01:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
if ((check_table_access(thd, SELECT_ACL, subselects_tables, 0)))
|
2006-05-26 10:47:53 +02:00
|
|
|
goto deny;
|
2003-12-19 15:25:50 +01:00
|
|
|
}
|
|
|
|
return 0;
|
2006-05-26 10:47:53 +02:00
|
|
|
|
|
|
|
deny:
|
|
|
|
thd->security_ctx= backup_ctx;
|
|
|
|
return 1;
|
2003-05-28 01:06:56 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
/****************************************************************************
|
2002-06-12 14:04:18 +02:00
|
|
|
Get the user (global) and database privileges for all used tables
|
2003-04-28 09:32:56 +02:00
|
|
|
|
|
|
|
NOTES
|
|
|
|
The idea of EXTRA_ACL is that one will be granted access to the table if
|
|
|
|
one has the asked privilege on any column combination of the table; For
|
|
|
|
example to be able to check a table one needs to have SELECT privilege on
|
|
|
|
any column of the table.
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
0 ok
|
|
|
|
1 If we can't get the privileges and we don't use table/column grants.
|
|
|
|
|
|
|
|
save_priv In this we store global and db level grants for the table
|
|
|
|
Note that we don't store db level grants if the global grants
|
2003-05-07 22:15:46 +02:00
|
|
|
is enough to satisfy the request and the global grants contains
|
|
|
|
a SELECT grant.
|
2000-07-31 21:29:14 +02:00
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
bool
|
2002-06-12 14:04:18 +02:00
|
|
|
check_access(THD *thd, ulong want_access, const char *db, ulong *save_priv,
|
2005-09-13 13:07:38 +02:00
|
|
|
bool dont_check_global_grants, bool no_errors, bool schema_db)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2005-09-20 20:20:38 +02:00
|
|
|
Security_context *sctx= thd->security_ctx;
|
2004-04-05 12:56:05 +02:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
|
|
|
ulong db_access;
|
2005-03-30 10:43:24 +02:00
|
|
|
bool db_is_pattern= test(want_access & GRANT_ACL);
|
2004-04-05 12:56:05 +02:00
|
|
|
#endif
|
|
|
|
ulong dummy;
|
2004-09-03 20:43:04 +02:00
|
|
|
DBUG_ENTER("check_access");
|
|
|
|
DBUG_PRINT("enter",("db: %s want_access: %lu master_access: %lu",
|
2005-09-15 21:29:07 +02:00
|
|
|
db ? db : "", want_access, sctx->master_access));
|
2000-07-31 21:29:14 +02:00
|
|
|
if (save_priv)
|
|
|
|
*save_priv=0;
|
|
|
|
else
|
|
|
|
save_priv= &dummy;
|
|
|
|
|
2001-09-18 05:05:55 +02:00
|
|
|
if ((!db || !db[0]) && !thd->db && !dont_check_global_grants)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-09-03 20:43:04 +02:00
|
|
|
DBUG_PRINT("error",("No database"));
|
2001-12-22 14:13:31 +01:00
|
|
|
if (!no_errors)
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_NO_DB_ERROR, ER(ER_NO_DB_ERROR),
|
|
|
|
MYF(0)); /* purecov: tested */
|
2002-06-12 14:04:18 +02:00
|
|
|
DBUG_RETURN(TRUE); /* purecov: tested */
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
2005-09-13 13:07:38 +02:00
|
|
|
if (schema_db)
|
|
|
|
{
|
|
|
|
if (want_access & ~(SELECT_ACL | EXTRA_ACL))
|
|
|
|
{
|
|
|
|
if (!no_errors)
|
2005-10-08 02:37:23 +02:00
|
|
|
{
|
|
|
|
const char *db_name= db ? db : thd->db;
|
2005-09-13 13:07:38 +02:00
|
|
|
my_error(ER_DBACCESS_DENIED_ERROR, MYF(0),
|
2005-10-08 02:37:23 +02:00
|
|
|
sctx->priv_user, sctx->priv_host, db_name);
|
|
|
|
}
|
2005-09-13 13:07:38 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
*save_priv= SELECT_ACL;
|
|
|
|
DBUG_RETURN(FALSE);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2003-12-19 15:25:50 +01:00
|
|
|
#ifdef NO_EMBEDDED_ACCESS_CHECKS
|
|
|
|
DBUG_RETURN(0);
|
|
|
|
#else
|
2005-09-15 21:29:07 +02:00
|
|
|
if ((sctx->master_access & want_access) == want_access)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2003-05-07 22:59:24 +02:00
|
|
|
/*
|
|
|
|
If we don't have a global SELECT privilege, we have to get the database
|
|
|
|
specific access rights to be able to handle queries of type
|
|
|
|
UPDATE t1 SET a=1 WHERE b > 0
|
|
|
|
*/
|
2005-09-15 21:29:07 +02:00
|
|
|
db_access= sctx->db_access;
|
|
|
|
if (!(sctx->master_access & SELECT_ACL) &&
|
2005-03-30 10:43:24 +02:00
|
|
|
(db && (!thd->db || db_is_pattern || strcmp(db,thd->db))))
|
2005-09-15 21:29:07 +02:00
|
|
|
db_access=acl_get(sctx->host, sctx->ip, sctx->priv_user, db,
|
|
|
|
db_is_pattern);
|
|
|
|
*save_priv=sctx->master_access | db_access;
|
2002-06-12 14:04:18 +02:00
|
|
|
DBUG_RETURN(FALSE);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2005-09-15 21:29:07 +02:00
|
|
|
if (((want_access & ~sctx->master_access) & ~(DB_ACLS | EXTRA_ACL)) ||
|
2001-01-14 11:25:30 +01:00
|
|
|
! db && dont_check_global_grants)
|
2000-07-31 21:29:14 +02:00
|
|
|
{ // We can never grant this
|
2004-09-03 20:43:04 +02:00
|
|
|
DBUG_PRINT("error",("No possible access"));
|
2001-12-22 14:13:31 +01:00
|
|
|
if (!no_errors)
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_ACCESS_DENIED_ERROR, MYF(0),
|
2005-09-15 21:29:07 +02:00
|
|
|
sctx->priv_user,
|
|
|
|
sctx->priv_host,
|
2004-11-13 18:35:51 +01:00
|
|
|
(thd->password ?
|
|
|
|
ER(ER_YES) :
|
|
|
|
ER(ER_NO))); /* purecov: tested */
|
2002-06-12 14:04:18 +02:00
|
|
|
DBUG_RETURN(TRUE); /* purecov: tested */
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
if (db == any_db)
|
2002-06-12 14:04:18 +02:00
|
|
|
DBUG_RETURN(FALSE); // Allow select on anything
|
2001-12-06 13:10:51 +01:00
|
|
|
|
2005-03-30 10:43:24 +02:00
|
|
|
if (db && (!thd->db || db_is_pattern || strcmp(db,thd->db)))
|
2005-09-15 21:29:07 +02:00
|
|
|
db_access= acl_get(sctx->host, sctx->ip, sctx->priv_user, db,
|
|
|
|
db_is_pattern);
|
2000-07-31 21:29:14 +02:00
|
|
|
else
|
2005-09-15 21:29:07 +02:00
|
|
|
db_access= sctx->db_access;
|
2004-10-26 18:30:01 +02:00
|
|
|
DBUG_PRINT("info",("db_access: %lu", db_access));
|
2004-07-07 10:29:39 +02:00
|
|
|
/* Remove SHOW attribute and access rights we already have */
|
2005-09-15 21:29:07 +02:00
|
|
|
want_access &= ~(sctx->master_access | EXTRA_ACL);
|
2004-09-03 20:43:04 +02:00
|
|
|
DBUG_PRINT("info",("db_access: %lu want_access: %lu",
|
|
|
|
db_access, want_access));
|
2005-09-15 21:29:07 +02:00
|
|
|
db_access= ((*save_priv=(db_access | sctx->master_access)) & want_access);
|
2001-01-14 11:25:30 +01:00
|
|
|
|
|
|
|
/* grant_option is set if there exists a single table or column grant */
|
2000-07-31 21:29:14 +02:00
|
|
|
if (db_access == want_access ||
|
2004-09-03 20:43:04 +02:00
|
|
|
(grant_option && !dont_check_global_grants &&
|
2004-12-23 11:46:24 +01:00
|
|
|
!(want_access & ~(db_access | TABLE_ACLS | PROC_ACLS))))
|
2002-06-12 14:04:18 +02:00
|
|
|
DBUG_RETURN(FALSE); /* Ok */
|
2004-09-03 20:43:04 +02:00
|
|
|
|
|
|
|
DBUG_PRINT("error",("Access denied"));
|
2001-12-22 14:13:31 +01:00
|
|
|
if (!no_errors)
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_DBACCESS_DENIED_ERROR, MYF(0),
|
2005-09-15 21:29:07 +02:00
|
|
|
sctx->priv_user, sctx->priv_host,
|
2004-11-13 18:35:51 +01:00
|
|
|
(db ? db : (thd->db ?
|
|
|
|
thd->db :
|
|
|
|
"unknown"))); /* purecov: tested */
|
2002-06-12 14:04:18 +02:00
|
|
|
DBUG_RETURN(TRUE); /* purecov: tested */
|
2003-12-19 15:25:50 +01:00
|
|
|
#endif /* NO_EMBEDDED_ACCESS_CHECKS */
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2003-05-14 00:55:23 +02:00
|
|
|
/*
|
|
|
|
check for global access and give descriptive error message if it fails
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
check_global_access()
|
|
|
|
thd Thread handler
|
|
|
|
want_access Use should have any of these global rights
|
|
|
|
|
|
|
|
WARNING
|
2004-10-07 00:45:06 +02:00
|
|
|
One gets access right if one has ANY of the rights in want_access
|
2003-05-14 00:55:23 +02:00
|
|
|
This is useful as one in most cases only need one global right,
|
|
|
|
but in some case we want to check if the user has SUPER or
|
|
|
|
REPL_CLIENT_ACL rights.
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
0 ok
|
|
|
|
1 Access denied. In this case an error is sent to the client
|
|
|
|
*/
|
2002-06-12 14:04:18 +02:00
|
|
|
|
|
|
|
bool check_global_access(THD *thd, ulong want_access)
|
2001-03-21 00:02:22 +01:00
|
|
|
{
|
2003-12-19 15:25:50 +01:00
|
|
|
#ifdef NO_EMBEDDED_ACCESS_CHECKS
|
|
|
|
return 0;
|
|
|
|
#else
|
2002-06-12 14:04:18 +02:00
|
|
|
char command[128];
|
2005-09-15 21:29:07 +02:00
|
|
|
if ((thd->security_ctx->master_access & want_access))
|
2002-06-12 14:04:18 +02:00
|
|
|
return 0;
|
|
|
|
get_privilege_desc(command, sizeof(command), want_access);
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_SPECIFIC_ACCESS_DENIED_ERROR, MYF(0), command);
|
2002-06-12 14:04:18 +02:00
|
|
|
return 1;
|
2003-12-19 15:25:50 +01:00
|
|
|
#endif /* NO_EMBEDDED_ACCESS_CHECKS */
|
2001-03-21 00:02:22 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
/*
|
2002-06-12 14:04:18 +02:00
|
|
|
Check the privilege for all used tables. Table privileges are cached
|
|
|
|
in the table list for GRANT checking
|
2000-07-31 21:29:14 +02:00
|
|
|
*/
|
|
|
|
|
2001-04-11 13:04:03 +02:00
|
|
|
bool
|
2002-06-12 14:04:18 +02:00
|
|
|
check_table_access(THD *thd, ulong want_access,TABLE_LIST *tables,
|
2001-12-22 14:13:31 +01:00
|
|
|
bool no_errors)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2002-06-12 14:04:18 +02:00
|
|
|
uint found=0;
|
|
|
|
ulong found_access=0;
|
2005-10-27 23:18:23 +02:00
|
|
|
TABLE_LIST *org_tables= tables;
|
|
|
|
TABLE_LIST *first_not_own_table= thd->lex->first_not_own_table();
|
2006-05-26 10:47:53 +02:00
|
|
|
Security_context *sctx= thd->security_ctx, *backup_ctx= thd->security_ctx;
|
2005-10-27 23:18:23 +02:00
|
|
|
/*
|
2005-10-27 23:56:44 +02:00
|
|
|
The check that first_not_own_table is not reached is for the case when
|
|
|
|
the given table list refers to the list for prelocking (contains tables
|
|
|
|
of other queries). For simple queries first_not_own_table is 0.
|
2005-10-27 23:18:23 +02:00
|
|
|
*/
|
2006-01-05 23:47:49 +01:00
|
|
|
for (; tables != first_not_own_table; tables= tables->next_global)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2006-05-26 10:47:53 +02:00
|
|
|
if (tables->security_ctx)
|
|
|
|
sctx= tables->security_ctx;
|
|
|
|
else
|
|
|
|
sctx= backup_ctx;
|
|
|
|
|
2005-09-13 13:07:38 +02:00
|
|
|
if (tables->schema_table &&
|
2005-10-07 11:16:44 +02:00
|
|
|
(want_access & ~(SELECT_ACL | EXTRA_ACL | FILE_ACL)))
|
2005-09-13 13:07:38 +02:00
|
|
|
{
|
|
|
|
if (!no_errors)
|
|
|
|
my_error(ER_DBACCESS_DENIED_ERROR, MYF(0),
|
2006-05-26 10:47:53 +02:00
|
|
|
sctx->priv_user, sctx->priv_host,
|
2005-09-13 13:07:38 +02:00
|
|
|
information_schema_name.str);
|
|
|
|
return TRUE;
|
|
|
|
}
|
2005-10-27 23:18:23 +02:00
|
|
|
/*
|
|
|
|
Register access for view underlying table.
|
|
|
|
Remove SHOW_VIEW_ACL, because it will be checked during making view
|
|
|
|
*/
|
|
|
|
tables->grant.orig_want_privilege= (want_access & ~SHOW_VIEW_ACL);
|
2006-05-26 10:47:53 +02:00
|
|
|
if (tables->derived || tables->schema_table ||
|
2005-01-06 12:00:13 +01:00
|
|
|
(tables->table && (int)tables->table->s->tmp_table) ||
|
2004-10-21 20:18:00 +02:00
|
|
|
my_tz_check_n_skip_implicit_tables(&tables,
|
|
|
|
thd->lex->time_zone_tables_used))
|
2002-11-11 13:04:50 +01:00
|
|
|
continue;
|
2006-05-26 10:47:53 +02:00
|
|
|
thd->security_ctx= sctx;
|
|
|
|
if ((sctx->master_access & want_access) ==
|
2005-09-15 21:29:07 +02:00
|
|
|
(want_access & ~EXTRA_ACL) &&
|
2000-08-29 11:31:01 +02:00
|
|
|
thd->db)
|
2000-07-31 21:29:14 +02:00
|
|
|
tables->grant.privilege= want_access;
|
2000-08-21 02:07:54 +02:00
|
|
|
else if (tables->db && tables->db == thd->db)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
if (found && !grant_option) // db already checked
|
|
|
|
tables->grant.privilege=found_access;
|
|
|
|
else
|
|
|
|
{
|
2001-12-22 14:13:31 +01:00
|
|
|
if (check_access(thd,want_access,tables->db,&tables->grant.privilege,
|
2005-09-13 13:07:38 +02:00
|
|
|
0, no_errors, test(tables->schema_table)))
|
2006-05-26 10:47:53 +02:00
|
|
|
goto deny; // Access denied
|
2000-07-31 21:29:14 +02:00
|
|
|
found_access=tables->grant.privilege;
|
2000-08-21 02:07:54 +02:00
|
|
|
found=1;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
}
|
2002-11-12 13:40:32 +01:00
|
|
|
else if (check_access(thd,want_access,tables->db,&tables->grant.privilege,
|
2005-09-13 13:07:38 +02:00
|
|
|
0, no_errors, test(tables->schema_table)))
|
2006-05-26 10:47:53 +02:00
|
|
|
goto deny;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2006-05-26 10:47:53 +02:00
|
|
|
thd->security_ctx= backup_ctx;
|
2000-07-31 21:29:14 +02:00
|
|
|
if (grant_option)
|
2001-04-11 13:04:03 +02:00
|
|
|
return check_grant(thd,want_access & ~EXTRA_ACL,org_tables,
|
2004-04-10 00:14:32 +02:00
|
|
|
test(want_access & EXTRA_ACL), UINT_MAX, no_errors);
|
2000-07-31 21:29:14 +02:00
|
|
|
return FALSE;
|
2006-05-26 10:47:53 +02:00
|
|
|
deny:
|
|
|
|
thd->security_ctx= backup_ctx;
|
|
|
|
return TRUE;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
2004-09-03 20:43:04 +02:00
|
|
|
|
2004-12-23 11:46:24 +01:00
|
|
|
bool
|
2005-05-17 20:54:20 +02:00
|
|
|
check_routine_access(THD *thd, ulong want_access,char *db, char *name,
|
|
|
|
bool is_proc, bool no_errors)
|
2004-12-23 11:46:24 +01:00
|
|
|
{
|
|
|
|
TABLE_LIST tables[1];
|
|
|
|
|
|
|
|
bzero((char *)tables, sizeof(TABLE_LIST));
|
|
|
|
tables->db= db;
|
2005-01-06 12:00:13 +01:00
|
|
|
tables->table_name= tables->alias= name;
|
2004-12-23 11:46:24 +01:00
|
|
|
|
2005-10-11 23:58:22 +02:00
|
|
|
/*
|
|
|
|
The following test is just a shortcut for check_access() (to avoid
|
|
|
|
calculating db_access) under the assumption that it's common to
|
|
|
|
give persons global right to execute all stored SP (but not
|
|
|
|
necessary to create them).
|
|
|
|
*/
|
|
|
|
if ((thd->security_ctx->master_access & want_access) == want_access)
|
2004-12-23 11:46:24 +01:00
|
|
|
tables->grant.privilege= want_access;
|
|
|
|
else if (check_access(thd,want_access,db,&tables->grant.privilege,
|
2005-10-11 23:58:22 +02:00
|
|
|
0, no_errors, 0))
|
2004-12-23 11:46:24 +01:00
|
|
|
return TRUE;
|
|
|
|
|
2004-12-23 12:42:57 +01:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2004-12-23 11:46:24 +01:00
|
|
|
if (grant_option)
|
2005-05-17 20:54:20 +02:00
|
|
|
return check_grant_routine(thd, want_access, tables, is_proc, no_errors);
|
2004-12-23 12:42:57 +01:00
|
|
|
#endif
|
2004-12-23 11:46:24 +01:00
|
|
|
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
2005-03-05 12:35:32 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
Check if the routine has any of the routine privileges
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
check_some_routine_access()
|
|
|
|
thd Thread handler
|
|
|
|
db Database name
|
|
|
|
name Routine name
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
0 ok
|
|
|
|
1 error
|
|
|
|
*/
|
|
|
|
|
2005-05-17 20:54:20 +02:00
|
|
|
bool check_some_routine_access(THD *thd, const char *db, const char *name,
|
|
|
|
bool is_proc)
|
2005-03-05 12:35:32 +01:00
|
|
|
{
|
|
|
|
ulong save_priv;
|
2005-09-15 21:29:07 +02:00
|
|
|
if (thd->security_ctx->master_access & SHOW_PROC_ACLS)
|
2005-03-05 12:35:32 +01:00
|
|
|
return FALSE;
|
2005-09-13 13:07:38 +02:00
|
|
|
/*
|
|
|
|
There are no routines in information_schema db. So we can safely
|
|
|
|
pass zero to last paramter of check_access function
|
|
|
|
*/
|
|
|
|
if (!check_access(thd, SHOW_PROC_ACLS, db, &save_priv, 0, 1, 0) ||
|
2005-03-05 12:35:32 +01:00
|
|
|
(save_priv & SHOW_PROC_ACLS))
|
|
|
|
return FALSE;
|
2005-05-17 20:54:20 +02:00
|
|
|
return check_routine_level_acl(thd, db, name, is_proc);
|
2005-03-05 12:35:32 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2004-09-03 20:43:04 +02:00
|
|
|
/*
|
|
|
|
Check if the given table has any of the asked privileges
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
check_some_access()
|
|
|
|
thd Thread handler
|
|
|
|
want_access Bitmap of possible privileges to check for
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
0 ok
|
|
|
|
1 error
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
|
|
bool check_some_access(THD *thd, ulong want_access, TABLE_LIST *table)
|
|
|
|
{
|
|
|
|
ulong access;
|
|
|
|
DBUG_ENTER("check_some_access");
|
|
|
|
|
|
|
|
/* This loop will work as long as we have less than 32 privileges */
|
|
|
|
for (access= 1; access < want_access ; access<<= 1)
|
|
|
|
{
|
|
|
|
if (access & want_access)
|
|
|
|
{
|
|
|
|
if (!check_access(thd, access, table->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&table->grant.privilege, 0, 1,
|
|
|
|
test(table->schema_table)) &&
|
2004-09-03 20:43:04 +02:00
|
|
|
!grant_option || !check_grant(thd, access, table, 0, 1, 1))
|
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
DBUG_PRINT("exit",("no matching access rights"));
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2004-04-07 23:16:17 +02:00
|
|
|
bool check_merge_table_access(THD *thd, char *db,
|
|
|
|
TABLE_LIST *table_list)
|
2000-09-14 01:39:07 +02:00
|
|
|
{
|
|
|
|
int error=0;
|
|
|
|
if (table_list)
|
|
|
|
{
|
2001-08-29 16:33:41 +02:00
|
|
|
/* Check that all tables use the current database */
|
2000-09-14 01:39:07 +02:00
|
|
|
TABLE_LIST *tmp;
|
2004-07-16 00:15:55 +02:00
|
|
|
for (tmp= table_list; tmp; tmp= tmp->next_local)
|
2001-08-29 16:33:41 +02:00
|
|
|
{
|
|
|
|
if (!tmp->db || !tmp->db[0])
|
|
|
|
tmp->db=db;
|
|
|
|
}
|
2000-09-14 01:39:07 +02:00
|
|
|
error=check_table_access(thd, SELECT_ACL | UPDATE_ACL | DELETE_ACL,
|
2003-09-26 12:33:13 +02:00
|
|
|
table_list,0);
|
2000-09-14 01:39:07 +02:00
|
|
|
}
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2003-09-26 12:33:13 +02:00
|
|
|
|
|
|
|
static bool check_db_used(THD *thd,TABLE_LIST *tables)
|
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
for (; tables; tables= tables->next_global)
|
2003-09-26 12:33:13 +02:00
|
|
|
{
|
|
|
|
if (!tables->db)
|
|
|
|
{
|
|
|
|
if (!(tables->db=thd->db))
|
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_NO_DB_ERROR, ER(ER_NO_DB_ERROR),
|
|
|
|
MYF(0)); /* purecov: tested */
|
2003-09-26 12:33:13 +02:00
|
|
|
return TRUE; /* purecov: tested */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return FALSE;
|
|
|
|
}
|
2000-09-14 01:39:07 +02:00
|
|
|
|
2004-10-22 20:29:06 +02:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
/****************************************************************************
|
|
|
|
Check stack size; Send error if there isn't enough stack to continue
|
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
#if STACK_DIRECTION < 0
|
|
|
|
#define used_stack(A,B) (long) (A - B)
|
|
|
|
#else
|
|
|
|
#define used_stack(A,B) (long) (B - A)
|
|
|
|
#endif
|
|
|
|
|
2004-04-07 03:33:58 +02:00
|
|
|
#ifndef DBUG_OFF
|
|
|
|
long max_stack_used;
|
|
|
|
#endif
|
|
|
|
|
2003-04-08 07:35:13 +02:00
|
|
|
#ifndef EMBEDDED_LIBRARY
|
2005-05-31 12:06:15 +02:00
|
|
|
/*
|
|
|
|
Note: The 'buf' parameter is necessary, even if it is unused here.
|
|
|
|
- fix_fields functions has a "dummy" buffer large enough for the
|
|
|
|
corresponding exec. (Thus we only have to check in fix_fields.)
|
|
|
|
- Passing to check_stack_overrun() prevents the compiler from removing it.
|
|
|
|
*/
|
|
|
|
bool check_stack_overrun(THD *thd, long margin,
|
|
|
|
char *buf __attribute__((unused)))
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
long stack_used;
|
2005-11-23 19:18:10 +01:00
|
|
|
DBUG_ASSERT(thd == current_thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
if ((stack_used=used_stack(thd->thread_stack,(char*) &stack_used)) >=
|
2005-06-07 04:43:59 +02:00
|
|
|
(long) (thread_stack - margin))
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2005-06-22 20:02:44 +02:00
|
|
|
sprintf(errbuff[0],ER(ER_STACK_OVERRUN_NEED_MORE),
|
|
|
|
stack_used,thread_stack,margin);
|
|
|
|
my_message(ER_STACK_OVERRUN_NEED_MORE,errbuff[0],MYF(0));
|
2003-01-30 21:15:44 +01:00
|
|
|
thd->fatal_error();
|
2000-07-31 21:29:14 +02:00
|
|
|
return 1;
|
|
|
|
}
|
2004-04-07 03:33:58 +02:00
|
|
|
#ifndef DBUG_OFF
|
|
|
|
max_stack_used= max(max_stack_used, stack_used);
|
|
|
|
#endif
|
2000-07-31 21:29:14 +02:00
|
|
|
return 0;
|
|
|
|
}
|
2003-04-08 07:35:13 +02:00
|
|
|
#endif /* EMBEDDED_LIBRARY */
|
2000-07-31 21:29:14 +02:00
|
|
|
|
|
|
|
#define MY_YACC_INIT 1000 // Start with big alloc
|
|
|
|
#define MY_YACC_MAX 32000 // Because of 'short'
|
|
|
|
|
2004-06-24 01:57:57 +02:00
|
|
|
bool my_yyoverflow(short **yyss, YYSTYPE **yyvs, ulong *yystacksize)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
LEX *lex=current_lex;
|
2004-06-24 01:57:57 +02:00
|
|
|
ulong old_info=0;
|
2000-07-31 21:29:14 +02:00
|
|
|
if ((uint) *yystacksize >= MY_YACC_MAX)
|
|
|
|
return 1;
|
|
|
|
if (!lex->yacc_yyvs)
|
|
|
|
old_info= *yystacksize;
|
|
|
|
*yystacksize= set_zone((*yystacksize)*2,MY_YACC_INIT,MY_YACC_MAX);
|
|
|
|
if (!(lex->yacc_yyvs= (char*)
|
|
|
|
my_realloc((gptr) lex->yacc_yyvs,
|
|
|
|
*yystacksize*sizeof(**yyvs),
|
|
|
|
MYF(MY_ALLOW_ZERO_PTR | MY_FREE_ON_ERROR))) ||
|
|
|
|
!(lex->yacc_yyss= (char*)
|
|
|
|
my_realloc((gptr) lex->yacc_yyss,
|
|
|
|
*yystacksize*sizeof(**yyss),
|
|
|
|
MYF(MY_ALLOW_ZERO_PTR | MY_FREE_ON_ERROR))))
|
|
|
|
return 1;
|
|
|
|
if (old_info)
|
|
|
|
{ // Copy old info from stack
|
|
|
|
memcpy(lex->yacc_yyss, (gptr) *yyss, old_info*sizeof(**yyss));
|
|
|
|
memcpy(lex->yacc_yyvs, (gptr) *yyvs, old_info*sizeof(**yyvs));
|
|
|
|
}
|
|
|
|
*yyss=(short*) lex->yacc_yyss;
|
|
|
|
*yyvs=(YYSTYPE*) lex->yacc_yyvs;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
2003-02-12 20:55:37 +01:00
|
|
|
Initialize global thd variables needed for query
|
2000-07-31 21:29:14 +02:00
|
|
|
****************************************************************************/
|
|
|
|
|
2002-05-27 19:52:54 +02:00
|
|
|
void
|
2004-07-21 21:44:12 +02:00
|
|
|
mysql_init_query(THD *thd, uchar *buf, uint length)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
DBUG_ENTER("mysql_init_query");
|
2004-07-21 21:44:12 +02:00
|
|
|
lex_start(thd, buf, length);
|
2004-10-14 00:53:59 +02:00
|
|
|
mysql_reset_thd_for_next_command(thd);
|
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Reset THD part responsible for command processing state.
|
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
This needs to be called before execution of every statement
|
|
|
|
(prepared or conventional).
|
|
|
|
|
|
|
|
TODO
|
|
|
|
Make it a method of THD and align its name with the rest of
|
|
|
|
reset/end/start/init methods.
|
|
|
|
Call it after we use THD for queries, not before.
|
|
|
|
*/
|
|
|
|
|
|
|
|
void mysql_reset_thd_for_next_command(THD *thd)
|
|
|
|
{
|
|
|
|
DBUG_ENTER("mysql_reset_thd_for_next_command");
|
2004-10-14 01:37:20 +02:00
|
|
|
thd->free_list= 0;
|
2004-10-14 00:53:59 +02:00
|
|
|
thd->select_number= 1;
|
2002-05-06 23:04:16 +02:00
|
|
|
thd->last_insert_id_used= thd->query_start_used= thd->insert_id_used=0;
|
2005-08-15 17:15:12 +02:00
|
|
|
thd->is_fatal_error= thd->time_zone_used= 0;
|
2003-12-17 16:35:34 +01:00
|
|
|
thd->server_status&= ~ (SERVER_MORE_RESULTS_EXISTS |
|
2004-10-29 18:26:52 +02:00
|
|
|
SERVER_QUERY_NO_INDEX_USED |
|
|
|
|
SERVER_QUERY_NO_GOOD_INDEX_USED);
|
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes
Changes that requires code changes in other code of other storage engines.
(Note that all changes are very straightforward and one should find all issues
by compiling a --debug build and fixing all compiler errors and all
asserts in field.cc while running the test suite),
- New optional handler function introduced: reset()
This is called after every DML statement to make it easy for a handler to
statement specific cleanups.
(The only case it's not called is if force the file to be closed)
- handler::extra(HA_EXTRA_RESET) is removed. Code that was there before
should be moved to handler::reset()
- table->read_set contains a bitmap over all columns that are needed
in the query. read_row() and similar functions only needs to read these
columns
- table->write_set contains a bitmap over all columns that will be updated
in the query. write_row() and update_row() only needs to update these
columns.
The above bitmaps should now be up to date in all context
(including ALTER TABLE, filesort()).
The handler is informed of any changes to the bitmap after
fix_fields() by calling the virtual function
handler::column_bitmaps_signal(). If the handler does caching of
these bitmaps (instead of using table->read_set, table->write_set),
it should redo the caching in this code. as the signal() may be sent
several times, it's probably best to set a variable in the signal
and redo the caching on read_row() / write_row() if the variable was
set.
- Removed the read_set and write_set bitmap objects from the handler class
- Removed all column bit handling functions from the handler class.
(Now one instead uses the normal bitmap functions in my_bitmap.c instead
of handler dedicated bitmap functions)
- field->query_id is removed. One should instead instead check
table->read_set and table->write_set if a field is used in the query.
- handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and
handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now
instead use table->read_set to check for which columns to retrieve.
- If a handler needs to call Field->val() or Field->store() on columns
that are not used in the query, one should install a temporary
all-columns-used map while doing so. For this, we provide the following
functions:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set);
field->val();
dbug_tmp_restore_column_map(table->read_set, old_map);
and similar for the write map:
my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set);
field->val();
dbug_tmp_restore_column_map(table->write_set, old_map);
If this is not done, you will sooner or later hit a DBUG_ASSERT
in the field store() / val() functions.
(For not DBUG binaries, the dbug_tmp_restore_column_map() and
dbug_tmp_restore_column_map() are inline dummy functions and should
be optimized away be the compiler).
- If one needs to temporary set the column map for all binaries (and not
just to avoid the DBUG_ASSERT() in the Field::store() / Field::val()
methods) one should use the functions tmp_use_all_columns() and
tmp_restore_column_map() instead of the above dbug_ variants.
- All 'status' fields in the handler base class (like records,
data_file_length etc) are now stored in a 'stats' struct. This makes
it easier to know what status variables are provided by the base
handler. This requires some trivial variable names in the extra()
function.
- New virtual function handler::records(). This is called to optimize
COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true.
(stats.records is not supposed to be an exact value. It's only has to
be 'reasonable enough' for the optimizer to be able to choose a good
optimization path).
- Non virtual handler::init() function added for caching of virtual
constants from engine.
- Removed has_transactions() virtual method. Now one should instead return
HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support
transactions.
- The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument
that is to be used with 'new handler_name()' to allocate the handler
in the right area. The xxxx_create_handler() function is also
responsible for any initialization of the object before returning.
For example, one should change:
static handler *myisam_create_handler(TABLE_SHARE *table)
{
return new ha_myisam(table);
}
->
static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root)
{
return new (mem_root) ha_myisam(table);
}
- New optional virtual function: use_hidden_primary_key().
This is called in case of an update/delete when
(table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined
but we don't have a primary key. This allows the handler to take precisions
in remembering any hidden primary key to able to update/delete any
found row. The default handler marks all columns to be read.
- handler::table_flags() now returns a ulonglong (to allow for more flags).
- New/changed table_flags()
- HA_HAS_RECORDS Set if ::records() is supported
- HA_NO_TRANSACTIONS Set if engine doesn't support transactions
- HA_PRIMARY_KEY_REQUIRED_FOR_DELETE
Set if we should mark all primary key columns for
read when reading rows as part of a DELETE
statement. If there is no primary key,
all columns are marked for read.
- HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some
cases (based on table->read_set)
- HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS
Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION.
- HA_DUPP_POS Renamed to HA_DUPLICATE_POS
- HA_REQUIRES_KEY_COLUMNS_FOR_DELETE
Set this if we should mark ALL key columns for
read when when reading rows as part of a DELETE
statement. In case of an update we will mark
all keys for read for which key part changed
value.
- HA_STATS_RECORDS_IS_EXACT
Set this if stats.records is exact.
(This saves us some extra records() calls
when optimizing COUNT(*))
- Removed table_flags()
- HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if
handler::records() gives an exact count() and
HA_STATS_RECORDS_IS_EXACT if stats.records is exact.
- HA_READ_RND_SAME Removed (no one supported this one)
- Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk()
- Renamed handler::dupp_pos to handler::dup_pos
- Removed not used variable handler::sortkey
Upper level handler changes:
- ha_reset() now does some overall checks and calls ::reset()
- ha_table_flags() added. This is a cached version of table_flags(). The
cache is updated on engine creation time and updated on open.
MySQL level changes (not obvious from the above):
- DBUG_ASSERT() added to check that column usage matches what is set
in the column usage bit maps. (This found a LOT of bugs in current
column marking code).
- In 5.1 before, all used columns was marked in read_set and only updated
columns was marked in write_set. Now we only mark columns for which we
need a value in read_set.
- Column bitmaps are created in open_binary_frm() and open_table_from_share().
(Before this was in table.cc)
- handler::table_flags() calls are replaced with handler::ha_table_flags()
- For calling field->val() you must have the corresponding bit set in
table->read_set. For calling field->store() you must have the
corresponding bit set in table->write_set. (There are asserts in
all store()/val() functions to catch wrong usage)
- thd->set_query_id is renamed to thd->mark_used_columns and instead
of setting this to an integer value, this has now the values:
MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE
Changed also all variables named 'set_query_id' to mark_used_columns.
- In filesort() we now inform the handler of exactly which columns are needed
doing the sort and choosing the rows.
- The TABLE_SHARE object has a 'all_set' column bitmap one can use
when one needs a column bitmap with all columns set.
(This is used for table->use_all_columns() and other places)
- The TABLE object has 3 column bitmaps:
- def_read_set Default bitmap for columns to be read
- def_write_set Default bitmap for columns to be written
- tmp_set Can be used as a temporary bitmap when needed.
The table object has also two pointer to bitmaps read_set and write_set
that the handler should use to find out which columns are used in which way.
- count() optimization now calls handler::records() instead of using
handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true).
- Added extra argument to Item::walk() to indicate if we should also
traverse sub queries.
- Added TABLE parameter to cp_buffer_from_ref()
- Don't close tables created with CREATE ... SELECT but keep them in
the table cache. (Faster usage of newly created tables).
New interfaces:
- table->clear_column_bitmaps() to initialize the bitmaps for tables
at start of new statements.
- table->column_bitmaps_set() to set up new column bitmaps and signal
the handler about this.
- table->column_bitmaps_set_no_signal() for some few cases where we need
to setup new column bitmaps but don't signal the handler (as the handler
has already been signaled about these before). Used for the momement
only in opt_range.cc when doing ROR scans.
- table->use_all_columns() to install a bitmap where all columns are marked
as use in the read and the write set.
- table->default_column_bitmaps() to install the normal read and write
column bitmaps, but not signaling the handler about this.
This is mainly used when creating TABLE instances.
- table->mark_columns_needed_for_delete(),
table->mark_columns_needed_for_delete() and
table->mark_columns_needed_for_insert() to allow us to put additional
columns in column usage maps if handler so requires.
(The handler indicates what it neads in handler->table_flags())
- table->prepare_for_position() to allow us to tell handler that it
needs to read primary key parts to be able to store them in
future table->position() calls.
(This replaces the table->file->ha_retrieve_all_pk function)
- table->mark_auto_increment_column() to tell handler are going to update
columns part of any auto_increment key.
- table->mark_columns_used_by_index() to mark all columns that is part of
an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow
it to quickly know that it only needs to read colums that are part
of the key. (The handler can also use the column map for detecting this,
but simpler/faster handler can just monitor the extra() call).
- table->mark_columns_used_by_index_no_reset() to in addition to other columns,
also mark all columns that is used by the given key.
- table->restore_column_maps_after_mark_index() to restore to default
column maps after a call to table->mark_columns_used_by_index().
- New item function register_field_in_read_map(), for marking used columns
in table->read_map. Used by filesort() to mark all used columns
- Maintain in TABLE->merge_keys set of all keys that are used in query.
(Simplices some optimization loops)
- Maintain Field->part_of_key_not_clustered which is like Field->part_of_key
but the field in the clustered key is not assumed to be part of all index.
(used in opt_range.cc for faster loops)
- dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map()
tmp_use_all_columns() and tmp_restore_column_map() functions to temporally
mark all columns as usable. The 'dbug_' version is primarily intended
inside a handler when it wants to just call Field:store() & Field::val()
functions, but don't need the column maps set for any other usage.
(ie:: bitmap_is_set() is never called)
- We can't use compare_records() to skip updates for handlers that returns
a partial column set and the read_set doesn't cover all columns in the
write set. The reason for this is that if we have a column marked only for
write we can't in the MySQL level know if the value changed or not.
The reason this worked before was that MySQL marked all to be written
columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden
bug'.
- open_table_from_share() does not anymore setup temporary MEM_ROOT
object as a thread specific variable for the handler. Instead we
send the to-be-used MEMROOT to get_new_handler().
(Simpler, faster code)
Bugs fixed:
- Column marking was not done correctly in a lot of cases.
(ALTER TABLE, when using triggers, auto_increment fields etc)
(Could potentially result in wrong values inserted in table handlers
relying on that the old column maps or field->set_query_id was correct)
Especially when it comes to triggers, there may be cases where the
old code would cause lost/wrong values for NDB and/or InnoDB tables.
- Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags:
OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG.
This allowed me to remove some wrong warnings about:
"Some non-transactional changed tables couldn't be rolled back"
- Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset
(thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose
some warnings about
"Some non-transactional changed tables couldn't be rolled back")
- Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table()
which could cause delete_table to report random failures.
- Fixed core dumps for some tests when running with --debug
- Added missing FN_LIBCHAR in mysql_rm_tmp_tables()
(This has probably caused us to not properly remove temporary files after
crash)
- slow_logs was not properly initialized, which could maybe cause
extra/lost entries in slow log.
- If we get an duplicate row on insert, change column map to read and
write all columns while retrying the operation. This is required by
the definition of REPLACE and also ensures that fields that are only
part of UPDATE are properly handled. This fixed a bug in NDB and
REPLACE where REPLACE wrongly copied some column values from the replaced
row.
- For table handler that doesn't support NULL in keys, we would give an error
when creating a primary key with NULL fields, even after the fields has been
automaticly converted to NOT NULL.
- Creating a primary key on a SPATIAL key, would fail if field was not
declared as NOT NULL.
Cleanups:
- Removed not used condition argument to setup_tables
- Removed not needed item function reset_query_id_processor().
- Field->add_index is removed. Now this is instead maintained in
(field->flags & FIELD_IN_ADD_INDEX)
- Field->fieldnr is removed (use field->field_index instead)
- New argument to filesort() to indicate that it should return a set of
row pointers (not used columns). This allowed me to remove some references
to sql_command in filesort and should also enable us to return column
results in some cases where we couldn't before.
- Changed column bitmap handling in opt_range.cc to be aligned with TABLE
bitmap, which allowed me to use bitmap functions instead of looping over
all fields to create some needed bitmaps. (Faster and smaller code)
- Broke up found too long lines
- Moved some variable declaration at start of function for better code
readability.
- Removed some not used arguments from functions.
(setup_fields(), mysql_prepare_insert_check_table())
- setup_fields() now takes an enum instead of an int for marking columns
usage.
- For internal temporary tables, use handler::write_row(),
handler::delete_row() and handler::update_row() instead of
handler::ha_xxxx() for faster execution.
- Changed some constants to enum's and define's.
- Using separate column read and write sets allows for easier checking
of timestamp field was set by statement.
- Remove calls to free_io_cache() as this is now done automaticly in ha_reset()
- Don't build table->normalized_path as this is now identical to table->path
(after bar's fixes to convert filenames)
- Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to
do comparision with the 'convert-dbug-for-diff' tool.
Things left to do in 5.1:
- We wrongly log failed CREATE TABLE ... SELECT in some cases when using
row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result)
Mats has promised to look into this.
- Test that my fix for CREATE TABLE ... SELECT is indeed correct.
(I added several test cases for this, but in this case it's better that
someone else also tests this throughly).
Lars has promosed to do this.
2006-06-04 17:52:22 +02:00
|
|
|
/*
|
|
|
|
If in autocommit mode and not in a transaction, reset
|
|
|
|
OPTION_STATUS_NO_TRANS_UPDATE | OPTION_KEEP_LOG to not get warnings
|
|
|
|
in ha_rollback_trans() about some tables couldn't be rolled back.
|
|
|
|
*/
|
|
|
|
if (!(thd->options & (OPTION_NOT_AUTOCOMMIT | OPTION_BEGIN)))
|
|
|
|
thd->options&= ~(OPTION_STATUS_NO_TRANS_UPDATE | OPTION_KEEP_LOG);
|
|
|
|
|
2005-09-20 20:20:38 +02:00
|
|
|
DBUG_ASSERT(thd->security_ctx== &thd->main_security_ctx);
|
2003-02-10 16:59:16 +01:00
|
|
|
thd->tmp_table_used= 0;
|
2005-08-15 17:15:12 +02:00
|
|
|
if (!thd->in_sub_stmt)
|
|
|
|
{
|
2005-08-15 19:08:58 +02:00
|
|
|
if (opt_bin_log)
|
2005-09-07 17:39:47 +02:00
|
|
|
{
|
2005-08-15 19:08:58 +02:00
|
|
|
reset_dynamic(&thd->user_var_events);
|
2005-09-07 17:39:47 +02:00
|
|
|
thd->user_var_events_alloc= thd->mem_root;
|
|
|
|
}
|
2005-08-15 19:08:58 +02:00
|
|
|
thd->clear_error();
|
2005-08-15 17:15:12 +02:00
|
|
|
thd->total_warn_count=0; // Warnings for this query
|
|
|
|
thd->rand_used= 0;
|
|
|
|
thd->sent_row_count= thd->examined_row_count= 0;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
2002-12-11 08:17:51 +01:00
|
|
|
|
2000-08-15 19:09:37 +02:00
|
|
|
void
|
|
|
|
mysql_init_select(LEX *lex)
|
|
|
|
{
|
2003-07-03 01:30:52 +02:00
|
|
|
SELECT_LEX *select_lex= lex->current_select;
|
2002-05-06 23:04:16 +02:00
|
|
|
select_lex->init_select();
|
2004-11-13 11:56:39 +01:00
|
|
|
lex->wild= 0;
|
2002-12-05 23:40:28 +01:00
|
|
|
if (select_lex == &lex->select_lex)
|
|
|
|
{
|
2004-08-24 18:17:11 +02:00
|
|
|
DBUG_ASSERT(lex->result == 0);
|
2002-12-05 23:40:28 +01:00
|
|
|
lex->exchange= 0;
|
|
|
|
}
|
2000-08-15 19:09:37 +02:00
|
|
|
}
|
|
|
|
|
2002-06-16 16:06:12 +02:00
|
|
|
|
2001-12-13 14:53:18 +01:00
|
|
|
bool
|
2002-05-06 23:04:16 +02:00
|
|
|
mysql_new_select(LEX *lex, bool move_down)
|
2001-06-13 12:36:53 +02:00
|
|
|
{
|
2004-11-08 00:13:54 +01:00
|
|
|
SELECT_LEX *select_lex;
|
2005-06-07 22:34:53 +02:00
|
|
|
THD *thd= lex->thd;
|
2005-02-05 15:05:46 +01:00
|
|
|
DBUG_ENTER("mysql_new_select");
|
|
|
|
|
2005-06-07 22:34:53 +02:00
|
|
|
if (!(select_lex= new (thd->mem_root) SELECT_LEX()))
|
2005-02-05 15:05:46 +01:00
|
|
|
DBUG_RETURN(1);
|
2005-06-07 22:34:53 +02:00
|
|
|
select_lex->select_number= ++thd->select_number;
|
2005-08-12 16:57:19 +02:00
|
|
|
select_lex->parent_lex= lex; /* Used in init_query. */
|
2002-05-06 23:04:16 +02:00
|
|
|
select_lex->init_query();
|
|
|
|
select_lex->init_select();
|
2005-10-15 23:32:37 +02:00
|
|
|
lex->nest_level++;
|
|
|
|
select_lex->nest_level= lex->nest_level;
|
2005-07-13 15:38:55 +02:00
|
|
|
/*
|
|
|
|
Don't evaluate this subquery during statement prepare even if
|
|
|
|
it's a constant one. The flag is switched off in the end of
|
|
|
|
mysql_stmt_prepare.
|
|
|
|
*/
|
2005-09-02 15:21:19 +02:00
|
|
|
if (thd->stmt_arena->is_stmt_prepare())
|
2005-06-07 22:34:53 +02:00
|
|
|
select_lex->uncacheable|= UNCACHEABLE_PREPARE;
|
2002-05-06 23:04:16 +02:00
|
|
|
if (move_down)
|
|
|
|
{
|
2005-02-05 15:05:46 +01:00
|
|
|
SELECT_LEX_UNIT *unit;
|
2004-11-30 20:41:12 +01:00
|
|
|
lex->subqueries= TRUE;
|
2002-05-06 23:04:16 +02:00
|
|
|
/* first select_lex of subselect or derived table */
|
2005-06-07 22:34:53 +02:00
|
|
|
if (!(unit= new (thd->mem_root) SELECT_LEX_UNIT()))
|
2005-02-05 15:05:46 +01:00
|
|
|
DBUG_RETURN(1);
|
2004-11-08 00:13:54 +01:00
|
|
|
|
2002-05-06 23:04:16 +02:00
|
|
|
unit->init_query();
|
|
|
|
unit->init_select();
|
2005-06-07 22:34:53 +02:00
|
|
|
unit->thd= thd;
|
2003-07-03 01:30:52 +02:00
|
|
|
unit->include_down(lex->current_select);
|
2002-12-15 21:01:09 +01:00
|
|
|
unit->link_next= 0;
|
|
|
|
unit->link_prev= 0;
|
2003-03-11 00:06:28 +01:00
|
|
|
unit->return_to= lex->current_select;
|
2002-05-06 23:04:16 +02:00
|
|
|
select_lex->include_down(unit);
|
2005-07-01 06:05:42 +02:00
|
|
|
/*
|
|
|
|
By default we assume that it is usual subselect and we have outer name
|
|
|
|
resolution context, if no we will assign it to 0 later
|
|
|
|
*/
|
|
|
|
select_lex->context.outer_context= &select_lex->outer_select()->context;
|
2002-05-06 23:04:16 +02:00
|
|
|
}
|
|
|
|
else
|
2003-07-03 01:30:52 +02:00
|
|
|
{
|
2004-07-16 00:15:55 +02:00
|
|
|
if (lex->current_select->order_list.first && !lex->current_select->braces)
|
|
|
|
{
|
2004-10-20 03:04:37 +02:00
|
|
|
my_error(ER_WRONG_USAGE, MYF(0), "UNION", "ORDER BY");
|
2005-02-05 15:05:46 +01:00
|
|
|
DBUG_RETURN(1);
|
2004-07-16 00:15:55 +02:00
|
|
|
}
|
2002-10-30 12:18:52 +01:00
|
|
|
select_lex->include_neighbour(lex->current_select);
|
2006-04-21 09:36:20 +02:00
|
|
|
SELECT_LEX_UNIT *unit= select_lex->master_unit();
|
|
|
|
if (!unit->fake_select_lex && unit->add_fake_select_lex(lex->thd))
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
select_lex->context.outer_context=
|
|
|
|
unit->first_select()->context.outer_context;
|
2003-07-03 01:30:52 +02:00
|
|
|
}
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2002-05-27 19:52:54 +02:00
|
|
|
select_lex->master_unit()->global_parameters= select_lex;
|
2002-11-27 00:12:16 +01:00
|
|
|
select_lex->include_global((st_select_lex_node**)&lex->all_selects_list);
|
2002-10-30 12:18:52 +01:00
|
|
|
lex->current_select= select_lex;
|
2005-07-01 06:05:42 +02:00
|
|
|
/*
|
|
|
|
in subquery is SELECT query and we allow resolution of names in SELECT
|
|
|
|
list
|
|
|
|
*/
|
|
|
|
select_lex->context.resolve_in_select_list= TRUE;
|
2005-02-05 15:05:46 +01:00
|
|
|
DBUG_RETURN(0);
|
2001-06-13 12:36:53 +02:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2002-10-02 12:33:08 +02:00
|
|
|
/*
|
|
|
|
Create a select to return the same output as 'SELECT @@var_name'.
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
create_select_for_variable()
|
|
|
|
var_name Variable name
|
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
Used for SHOW COUNT(*) [ WARNINGS | ERROR]
|
|
|
|
|
|
|
|
This will crash with a core dump if the variable doesn't exists
|
|
|
|
*/
|
|
|
|
|
|
|
|
void create_select_for_variable(const char *var_name)
|
|
|
|
{
|
2003-07-06 18:09:57 +02:00
|
|
|
THD *thd;
|
2002-10-02 12:33:08 +02:00
|
|
|
LEX *lex;
|
2003-07-06 18:09:57 +02:00
|
|
|
LEX_STRING tmp, null_lex_string;
|
2005-07-25 20:25:28 +02:00
|
|
|
Item *var;
|
|
|
|
char buff[MAX_SYS_VAR_LENGTH*2+4+8], *end;
|
2002-10-02 12:33:08 +02:00
|
|
|
DBUG_ENTER("create_select_for_variable");
|
2003-07-06 18:09:57 +02:00
|
|
|
|
|
|
|
thd= current_thd;
|
2003-08-26 17:41:40 +02:00
|
|
|
lex= thd->lex;
|
2002-10-02 12:33:08 +02:00
|
|
|
mysql_init_select(lex);
|
|
|
|
lex->sql_command= SQLCOM_SELECT;
|
|
|
|
tmp.str= (char*) var_name;
|
|
|
|
tmp.length=strlen(var_name);
|
2003-07-06 18:09:57 +02:00
|
|
|
bzero((char*) &null_lex_string.str, sizeof(null_lex_string));
|
2005-07-25 20:25:28 +02:00
|
|
|
/*
|
|
|
|
We set the name of Item to @@session.var_name because that then is used
|
|
|
|
as the column name in the output.
|
|
|
|
*/
|
2005-08-08 23:13:49 +02:00
|
|
|
if ((var= get_system_var(thd, OPT_SESSION, tmp, null_lex_string)))
|
|
|
|
{
|
|
|
|
end= strxmov(buff, "@@session.", var_name, NullS);
|
|
|
|
var->set_name(buff, end-buff, system_charset_info);
|
|
|
|
add_item_to_list(thd, var);
|
|
|
|
}
|
2002-10-02 12:33:08 +02:00
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
2002-06-16 16:06:12 +02:00
|
|
|
|
2002-01-02 23:46:43 +01:00
|
|
|
void mysql_init_multi_delete(LEX *lex)
|
|
|
|
{
|
2002-05-06 23:04:16 +02:00
|
|
|
lex->sql_command= SQLCOM_DELETE_MULTI;
|
2002-01-02 23:46:43 +01:00
|
|
|
mysql_init_select(lex);
|
2005-06-07 12:11:36 +02:00
|
|
|
lex->select_lex.select_limit= 0;
|
|
|
|
lex->unit.select_limit_cnt= HA_POS_ERROR;
|
2003-08-18 23:08:08 +02:00
|
|
|
lex->select_lex.table_list.save_and_clear(&lex->auxilliary_table_list);
|
2005-05-24 12:44:34 +02:00
|
|
|
lex->lock_option= using_update_log ? TL_READ_NO_INSERT : TL_READ;
|
2004-07-16 00:15:55 +02:00
|
|
|
lex->query_tables= 0;
|
|
|
|
lex->query_tables_last= &lex->query_tables;
|
2002-01-02 23:46:43 +01:00
|
|
|
}
|
2001-12-13 14:53:18 +01:00
|
|
|
|
2004-03-11 17:38:19 +01:00
|
|
|
/*
|
|
|
|
When you modify mysql_parse(), you may need to mofify
|
|
|
|
mysql_test_parse_for_slave() in this same file.
|
|
|
|
*/
|
2004-04-12 02:26:32 +02:00
|
|
|
|
2004-03-11 17:38:19 +01:00
|
|
|
void mysql_parse(THD *thd, char *inBuf, uint length)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
DBUG_ENTER("mysql_parse");
|
2004-07-21 21:44:12 +02:00
|
|
|
mysql_init_query(thd, (uchar*) inBuf, length);
|
2002-03-22 21:55:08 +01:00
|
|
|
if (query_cache_send_result_to_client(thd, inBuf, length) <= 0)
|
2001-12-02 13:34:01 +01:00
|
|
|
{
|
2004-07-21 21:44:12 +02:00
|
|
|
LEX *lex= thd->lex;
|
2005-08-25 15:34:34 +02:00
|
|
|
|
2005-08-10 23:17:02 +02:00
|
|
|
sp_cache_flush_obsolete(&thd->sp_proc_cache);
|
|
|
|
sp_cache_flush_obsolete(&thd->sp_func_cache);
|
2005-08-25 15:34:34 +02:00
|
|
|
|
2006-03-10 01:44:08 +01:00
|
|
|
if (!MYSQLparse((void *)thd) && ! thd->is_fatal_error)
|
2001-12-02 13:34:01 +01:00
|
|
|
{
|
2003-09-26 12:33:13 +02:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2002-06-16 16:06:12 +02:00
|
|
|
if (mqh_used && thd->user_connect &&
|
2002-11-09 14:40:46 +01:00
|
|
|
check_mqh(thd, lex->sql_command))
|
2002-05-15 12:50:38 +02:00
|
|
|
{
|
|
|
|
thd->net.error = 0;
|
|
|
|
}
|
|
|
|
else
|
2003-09-26 12:33:13 +02:00
|
|
|
#endif
|
2002-05-15 12:50:38 +02:00
|
|
|
{
|
2002-11-28 18:29:26 +01:00
|
|
|
if (thd->net.report_error)
|
2003-05-23 15:32:31 +02:00
|
|
|
{
|
2006-05-02 19:56:43 +02:00
|
|
|
delete lex->sphead;
|
|
|
|
lex->sphead= NULL;
|
|
|
|
if (lex->et)
|
|
|
|
{
|
|
|
|
lex->et->free_sphead_on_delete= true;
|
|
|
|
/* alloced on thd->mem_root so no real memory free but dtor call */
|
|
|
|
lex->et->free_sp();
|
|
|
|
lex->et->deinit_mutexes();
|
|
|
|
lex->et= NULL;
|
|
|
|
}
|
2003-05-23 15:32:31 +02:00
|
|
|
}
|
2002-11-28 18:29:26 +01:00
|
|
|
else
|
|
|
|
{
|
2005-02-14 23:47:17 +01:00
|
|
|
/*
|
|
|
|
Binlog logs a string starting from thd->query and having length
|
|
|
|
thd->query_length; so we set thd->query_length correctly (to not
|
|
|
|
log several statements in one event, when we executed only first).
|
|
|
|
We set it to not see the ';' (otherwise it would get into binlog
|
|
|
|
and Query_log_event::print() would give ';;' output).
|
|
|
|
This also helps display only the current query in SHOW
|
|
|
|
PROCESSLIST.
|
|
|
|
Note that we don't need LOCK_thread_count to modify query_length.
|
|
|
|
*/
|
2005-02-18 14:42:14 +01:00
|
|
|
if (lex->found_semicolon &&
|
|
|
|
(thd->query_length= (ulong)(lex->found_semicolon - thd->query)))
|
2005-02-14 23:47:17 +01:00
|
|
|
thd->query_length--;
|
|
|
|
/* Actually execute the query */
|
2002-11-28 18:29:26 +01:00
|
|
|
mysql_execute_command(thd);
|
2003-12-01 14:19:10 +01:00
|
|
|
query_cache_end_of_result(thd);
|
2002-11-28 18:29:26 +01:00
|
|
|
}
|
2002-05-15 12:50:38 +02:00
|
|
|
}
|
2004-05-20 01:02:49 +02:00
|
|
|
lex->unit.cleanup();
|
2001-12-02 13:34:01 +01:00
|
|
|
}
|
|
|
|
else
|
2002-07-24 18:55:08 +02:00
|
|
|
{
|
2006-05-03 16:02:43 +02:00
|
|
|
DBUG_ASSERT(thd->net.report_error);
|
2002-07-24 18:55:08 +02:00
|
|
|
DBUG_PRINT("info",("Command aborted. Fatal_error: %d",
|
2003-01-30 21:15:44 +01:00
|
|
|
thd->is_fatal_error));
|
2001-12-02 13:34:01 +01:00
|
|
|
query_cache_abort(&thd->net);
|
2004-05-20 01:02:49 +02:00
|
|
|
lex->unit.cleanup();
|
2006-05-02 19:56:43 +02:00
|
|
|
if (lex->sphead)
|
2003-06-29 18:15:17 +02:00
|
|
|
{
|
2003-12-21 11:48:03 +01:00
|
|
|
/* Clean up after failed stored procedure/function */
|
2006-05-02 19:56:43 +02:00
|
|
|
delete lex->sphead;
|
|
|
|
lex->sphead= NULL;
|
|
|
|
}
|
|
|
|
if (lex->et)
|
|
|
|
{
|
|
|
|
lex->et->free_sphead_on_delete= true;
|
|
|
|
lex->et->free_sp();
|
|
|
|
lex->et->deinit_mutexes();
|
|
|
|
lex->et= NULL;
|
2003-06-29 18:15:17 +02:00
|
|
|
}
|
2002-07-24 18:55:08 +02:00
|
|
|
}
|
2001-12-02 13:34:01 +01:00
|
|
|
thd->proc_info="freeing items";
|
2004-08-24 18:17:11 +02:00
|
|
|
thd->end_statement();
|
2004-09-15 21:10:31 +02:00
|
|
|
thd->cleanup_after_query();
|
2004-10-08 00:21:19 +02:00
|
|
|
DBUG_ASSERT(thd->change_list.is_empty());
|
2001-12-02 13:34:01 +01:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2004-03-17 23:09:13 +01:00
|
|
|
#ifdef HAVE_REPLICATION
|
2004-03-11 17:38:19 +01:00
|
|
|
/*
|
|
|
|
Usable by the replication SQL thread only: just parse a query to know if it
|
|
|
|
can be ignored because of replicate-*-table rules.
|
|
|
|
|
|
|
|
RETURN VALUES
|
|
|
|
0 cannot be ignored
|
|
|
|
1 can be ignored
|
|
|
|
*/
|
|
|
|
|
|
|
|
bool mysql_test_parse_for_slave(THD *thd, char *inBuf, uint length)
|
|
|
|
{
|
2004-07-21 21:44:12 +02:00
|
|
|
LEX *lex= thd->lex;
|
2004-03-11 17:38:19 +01:00
|
|
|
bool error= 0;
|
2004-07-15 03:19:07 +02:00
|
|
|
DBUG_ENTER("mysql_test_parse_for_slave");
|
2004-03-11 17:38:19 +01:00
|
|
|
|
2004-07-21 21:44:12 +02:00
|
|
|
mysql_init_query(thd, (uchar*) inBuf, length);
|
2006-03-10 01:44:08 +01:00
|
|
|
if (!MYSQLparse((void*) thd) && ! thd->is_fatal_error &&
|
2004-03-11 17:38:19 +01:00
|
|
|
all_tables_not_ok(thd,(TABLE_LIST*) lex->select_lex.table_list.first))
|
2004-07-15 03:19:07 +02:00
|
|
|
error= 1; /* Ignore question */
|
2004-08-24 18:17:11 +02:00
|
|
|
thd->end_statement();
|
2004-09-15 21:10:31 +02:00
|
|
|
thd->cleanup_after_query();
|
2004-07-15 03:19:07 +02:00
|
|
|
DBUG_RETURN(error);
|
2004-03-11 17:38:19 +01:00
|
|
|
}
|
2004-03-17 23:09:13 +01:00
|
|
|
#endif
|
2000-07-31 21:29:14 +02:00
|
|
|
|
2004-09-15 21:10:31 +02:00
|
|
|
|
2004-10-29 18:26:52 +02:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
/*****************************************************************************
|
|
|
|
** Store field definition for create
|
|
|
|
** Return 0 if ok
|
|
|
|
******************************************************************************/
|
|
|
|
|
2002-12-06 20:11:27 +01:00
|
|
|
bool add_field_to_list(THD *thd, char *field_name, enum_field_types type,
|
2000-07-31 21:29:14 +02:00
|
|
|
char *length, char *decimals,
|
2002-06-02 20:22:20 +02:00
|
|
|
uint type_modifier,
|
2004-04-02 08:12:53 +02:00
|
|
|
Item *default_value, Item *on_update_value,
|
|
|
|
LEX_STRING *comment,
|
2004-12-02 09:48:43 +01:00
|
|
|
char *change,
|
|
|
|
List<String> *interval_list, CHARSET_INFO *cs,
|
2003-03-27 10:09:09 +01:00
|
|
|
uint uint_geom_type)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
register create_field *new_field;
|
2003-05-05 20:54:37 +02:00
|
|
|
LEX *lex= thd->lex;
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_ENTER("add_field_to_list");
|
|
|
|
|
|
|
|
if (strlen(field_name) > NAME_LEN)
|
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_TOO_LONG_IDENT, MYF(0), field_name); /* purecov: inspected */
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_RETURN(1); /* purecov: inspected */
|
|
|
|
}
|
|
|
|
if (type_modifier & PRI_KEY_FLAG)
|
|
|
|
{
|
|
|
|
lex->col_list.push_back(new key_part_spec(field_name,0));
|
2006-05-03 14:59:17 +02:00
|
|
|
lex->key_list.push_back(new Key(Key::PRIMARY, NullS,
|
|
|
|
&default_key_create_info,
|
2004-05-11 23:29:52 +02:00
|
|
|
0, lex->col_list));
|
2000-07-31 21:29:14 +02:00
|
|
|
lex->col_list.empty();
|
|
|
|
}
|
|
|
|
if (type_modifier & (UNIQUE_FLAG | UNIQUE_KEY_FLAG))
|
|
|
|
{
|
|
|
|
lex->col_list.push_back(new key_part_spec(field_name,0));
|
2006-05-03 14:59:17 +02:00
|
|
|
lex->key_list.push_back(new Key(Key::UNIQUE, NullS,
|
|
|
|
&default_key_create_info, 0,
|
2000-07-31 21:29:14 +02:00
|
|
|
lex->col_list));
|
|
|
|
lex->col_list.empty();
|
|
|
|
}
|
|
|
|
|
2003-03-22 18:55:09 +01:00
|
|
|
if (default_value)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-04-02 08:12:53 +02:00
|
|
|
/*
|
2004-03-20 12:36:26 +01:00
|
|
|
Default value should be literal => basic constants =>
|
|
|
|
no need fix_fields()
|
2004-04-02 08:12:53 +02:00
|
|
|
|
|
|
|
We allow only one function as part of default value -
|
|
|
|
NOW() as default for TIMESTAMP type.
|
2004-01-30 13:13:19 +01:00
|
|
|
*/
|
2004-04-02 08:12:53 +02:00
|
|
|
if (default_value->type() == Item::FUNC_ITEM &&
|
|
|
|
!(((Item_func*)default_value)->functype() == Item_func::NOW_FUNC &&
|
|
|
|
type == FIELD_TYPE_TIMESTAMP))
|
2003-12-09 09:36:57 +01:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_INVALID_DEFAULT, MYF(0), field_name);
|
2003-12-09 09:36:57 +01:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
else if (default_value->type() == Item::NULL_ITEM)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-10-04 15:13:31 +02:00
|
|
|
default_value= 0;
|
2003-03-22 18:55:09 +01:00
|
|
|
if ((type_modifier & (NOT_NULL_FLAG | AUTO_INCREMENT_FLAG)) ==
|
|
|
|
NOT_NULL_FLAG)
|
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_INVALID_DEFAULT, MYF(0), field_name);
|
2003-03-22 18:55:09 +01:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else if (type_modifier & AUTO_INCREMENT_FLAG)
|
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_INVALID_DEFAULT, MYF(0), field_name);
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
|
|
|
}
|
2004-04-02 08:12:53 +02:00
|
|
|
|
|
|
|
if (on_update_value && type != FIELD_TYPE_TIMESTAMP)
|
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_INVALID_ON_UPDATE, MYF(0), field_name);
|
2004-04-02 08:12:53 +02:00
|
|
|
DBUG_RETURN(1);
|
|
|
|
}
|
2005-03-04 22:14:35 +01:00
|
|
|
|
2005-06-20 12:09:00 +02:00
|
|
|
if (type == FIELD_TYPE_TIMESTAMP && length)
|
|
|
|
{
|
|
|
|
/* Display widths are no longer supported for TIMSTAMP as of MySQL 4.1.
|
|
|
|
In other words, for declarations such as TIMESTAMP(2), TIMESTAMP(4),
|
|
|
|
and so on, the display width is ignored.
|
|
|
|
*/
|
2005-06-27 14:26:07 +02:00
|
|
|
char buf[32];
|
2005-09-23 15:47:08 +02:00
|
|
|
my_snprintf(buf, sizeof(buf), "TIMESTAMP(%s)", length);
|
2006-05-02 19:56:43 +02:00
|
|
|
WARN_DEPRECATED(thd, "5.2", buf, "'TIMESTAMP'");
|
2005-06-20 12:09:00 +02:00
|
|
|
}
|
|
|
|
|
2005-12-07 15:01:17 +01:00
|
|
|
if (!(new_field= new create_field()) ||
|
|
|
|
new_field->init(thd, field_name, type, length, decimals, type_modifier,
|
|
|
|
default_value, on_update_value, comment, change,
|
|
|
|
interval_list, cs, uint_geom_type))
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_RETURN(1);
|
2005-03-04 22:14:35 +01:00
|
|
|
|
|
|
|
lex->create_list.push_back(new_field);
|
|
|
|
lex->last_field=new_field;
|
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
|
|
|
|
2004-12-18 04:19:21 +01:00
|
|
|
|
2000-07-31 21:29:14 +02:00
|
|
|
/* Store position for column in ALTER TABLE .. ADD column */
|
|
|
|
|
|
|
|
void store_position_for_column(const char *name)
|
|
|
|
{
|
|
|
|
current_lex->last_field->after=my_const_cast(char*) (name);
|
|
|
|
}
|
|
|
|
|
|
|
|
bool
|
2002-11-29 15:40:18 +01:00
|
|
|
add_proc_to_list(THD* thd, Item *item)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
ORDER *order;
|
|
|
|
Item **item_ptr;
|
|
|
|
|
2002-12-06 20:11:27 +01:00
|
|
|
if (!(order = (ORDER *) thd->alloc(sizeof(ORDER)+sizeof(Item*))))
|
2000-07-31 21:29:14 +02:00
|
|
|
return 1;
|
|
|
|
item_ptr = (Item**) (order+1);
|
|
|
|
*item_ptr= item;
|
|
|
|
order->item=item_ptr;
|
|
|
|
order->free_me=0;
|
2003-05-05 20:54:37 +02:00
|
|
|
thd->lex->proc_list.link_in_list((byte*) order,(byte**) &order->next);
|
2000-07-31 21:29:14 +02:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* Fix escaping of _, % and \ in database and table names (for ODBC) */
|
|
|
|
|
|
|
|
static void remove_escape(char *name)
|
|
|
|
{
|
2001-08-15 19:09:20 +02:00
|
|
|
if (!*name) // For empty DB names
|
|
|
|
return;
|
2000-07-31 21:29:14 +02:00
|
|
|
char *to;
|
|
|
|
#ifdef USE_MB
|
2000-08-21 23:18:32 +02:00
|
|
|
char *strend=name+(uint) strlen(name);
|
2000-07-31 21:29:14 +02:00
|
|
|
#endif
|
|
|
|
for (to=name; *name ; name++)
|
|
|
|
{
|
|
|
|
#ifdef USE_MB
|
|
|
|
int l;
|
2002-03-14 17:52:48 +01:00
|
|
|
if (use_mb(system_charset_info) &&
|
|
|
|
(l = my_ismbchar(system_charset_info, name, strend)))
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
while (l--)
|
|
|
|
*to++ = *name++;
|
|
|
|
name--;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
if (*name == '\\' && name[1])
|
2001-11-06 23:13:29 +01:00
|
|
|
name++; // Skip '\\'
|
2000-07-31 21:29:14 +02:00
|
|
|
*to++= *name;
|
|
|
|
}
|
|
|
|
*to=0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/****************************************************************************
|
|
|
|
** save order by and tables in own lists
|
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
2002-12-06 20:11:27 +01:00
|
|
|
bool add_to_list(THD *thd, SQL_LIST &list,Item *item,bool asc)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
ORDER *order;
|
|
|
|
DBUG_ENTER("add_to_list");
|
2004-04-03 10:13:51 +02:00
|
|
|
if (!(order = (ORDER *) thd->alloc(sizeof(ORDER))))
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_RETURN(1);
|
2004-04-03 10:13:51 +02:00
|
|
|
order->item_ptr= item;
|
|
|
|
order->item= &order->item_ptr;
|
2000-07-31 21:29:14 +02:00
|
|
|
order->asc = asc;
|
|
|
|
order->free_me=0;
|
|
|
|
order->used=0;
|
2004-08-31 10:58:45 +02:00
|
|
|
order->counter_used= 0;
|
2002-11-29 15:40:18 +01:00
|
|
|
list.link_in_list((byte*) order,(byte**) &order->next);
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2003-01-09 21:42:31 +01:00
|
|
|
/*
|
|
|
|
Add a table to list of used tables
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
add_table_to_list()
|
|
|
|
table Table to add
|
|
|
|
alias alias for table (or null if no alias)
|
|
|
|
table_options A set of the following bits:
|
|
|
|
TL_OPTION_UPDATING Table will be updated
|
|
|
|
TL_OPTION_FORCE_INDEX Force usage of index
|
|
|
|
lock_type How table should be locked
|
|
|
|
use_index List of indexed used in USE INDEX
|
|
|
|
ignore_index List of indexed used in IGNORE INDEX
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
0 Error
|
|
|
|
# Pointer to TABLE_LIST element added to the total table list
|
|
|
|
*/
|
|
|
|
|
2002-12-06 20:11:27 +01:00
|
|
|
TABLE_LIST *st_select_lex::add_table_to_list(THD *thd,
|
|
|
|
Table_ident *table,
|
2002-10-30 12:18:52 +01:00
|
|
|
LEX_STRING *alias,
|
2003-01-09 21:42:31 +01:00
|
|
|
ulong table_options,
|
|
|
|
thr_lock_type lock_type,
|
2003-11-28 11:18:13 +01:00
|
|
|
List<String> *use_index_arg,
|
|
|
|
List<String> *ignore_index_arg,
|
2003-07-16 21:30:49 +02:00
|
|
|
LEX_STRING *option)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
register TABLE_LIST *ptr;
|
2005-08-12 16:57:19 +02:00
|
|
|
TABLE_LIST *previous_table_ref; /* The table preceding the current one. */
|
2000-07-31 21:29:14 +02:00
|
|
|
char *alias_str;
|
2004-09-03 20:43:04 +02:00
|
|
|
LEX *lex= thd->lex;
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_ENTER("add_table_to_list");
|
2005-08-18 02:12:42 +02:00
|
|
|
LINT_INIT(previous_table_ref);
|
2000-07-31 21:29:14 +02:00
|
|
|
|
|
|
|
if (!table)
|
|
|
|
DBUG_RETURN(0); // End of memory
|
|
|
|
alias_str= alias ? alias->str : table->table.str;
|
2006-05-02 19:56:43 +02:00
|
|
|
if (check_table_name(table->table.str,table->table.length))
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_WRONG_TABLE_NAME, MYF(0), table->table.str);
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
2006-05-02 19:56:43 +02:00
|
|
|
if (table->db.str && check_db_name(table->db.str))
|
|
|
|
{
|
|
|
|
my_error(ER_WRONG_DB_NAME, MYF(0), table->db.str);
|
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
|
|
|
|
if (!alias) /* Alias is case sensitive */
|
2002-11-28 18:19:21 +01:00
|
|
|
{
|
|
|
|
if (table->sel)
|
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_DERIVED_MUST_HAVE_ALIAS,
|
|
|
|
ER(ER_DERIVED_MUST_HAVE_ALIAS), MYF(0));
|
2002-11-28 18:19:21 +01:00
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
2001-08-14 19:33:49 +02:00
|
|
|
if (!(alias_str=thd->memdup(alias_str,table->table.length+1)))
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_RETURN(0);
|
2002-11-28 18:19:21 +01:00
|
|
|
}
|
2000-08-21 02:07:54 +02:00
|
|
|
if (!(ptr = (TABLE_LIST *) thd->calloc(sizeof(TABLE_LIST))))
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_RETURN(0); /* purecov: inspected */
|
2002-11-30 18:58:53 +01:00
|
|
|
if (table->db.str)
|
2002-03-15 22:57:31 +01:00
|
|
|
{
|
|
|
|
ptr->db= table->db.str;
|
|
|
|
ptr->db_length= table->db.length;
|
|
|
|
}
|
|
|
|
else if (thd->db)
|
|
|
|
{
|
|
|
|
ptr->db= thd->db;
|
|
|
|
ptr->db_length= thd->db_length;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2002-10-02 12:33:08 +02:00
|
|
|
/* The following can't be "" as we may do 'casedn_str()' on it */
|
|
|
|
ptr->db= empty_c_string;
|
2002-03-15 22:57:31 +01:00
|
|
|
ptr->db_length= 0;
|
|
|
|
}
|
2005-09-02 15:21:19 +02:00
|
|
|
if (thd->stmt_arena->is_stmt_prepare_or_first_sp_execute())
|
2004-10-26 18:30:01 +02:00
|
|
|
ptr->db= thd->strdup(ptr->db);
|
2002-11-30 18:58:53 +01:00
|
|
|
|
2002-09-20 13:05:18 +02:00
|
|
|
ptr->alias= alias_str;
|
2003-02-27 11:45:00 +01:00
|
|
|
if (lower_case_table_names && table->table.length)
|
|
|
|
my_casedn_str(files_charset_info, table->table.str);
|
2005-01-06 12:00:13 +01:00
|
|
|
ptr->table_name=table->table.str;
|
|
|
|
ptr->table_name_length=table->table.length;
|
2003-02-12 20:55:37 +01:00
|
|
|
ptr->lock_type= lock_type;
|
2003-01-09 21:42:31 +01:00
|
|
|
ptr->updating= test(table_options & TL_OPTION_UPDATING);
|
|
|
|
ptr->force_index= test(table_options & TL_OPTION_FORCE_INDEX);
|
2003-06-12 13:29:02 +02:00
|
|
|
ptr->ignore_leaves= test(table_options & TL_OPTION_IGNORE_LEAVES);
|
2003-02-12 20:55:37 +01:00
|
|
|
ptr->derived= table->sel;
|
2005-10-21 10:14:54 +02:00
|
|
|
if (!ptr->derived && !my_strcasecmp(system_charset_info, ptr->db,
|
|
|
|
information_schema_name.str))
|
2004-11-13 11:56:39 +01:00
|
|
|
{
|
2005-01-06 12:00:13 +01:00
|
|
|
ST_SCHEMA_TABLE *schema_table= find_schema_table(thd, ptr->table_name);
|
2004-12-18 11:49:13 +01:00
|
|
|
if (!schema_table ||
|
|
|
|
(schema_table->hidden &&
|
|
|
|
lex->orig_sql_command == SQLCOM_END)) // not a 'show' command
|
2004-11-13 11:56:39 +01:00
|
|
|
{
|
2004-11-13 22:26:15 +01:00
|
|
|
my_error(ER_UNKNOWN_TABLE, MYF(0),
|
2005-01-06 12:00:13 +01:00
|
|
|
ptr->table_name, information_schema_name.str);
|
2004-11-13 11:56:39 +01:00
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
2005-01-27 11:16:51 +01:00
|
|
|
ptr->schema_table_name= ptr->table_name;
|
2004-11-13 11:56:39 +01:00
|
|
|
ptr->schema_table= schema_table;
|
|
|
|
}
|
2004-09-03 20:43:04 +02:00
|
|
|
ptr->select_lex= lex->current_select;
|
2004-04-07 12:25:24 +02:00
|
|
|
ptr->cacheable_table= 1;
|
2003-11-28 11:18:13 +01:00
|
|
|
if (use_index_arg)
|
|
|
|
ptr->use_index=(List<String> *) thd->memdup((gptr) use_index_arg,
|
|
|
|
sizeof(*use_index_arg));
|
|
|
|
if (ignore_index_arg)
|
|
|
|
ptr->ignore_index=(List<String> *) thd->memdup((gptr) ignore_index_arg,
|
|
|
|
sizeof(*ignore_index_arg));
|
2003-07-16 21:30:49 +02:00
|
|
|
ptr->option= option ? option->str : 0;
|
2000-07-31 21:29:14 +02:00
|
|
|
/* check that used name is unique */
|
2003-01-09 01:19:14 +01:00
|
|
|
if (lock_type != TL_IGNORE)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2005-09-12 16:01:17 +02:00
|
|
|
TABLE_LIST *first_table= (TABLE_LIST*) table_list.first;
|
|
|
|
if (lex->sql_command == SQLCOM_CREATE_VIEW)
|
|
|
|
first_table= first_table ? first_table->next_local : NULL;
|
|
|
|
for (TABLE_LIST *tables= first_table ;
|
2001-12-13 01:31:19 +01:00
|
|
|
tables ;
|
2004-07-16 00:15:55 +02:00
|
|
|
tables=tables->next_local)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2004-03-22 14:43:13 +01:00
|
|
|
if (!my_strcasecmp(table_alias_charset, alias_str, tables->alias) &&
|
|
|
|
!strcmp(ptr->db, tables->db))
|
2000-08-21 02:07:54 +02:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_NONUNIQ_TABLE, MYF(0), alias_str); /* purecov: tested */
|
2000-08-21 02:07:54 +02:00
|
|
|
DBUG_RETURN(0); /* purecov: tested */
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
}
|
2005-08-12 16:57:19 +02:00
|
|
|
/* Store the table reference preceding the current one. */
|
|
|
|
if (table_list.elements > 0)
|
|
|
|
{
|
2005-08-18 02:12:42 +02:00
|
|
|
/*
|
|
|
|
table_list.next points to the last inserted TABLE_LIST->next_local'
|
|
|
|
element
|
2006-02-25 16:46:30 +01:00
|
|
|
We don't use the offsetof() macro here to avoid warnings from gcc
|
2005-08-18 02:12:42 +02:00
|
|
|
*/
|
2006-02-25 16:46:30 +01:00
|
|
|
previous_table_ref= (TABLE_LIST*) ((char*) table_list.next -
|
|
|
|
((char*) &(ptr->next_local) -
|
|
|
|
(char*) ptr));
|
2005-08-18 02:12:42 +02:00
|
|
|
/*
|
|
|
|
Set next_name_resolution_table of the previous table reference to point
|
|
|
|
to the current table reference. In effect the list
|
|
|
|
TABLE_LIST::next_name_resolution_table coincides with
|
|
|
|
TABLE_LIST::next_local. Later this may be changed in
|
|
|
|
store_top_level_join_columns() for NATURAL/USING joins.
|
|
|
|
*/
|
|
|
|
previous_table_ref->next_name_resolution_table= ptr;
|
2005-08-12 16:57:19 +02:00
|
|
|
}
|
2005-08-18 02:12:42 +02:00
|
|
|
|
2005-08-12 16:57:19 +02:00
|
|
|
/*
|
|
|
|
Link the current table reference in a local list (list for current select).
|
|
|
|
Notice that as a side effect here we set the next_local field of the
|
|
|
|
previous table reference to 'ptr'. Here we also add one element to the
|
|
|
|
list 'table_list'.
|
|
|
|
*/
|
2004-07-16 00:15:55 +02:00
|
|
|
table_list.link_in_list((byte*) ptr, (byte**) &ptr->next_local);
|
2005-08-12 16:57:19 +02:00
|
|
|
ptr->next_name_resolution_table= NULL;
|
2004-09-03 20:43:04 +02:00
|
|
|
/* Link table in global list (all used tables) */
|
2004-09-11 22:52:55 +02:00
|
|
|
lex->add_to_query_tables(ptr);
|
2000-07-31 21:29:14 +02:00
|
|
|
DBUG_RETURN(ptr);
|
|
|
|
}
|
|
|
|
|
2001-08-02 05:29:50 +02:00
|
|
|
|
2004-06-11 07:27:21 +02:00
|
|
|
/*
|
|
|
|
Initialize a new table list for a nested join
|
|
|
|
|
|
|
|
SYNOPSIS
|
2005-01-25 15:25:46 +01:00
|
|
|
init_nested_join()
|
2004-06-11 07:27:21 +02:00
|
|
|
thd current thread
|
2004-09-11 22:52:55 +02:00
|
|
|
|
2004-06-11 07:27:21 +02:00
|
|
|
DESCRIPTION
|
|
|
|
The function initializes a structure of the TABLE_LIST type
|
|
|
|
for a nested join. It sets up its nested join list as empty.
|
|
|
|
The created structure is added to the front of the current
|
|
|
|
join list in the st_select_lex object. Then the function
|
|
|
|
changes the current nest level for joins to refer to the newly
|
|
|
|
created empty list after having saved the info on the old level
|
|
|
|
in the initialized structure.
|
|
|
|
|
|
|
|
RETURN VALUE
|
|
|
|
0, if success
|
|
|
|
1, otherwise
|
|
|
|
*/
|
|
|
|
|
|
|
|
bool st_select_lex::init_nested_join(THD *thd)
|
|
|
|
{
|
|
|
|
TABLE_LIST *ptr;
|
|
|
|
NESTED_JOIN *nested_join;
|
|
|
|
DBUG_ENTER("init_nested_join");
|
2004-09-11 22:52:55 +02:00
|
|
|
|
2005-08-18 02:12:42 +02:00
|
|
|
if (!(ptr= (TABLE_LIST*) thd->calloc(ALIGN_SIZE(sizeof(TABLE_LIST))+
|
|
|
|
sizeof(NESTED_JOIN))))
|
2004-06-11 07:27:21 +02:00
|
|
|
DBUG_RETURN(1);
|
2005-08-18 02:12:42 +02:00
|
|
|
nested_join= ptr->nested_join=
|
|
|
|
((NESTED_JOIN*) ((byte*) ptr + ALIGN_SIZE(sizeof(TABLE_LIST))));
|
|
|
|
|
2004-06-11 07:27:21 +02:00
|
|
|
join_list->push_front(ptr);
|
|
|
|
ptr->embedding= embedding;
|
|
|
|
ptr->join_list= join_list;
|
|
|
|
embedding= ptr;
|
|
|
|
join_list= &nested_join->join_list;
|
|
|
|
join_list->empty();
|
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
End a nested join table list
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
end_nested_join()
|
|
|
|
thd current thread
|
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
The function returns to the previous join nest level.
|
|
|
|
If the current level contains only one member, the function
|
2004-09-11 22:52:55 +02:00
|
|
|
moves it one level up, eliminating the nest.
|
2004-06-11 07:27:21 +02:00
|
|
|
|
|
|
|
RETURN VALUE
|
|
|
|
Pointer to TABLE_LIST element added to the total table list, if success
|
|
|
|
0, otherwise
|
|
|
|
*/
|
|
|
|
|
|
|
|
TABLE_LIST *st_select_lex::end_nested_join(THD *thd)
|
|
|
|
{
|
|
|
|
TABLE_LIST *ptr;
|
2005-03-17 00:22:12 +01:00
|
|
|
NESTED_JOIN *nested_join;
|
2004-06-11 07:27:21 +02:00
|
|
|
DBUG_ENTER("end_nested_join");
|
2005-03-17 00:22:12 +01:00
|
|
|
|
2005-03-16 01:13:23 +01:00
|
|
|
DBUG_ASSERT(embedding);
|
2004-06-11 07:27:21 +02:00
|
|
|
ptr= embedding;
|
|
|
|
join_list= ptr->join_list;
|
|
|
|
embedding= ptr->embedding;
|
2005-03-17 00:22:12 +01:00
|
|
|
nested_join= ptr->nested_join;
|
2004-06-11 07:27:21 +02:00
|
|
|
if (nested_join->join_list.elements == 1)
|
|
|
|
{
|
|
|
|
TABLE_LIST *embedded= nested_join->join_list.head();
|
|
|
|
join_list->pop();
|
|
|
|
embedded->join_list= join_list;
|
|
|
|
embedded->embedding= embedding;
|
|
|
|
join_list->push_front(embedded);
|
|
|
|
ptr= embedded;
|
|
|
|
}
|
2005-03-17 00:22:12 +01:00
|
|
|
else if (nested_join->join_list.elements == 0)
|
2005-03-16 01:13:23 +01:00
|
|
|
{
|
|
|
|
join_list->pop();
|
2005-03-17 00:22:12 +01:00
|
|
|
ptr= 0; // return value
|
2005-03-16 01:13:23 +01:00
|
|
|
}
|
2004-06-11 07:27:21 +02:00
|
|
|
DBUG_RETURN(ptr);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Nest last join operation
|
|
|
|
|
|
|
|
SYNOPSIS
|
2004-09-11 22:52:55 +02:00
|
|
|
nest_last_join()
|
2004-06-11 07:27:21 +02:00
|
|
|
thd current thread
|
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
The function nest last join operation as if it was enclosed in braces.
|
|
|
|
|
|
|
|
RETURN VALUE
|
2005-08-18 02:12:42 +02:00
|
|
|
0 Error
|
|
|
|
# Pointer to TABLE_LIST element created for the new nested join
|
|
|
|
|
2004-06-11 07:27:21 +02:00
|
|
|
*/
|
|
|
|
|
|
|
|
TABLE_LIST *st_select_lex::nest_last_join(THD *thd)
|
|
|
|
{
|
|
|
|
TABLE_LIST *ptr;
|
|
|
|
NESTED_JOIN *nested_join;
|
2005-08-18 02:12:42 +02:00
|
|
|
List<TABLE_LIST> *embedded_list;
|
2004-06-11 07:27:21 +02:00
|
|
|
DBUG_ENTER("nest_last_join");
|
2004-09-11 22:52:55 +02:00
|
|
|
|
2005-08-18 02:12:42 +02:00
|
|
|
if (!(ptr= (TABLE_LIST*) thd->calloc(ALIGN_SIZE(sizeof(TABLE_LIST))+
|
|
|
|
sizeof(NESTED_JOIN))))
|
2004-06-11 07:27:21 +02:00
|
|
|
DBUG_RETURN(0);
|
2005-08-18 02:12:42 +02:00
|
|
|
nested_join= ptr->nested_join=
|
|
|
|
((NESTED_JOIN*) ((byte*) ptr + ALIGN_SIZE(sizeof(TABLE_LIST))));
|
|
|
|
|
2004-06-11 07:27:21 +02:00
|
|
|
ptr->embedding= embedding;
|
|
|
|
ptr->join_list= join_list;
|
2005-08-18 02:12:42 +02:00
|
|
|
embedded_list= &nested_join->join_list;
|
2004-06-11 07:27:21 +02:00
|
|
|
embedded_list->empty();
|
2005-08-18 02:12:42 +02:00
|
|
|
|
|
|
|
for (uint i=0; i < 2; i++)
|
2004-06-11 07:27:21 +02:00
|
|
|
{
|
|
|
|
TABLE_LIST *table= join_list->pop();
|
|
|
|
table->join_list= embedded_list;
|
|
|
|
table->embedding= ptr;
|
|
|
|
embedded_list->push_back(table);
|
2005-08-12 16:57:19 +02:00
|
|
|
if (table->natural_join)
|
|
|
|
{
|
|
|
|
ptr->is_natural_join= TRUE;
|
|
|
|
/*
|
|
|
|
If this is a JOIN ... USING, move the list of joined fields to the
|
|
|
|
table reference that describes the join.
|
|
|
|
*/
|
|
|
|
if (table->join_using_fields)
|
|
|
|
{
|
|
|
|
ptr->join_using_fields= table->join_using_fields;
|
|
|
|
table->join_using_fields= NULL;
|
|
|
|
}
|
|
|
|
}
|
2004-06-11 07:27:21 +02:00
|
|
|
}
|
|
|
|
join_list->push_front(ptr);
|
|
|
|
nested_join->used_tables= nested_join->not_null_tables= (table_map) 0;
|
|
|
|
DBUG_RETURN(ptr);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Add a table to the current join list
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
add_joined_table()
|
|
|
|
table the table to add
|
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
The function puts a table in front of the current join list
|
|
|
|
of st_select_lex object.
|
|
|
|
Thus, joined tables are put into this list in the reverse order
|
|
|
|
(the most outer join operation follows first).
|
|
|
|
|
|
|
|
RETURN VALUE
|
|
|
|
None
|
|
|
|
*/
|
|
|
|
|
|
|
|
void st_select_lex::add_joined_table(TABLE_LIST *table)
|
|
|
|
{
|
|
|
|
DBUG_ENTER("add_joined_table");
|
|
|
|
join_list->push_front(table);
|
|
|
|
table->join_list= join_list;
|
|
|
|
table->embedding= embedding;
|
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Convert a right join into equivalent left join
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
convert_right_join()
|
|
|
|
thd current thread
|
2004-09-11 22:52:55 +02:00
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
The function takes the current join list t[0],t[1] ... and
|
2004-06-11 07:27:21 +02:00
|
|
|
effectively converts it into the list t[1],t[0] ...
|
|
|
|
Although the outer_join flag for the new nested table contains
|
|
|
|
JOIN_TYPE_RIGHT, it will be handled as the inner table of a left join
|
|
|
|
operation.
|
|
|
|
|
|
|
|
EXAMPLES
|
|
|
|
SELECT * FROM t1 RIGHT JOIN t2 ON on_expr =>
|
|
|
|
SELECT * FROM t2 LEFT JOIN t1 ON on_expr
|
|
|
|
|
|
|
|
SELECT * FROM t1,t2 RIGHT JOIN t3 ON on_expr =>
|
|
|
|
SELECT * FROM t1,t3 LEFT JOIN t2 ON on_expr
|
|
|
|
|
|
|
|
SELECT * FROM t1,t2 RIGHT JOIN (t3,t4) ON on_expr =>
|
|
|
|
SELECT * FROM t1,(t3,t4) LEFT JOIN t2 ON on_expr
|
|
|
|
|
|
|
|
SELECT * FROM t1 LEFT JOIN t2 ON on_expr1 RIGHT JOIN t3 ON on_expr2 =>
|
|
|
|
SELECT * FROM t3 LEFT JOIN (t1 LEFT JOIN t2 ON on_expr2) ON on_expr1
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
Pointer to the table representing the inner table, if success
|
|
|
|
0, otherwise
|
|
|
|
*/
|
|
|
|
|
2004-09-11 22:52:55 +02:00
|
|
|
TABLE_LIST *st_select_lex::convert_right_join()
|
2004-06-11 07:27:21 +02:00
|
|
|
{
|
|
|
|
TABLE_LIST *tab2= join_list->pop();
|
2004-09-11 22:52:55 +02:00
|
|
|
TABLE_LIST *tab1= join_list->pop();
|
2004-06-11 07:27:21 +02:00
|
|
|
DBUG_ENTER("convert_right_join");
|
|
|
|
|
|
|
|
join_list->push_front(tab2);
|
|
|
|
join_list->push_front(tab1);
|
|
|
|
tab1->outer_join|= JOIN_TYPE_RIGHT;
|
|
|
|
|
|
|
|
DBUG_RETURN(tab1);
|
|
|
|
}
|
|
|
|
|
2002-11-16 19:19:10 +01:00
|
|
|
/*
|
|
|
|
Set lock for all tables in current select level
|
|
|
|
|
|
|
|
SYNOPSIS:
|
|
|
|
set_lock_for_tables()
|
|
|
|
lock_type Lock to set for tables
|
|
|
|
|
|
|
|
NOTE:
|
|
|
|
If lock is a write lock, then tables->updating is set 1
|
|
|
|
This is to get tables_ok to know that the table is updated by the
|
|
|
|
query
|
|
|
|
*/
|
|
|
|
|
2002-11-21 21:25:53 +01:00
|
|
|
void st_select_lex::set_lock_for_tables(thr_lock_type lock_type)
|
2002-11-16 19:19:10 +01:00
|
|
|
{
|
|
|
|
bool for_update= lock_type >= TL_READ_NO_INSERT;
|
|
|
|
DBUG_ENTER("set_lock_for_tables");
|
|
|
|
DBUG_PRINT("enter", ("lock_type: %d for_update: %d", lock_type,
|
|
|
|
for_update));
|
|
|
|
|
2004-07-16 00:15:55 +02:00
|
|
|
for (TABLE_LIST *tables= (TABLE_LIST*) table_list.first;
|
|
|
|
tables;
|
|
|
|
tables= tables->next_local)
|
2002-11-16 19:19:10 +01:00
|
|
|
{
|
|
|
|
tables->lock_type= lock_type;
|
|
|
|
tables->updating= for_update;
|
|
|
|
}
|
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
2001-08-02 05:29:50 +02:00
|
|
|
|
2006-04-21 07:15:38 +02:00
|
|
|
/*
|
|
|
|
Create a fake SELECT_LEX for a unit
|
|
|
|
|
|
|
|
SYNOPSIS:
|
|
|
|
add_fake_select_lex()
|
|
|
|
thd thread handle
|
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
The method create a fake SELECT_LEX object for a unit.
|
|
|
|
This object is created for any union construct containing a union
|
|
|
|
operation and also for any single select union construct of the form
|
|
|
|
(SELECT ... ORDER BY order_list [LIMIT n]) ORDER BY ...
|
|
|
|
or of the form
|
|
|
|
(SELECT ... ORDER BY LIMIT n) ORDER BY ...
|
|
|
|
|
|
|
|
NOTES
|
|
|
|
The object is used to retrieve rows from the temporary table
|
|
|
|
where the result on the union is obtained.
|
|
|
|
|
|
|
|
RETURN VALUES
|
|
|
|
1 on failure to create the object
|
|
|
|
0 on success
|
|
|
|
*/
|
|
|
|
|
|
|
|
bool st_select_lex_unit::add_fake_select_lex(THD *thd)
|
|
|
|
{
|
|
|
|
SELECT_LEX *first_sl= first_select();
|
|
|
|
DBUG_ENTER("add_fake_select_lex");
|
|
|
|
DBUG_ASSERT(!fake_select_lex);
|
2006-04-21 17:19:38 +02:00
|
|
|
|
2006-04-21 07:15:38 +02:00
|
|
|
if (!(fake_select_lex= new (thd->mem_root) SELECT_LEX()))
|
|
|
|
DBUG_RETURN(1);
|
|
|
|
fake_select_lex->include_standalone(this,
|
|
|
|
(SELECT_LEX_NODE**)&fake_select_lex);
|
|
|
|
fake_select_lex->select_number= INT_MAX;
|
2006-04-21 17:19:38 +02:00
|
|
|
fake_select_lex->parent_lex= thd->lex; /* Used in init_query. */
|
2006-04-21 07:15:38 +02:00
|
|
|
fake_select_lex->make_empty_select();
|
|
|
|
fake_select_lex->linkage= GLOBAL_OPTIONS_TYPE;
|
2006-04-21 09:36:20 +02:00
|
|
|
fake_select_lex->select_limit= 0;
|
|
|
|
|
2006-04-21 17:19:38 +02:00
|
|
|
fake_select_lex->context.outer_context=first_sl->context.outer_context;
|
2006-04-21 09:36:20 +02:00
|
|
|
/* allow item list resolving in fake select for ORDER BY */
|
|
|
|
fake_select_lex->context.resolve_in_select_list= TRUE;
|
|
|
|
fake_select_lex->context.select_lex= fake_select_lex;
|
2006-04-21 07:15:38 +02:00
|
|
|
|
|
|
|
if (!first_sl->next_select())
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
This works only for
|
|
|
|
(SELECT ... ORDER BY list [LIMIT n]) ORDER BY order_list [LIMIT m],
|
|
|
|
(SELECT ... LIMIT n) ORDER BY order_list [LIMIT m]
|
|
|
|
just before the parser starts processing order_list
|
|
|
|
*/
|
|
|
|
global_parameters= fake_select_lex;
|
|
|
|
fake_select_lex->no_table_names_allowed= 1;
|
|
|
|
thd->lex->current_select= fake_select_lex;
|
|
|
|
}
|
2006-04-21 09:36:20 +02:00
|
|
|
thd->lex->pop_context();
|
2006-04-21 07:15:38 +02:00
|
|
|
DBUG_RETURN(0);
|
|
|
|
}
|
|
|
|
|
2006-04-21 09:36:20 +02:00
|
|
|
|
2005-08-12 16:57:19 +02:00
|
|
|
/*
|
2005-11-28 20:57:50 +01:00
|
|
|
Push a new name resolution context for a JOIN ... ON clause to the
|
|
|
|
context stack of a query block.
|
2005-08-12 16:57:19 +02:00
|
|
|
|
|
|
|
SYNOPSIS
|
2005-11-28 20:57:50 +01:00
|
|
|
push_new_name_resolution_context()
|
2005-08-12 16:57:19 +02:00
|
|
|
thd pointer to current thread
|
2005-09-10 14:01:54 +02:00
|
|
|
left_op left operand of the JOIN
|
2005-08-12 16:57:19 +02:00
|
|
|
right_op rigth operand of the JOIN
|
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
Create a new name resolution context for a JOIN ... ON clause,
|
2005-11-28 20:57:50 +01:00
|
|
|
set the first and last leaves of the list of table references
|
|
|
|
to be used for name resolution, and push the newly created
|
|
|
|
context to the stack of contexts of the query.
|
2005-08-12 16:57:19 +02:00
|
|
|
|
|
|
|
RETURN
|
2005-11-28 20:57:50 +01:00
|
|
|
FALSE if all is OK
|
|
|
|
TRUE if a memory allocation error occured
|
2005-08-12 16:57:19 +02:00
|
|
|
*/
|
|
|
|
|
2005-11-28 20:57:50 +01:00
|
|
|
bool
|
|
|
|
push_new_name_resolution_context(THD *thd,
|
|
|
|
TABLE_LIST *left_op, TABLE_LIST *right_op)
|
2005-08-12 16:57:19 +02:00
|
|
|
{
|
|
|
|
Name_resolution_context *on_context;
|
2005-08-22 00:13:37 +02:00
|
|
|
if (!(on_context= new (thd->mem_root) Name_resolution_context))
|
2005-11-28 20:57:50 +01:00
|
|
|
return TRUE;
|
2005-08-12 16:57:19 +02:00
|
|
|
on_context->init();
|
|
|
|
on_context->first_name_resolution_table=
|
|
|
|
left_op->first_leaf_for_name_resolution();
|
|
|
|
on_context->last_name_resolution_table=
|
|
|
|
right_op->last_leaf_for_name_resolution();
|
2005-11-28 20:57:50 +01:00
|
|
|
return thd->lex->push_context(on_context);
|
2005-08-12 16:57:19 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Add an ON condition to the second operand of a JOIN ... ON.
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
add_join_on
|
|
|
|
b the second operand of a JOIN ... ON
|
|
|
|
expr the condition to be added to the ON clause
|
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
Add an ON condition to the right operand of a JOIN ... ON clause.
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
FALSE if there was some error
|
|
|
|
TRUE if all is OK
|
|
|
|
*/
|
|
|
|
|
|
|
|
void add_join_on(TABLE_LIST *b, Item *expr)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2002-11-11 14:57:35 +01:00
|
|
|
if (expr)
|
2000-09-25 23:33:25 +02:00
|
|
|
{
|
2002-11-11 14:57:35 +01:00
|
|
|
if (!b->on_expr)
|
2005-08-12 16:57:19 +02:00
|
|
|
b->on_expr= expr;
|
2002-11-11 14:57:35 +01:00
|
|
|
else
|
|
|
|
{
|
2005-08-12 16:57:19 +02:00
|
|
|
/*
|
|
|
|
If called from the parser, this happens if you have both a
|
|
|
|
right and left join. If called later, it happens if we add more
|
|
|
|
than one condition to the ON clause.
|
|
|
|
*/
|
|
|
|
b->on_expr= new Item_cond_and(b->on_expr,expr);
|
2002-11-11 14:57:35 +01:00
|
|
|
}
|
|
|
|
b->on_expr->top_level_item();
|
2000-09-25 23:33:25 +02:00
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2003-04-26 19:43:28 +02:00
|
|
|
/*
|
2005-08-12 16:57:19 +02:00
|
|
|
Mark that there is a NATURAL JOIN or JOIN ... USING between two
|
|
|
|
tables.
|
2003-04-26 19:43:28 +02:00
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
add_join_natural()
|
2005-08-12 16:57:19 +02:00
|
|
|
a Left join argument
|
|
|
|
b Right join argument
|
|
|
|
using_fields Field names from USING clause
|
|
|
|
|
2003-04-26 19:43:28 +02:00
|
|
|
IMPLEMENTATION
|
2005-08-12 16:57:19 +02:00
|
|
|
This function marks that table b should be joined with a either via
|
|
|
|
a NATURAL JOIN or via JOIN ... USING. Both join types are special
|
|
|
|
cases of each other, so we treat them together. The function
|
|
|
|
setup_conds() creates a list of equal condition between all fields
|
|
|
|
of the same name for NATURAL JOIN or the fields in 'using_fields'
|
|
|
|
for JOIN ... USING. The list of equality conditions is stored
|
|
|
|
either in b->on_expr, or in JOIN::conds, depending on whether there
|
|
|
|
was an outer join.
|
|
|
|
|
|
|
|
EXAMPLE
|
2003-04-26 19:43:28 +02:00
|
|
|
SELECT * FROM t1 NATURAL LEFT JOIN t2
|
|
|
|
<=>
|
|
|
|
SELECT * FROM t1 LEFT JOIN t2 ON (t1.i=t2.i and t1.j=t2.j ... )
|
2005-08-12 16:57:19 +02:00
|
|
|
|
|
|
|
SELECT * FROM t1 NATURAL JOIN t2 WHERE <some_cond>
|
|
|
|
<=>
|
|
|
|
SELECT * FROM t1, t2 WHERE (t1.i=t2.i and t1.j=t2.j and <some_cond>)
|
|
|
|
|
|
|
|
SELECT * FROM t1 JOIN t2 USING(j) WHERE <some_cond>
|
|
|
|
<=>
|
|
|
|
SELECT * FROM t1, t2 WHERE (t1.j=t2.j and <some_cond>)
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
None
|
2003-04-26 19:43:28 +02:00
|
|
|
*/
|
|
|
|
|
2005-08-12 16:57:19 +02:00
|
|
|
void add_join_natural(TABLE_LIST *a, TABLE_LIST *b, List<String> *using_fields)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2005-08-12 16:57:19 +02:00
|
|
|
b->natural_join= a;
|
|
|
|
b->join_using_fields= using_fields;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
2005-08-12 16:57:19 +02:00
|
|
|
|
2002-12-13 11:05:24 +01:00
|
|
|
/*
|
2003-05-15 18:35:39 +02:00
|
|
|
Reload/resets privileges and the different caches.
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
reload_acl_and_cache()
|
2005-08-24 21:44:42 +02:00
|
|
|
thd Thread handler (can be NULL!)
|
2003-05-15 18:35:39 +02:00
|
|
|
options What should be reset/reloaded (tables, privileges,
|
|
|
|
slave...)
|
|
|
|
tables Tables to flush (if any)
|
|
|
|
write_to_binlog Depending on 'options', it may be very bad to write the
|
|
|
|
query to the binlog (e.g. FLUSH SLAVE); this is a
|
2005-08-11 14:58:15 +02:00
|
|
|
pointer where reload_acl_and_cache() will put 0 if
|
|
|
|
it thinks we really should not write to the binlog.
|
|
|
|
Otherwise it will put 1.
|
2003-05-15 18:35:39 +02:00
|
|
|
|
|
|
|
RETURN
|
|
|
|
0 ok
|
2005-08-11 14:58:15 +02:00
|
|
|
!=0 error. thd->killed or thd->net.report_error is set
|
2002-12-13 11:05:24 +01:00
|
|
|
*/
|
|
|
|
|
2003-05-15 18:35:39 +02:00
|
|
|
bool reload_acl_and_cache(THD *thd, ulong options, TABLE_LIST *tables,
|
|
|
|
bool *write_to_binlog)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
bool result=0;
|
|
|
|
select_errors=0; /* Write if more errors */
|
2003-05-15 18:35:39 +02:00
|
|
|
bool tmp_write_to_binlog= 1;
|
2005-08-10 08:31:32 +02:00
|
|
|
|
2005-08-24 21:44:42 +02:00
|
|
|
if (thd && thd->in_sub_stmt)
|
2005-08-10 08:31:32 +02:00
|
|
|
{
|
|
|
|
my_error(ER_STMT_NOT_ALLOWED_IN_SF_OR_TRG, MYF(0), "FLUSH");
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2003-09-29 08:47:37 +02:00
|
|
|
#ifndef NO_EMBEDDED_ACCESS_CHECKS
|
2000-07-31 21:29:14 +02:00
|
|
|
if (options & REFRESH_GRANT)
|
|
|
|
{
|
2005-09-01 14:52:59 +02:00
|
|
|
THD *tmp_thd= 0;
|
|
|
|
/*
|
|
|
|
If reload_acl_and_cache() is called from SIGHUP handler we have to
|
|
|
|
allocate temporary THD for execution of acl_reload()/grant_reload().
|
|
|
|
*/
|
|
|
|
if (!thd && (thd= (tmp_thd= new THD)))
|
2005-11-23 19:18:10 +01:00
|
|
|
{
|
|
|
|
thd->thread_stack= (char*) &tmp_thd;
|
2005-09-01 14:52:59 +02:00
|
|
|
thd->store_globals();
|
2005-11-23 19:18:10 +01:00
|
|
|
}
|
2005-09-01 14:52:59 +02:00
|
|
|
if (thd)
|
|
|
|
{
|
|
|
|
(void)acl_reload(thd);
|
|
|
|
(void)grant_reload(thd);
|
|
|
|
}
|
|
|
|
if (tmp_thd)
|
|
|
|
{
|
|
|
|
delete tmp_thd;
|
|
|
|
/* Remember that we don't have a THD */
|
|
|
|
my_pthread_setspecific_ptr(THR_THD, 0);
|
|
|
|
thd= 0;
|
|
|
|
}
|
2004-12-29 20:35:16 +01:00
|
|
|
reset_mqh((LEX_USER *)NULL, TRUE);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2003-09-29 08:47:37 +02:00
|
|
|
#endif
|
2000-07-31 21:29:14 +02:00
|
|
|
if (options & REFRESH_LOG)
|
|
|
|
{
|
2003-07-06 17:59:54 +02:00
|
|
|
/*
|
2003-08-11 21:44:43 +02:00
|
|
|
Flush the normal query log, the update log, the binary log,
|
2006-05-02 19:56:43 +02:00
|
|
|
the slow query log, the relay log (if it exists) and the log
|
|
|
|
tables.
|
2003-07-06 17:59:54 +02:00
|
|
|
*/
|
2003-08-11 21:44:43 +02:00
|
|
|
|
2005-02-22 00:15:31 +01:00
|
|
|
/*
|
2005-07-03 13:17:52 +02:00
|
|
|
Writing this command to the binlog may result in infinite loops
|
|
|
|
when doing mysqlbinlog|mysql, and anyway it does not really make
|
|
|
|
sense to log it automatically (would cause more trouble to users
|
|
|
|
than it would help them)
|
2003-05-15 18:35:39 +02:00
|
|
|
*/
|
|
|
|
tmp_write_to_binlog= 0;
|
2005-02-22 00:15:31 +01:00
|
|
|
mysql_bin_log.rotate_and_purge(RP_FORCE_ROTATE);
|
2003-03-13 11:12:28 +01:00
|
|
|
#ifdef HAVE_REPLICATION
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_lock(&LOCK_active_mi);
|
2003-07-06 17:59:54 +02:00
|
|
|
rotate_relay_log(active_mi);
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_unlock(&LOCK_active_mi);
|
2003-03-13 11:12:28 +01:00
|
|
|
#endif
|
2006-05-02 19:56:43 +02:00
|
|
|
|
|
|
|
/* flush slow and general logs */
|
|
|
|
logger.flush_logs(thd);
|
|
|
|
|
|
|
|
if (ha_flush_logs(NULL))
|
2000-07-31 21:29:14 +02:00
|
|
|
result=1;
|
2003-01-27 18:37:25 +01:00
|
|
|
if (flush_error_log())
|
|
|
|
result=1;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2002-03-22 21:55:08 +01:00
|
|
|
#ifdef HAVE_QUERY_CACHE
|
2001-12-02 13:34:01 +01:00
|
|
|
if (options & REFRESH_QUERY_CACHE_FREE)
|
|
|
|
{
|
2001-12-06 00:05:30 +01:00
|
|
|
query_cache.pack(); // FLUSH QUERY CACHE
|
2005-02-22 00:15:31 +01:00
|
|
|
options &= ~REFRESH_QUERY_CACHE; // Don't flush cache, just free memory
|
2001-12-02 13:34:01 +01:00
|
|
|
}
|
|
|
|
if (options & (REFRESH_TABLES | REFRESH_QUERY_CACHE))
|
|
|
|
{
|
2001-12-06 00:05:30 +01:00
|
|
|
query_cache.flush(); // RESET QUERY CACHE
|
2001-12-02 13:34:01 +01:00
|
|
|
}
|
2002-03-22 21:55:08 +01:00
|
|
|
#endif /*HAVE_QUERY_CACHE*/
|
2003-05-15 18:35:39 +02:00
|
|
|
/*
|
|
|
|
Note that if REFRESH_READ_LOCK bit is set then REFRESH_TABLES is set too
|
|
|
|
(see sql_yacc.yy)
|
|
|
|
*/
|
|
|
|
if (options & (REFRESH_TABLES | REFRESH_READ_LOCK))
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2001-08-14 19:33:49 +02:00
|
|
|
if ((options & REFRESH_READ_LOCK) && thd)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
2005-08-05 15:37:24 +02:00
|
|
|
/*
|
|
|
|
We must not try to aspire a global read lock if we have a write
|
|
|
|
locked table. This would lead to a deadlock when trying to
|
|
|
|
reopen (and re-lock) the table after the flush.
|
|
|
|
*/
|
|
|
|
if (thd->locked_tables)
|
|
|
|
{
|
|
|
|
THR_LOCK_DATA **lock_p= thd->locked_tables->locks;
|
|
|
|
THR_LOCK_DATA **end_p= lock_p + thd->locked_tables->lock_count;
|
|
|
|
|
|
|
|
for (; lock_p < end_p; lock_p++)
|
2005-08-08 23:13:49 +02:00
|
|
|
{
|
2005-08-05 15:37:24 +02:00
|
|
|
if ((*lock_p)->type == TL_WRITE)
|
|
|
|
{
|
|
|
|
my_error(ER_LOCK_OR_ACTIVE_TRANSACTION, MYF(0));
|
|
|
|
return 1;
|
|
|
|
}
|
2005-08-08 23:13:49 +02:00
|
|
|
}
|
2005-08-05 15:37:24 +02:00
|
|
|
}
|
2003-10-08 20:50:05 +02:00
|
|
|
/*
|
|
|
|
Writing to the binlog could cause deadlocks, as we don't log
|
|
|
|
UNLOCK TABLES
|
|
|
|
*/
|
2003-05-15 18:35:39 +02:00
|
|
|
tmp_write_to_binlog= 0;
|
2001-08-14 19:33:49 +02:00
|
|
|
if (lock_global_read_lock(thd))
|
2005-08-11 14:58:15 +02:00
|
|
|
return 1; // Killed
|
2004-08-20 16:35:23 +02:00
|
|
|
result=close_cached_tables(thd,(options & REFRESH_FAST) ? 0 : 1,
|
|
|
|
tables);
|
2005-08-11 14:58:15 +02:00
|
|
|
if (make_global_read_lock_block_commit(thd)) // Killed
|
2004-12-02 23:02:38 +01:00
|
|
|
{
|
|
|
|
/* Don't leave things in a half-locked state */
|
|
|
|
unlock_global_read_lock(thd);
|
|
|
|
return 1;
|
|
|
|
}
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
2004-08-20 16:35:23 +02:00
|
|
|
else
|
|
|
|
result=close_cached_tables(thd,(options & REFRESH_FAST) ? 0 : 1, tables);
|
2004-07-09 09:55:16 +02:00
|
|
|
my_dbopt_cleanup();
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
if (options & REFRESH_HOSTS)
|
|
|
|
hostname_cache_refresh();
|
2004-12-06 16:15:54 +01:00
|
|
|
if (thd && (options & REFRESH_STATUS))
|
2006-01-10 17:56:23 +01:00
|
|
|
refresh_status(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
if (options & REFRESH_THREADS)
|
|
|
|
flush_thread_cache();
|
2003-09-08 12:08:53 +02:00
|
|
|
#ifdef HAVE_REPLICATION
|
2000-07-31 21:29:14 +02:00
|
|
|
if (options & REFRESH_MASTER)
|
2003-05-15 18:35:39 +02:00
|
|
|
{
|
2006-01-03 17:54:54 +01:00
|
|
|
DBUG_ASSERT(thd);
|
2003-05-15 18:35:39 +02:00
|
|
|
tmp_write_to_binlog= 0;
|
2002-01-20 03:16:52 +01:00
|
|
|
if (reset_master(thd))
|
2005-08-11 14:58:15 +02:00
|
|
|
{
|
2002-01-20 03:16:52 +01:00
|
|
|
result=1;
|
2005-08-11 14:58:15 +02:00
|
|
|
thd->fatal_error(); // Ensure client get error
|
|
|
|
}
|
2003-05-15 18:35:39 +02:00
|
|
|
}
|
2002-12-16 14:33:29 +01:00
|
|
|
#endif
|
2001-12-13 14:53:18 +01:00
|
|
|
#ifdef OPENSSL
|
2002-01-20 06:46:25 +01:00
|
|
|
if (options & REFRESH_DES_KEY_FILE)
|
|
|
|
{
|
|
|
|
if (des_key_file)
|
|
|
|
result=load_des_key_file(des_key_file);
|
|
|
|
}
|
|
|
|
#endif
|
2003-09-08 12:08:53 +02:00
|
|
|
#ifdef HAVE_REPLICATION
|
2002-05-15 12:50:38 +02:00
|
|
|
if (options & REFRESH_SLAVE)
|
|
|
|
{
|
2003-05-15 18:35:39 +02:00
|
|
|
tmp_write_to_binlog= 0;
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_lock(&LOCK_active_mi);
|
2002-08-08 02:12:02 +02:00
|
|
|
if (reset_slave(thd, active_mi))
|
2002-05-15 12:50:38 +02:00
|
|
|
result=1;
|
2004-03-11 16:23:35 +01:00
|
|
|
pthread_mutex_unlock(&LOCK_active_mi);
|
2002-05-15 12:50:38 +02:00
|
|
|
}
|
2002-12-16 14:33:29 +01:00
|
|
|
#endif
|
2002-05-15 12:50:38 +02:00
|
|
|
if (options & REFRESH_USER_RESOURCES)
|
2004-12-06 16:15:54 +01:00
|
|
|
reset_mqh((LEX_USER *) NULL);
|
2005-08-11 14:58:15 +02:00
|
|
|
*write_to_binlog= tmp_write_to_binlog;
|
2002-05-15 12:50:38 +02:00
|
|
|
return result;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
2006-05-22 20:46:13 +02:00
|
|
|
|
2002-08-22 15:50:58 +02:00
|
|
|
/*
|
2006-05-22 20:46:13 +02:00
|
|
|
kills a thread
|
2002-08-22 15:50:58 +02:00
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
kill_one_thread()
|
|
|
|
thd Thread class
|
|
|
|
id Thread id
|
2006-05-22 20:46:13 +02:00
|
|
|
only_kill_query Should it kill the query or the connection
|
2002-08-22 15:50:58 +02:00
|
|
|
|
|
|
|
NOTES
|
|
|
|
This is written such that we have a short lock on LOCK_thread_count
|
|
|
|
*/
|
|
|
|
|
2006-05-22 20:46:13 +02:00
|
|
|
uint kill_one_thread(THD *thd, ulong id, bool only_kill_query)
|
2000-07-31 21:29:14 +02:00
|
|
|
{
|
|
|
|
THD *tmp;
|
|
|
|
uint error=ER_NO_SUCH_THREAD;
|
2006-05-22 20:46:13 +02:00
|
|
|
DBUG_ENTER("kill_one_thread");
|
|
|
|
DBUG_PRINT("enter", ("id=%lu only_kill=%d", id, only_kill_query));
|
2002-08-22 15:50:58 +02:00
|
|
|
VOID(pthread_mutex_lock(&LOCK_thread_count)); // For unlink from list
|
|
|
|
I_List_iterator<THD> it(threads);
|
2000-07-31 21:29:14 +02:00
|
|
|
while ((tmp=it++))
|
|
|
|
{
|
2006-05-02 19:56:43 +02:00
|
|
|
if (tmp->command == COM_DAEMON)
|
|
|
|
continue;
|
2000-07-31 21:29:14 +02:00
|
|
|
if (tmp->thread_id == id)
|
|
|
|
{
|
2002-08-22 15:50:58 +02:00
|
|
|
pthread_mutex_lock(&tmp->LOCK_delete); // Lock from delete
|
|
|
|
break;
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
VOID(pthread_mutex_unlock(&LOCK_thread_count));
|
2002-08-22 15:50:58 +02:00
|
|
|
if (tmp)
|
|
|
|
{
|
2005-09-15 21:29:07 +02:00
|
|
|
if ((thd->security_ctx->master_access & SUPER_ACL) ||
|
|
|
|
!strcmp(thd->security_ctx->user, tmp->security_ctx->user))
|
2002-08-22 15:50:58 +02:00
|
|
|
{
|
2003-03-31 10:39:46 +02:00
|
|
|
tmp->awake(only_kill_query ? THD::KILL_QUERY : THD::KILL_CONNECTION);
|
2002-08-22 15:50:58 +02:00
|
|
|
error=0;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
error=ER_KILL_DENIED_ERROR;
|
|
|
|
pthread_mutex_unlock(&tmp->LOCK_delete);
|
|
|
|
}
|
2006-05-22 20:46:13 +02:00
|
|
|
DBUG_PRINT("exit", ("%d", error));
|
|
|
|
DBUG_RETURN(error);
|
|
|
|
}
|
|
|
|
|
2002-08-22 15:50:58 +02:00
|
|
|
|
2006-05-22 20:46:13 +02:00
|
|
|
/*
|
|
|
|
kills a thread and sends response
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
sql_kill()
|
|
|
|
thd Thread class
|
|
|
|
id Thread id
|
|
|
|
only_kill_query Should it kill the query or the connection
|
|
|
|
*/
|
|
|
|
|
|
|
|
void sql_kill(THD *thd, ulong id, bool only_kill_query)
|
|
|
|
{
|
|
|
|
uint error;
|
|
|
|
if (!(error= kill_one_thread(thd, id, only_kill_query)))
|
2002-10-02 12:33:08 +02:00
|
|
|
send_ok(thd);
|
2000-07-31 21:29:14 +02:00
|
|
|
else
|
2004-10-20 03:04:37 +02:00
|
|
|
my_error(error, MYF(0), id);
|
2000-07-31 21:29:14 +02:00
|
|
|
}
|
|
|
|
|
2005-01-19 11:55:54 +01:00
|
|
|
|
2001-06-01 03:27:59 +02:00
|
|
|
/* If pointer is not a null pointer, append filename to it */
|
|
|
|
|
2006-05-02 19:56:43 +02:00
|
|
|
bool append_file_to_dir(THD *thd, const char **filename_ptr,
|
|
|
|
const char *table_name)
|
2001-06-01 03:27:59 +02:00
|
|
|
{
|
2001-10-08 03:58:07 +02:00
|
|
|
char buff[FN_REFLEN],*ptr, *end;
|
2001-06-01 03:27:59 +02:00
|
|
|
if (!*filename_ptr)
|
|
|
|
return 0; // nothing to do
|
|
|
|
|
|
|
|
/* Check that the filename is not too long and it's a hard path */
|
|
|
|
if (strlen(*filename_ptr)+strlen(table_name) >= FN_REFLEN-1 ||
|
|
|
|
!test_if_hard_path(*filename_ptr))
|
|
|
|
{
|
2003-11-18 16:28:00 +01:00
|
|
|
my_error(ER_WRONG_TABLE_NAME, MYF(0), *filename_ptr);
|
2001-06-01 03:27:59 +02:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
/* Fix is using unix filename format on dos */
|
|
|
|
strmov(buff,*filename_ptr);
|
2001-10-08 03:58:07 +02:00
|
|
|
end=convert_dirname(buff, *filename_ptr, NullS);
|
2001-12-05 12:03:00 +01:00
|
|
|
if (!(ptr=thd->alloc((uint) (end-buff)+(uint) strlen(table_name)+1)))
|
2001-06-01 03:27:59 +02:00
|
|
|
return 1; // End of memory
|
|
|
|
*filename_ptr=ptr;
|
2001-06-05 02:38:10 +02:00
|
|
|
strxmov(ptr,buff,table_name,NullS);
|
2001-06-01 03:27:59 +02:00
|
|
|
return 0;
|
|
|
|
}
|
2002-07-24 18:55:08 +02:00
|
|
|
|
2003-06-27 15:29:10 +02:00
|
|
|
|
2002-07-24 18:55:08 +02:00
|
|
|
/*
|
|
|
|
Check if the select is a simple select (not an union)
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
check_simple_select()
|
|
|
|
|
|
|
|
RETURN VALUES
|
|
|
|
0 ok
|
|
|
|
1 error ; In this case the error messege is sent to the client
|
|
|
|
*/
|
|
|
|
|
|
|
|
bool check_simple_select()
|
|
|
|
{
|
|
|
|
THD *thd= current_thd;
|
2004-11-11 20:37:48 +01:00
|
|
|
LEX *lex= thd->lex;
|
|
|
|
if (lex->current_select != &lex->select_lex)
|
2002-07-24 18:55:08 +02:00
|
|
|
{
|
|
|
|
char command[80];
|
2004-11-11 20:37:48 +01:00
|
|
|
strmake(command, lex->yylval->symbol.str,
|
|
|
|
min(lex->yylval->symbol.length, sizeof(command)-1));
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_CANT_USE_OPTION_HERE, MYF(0), command);
|
2002-07-24 18:55:08 +02:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
2002-11-07 22:45:19 +01:00
|
|
|
|
2003-10-08 16:40:54 +02:00
|
|
|
|
2003-11-03 11:28:36 +01:00
|
|
|
Comp_creator *comp_eq_creator(bool invert)
|
2002-11-07 22:45:19 +01:00
|
|
|
{
|
2003-11-03 11:28:36 +01:00
|
|
|
return invert?(Comp_creator *)&ne_creator:(Comp_creator *)&eq_creator;
|
2002-11-07 22:45:19 +01:00
|
|
|
}
|
|
|
|
|
2003-10-08 16:40:54 +02:00
|
|
|
|
2003-11-03 11:28:36 +01:00
|
|
|
Comp_creator *comp_ge_creator(bool invert)
|
2002-11-07 22:45:19 +01:00
|
|
|
{
|
2003-11-03 11:28:36 +01:00
|
|
|
return invert?(Comp_creator *)<_creator:(Comp_creator *)&ge_creator;
|
2002-11-07 22:45:19 +01:00
|
|
|
}
|
|
|
|
|
2003-10-08 16:40:54 +02:00
|
|
|
|
2003-11-03 11:28:36 +01:00
|
|
|
Comp_creator *comp_gt_creator(bool invert)
|
2002-11-07 22:45:19 +01:00
|
|
|
{
|
2003-11-03 11:28:36 +01:00
|
|
|
return invert?(Comp_creator *)&le_creator:(Comp_creator *)>_creator;
|
2002-11-07 22:45:19 +01:00
|
|
|
}
|
|
|
|
|
2003-10-08 16:40:54 +02:00
|
|
|
|
2003-11-03 11:28:36 +01:00
|
|
|
Comp_creator *comp_le_creator(bool invert)
|
2002-11-07 22:45:19 +01:00
|
|
|
{
|
2003-11-03 11:28:36 +01:00
|
|
|
return invert?(Comp_creator *)>_creator:(Comp_creator *)&le_creator;
|
2002-11-07 22:45:19 +01:00
|
|
|
}
|
|
|
|
|
2003-10-08 16:40:54 +02:00
|
|
|
|
2003-11-03 11:28:36 +01:00
|
|
|
Comp_creator *comp_lt_creator(bool invert)
|
2002-11-07 22:45:19 +01:00
|
|
|
{
|
2003-11-03 11:28:36 +01:00
|
|
|
return invert?(Comp_creator *)&ge_creator:(Comp_creator *)<_creator;
|
2002-11-07 22:45:19 +01:00
|
|
|
}
|
|
|
|
|
2003-10-08 16:40:54 +02:00
|
|
|
|
2003-11-03 11:28:36 +01:00
|
|
|
Comp_creator *comp_ne_creator(bool invert)
|
2002-11-07 22:45:19 +01:00
|
|
|
{
|
2003-11-03 11:28:36 +01:00
|
|
|
return invert?(Comp_creator *)&eq_creator:(Comp_creator *)&ne_creator;
|
2002-11-07 22:45:19 +01:00
|
|
|
}
|
2003-10-08 16:40:54 +02:00
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Construct ALL/ANY/SOME subquery Item
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
all_any_subquery_creator()
|
|
|
|
left_expr - pointer to left expression
|
|
|
|
cmp - compare function creator
|
|
|
|
all - true if we create ALL subquery
|
|
|
|
select_lex - pointer on parsed subquery structure
|
|
|
|
|
|
|
|
RETURN VALUE
|
|
|
|
constructed Item (or 0 if out of memory)
|
|
|
|
*/
|
|
|
|
Item * all_any_subquery_creator(Item *left_expr,
|
|
|
|
chooser_compare_func_creator cmp,
|
|
|
|
bool all,
|
|
|
|
SELECT_LEX *select_lex)
|
|
|
|
{
|
2003-10-25 12:00:53 +02:00
|
|
|
if ((cmp == &comp_eq_creator) && !all) // = ANY <=> IN
|
2003-10-08 16:40:54 +02:00
|
|
|
return new Item_in_subselect(left_expr, select_lex);
|
2003-10-25 12:00:53 +02:00
|
|
|
|
|
|
|
if ((cmp == &comp_ne_creator) && all) // <> ALL <=> NOT IN
|
2003-10-08 16:40:54 +02:00
|
|
|
return new Item_func_not(new Item_in_subselect(left_expr, select_lex));
|
|
|
|
|
|
|
|
Item_allany_subselect *it=
|
2003-10-30 11:57:26 +01:00
|
|
|
new Item_allany_subselect(left_expr, (*cmp)(all), select_lex, all);
|
2003-10-08 16:40:54 +02:00
|
|
|
if (all)
|
2004-11-18 17:10:07 +01:00
|
|
|
return it->upper_item= new Item_func_not_all(it); /* ALL */
|
2003-10-08 16:40:54 +02:00
|
|
|
|
2004-11-18 17:10:07 +01:00
|
|
|
return it->upper_item= new Item_func_nop_all(it); /* ANY/SOME */
|
2003-10-08 16:40:54 +02:00
|
|
|
}
|
2004-04-07 23:16:17 +02:00
|
|
|
|
|
|
|
|
2004-04-08 16:56:45 +02:00
|
|
|
/*
|
|
|
|
CREATE INDEX and DROP INDEX are implemented by calling ALTER TABLE with
|
|
|
|
the proper arguments. This isn't very fast but it should work for most
|
|
|
|
cases.
|
|
|
|
|
|
|
|
In the future ALTER TABLE will notice that only added indexes
|
|
|
|
and create these one by one for the existing table without having to do
|
|
|
|
a full rebuild.
|
|
|
|
|
|
|
|
One should normally create all indexes with CREATE TABLE or ALTER TABLE.
|
|
|
|
*/
|
|
|
|
|
2004-10-20 03:04:37 +02:00
|
|
|
bool mysql_create_index(THD *thd, TABLE_LIST *table_list, List<Key> &keys)
|
2004-04-08 16:56:45 +02:00
|
|
|
{
|
|
|
|
List<create_field> fields;
|
2004-05-21 16:57:03 +02:00
|
|
|
ALTER_INFO alter_info;
|
|
|
|
alter_info.flags= ALTER_ADD_INDEX;
|
2004-04-08 16:56:45 +02:00
|
|
|
HA_CREATE_INFO create_info;
|
|
|
|
DBUG_ENTER("mysql_create_index");
|
|
|
|
bzero((char*) &create_info,sizeof(create_info));
|
2006-05-28 14:51:01 +02:00
|
|
|
create_info.db_type= 0;
|
2004-04-08 16:56:45 +02:00
|
|
|
create_info.default_table_charset= thd->variables.collation_database;
|
2005-01-06 12:00:13 +01:00
|
|
|
DBUG_RETURN(mysql_alter_table(thd,table_list->db,table_list->table_name,
|
2004-04-08 16:56:45 +02:00
|
|
|
&create_info, table_list,
|
2004-05-21 16:57:03 +02:00
|
|
|
fields, keys, 0, (ORDER*)0,
|
2005-03-16 15:11:01 +01:00
|
|
|
DUP_ERROR, 0, &alter_info, 1));
|
2004-04-08 16:56:45 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2004-10-20 03:04:37 +02:00
|
|
|
bool mysql_drop_index(THD *thd, TABLE_LIST *table_list, ALTER_INFO *alter_info)
|
2004-04-08 16:56:45 +02:00
|
|
|
{
|
|
|
|
List<create_field> fields;
|
|
|
|
List<Key> keys;
|
|
|
|
HA_CREATE_INFO create_info;
|
|
|
|
DBUG_ENTER("mysql_drop_index");
|
|
|
|
bzero((char*) &create_info,sizeof(create_info));
|
2006-05-28 14:51:01 +02:00
|
|
|
create_info.db_type= 0;
|
2004-04-08 16:56:45 +02:00
|
|
|
create_info.default_table_charset= thd->variables.collation_database;
|
2004-05-21 16:57:03 +02:00
|
|
|
alter_info->clear();
|
|
|
|
alter_info->flags= ALTER_DROP_INDEX;
|
2005-01-06 12:00:13 +01:00
|
|
|
DBUG_RETURN(mysql_alter_table(thd,table_list->db,table_list->table_name,
|
2004-04-08 16:56:45 +02:00
|
|
|
&create_info, table_list,
|
2004-05-21 16:57:03 +02:00
|
|
|
fields, keys, 0, (ORDER*)0,
|
2005-03-16 15:11:01 +01:00
|
|
|
DUP_ERROR, 0, alter_info, 1));
|
2004-04-08 16:56:45 +02:00
|
|
|
}
|
2004-04-08 22:50:10 +02:00
|
|
|
|
|
|
|
|
2004-04-07 23:16:17 +02:00
|
|
|
/*
|
|
|
|
Multi update query pre-check
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
multi_update_precheck()
|
2004-04-12 02:26:32 +02:00
|
|
|
thd Thread handler
|
2004-07-16 00:15:55 +02:00
|
|
|
tables Global/local table list (have to be the same)
|
2004-04-07 23:16:17 +02:00
|
|
|
|
2004-04-10 00:14:32 +02:00
|
|
|
RETURN VALUE
|
2004-10-20 03:04:37 +02:00
|
|
|
FALSE OK
|
|
|
|
TRUE Error
|
2004-04-07 23:16:17 +02:00
|
|
|
*/
|
2004-04-12 02:26:32 +02:00
|
|
|
|
2004-10-20 03:04:37 +02:00
|
|
|
bool multi_update_precheck(THD *thd, TABLE_LIST *tables)
|
2004-04-07 23:16:17 +02:00
|
|
|
{
|
|
|
|
const char *msg= 0;
|
|
|
|
TABLE_LIST *table;
|
|
|
|
LEX *lex= thd->lex;
|
|
|
|
SELECT_LEX *select_lex= &lex->select_lex;
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ENTER("multi_update_precheck");
|
2004-04-07 23:16:17 +02:00
|
|
|
|
|
|
|
if (select_lex->item_list.elements != lex->value_list.elements)
|
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_message(ER_WRONG_VALUE_COUNT, ER(ER_WRONG_VALUE_COUNT), MYF(0));
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
2004-04-07 23:16:17 +02:00
|
|
|
}
|
|
|
|
/*
|
|
|
|
Ensure that we have UPDATE or SELECT privilege for each table
|
|
|
|
The exact privilege is checked in mysql_multi_update()
|
|
|
|
*/
|
2004-07-16 00:15:55 +02:00
|
|
|
for (table= tables; table; table= table->next_local)
|
2004-04-07 23:16:17 +02:00
|
|
|
{
|
2004-10-26 18:30:01 +02:00
|
|
|
if (table->derived)
|
|
|
|
table->grant.privilege= SELECT_ACL;
|
|
|
|
else if ((check_access(thd, UPDATE_ACL, table->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&table->grant.privilege, 0, 1,
|
|
|
|
test(table->schema_table)) ||
|
2004-10-26 18:30:01 +02:00
|
|
|
grant_option &&
|
|
|
|
check_grant(thd, UPDATE_ACL, table, 0, 1, 1)) &&
|
2004-10-29 18:26:52 +02:00
|
|
|
(check_access(thd, SELECT_ACL, table->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&table->grant.privilege, 0, 0,
|
|
|
|
test(table->schema_table)) ||
|
2004-10-29 18:26:52 +02:00
|
|
|
grant_option && check_grant(thd, SELECT_ACL, table, 0, 1, 0)))
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
2004-04-10 00:14:32 +02:00
|
|
|
|
2004-07-16 00:15:55 +02:00
|
|
|
table->table_in_first_from_clause= 1;
|
2004-04-07 23:16:17 +02:00
|
|
|
}
|
2004-04-10 00:14:32 +02:00
|
|
|
/*
|
|
|
|
Is there tables of subqueries?
|
|
|
|
*/
|
2005-06-17 17:56:04 +02:00
|
|
|
if (&lex->select_lex != lex->all_selects_list || lex->time_zone_tables_used)
|
2004-04-07 23:16:17 +02:00
|
|
|
{
|
2004-10-26 18:30:01 +02:00
|
|
|
DBUG_PRINT("info",("Checking sub query list"));
|
2004-07-16 00:15:55 +02:00
|
|
|
for (table= tables; table; table= table->next_global)
|
2004-04-07 23:16:17 +02:00
|
|
|
{
|
2004-12-30 23:44:00 +01:00
|
|
|
if (!my_tz_check_n_skip_implicit_tables(&table,
|
|
|
|
lex->time_zone_tables_used) &&
|
|
|
|
!table->table_in_first_from_clause)
|
2004-04-07 23:16:17 +02:00
|
|
|
{
|
|
|
|
if (check_access(thd, SELECT_ACL, table->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&table->grant.privilege, 0, 0,
|
|
|
|
test(table->schema_table)) ||
|
2004-04-10 00:14:32 +02:00
|
|
|
grant_option && check_grant(thd, SELECT_ACL, table, 0, 1, 0))
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
2004-04-07 23:16:17 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (select_lex->order_list.elements)
|
|
|
|
msg= "ORDER BY";
|
2005-06-07 12:11:36 +02:00
|
|
|
else if (select_lex->select_limit)
|
2004-04-07 23:16:17 +02:00
|
|
|
msg= "LIMIT";
|
|
|
|
if (msg)
|
|
|
|
{
|
|
|
|
my_error(ER_WRONG_USAGE, MYF(0), "UPDATE", msg);
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
2004-04-07 23:16:17 +02:00
|
|
|
}
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(FALSE);
|
2004-04-07 23:16:17 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
Multi delete query pre-check
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
multi_delete_precheck()
|
2004-04-12 02:26:32 +02:00
|
|
|
thd Thread handler
|
2004-07-16 00:15:55 +02:00
|
|
|
tables Global/local table list
|
2004-04-07 23:16:17 +02:00
|
|
|
|
2004-04-10 00:14:32 +02:00
|
|
|
RETURN VALUE
|
2004-10-20 03:04:37 +02:00
|
|
|
FALSE OK
|
|
|
|
TRUE error
|
2004-04-07 23:16:17 +02:00
|
|
|
*/
|
2004-10-20 03:04:37 +02:00
|
|
|
|
2005-06-08 23:07:52 +02:00
|
|
|
bool multi_delete_precheck(THD *thd, TABLE_LIST *tables)
|
2004-04-07 23:16:17 +02:00
|
|
|
{
|
|
|
|
SELECT_LEX *select_lex= &thd->lex->select_lex;
|
|
|
|
TABLE_LIST *aux_tables=
|
|
|
|
(TABLE_LIST *)thd->lex->auxilliary_table_list.first;
|
2004-07-16 00:15:55 +02:00
|
|
|
DBUG_ENTER("multi_delete_precheck");
|
2004-04-10 00:14:32 +02:00
|
|
|
|
2004-04-07 23:16:17 +02:00
|
|
|
/* sql_yacc guarantees that tables and aux_tables are not zero */
|
|
|
|
DBUG_ASSERT(aux_tables != 0);
|
|
|
|
if (check_db_used(thd, tables) || check_db_used(thd,aux_tables) ||
|
|
|
|
check_table_access(thd,SELECT_ACL, tables,0) ||
|
|
|
|
check_table_access(thd,DELETE_ACL, aux_tables,0))
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
2004-04-07 23:16:17 +02:00
|
|
|
if ((thd->options & OPTION_SAFE_UPDATES) && !select_lex->where)
|
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_UPDATE_WITHOUT_KEY_IN_SAFE_MODE,
|
|
|
|
ER(ER_UPDATE_WITHOUT_KEY_IN_SAFE_MODE), MYF(0));
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
2004-04-07 23:16:17 +02:00
|
|
|
}
|
2005-06-08 23:07:52 +02:00
|
|
|
DBUG_RETURN(FALSE);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Link tables in auxilary table list of multi-delete with corresponding
|
|
|
|
elements in main table list, and set proper locks for them.
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
multi_delete_set_locks_and_link_aux_tables()
|
|
|
|
lex - pointer to LEX representing multi-delete
|
|
|
|
|
|
|
|
RETURN VALUE
|
|
|
|
FALSE - success
|
|
|
|
TRUE - error
|
|
|
|
*/
|
|
|
|
|
|
|
|
bool multi_delete_set_locks_and_link_aux_tables(LEX *lex)
|
|
|
|
{
|
|
|
|
TABLE_LIST *tables= (TABLE_LIST*)lex->select_lex.table_list.first;
|
|
|
|
TABLE_LIST *target_tbl;
|
|
|
|
DBUG_ENTER("multi_delete_set_locks_and_link_aux_tables");
|
|
|
|
|
|
|
|
lex->table_count= 0;
|
|
|
|
|
|
|
|
for (target_tbl= (TABLE_LIST *)lex->auxilliary_table_list.first;
|
|
|
|
target_tbl; target_tbl= target_tbl->next_local)
|
2004-04-07 23:16:17 +02:00
|
|
|
{
|
2005-06-08 23:07:52 +02:00
|
|
|
lex->table_count++;
|
2004-04-07 23:16:17 +02:00
|
|
|
/* All tables in aux_tables must be found in FROM PART */
|
|
|
|
TABLE_LIST *walk;
|
2004-07-16 00:15:55 +02:00
|
|
|
for (walk= tables; walk; walk= walk->next_local)
|
2004-04-07 23:16:17 +02:00
|
|
|
{
|
2004-04-10 00:14:32 +02:00
|
|
|
if (!my_strcasecmp(table_alias_charset,
|
|
|
|
target_tbl->alias, walk->alias) &&
|
|
|
|
!strcmp(walk->db, target_tbl->db))
|
2004-04-07 23:16:17 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (!walk)
|
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
my_error(ER_UNKNOWN_TABLE, MYF(0),
|
2005-01-06 12:00:13 +01:00
|
|
|
target_tbl->table_name, "MULTI DELETE");
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
2004-04-07 23:16:17 +02:00
|
|
|
}
|
2005-09-21 17:38:26 +02:00
|
|
|
if (!walk->derived)
|
|
|
|
{
|
|
|
|
target_tbl->table_name= walk->table_name;
|
|
|
|
target_tbl->table_name_length= walk->table_name_length;
|
|
|
|
}
|
2005-09-21 12:37:51 +02:00
|
|
|
walk->updating= target_tbl->updating;
|
2004-04-10 00:14:32 +02:00
|
|
|
walk->lock_type= target_tbl->lock_type;
|
2004-07-16 00:15:55 +02:00
|
|
|
target_tbl->correspondent_table= walk; // Remember corresponding table
|
2004-04-07 23:16:17 +02:00
|
|
|
}
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(FALSE);
|
2004-04-07 23:16:17 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2004-04-10 00:14:32 +02:00
|
|
|
/*
|
|
|
|
simple UPDATE query pre-check
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
update_precheck()
|
2004-04-12 02:26:32 +02:00
|
|
|
thd Thread handler
|
|
|
|
tables Global table list
|
2004-04-10 00:14:32 +02:00
|
|
|
|
|
|
|
RETURN VALUE
|
2004-10-20 03:04:37 +02:00
|
|
|
FALSE OK
|
|
|
|
TRUE Error
|
2004-04-10 00:14:32 +02:00
|
|
|
*/
|
2004-04-12 02:26:32 +02:00
|
|
|
|
2004-10-20 03:04:37 +02:00
|
|
|
bool update_precheck(THD *thd, TABLE_LIST *tables)
|
2004-04-10 00:14:32 +02:00
|
|
|
{
|
|
|
|
DBUG_ENTER("update_precheck");
|
|
|
|
if (thd->lex->select_lex.item_list.elements != thd->lex->value_list.elements)
|
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_WRONG_VALUE_COUNT, ER(ER_WRONG_VALUE_COUNT), MYF(0));
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
2004-04-10 00:14:32 +02:00
|
|
|
}
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(check_db_used(thd, tables) ||
|
|
|
|
check_one_table_access(thd, UPDATE_ACL, tables));
|
2004-04-10 00:14:32 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
simple DELETE query pre-check
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
delete_precheck()
|
2004-04-12 02:26:32 +02:00
|
|
|
thd Thread handler
|
|
|
|
tables Global table list
|
2004-04-10 00:14:32 +02:00
|
|
|
|
|
|
|
RETURN VALUE
|
2004-10-20 03:04:37 +02:00
|
|
|
FALSE OK
|
|
|
|
TRUE error
|
2004-04-10 00:14:32 +02:00
|
|
|
*/
|
2004-04-12 02:26:32 +02:00
|
|
|
|
2004-10-20 03:04:37 +02:00
|
|
|
bool delete_precheck(THD *thd, TABLE_LIST *tables)
|
2004-04-10 00:14:32 +02:00
|
|
|
{
|
|
|
|
DBUG_ENTER("delete_precheck");
|
|
|
|
if (check_one_table_access(thd, DELETE_ACL, tables))
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
2004-04-12 02:26:32 +02:00
|
|
|
/* Set privilege for the WHERE clause */
|
2004-04-10 00:14:32 +02:00
|
|
|
tables->grant.want_privilege=(SELECT_ACL & ~tables->grant.privilege);
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(FALSE);
|
2004-04-10 00:14:32 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
simple INSERT query pre-check
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
insert_precheck()
|
2004-04-12 02:26:32 +02:00
|
|
|
thd Thread handler
|
|
|
|
tables Global table list
|
2004-04-10 00:14:32 +02:00
|
|
|
|
|
|
|
RETURN VALUE
|
2004-10-20 03:04:37 +02:00
|
|
|
FALSE OK
|
|
|
|
TRUE error
|
2004-04-10 00:14:32 +02:00
|
|
|
*/
|
2004-04-12 02:26:32 +02:00
|
|
|
|
2004-11-12 14:36:31 +01:00
|
|
|
bool insert_precheck(THD *thd, TABLE_LIST *tables)
|
2004-04-10 00:14:32 +02:00
|
|
|
{
|
|
|
|
LEX *lex= thd->lex;
|
|
|
|
DBUG_ENTER("insert_precheck");
|
|
|
|
|
2004-12-13 13:26:28 +01:00
|
|
|
/*
|
|
|
|
Check that we have modify privileges for the first table and
|
|
|
|
select privileges for the rest
|
|
|
|
*/
|
2004-11-03 11:39:38 +01:00
|
|
|
ulong privilege= (INSERT_ACL |
|
|
|
|
(lex->duplicates == DUP_REPLACE ? DELETE_ACL : 0) |
|
|
|
|
(lex->value_list.elements ? UPDATE_ACL : 0));
|
2004-04-10 00:14:32 +02:00
|
|
|
|
|
|
|
if (check_one_table_access(thd, privilege, tables))
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
2004-04-10 00:14:32 +02:00
|
|
|
|
2004-12-13 13:26:28 +01:00
|
|
|
if (lex->update_list.elements != lex->value_list.elements)
|
2004-04-10 00:14:32 +02:00
|
|
|
{
|
2004-11-12 13:34:00 +01:00
|
|
|
my_message(ER_WRONG_VALUE_COUNT, ER(ER_WRONG_VALUE_COUNT), MYF(0));
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(TRUE);
|
2004-04-10 00:14:32 +02:00
|
|
|
}
|
2004-10-20 03:04:37 +02:00
|
|
|
DBUG_RETURN(FALSE);
|
2004-04-07 23:16:17 +02:00
|
|
|
}
|
2004-04-10 00:14:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
CREATE TABLE query pre-check
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
create_table_precheck()
|
2004-04-12 02:26:32 +02:00
|
|
|
thd Thread handler
|
|
|
|
tables Global table list
|
|
|
|
create_table Table which will be created
|
2004-04-10 00:14:32 +02:00
|
|
|
|
|
|
|
RETURN VALUE
|
2004-10-20 03:04:37 +02:00
|
|
|
FALSE OK
|
|
|
|
TRUE Error
|
2004-04-10 00:14:32 +02:00
|
|
|
*/
|
2004-04-12 02:26:32 +02:00
|
|
|
|
2004-10-20 03:04:37 +02:00
|
|
|
bool create_table_precheck(THD *thd, TABLE_LIST *tables,
|
|
|
|
TABLE_LIST *create_table)
|
2004-04-10 00:14:32 +02:00
|
|
|
{
|
|
|
|
LEX *lex= thd->lex;
|
2004-10-22 17:44:51 +02:00
|
|
|
SELECT_LEX *select_lex= &lex->select_lex;
|
|
|
|
ulong want_priv;
|
2004-11-12 14:36:31 +01:00
|
|
|
bool error= TRUE; // Error message is given
|
2004-04-10 00:14:32 +02:00
|
|
|
DBUG_ENTER("create_table_precheck");
|
2004-10-22 17:44:51 +02:00
|
|
|
|
|
|
|
want_priv= ((lex->create_info.options & HA_LEX_CREATE_TMP_TABLE) ?
|
|
|
|
CREATE_TMP_ACL : CREATE_ACL);
|
2004-04-10 00:14:32 +02:00
|
|
|
lex->create_info.alias= create_table->alias;
|
|
|
|
if (check_access(thd, want_priv, create_table->db,
|
2005-09-13 13:07:38 +02:00
|
|
|
&create_table->grant.privilege, 0, 0,
|
|
|
|
test(create_table->schema_table)) ||
|
2004-04-10 00:14:32 +02:00
|
|
|
check_merge_table_access(thd, create_table->db,
|
|
|
|
(TABLE_LIST *)
|
|
|
|
lex->create_info.merge_list.first))
|
2004-10-22 17:44:51 +02:00
|
|
|
goto err;
|
|
|
|
if (grant_option && want_priv != CREATE_TMP_ACL &&
|
2006-05-10 12:12:32 +02:00
|
|
|
check_grant(thd, want_priv, create_table, 0, 1, 0))
|
2004-10-22 17:44:51 +02:00
|
|
|
goto err;
|
|
|
|
|
|
|
|
if (select_lex->item_list.elements)
|
|
|
|
{
|
|
|
|
/* Check permissions for used tables in CREATE TABLE ... SELECT */
|
|
|
|
|
2005-06-01 13:22:17 +02:00
|
|
|
#ifdef NOT_NECESSARY_TO_CHECK_CREATE_TABLE_EXIST_WHEN_PREPARING_STATEMENT
|
|
|
|
/* This code throws an ill error for CREATE TABLE t1 SELECT * FROM t1 */
|
2004-10-22 17:44:51 +02:00
|
|
|
/*
|
2004-10-29 18:26:52 +02:00
|
|
|
Only do the check for PS, becasue we on execute we have to check that
|
|
|
|
against the opened tables to ensure we don't use a table that is part
|
|
|
|
of the view (which can only be done after the table has been opened).
|
2004-10-22 17:44:51 +02:00
|
|
|
*/
|
2005-09-02 15:21:19 +02:00
|
|
|
if (thd->stmt_arena->is_stmt_prepare_or_first_sp_execute())
|
2004-10-22 17:44:51 +02:00
|
|
|
{
|
2004-10-29 18:26:52 +02:00
|
|
|
/*
|
|
|
|
For temporary tables we don't have to check if the created table exists
|
|
|
|
*/
|
|
|
|
if (!(lex->create_info.options & HA_LEX_CREATE_TMP_TABLE) &&
|
2004-11-03 11:39:38 +01:00
|
|
|
find_table_in_global_list(tables, create_table->db,
|
2005-01-06 12:00:13 +01:00
|
|
|
create_table->table_name))
|
2004-10-29 18:26:52 +02:00
|
|
|
{
|
2004-11-13 18:35:51 +01:00
|
|
|
error= FALSE;
|
2004-10-29 18:26:52 +02:00
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
}
|
2005-06-01 13:22:17 +02:00
|
|
|
#endif
|
2004-10-22 17:44:51 +02:00
|
|
|
if (tables && check_table_access(thd, SELECT_ACL, tables,0))
|
|
|
|
goto err;
|
|
|
|
}
|
2004-11-12 14:36:31 +01:00
|
|
|
error= FALSE;
|
2004-10-22 17:44:51 +02:00
|
|
|
|
|
|
|
err:
|
|
|
|
DBUG_RETURN(error);
|
2004-04-10 00:14:32 +02:00
|
|
|
}
|
2004-08-31 20:10:57 +02:00
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
negate given expression
|
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
negate_expression()
|
2004-10-07 00:45:06 +02:00
|
|
|
thd thread handler
|
2004-08-31 20:10:57 +02:00
|
|
|
expr expression for negation
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
negated expression
|
|
|
|
*/
|
|
|
|
|
|
|
|
Item *negate_expression(THD *thd, Item *expr)
|
|
|
|
{
|
|
|
|
Item *negated;
|
|
|
|
if (expr->type() == Item::FUNC_ITEM &&
|
|
|
|
((Item_func *) expr)->functype() == Item_func::NOT_FUNC)
|
|
|
|
{
|
|
|
|
/* it is NOT(NOT( ... )) */
|
|
|
|
Item *arg= ((Item_func *) expr)->arguments()[0];
|
|
|
|
enum_parsing_place place= thd->lex->current_select->parsing_place;
|
|
|
|
if (arg->is_bool_func() || place == IN_WHERE || place == IN_HAVING)
|
|
|
|
return arg;
|
|
|
|
/*
|
|
|
|
if it is not boolean function then we have to emulate value of
|
|
|
|
not(not(a)), it will be a != 0
|
|
|
|
*/
|
|
|
|
return new Item_func_ne(arg, new Item_int((char*) "0", 0, 1));
|
|
|
|
}
|
|
|
|
|
|
|
|
if ((negated= expr->neg_transformer(thd)) != 0)
|
|
|
|
return negated;
|
|
|
|
return new Item_func_not(expr);
|
|
|
|
}
|
2005-09-14 09:53:09 +02:00
|
|
|
|
2005-11-10 20:25:03 +01:00
|
|
|
/*
|
|
|
|
Set the specified definer to the default value, which is the current user in
|
2006-03-02 12:17:13 +01:00
|
|
|
the thread.
|
2005-11-10 20:25:03 +01:00
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
get_default_definer()
|
|
|
|
thd [in] thread handler
|
|
|
|
definer [out] definer
|
|
|
|
*/
|
|
|
|
|
2006-03-02 12:17:13 +01:00
|
|
|
void get_default_definer(THD *thd, LEX_USER *definer)
|
2005-11-10 20:25:03 +01:00
|
|
|
{
|
|
|
|
const Security_context *sctx= thd->security_ctx;
|
|
|
|
|
|
|
|
definer->user.str= (char *) sctx->priv_user;
|
|
|
|
definer->user.length= strlen(definer->user.str);
|
|
|
|
|
|
|
|
definer->host.str= (char *) sctx->priv_host;
|
|
|
|
definer->host.length= strlen(definer->host.str);
|
|
|
|
}
|
|
|
|
|
2005-09-14 09:53:09 +02:00
|
|
|
|
2006-03-01 12:13:07 +01:00
|
|
|
/*
|
2006-03-02 12:17:13 +01:00
|
|
|
Create default definer for the specified THD.
|
2006-03-01 12:13:07 +01:00
|
|
|
|
|
|
|
SYNOPSIS
|
|
|
|
create_default_definer()
|
|
|
|
thd [in] thread handler
|
|
|
|
|
|
|
|
RETURN
|
|
|
|
On success, return a valid pointer to the created and initialized
|
|
|
|
LEX_USER, which contains definer information.
|
|
|
|
On error, return 0.
|
|
|
|
*/
|
|
|
|
|
|
|
|
LEX_USER *create_default_definer(THD *thd)
|
|
|
|
{
|
|
|
|
LEX_USER *definer;
|
|
|
|
|
|
|
|
if (! (definer= (LEX_USER*) thd->alloc(sizeof(LEX_USER))))
|
|
|
|
return 0;
|
|
|
|
|
2006-03-02 12:17:13 +01:00
|
|
|
get_default_definer(thd, definer);
|
2006-03-01 12:13:07 +01:00
|
|
|
|
|
|
|
return definer;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2005-09-14 09:53:09 +02:00
|
|
|
/*
|
2006-03-10 12:40:15 +01:00
|
|
|
Create definer with the given user and host names.
|
2005-09-14 09:53:09 +02:00
|
|
|
|
|
|
|
SYNOPSIS
|
2005-11-10 20:25:03 +01:00
|
|
|
create_definer()
|
|
|
|
thd [in] thread handler
|
|
|
|
user_name [in] user name
|
|
|
|
host_name [in] host name
|
2005-09-14 09:53:09 +02:00
|
|
|
|
|
|
|
RETURN
|
2005-11-10 20:25:03 +01:00
|
|
|
On success, return a valid pointer to the created and initialized
|
2006-03-01 12:13:07 +01:00
|
|
|
LEX_USER, which contains definer information.
|
2005-11-10 20:25:03 +01:00
|
|
|
On error, return 0.
|
2005-09-14 09:53:09 +02:00
|
|
|
*/
|
|
|
|
|
2005-11-10 20:25:03 +01:00
|
|
|
LEX_USER *create_definer(THD *thd, LEX_STRING *user_name, LEX_STRING *host_name)
|
2005-09-14 09:53:09 +02:00
|
|
|
{
|
2005-11-10 20:25:03 +01:00
|
|
|
LEX_USER *definer;
|
|
|
|
|
|
|
|
/* Create and initialize. */
|
|
|
|
|
2006-01-05 23:47:49 +01:00
|
|
|
if (! (definer= (LEX_USER*) thd->alloc(sizeof(LEX_USER))))
|
2005-11-10 20:25:03 +01:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
definer->user= *user_name;
|
|
|
|
definer->host= *host_name;
|
|
|
|
|
|
|
|
return definer;
|
2005-09-14 09:53:09 +02:00
|
|
|
}
|