mariadb/sql/sql_view.cc

1877 lines
56 KiB
C++
Raw Normal View History

2004-07-16 00:15:55 +02:00
/* Copyright (C) 2004 MySQL AB
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; version 2 of the License.
2004-07-16 00:15:55 +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.
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
*/
#define MYSQL_LEX 1
2004-07-16 00:15:55 +02:00
#include "mysql_priv.h"
#include "sql_select.h"
#include "parse_file.h"
2004-07-16 14:20:51 +02:00
#include "sp.h"
#include "sp_head.h"
#include "sp_cache.h"
2004-07-16 00:15:55 +02:00
#define MD5_BUFF_LENGTH 33
const LEX_STRING view_type= { C_STRING_WITH_LEN("VIEW") };
2004-07-16 00:15:55 +02:00
static int mysql_register_view(THD *thd, TABLE_LIST *view,
enum_view_create_mode mode);
const char *updatable_views_with_limit_names[]= { "NO", "YES", NullS };
TYPELIB updatable_views_with_limit_typelib=
2004-07-16 00:15:55 +02:00
{
array_elements(updatable_views_with_limit_names)-1, "",
updatable_views_with_limit_names,
0
2004-07-16 00:15:55 +02:00
};
/*
Make a unique name for an anonymous view column
SYNOPSIS
target reference to the item for which a new name has to be made
item_list list of items within which we should check uniqueness of
the created name
last_element the last element of the list above
NOTE
Unique names are generated by adding 'My_exp_' to the old name of the
column. In case the name that was created this way already exists, we
add a numeric postfix to its end (i.e. "1") and increase the number
until the name becomes unique. If the generated name is longer than
NAME_LEN, it is truncated.
*/
static void make_unique_view_field_name(Item *target,
List<Item> &item_list,
Item *last_element)
{
char *name= (target->orig_name ?
target->orig_name :
target->name);
uint name_len, attempt;
char buff[NAME_LEN+1];
List_iterator_fast<Item> itc(item_list);
for (attempt= 0;; attempt++)
{
Item *check;
bool ok= TRUE;
if (attempt)
name_len= my_snprintf(buff, NAME_LEN, "My_exp_%d_%s", attempt, name);
else
name_len= my_snprintf(buff, NAME_LEN, "My_exp_%s", name);
do
{
check= itc++;
if (check != target &&
my_strcasecmp(system_charset_info, buff, check->name) == 0)
{
ok= FALSE;
break;
}
} while (check != last_element);
if (ok)
break;
itc.rewind();
}
target->orig_name= target->name;
target->set_name(buff, name_len, system_charset_info);
}
/*
Check if items with same names are present in list and possibly
generate unique names for them.
SYNOPSIS
item_list list of Items which should be checked for duplicates
gen_unique_view_name flag: generate unique name or return with error when
duplicate names are found.
DESCRIPTION
This function is used on view creation and preparation of derived tables.
It checks item_list for items with duplicate names. If it founds two
items with same name and conversion to unique names isn't allowed, or
names for both items are set by user - function fails.
Otherwise it generates unique name for one item with autogenerated name
using make_unique_view_field_name()
RETURN VALUE
FALSE no duplicate names found, or they are converted to unique ones
TRUE duplicate names are found and they can't be converted or conversion
isn't allowed
*/
bool check_duplicate_names(List<Item> &item_list, bool gen_unique_view_name)
{
Item *item;
List_iterator_fast<Item> it(item_list);
List_iterator_fast<Item> itc(item_list);
DBUG_ENTER("check_duplicate_names");
while ((item= it++))
{
Item *check;
/* treat underlying fields like set by user names */
if (item->real_item()->type() == Item::FIELD_ITEM)
item->is_autogenerated_name= FALSE;
itc.rewind();
while ((check= itc++) && check != item)
{
if (my_strcasecmp(system_charset_info, item->name, check->name) == 0)
{
if (!gen_unique_view_name)
goto err;
if (item->is_autogenerated_name)
make_unique_view_field_name(item, item_list, item);
else if (check->is_autogenerated_name)
make_unique_view_field_name(check, item_list, item);
else
goto err;
}
}
}
DBUG_RETURN(FALSE);
err:
my_error(ER_DUP_FIELDNAME, MYF(0), item->name);
DBUG_RETURN(TRUE);
}
/*
Fill defined view parts
SYNOPSIS
fill_defined_view_parts()
thd current thread.
view view to operate on
DESCRIPTION
This function will initialize the parts of the view
definition that are not specified in ALTER VIEW
to their values from CREATE VIEW.
The view must be opened to get its definition.
We use a copy of the view when opening because we want
to preserve the original view instance.
RETURN VALUE
TRUE can't open table
FALSE success
*/
static bool
fill_defined_view_parts (THD *thd, TABLE_LIST *view)
{
LEX *lex= thd->lex;
bool not_used;
TABLE_LIST decoy;
memcpy (&decoy, view, sizeof (TABLE_LIST));
if (!open_table(thd, &decoy, thd->mem_root, &not_used, OPEN_VIEW_NO_PARSE) &&
!decoy.view)
{
2006-08-06 22:06:03 +02:00
/* It's a table */
return TRUE;
}
2006-08-06 22:06:03 +02:00
if (!lex->definer)
{
view->definer.host= decoy.definer.host;
view->definer.user= decoy.definer.user;
lex->definer= &view->definer;
}
if (lex->create_view_algorithm == VIEW_ALGORITHM_UNDEFINED)
lex->create_view_algorithm= (uint8) decoy.algorithm;
if (lex->create_view_suid == VIEW_SUID_DEFAULT)
lex->create_view_suid= decoy.view_suid ?
VIEW_SUID_DEFINER : VIEW_SUID_INVOKER;
return FALSE;
}
/**
@brief Creating/altering VIEW procedure
2004-07-16 00:15:55 +02:00
@param thd thread handler
@param views views to create
@param mode VIEW_CREATE_NEW, VIEW_ALTER, VIEW_CREATE_OR_REPLACE
2004-07-16 00:15:55 +02:00
@note This function handles both create and alter view commands.
@retval FALSE Operation was a success.
@retval TRUE An error occured.
2004-07-16 00:15:55 +02:00
*/
bool mysql_create_view(THD *thd, TABLE_LIST *views,
enum_view_create_mode mode)
2004-07-16 00:15:55 +02:00
{
LEX *lex= thd->lex;
bool link_to_local;
2007-03-22 21:13:40 +01:00
#ifndef NO_EMBEDDED_ACCESS_CHECKS
bool definer_check_is_needed= mode != VIEW_ALTER || lex->definer;
2007-03-22 21:13:40 +01:00
#endif
2004-07-16 00:15:55 +02:00
/* first table in list is target VIEW name => cut off it */
TABLE_LIST *view= lex->unlink_first_table(&link_to_local);
TABLE_LIST *tables= lex->query_tables;
2004-07-16 22:44:14 +02:00
TABLE_LIST *tbl;
2004-11-17 13:17:09 +01:00
SELECT_LEX *select_lex= &lex->select_lex;
#ifndef NO_EMBEDDED_ACCESS_CHECKS
SELECT_LEX *sl;
#endif
2004-07-16 00:15:55 +02:00
SELECT_LEX_UNIT *unit= &lex->unit;
bool res= FALSE;
2004-07-16 00:15:55 +02:00
DBUG_ENTER("mysql_create_view");
/* This is ensured in the parser. */
DBUG_ASSERT(!lex->proc_list.first && !lex->result &&
!lex->param_list.elements && !lex->derived_tables);
2004-07-16 00:15:55 +02:00
if (mode != VIEW_CREATE_NEW)
{
if (mode == VIEW_ALTER &&
fill_defined_view_parts(thd, view))
{
res= TRUE;
goto err;
}
sp_cache_invalidate();
}
if (!lex->definer)
{
/*
DEFINER-clause is missing; we have to create default definer in
persistent arena to be PS/SP friendly.
If this is an ALTER VIEW then the current user should be set as
the definer.
*/
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)
goto err;
}
2004-07-16 00:15:55 +02:00
#ifndef NO_EMBEDDED_ACCESS_CHECKS
/*
check definer of view:
- same as current user
- current user has SUPER_ACL
*/
if (definer_check_is_needed &&
(strcmp(lex->definer->user.str, thd->security_ctx->priv_user) != 0 ||
my_strcasecmp(system_charset_info,
lex->definer->host.str,
thd->security_ctx->priv_host) != 0))
{
if (!(thd->security_ctx->master_access & SUPER_ACL))
{
my_error(ER_SPECIFIC_ACCESS_DENIED_ERROR, MYF(0), "SUPER");
res= TRUE;
goto err;
}
else
{
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);
}
}
}
/*
Privilege check for view creation:
2005-04-02 17:45:44 +02:00
- user has CREATE VIEW privilege on view table
- user has DROP privilege in case of ALTER VIEW or CREATE OR REPLACE
VIEW
2005-04-02 17:45:44 +02:00
- user has some (SELECT/UPDATE/INSERT/DELETE) privileges on columns of
underlying tables used on top of SELECT list (because it can be
(theoretically) updated, so it is enough to have UPDATE privilege on
them, for example)
2005-04-02 17:45:44 +02:00
- user has SELECT privilege on columns used in expressions of VIEW select
- for columns of underly tables used on top of SELECT list also will be
checked that we have not more privileges on correspondent column of view
table (i.e. user will not get some privileges by view creation)
*/
if ((check_access(thd, CREATE_VIEW_ACL, view->db, &view->grant.privilege,
0, 0, is_schema_db(view->db)) ||
check_grant(thd, CREATE_VIEW_ACL, view, 0, 1, 0)) ||
(mode != VIEW_CREATE_NEW &&
(check_access(thd, DROP_ACL, view->db, &view->grant.privilege,
0, 0, is_schema_db(view->db)) ||
check_grant(thd, DROP_ACL, view, 0, 1, 0))))
{
res= TRUE;
goto err;
}
for (sl= select_lex; sl; sl= sl->next_select())
2004-07-16 00:15:55 +02:00
{
for (tbl= sl->get_table_list(); tbl; tbl= tbl->next_local)
2004-07-16 00:15:55 +02:00
{
/*
Ensure that we have some privileges on this table, more strict check
will be done on column level after preparation,
*/
2004-09-03 21:38:45 +02:00
if (check_some_access(thd, VIEW_ANY_ACL, tbl))
{
my_error(ER_TABLEACCESS_DENIED_ERROR, MYF(0),
"ANY", thd->security_ctx->priv_user,
thd->security_ctx->priv_host, tbl->table_name);
res= TRUE;
goto err;
}
/*
Mark this table as a table which will be checked after the prepare
phase
*/
tbl->table_in_first_from_clause= 1;
2004-07-16 00:15:55 +02:00
/*
We need to check only SELECT_ACL for all normal fields, fields for
which we need "any" (SELECT/UPDATE/INSERT/DELETE) privilege will be
checked later
*/
tbl->grant.want_privilege= SELECT_ACL;
/*
Make sure that all rights are loaded to the TABLE::grant field.
tbl->table_name will be correct name of table because VIEWs are
not opened yet.
*/
fill_effective_table_privileges(thd, &tbl->grant, tbl->db,
tbl->table_name);
}
2004-07-16 00:15:55 +02:00
}
if (&lex->select_lex != lex->all_selects_list)
{
/* check tables of subqueries */
2004-07-16 22:44:14 +02:00
for (tbl= tables; tbl; tbl= tbl->next_global)
2004-07-16 00:15:55 +02:00
{
if (!tbl->table_in_first_from_clause)
{
if (check_access(thd, SELECT_ACL, tbl->db,
&tbl->grant.privilege, 0, 0, test(tbl->schema_table)) ||
check_grant(thd, SELECT_ACL, tbl, 0, 1, 0))
2004-07-16 00:15:55 +02:00
{
res= TRUE;
2004-07-16 00:15:55 +02:00
goto err;
}
}
}
}
/*
Mark fields for special privilege check ("any" privilege)
2004-07-16 00:15:55 +02:00
*/
for (sl= select_lex; sl; sl= sl->next_select())
2004-07-16 00:15:55 +02:00
{
List_iterator_fast<Item> it(sl->item_list);
2004-07-16 00:15:55 +02:00
Item *item;
while ((item= it++))
{
Item_field *field;
if ((field= item->filed_for_view_update()))
{
/*
any_privileges may be reset later by the Item_field::set_field
method in case of a system temporary table.
*/
field->any_privileges= 1;
}
2004-07-16 00:15:55 +02:00
}
}
#endif
if (open_and_lock_tables(thd, tables))
{
res= TRUE;
goto err;
}
2004-07-16 00:15:55 +02:00
/*
check that tables are not temporary and this VIEW do not used in query
(it is possible with ALTERing VIEW).
open_and_lock_tables can change the value of tables,
e.g. it may happen if before the function call tables was equal to 0.
*/
for (tbl= lex->query_tables; tbl; tbl= tbl->next_global)
2004-07-16 00:15:55 +02:00
{
/* is this table view and the same view which we creates now? */
if (tbl->view &&
strcmp(tbl->view_db.str, view->db) == 0 &&
strcmp(tbl->view_name.str, view->table_name) == 0)
{
my_error(ER_NO_SUCH_TABLE, MYF(0), tbl->view_db.str, tbl->view_name.str);
res= TRUE;
goto err;
}
2004-07-16 00:15:55 +02:00
/*
tbl->table can be NULL when tbl is a placeholder for a view
that is indirectly referenced via a stored function from the
view being created. We don't check these indirectly
referenced views in CREATE VIEW so they don't have table
object.
2004-07-16 00:15:55 +02:00
*/
if (tbl->table)
{
/* is this table temporary and is not view? */
if (tbl->table->s->tmp_table != NO_TMP_TABLE && !tbl->view &&
!tbl->schema_table)
{
my_error(ER_VIEW_SELECT_TMPTABLE, MYF(0), tbl->alias);
res= TRUE;
goto err;
}
/*
Copy the privileges of the underlying VIEWs which were filled by
fill_effective_table_privileges
(they were not copied at derived tables processing)
*/
tbl->table->grant.privilege= tbl->grant.privilege;
}
2004-07-16 00:15:55 +02:00
}
/* prepare select to resolve all fields */
2004-07-16 00:15:55 +02:00
lex->view_prepare_mode= 1;
if (unit->prepare(thd, 0, 0))
{
/*
some errors from prepare are reported to user, if is not then
it will be checked after err: label
*/
res= TRUE;
2004-07-16 00:15:55 +02:00
goto err;
}
2004-07-16 00:15:55 +02:00
/* view list (list of view fields names) */
if (lex->view_list.elements)
{
List_iterator_fast<Item> it(select_lex->item_list);
List_iterator_fast<LEX_STRING> nm(lex->view_list);
Item *item;
LEX_STRING *name;
if (lex->view_list.elements != select_lex->item_list.elements)
2004-07-16 00:15:55 +02:00
{
my_message(ER_VIEW_WRONG_LIST, ER(ER_VIEW_WRONG_LIST), MYF(0));
res= TRUE;
goto err;
2004-07-16 00:15:55 +02:00
}
while ((item= it++, name= nm++))
{
item->set_name(name->str, name->length, system_charset_info);
item->is_autogenerated_name= FALSE;
}
2004-07-16 00:15:55 +02:00
}
if (check_duplicate_names(select_lex->item_list, 1))
{
res= TRUE;
goto err;
}
2004-07-16 00:15:55 +02:00
#ifndef NO_EMBEDDED_ACCESS_CHECKS
/*
Compare/check grants on view with grants of underlying tables
2004-07-16 00:15:55 +02:00
*/
fill_effective_table_privileges(thd, &view->grant, view->db,
view->table_name);
{
Item *report_item= NULL;
uint final_priv= VIEW_ANY_ACL;
for (sl= select_lex; sl; sl= sl->next_select())
2004-07-16 00:15:55 +02:00
{
A fix and a test case for Bug#19022 "Memory bug when switching db during trigger execution" Bug#17199 "Problem when view calls function from another database." Bug#18444 "Fully qualified stored function names don't work correctly in SELECT statements" Documentation note: this patch introduces a change in behaviour of prepared statements. This patch adds a few new invariants with regard to how THD::db should be used. These invariants should be preserved in future: - one should never refer to THD::db by pointer and always make a deep copy (strmake, strdup) - one should never compare two databases by pointer, but use strncmp or my_strncasecmp - TABLE_LIST object table->db should be always initialized in the parser or by creator of the object. For prepared statements it means that if the current database is changed after a statement is prepared, the database that was current at prepare remains active. This also means that you can not prepare a statement that implicitly refers to the current database if the latter is not set. This is not documented, and therefore needs documentation. This is NOT a change in behavior for almost all SQL statements except: - ALTER TABLE t1 RENAME t2 - OPTIMIZE TABLE t1 - ANALYZE TABLE t1 - TRUNCATE TABLE t1 -- until this patch t1 or t2 could be evaluated at the first execution of prepared statement. CURRENT_DATABASE() still works OK and is evaluated at every execution of prepared statement. Note, that in stored routines this is not an issue as the default database is the database of the stored procedure and "use" statement is prohibited in stored routines. This patch makes obsolete the use of check_db_used (it was never used in the old code too) and all other places that check for table->db and assign it from THD::db if it's NULL, except the parser. How this patch was created: THD::{db,db_length} were replaced with a LEX_STRING, THD::db. All the places that refer to THD::{db,db_length} were manually checked and: - if the place uses thd->db by pointer, it was fixed to make a deep copy - if a place compared two db pointers, it was fixed to compare them by value (via strcmp/my_strcasecmp, whatever was approproate) Then this intermediate patch was used to write a smaller patch that does the same thing but without a rename. TODO in 5.1: - remove check_db_used - deploy THD::set_db in mysql_change_db See also comments to individual files.
2006-06-26 22:47:52 +02:00
DBUG_ASSERT(view->db); /* Must be set in the parser */
List_iterator_fast<Item> it(sl->item_list);
2004-07-16 00:15:55 +02:00
Item *item;
while ((item= it++))
2004-07-16 00:15:55 +02:00
{
Item_field *fld= item->filed_for_view_update();
A fix and a test case for Bug#19022 "Memory bug when switching db during trigger execution" Bug#17199 "Problem when view calls function from another database." Bug#18444 "Fully qualified stored function names don't work correctly in SELECT statements" Documentation note: this patch introduces a change in behaviour of prepared statements. This patch adds a few new invariants with regard to how THD::db should be used. These invariants should be preserved in future: - one should never refer to THD::db by pointer and always make a deep copy (strmake, strdup) - one should never compare two databases by pointer, but use strncmp or my_strncasecmp - TABLE_LIST object table->db should be always initialized in the parser or by creator of the object. For prepared statements it means that if the current database is changed after a statement is prepared, the database that was current at prepare remains active. This also means that you can not prepare a statement that implicitly refers to the current database if the latter is not set. This is not documented, and therefore needs documentation. This is NOT a change in behavior for almost all SQL statements except: - ALTER TABLE t1 RENAME t2 - OPTIMIZE TABLE t1 - ANALYZE TABLE t1 - TRUNCATE TABLE t1 -- until this patch t1 or t2 could be evaluated at the first execution of prepared statement. CURRENT_DATABASE() still works OK and is evaluated at every execution of prepared statement. Note, that in stored routines this is not an issue as the default database is the database of the stored procedure and "use" statement is prohibited in stored routines. This patch makes obsolete the use of check_db_used (it was never used in the old code too) and all other places that check for table->db and assign it from THD::db if it's NULL, except the parser. How this patch was created: THD::{db,db_length} were replaced with a LEX_STRING, THD::db. All the places that refer to THD::{db,db_length} were manually checked and: - if the place uses thd->db by pointer, it was fixed to make a deep copy - if a place compared two db pointers, it was fixed to compare them by value (via strcmp/my_strcasecmp, whatever was approproate) Then this intermediate patch was used to write a smaller patch that does the same thing but without a rename. TODO in 5.1: - remove check_db_used - deploy THD::set_db in mysql_change_db See also comments to individual files.
2006-06-26 22:47:52 +02:00
uint priv= (get_column_grant(thd, &view->grant, view->db,
view->table_name, item->name) &
2004-07-16 00:15:55 +02:00
VIEW_ANY_ACL);
if (fld && !fld->field->table->s->tmp_table)
2004-07-16 00:15:55 +02:00
{
final_priv&= fld->have_privileges;
if (~fld->have_privileges & priv)
report_item= item;
}
}
}
if (!final_priv)
2004-07-16 00:15:55 +02:00
{
DBUG_ASSERT(report_item);
my_error(ER_COLUMNACCESS_DENIED_ERROR, MYF(0),
"create view", thd->security_ctx->priv_user,
thd->security_ctx->priv_host, report_item->name,
view->table_name);
res= TRUE;
goto err;
2004-07-16 00:15:55 +02:00
}
}
#endif
if (wait_if_global_read_lock(thd, 0, 0))
2004-07-16 00:15:55 +02:00
{
res= TRUE;
2004-07-16 00:15:55 +02:00
goto err;
}
VOID(pthread_mutex_lock(&LOCK_open));
res= mysql_register_view(thd, view, mode);
if (mysql_bin_log.is_open())
{
String buff;
const LEX_STRING command[3]=
{{ C_STRING_WITH_LEN("CREATE ") },
{ C_STRING_WITH_LEN("ALTER ") },
{ C_STRING_WITH_LEN("CREATE OR REPLACE ") }};
buff.append(command[thd->lex->create_view_mode].str,
command[thd->lex->create_view_mode].length);
view_store_options(thd, views, &buff);
buff.append(STRING_WITH_LEN("VIEW "));
/* Test if user supplied a db (ie: we did not use thd->db) */
if (views->db && views->db[0] &&
(thd->db == NULL || strcmp(views->db, thd->db)))
{
append_identifier(thd, &buff, views->db,
views->db_length);
buff.append('.');
}
append_identifier(thd, &buff, views->table_name,
views->table_name_length);
if (lex->view_list.elements)
{
List_iterator_fast<LEX_STRING> names(lex->view_list);
LEX_STRING *name;
int i;
2006-11-07 14:25:57 +01:00
for (i= 0; (name= names++); i++)
{
buff.append(i ? ", " : "(");
append_identifier(thd, &buff, name->str, name->length);
}
buff.append(')');
}
buff.append(STRING_WITH_LEN(" AS "));
buff.append(views->source.str, views->source.length);
thd->binlog_query(THD::STMT_QUERY_TYPE,
buff.ptr(), buff.length(), FALSE, FALSE);
}
2004-07-16 00:15:55 +02:00
VOID(pthread_mutex_unlock(&LOCK_open));
if (view->revision != 1)
query_cache_invalidate3(thd, view, 0);
2004-07-16 00:15:55 +02:00
start_waiting_global_read_lock(thd);
if (res)
goto err;
2004-07-16 00:15:55 +02:00
send_ok(thd);
lex->link_first_table_back(view, link_to_local);
DBUG_RETURN(0);
2004-07-16 00:15:55 +02:00
err:
thd->proc_info= "end";
lex->link_first_table_back(view, link_to_local);
unit->cleanup();
DBUG_RETURN(res || thd->net.report_error);
2004-07-16 00:15:55 +02:00
}
/* index of revision number in following table */
static const int revision_number_position= 8;
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
/* number of required parameters for making view */
static const int required_view_parameters= 16;
2005-09-16 17:13:21 +02:00
/* number of backups */
static const int num_view_backups= 3;
2004-07-16 00:15:55 +02:00
/*
table of VIEW .frm field descriptors
Note that one should NOT change the order for this, as it's used by
parse()
*/
2004-07-16 00:15:55 +02:00
static File_option view_parameters[]=
{{{ C_STRING_WITH_LEN("query")},
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
my_offsetof(TABLE_LIST, select_stmt),
FILE_OPTIONS_ESTRING},
{{ C_STRING_WITH_LEN("md5")},
my_offsetof(TABLE_LIST, md5),
FILE_OPTIONS_STRING},
{{ C_STRING_WITH_LEN("updatable")},
my_offsetof(TABLE_LIST, updatable_view),
FILE_OPTIONS_ULONGLONG},
{{ C_STRING_WITH_LEN("algorithm")},
my_offsetof(TABLE_LIST, algorithm),
FILE_OPTIONS_ULONGLONG},
{{ C_STRING_WITH_LEN("definer_user")},
my_offsetof(TABLE_LIST, definer.user),
2004-07-16 00:15:55 +02:00
FILE_OPTIONS_STRING},
{{ C_STRING_WITH_LEN("definer_host")},
my_offsetof(TABLE_LIST, definer.host),
2004-07-16 00:15:55 +02:00
FILE_OPTIONS_STRING},
{{ C_STRING_WITH_LEN("suid")},
my_offsetof(TABLE_LIST, view_suid),
2004-07-16 00:15:55 +02:00
FILE_OPTIONS_ULONGLONG},
{{ C_STRING_WITH_LEN("with_check_option")},
my_offsetof(TABLE_LIST, with_check),
2004-07-16 00:15:55 +02:00
FILE_OPTIONS_ULONGLONG},
{{ C_STRING_WITH_LEN("revision")},
my_offsetof(TABLE_LIST, revision),
2004-07-16 00:15:55 +02:00
FILE_OPTIONS_REV},
{{ C_STRING_WITH_LEN("timestamp")},
my_offsetof(TABLE_LIST, timestamp),
2004-07-16 00:15:55 +02:00
FILE_OPTIONS_TIMESTAMP},
{{ C_STRING_WITH_LEN("create-version")},
my_offsetof(TABLE_LIST, file_version),
2004-07-16 00:15:55 +02:00
FILE_OPTIONS_ULONGLONG},
{{ C_STRING_WITH_LEN("source")},
my_offsetof(TABLE_LIST, source),
2004-07-16 00:15:55 +02:00
FILE_OPTIONS_ESTRING},
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
{{(char*) STRING_WITH_LEN("client_cs_name")},
my_offsetof(TABLE_LIST, view_client_cs_name),
FILE_OPTIONS_STRING},
{{(char*) STRING_WITH_LEN("connection_cl_name")},
my_offsetof(TABLE_LIST, view_connection_cl_name),
FILE_OPTIONS_STRING},
{{(char*) STRING_WITH_LEN("view_body_utf8")},
my_offsetof(TABLE_LIST, view_body_utf8),
FILE_OPTIONS_STRING},
{{NullS, 0}, 0,
2004-07-16 00:15:55 +02:00
FILE_OPTIONS_STRING}
};
static LEX_STRING view_file_type[]= {{(char*) STRING_WITH_LEN("VIEW") }};
2004-07-16 00:15:55 +02:00
/*
Register VIEW (write .frm & process .frm's history backups)
SYNOPSIS
mysql_register_view()
thd - thread handler
view - view description
mode - VIEW_CREATE_NEW, VIEW_ALTER, VIEW_CREATE_OR_REPLACE
RETURN
0 OK
-1 Error
1 Error and error message given
*/
2004-07-16 00:15:55 +02:00
static int mysql_register_view(THD *thd, TABLE_LIST *view,
enum_view_create_mode mode)
{
LEX *lex= thd->lex;
2004-07-16 00:15:55 +02:00
char buff[4096];
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
String view_query(buff, sizeof (buff), thd->charset());
char md5[MD5_BUFF_LENGTH];
2004-07-16 00:15:55 +02:00
bool can_be_merged;
char dir_buff[FN_REFLEN], path_buff[FN_REFLEN];
LEX_STRING dir, file, path;
int error= 0;
2004-07-16 00:15:55 +02:00
DBUG_ENTER("mysql_register_view");
/* print query */
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
view_query.length(0);
{
ulong sql_mode= thd->variables.sql_mode & MODE_ANSI_QUOTES;
thd->variables.sql_mode&= ~MODE_ANSI_QUOTES;
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
lex->unit.print(&view_query);
thd->variables.sql_mode|= sql_mode;
}
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
DBUG_PRINT("info", ("View: %s", view_query.ptr()));
2004-07-16 00:15:55 +02:00
/* fill structure */
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
view->select_stmt.str= view_query.c_ptr_safe();
view->select_stmt.length= view_query.length();
Bug#25411 (trigger code truncated), PART II Bug 28127 (Some valid identifiers names are not parsed correctly) Bug 26302 (MySQL server cuts off trailing "*/" from comments in SP/func) This patch is the second part of a major cleanup, required to fix Bug 25411 (trigger code truncated). The root cause of the issue stems from the function skip_rear_comments, which was a work around to remove "extra" "*/" characters from the query text, when parsing a query and reusing the text fragments to represent a view, trigger, function or stored procedure. The reason for this work around is that "special comments", like /*!50002 XXX */, were not parsed properly, so that a query like: AAA /*!50002 BBB */ CCC would be seen by the parser as "AAA BBB */ CCC" when the current version is greater or equal to 5.0.2 The root cause of this stems from how special comments are parsed. Special comments are really out-of-bound text that appear inside a query, that affects how the parser behave. In nature, /*!50002 XXX */ in MySQL is similar to the C concept of preprocessing : #if VERSION >= 50002 XXX #endif Depending on the current VERSION of the server, either the special comment should be expanded or it should be ignored, but in all cases the "text" of the query should be re-written to strip the "/*!50002" and "*/" markers, which does not belong to the SQL language itself. Prior to this fix, these markers would leak into : - the storage format for VIEW, - the storage format for FUNCTION, - the storage format for FUNCTION parameters, in mysql.proc (param_list), - the storage format for PROCEDURE, - the storage format for PROCEDURE parameters, in mysql.proc (param_list), - the storage format for TRIGGER, - the binary log used for replication. In all cases, not only this cause format corruption, but also provide a vector for dormant security issues, by allowing to tunnel code that will be activated after an upgrade. The proper solution is to deal with special comments strictly during parsing, when accepting a query from the outside world. Once a query is parsed and an object is created with a persistant representation, this object should not arbitrarily mutate after an upgrade. In short, special comments are a useful but limited feature for MYSQLdump, when used at an *interface* level to facilitate import/export, but bloating the server *internal* storage format is *not* the proper way to deal with configuration management of the user logic. With this fix: - the Lex_input_stream class now acts as a comment pre-processor, and either expands or ignore special comments on the fly. - MYSQLlex and sql_yacc.yy have been cleaned up to strictly use the public interface of Lex_input_stream. In particular, how the input stream accepts or rejects a character is private to Lex_input_stream, and the internal buffer pointers of that class are strictly private, and should not be tempered with during parsing. This caused many changes mostly in sql_lex.cc. During the code cleanup in case MY_LEX_NUMBER_IDENT, Bug 28127 (Some valid identifiers names are not parsed correctly) was found and fixed. By parsing special comments properly, and removing the function 'skip_rear_comments' [sic], Bug 26302 (MySQL server cuts off trailing "*/" from comments in SP/func) has been fixed as well.
2007-06-12 23:23:58 +02:00
view->source.str= (char*) thd->lex->create_view_select_start;
view->source.length= (thd->lex->create_view_select_end
- thd->lex->create_view_select_start);
trim_whitespace(thd->charset(), & view->source);
view->file_version= 1;
view->calc_md5(md5);
view->md5.str= md5;
view->md5.length= 32;
can_be_merged= lex->can_be_merged();
if (lex->create_view_algorithm == VIEW_ALGORITHM_MERGE &&
!lex->can_be_merged())
{
push_warning(thd, MYSQL_ERROR::WARN_LEVEL_WARN, ER_WARN_VIEW_MERGE,
ER(ER_WARN_VIEW_MERGE));
lex->create_view_algorithm= VIEW_ALGORITHM_UNDEFINED;
}
view->algorithm= lex->create_view_algorithm;
view->definer.user= lex->definer->user;
view->definer.host= lex->definer->host;
view->view_suid= lex->create_view_suid;
view->with_check= lex->create_view_check;
if ((view->updatable_view= (can_be_merged &&
view->algorithm != VIEW_ALGORITHM_TMPTABLE)))
{
/* TODO: change here when we will support UNIONs */
for (TABLE_LIST *tbl= (TABLE_LIST *)lex->select_lex.table_list.first;
tbl;
tbl= tbl->next_local)
{
if ((tbl->view && !tbl->updatable_view) || tbl->schema_table)
{
view->updatable_view= 0;
break;
}
for (TABLE_LIST *up= tbl; up; up= up->embedding)
{
if (up->outer_join)
{
view->updatable_view= 0;
goto loop_out;
}
}
}
}
loop_out:
/* print file name */
dir.length= build_table_filename(dir_buff, sizeof(dir_buff),
Bug#18775 - Temporary table from alter table visible to other threads Continued implementation of WL#1324 (table name to filename encoding) The intermediate (not temporary) files of the new table during ALTER TABLE was visible for SHOW TABLES. These intermediate files are copies of the original table with the changes done by ALTER TABLE. After all the data is copied over from the original table, these files are renamed to the original tables file names. So they are not temporary files. They persist after ALTER TABLE, but just with another name. In 5.0 the intermediate files are invisible for SHOW TABLES because all file names beginning with "#sql" were suppressed. This failed since 5.1.6 because even temporary table names were converted when making file names from them. The prefix became converted to "@0023sql". Converting the prefix during SHOW TABLES would suppress the listing of user tables that start with "#sql". The solution of the problem is to continue the implementation of the table name to file name conversion feature. One requirement is to suppress the conversion for temporary table names. This change is straightforward for real temporary tables as there is a function that creates temporary file names. But the generated path names are located in TMPDIR and have no relation to the internal table name. This cannot be used for ALTER TABLE. Its intermediate files need to be in the same directory as the old table files. And it is necessary to be able to deduce the same path from the same table name repeatedly. Consequently the intermediate table files must be handled like normal tables. Their internal names shall start with tmp_file_prefix (#sql) and they shall not be converted like normal table names. I added a flags parameter to all relevant functions that are called from ALTER TABLE. It is used to suppress the conversion for the intermediate table files. The outcome is that the suppression of #sql in SHOW TABLES works again. It does not suppress user tables as these are converted to @0023sql on file level. This patch does also fix ALTER TABLE ... RENAME, which could not rename a table with non-ASCII characters in its name. It does also fix the problem that a user could create a table like `#sql-xxxx-yyyy`, where xxxx is mysqld's pid and yyyy is the thread ID of some other thread, which prevented this thread from running ALTER TABLE. Some of the above problems are mentioned in Bug 1405, which can be closed with this patch. This patch does also contain some minor fixes for other forgotten conversions. Still known problems are reported as bugs 21370, 21373, and 21387.
2006-08-02 17:57:06 +02:00
view->db, "", "", 0);
2004-07-16 00:15:55 +02:00
dir.str= dir_buff;
path.length= build_table_filename(path_buff, sizeof(path_buff),
Bug#18775 - Temporary table from alter table visible to other threads Continued implementation of WL#1324 (table name to filename encoding) The intermediate (not temporary) files of the new table during ALTER TABLE was visible for SHOW TABLES. These intermediate files are copies of the original table with the changes done by ALTER TABLE. After all the data is copied over from the original table, these files are renamed to the original tables file names. So they are not temporary files. They persist after ALTER TABLE, but just with another name. In 5.0 the intermediate files are invisible for SHOW TABLES because all file names beginning with "#sql" were suppressed. This failed since 5.1.6 because even temporary table names were converted when making file names from them. The prefix became converted to "@0023sql". Converting the prefix during SHOW TABLES would suppress the listing of user tables that start with "#sql". The solution of the problem is to continue the implementation of the table name to file name conversion feature. One requirement is to suppress the conversion for temporary table names. This change is straightforward for real temporary tables as there is a function that creates temporary file names. But the generated path names are located in TMPDIR and have no relation to the internal table name. This cannot be used for ALTER TABLE. Its intermediate files need to be in the same directory as the old table files. And it is necessary to be able to deduce the same path from the same table name repeatedly. Consequently the intermediate table files must be handled like normal tables. Their internal names shall start with tmp_file_prefix (#sql) and they shall not be converted like normal table names. I added a flags parameter to all relevant functions that are called from ALTER TABLE. It is used to suppress the conversion for the intermediate table files. The outcome is that the suppression of #sql in SHOW TABLES works again. It does not suppress user tables as these are converted to @0023sql on file level. This patch does also fix ALTER TABLE ... RENAME, which could not rename a table with non-ASCII characters in its name. It does also fix the problem that a user could create a table like `#sql-xxxx-yyyy`, where xxxx is mysqld's pid and yyyy is the thread ID of some other thread, which prevented this thread from running ALTER TABLE. Some of the above problems are mentioned in Bug 1405, which can be closed with this patch. This patch does also contain some minor fixes for other forgotten conversions. Still known problems are reported as bugs 21370, 21373, and 21387.
2006-08-02 17:57:06 +02:00
view->db, view->table_name, reg_ext, 0);
path.str= path_buff;
file.str= path.str + dir.length;
file.length= path.length - dir.length;
2004-07-16 00:15:55 +02:00
/* init timestamp */
if (!view->timestamp.str)
2004-07-16 00:15:55 +02:00
view->timestamp.str= view->timestamp_buffer;
/* check old .frm */
2004-07-16 00:15:55 +02:00
{
char path_buff[FN_REFLEN];
LEX_STRING path;
File_parser *parser;
2004-07-16 00:15:55 +02:00
path.str= path_buff;
Fixed compiler warnings Fixed compile-pentium64 scripts Fixed wrong estimate of update_with_key_prefix in sql-bench Merge bk-internal.mysql.com:/home/bk/mysql-5.1 into mysql.com:/home/my/mysql-5.1 Fixed unsafe define of uint4korr() Fixed that --extern works with mysql-test-run.pl Small trivial cleanups This also fixes a bug in counting number of rows that are updated when we have many simultanous queries Move all connection handling and command exectuion main loop from sql_parse.cc to sql_connection.cc Split handle_one_connection() into reusable sub functions. Split create_new_thread() into reusable sub functions. Added thread_scheduler; Preliminary interface code for future thread_handling code. Use 'my_thread_id' for internal thread id's Make thr_alarm_kill() to depend on thread_id instead of thread Make thr_abort_locks_for_thread() depend on thread_id instead of thread In store_globals(), set my_thread_var->id to be thd->thread_id. Use my_thread_var->id as basis for my_thread_name() The above changes makes the connection we have between THD and threads more soft. Added a lot of DBUG_PRINT() and DBUG_ASSERT() functions Fixed compiler warnings Fixed core dumps when running with --debug Removed setting of signal masks (was never used) Made event code call pthread_exit() (portability fix) Fixed that event code doesn't call DBUG_xxx functions before my_thread_init() is called. Made handling of thread_id and thd->variables.pseudo_thread_id uniform. Removed one common 'not freed memory' warning from mysqltest Fixed a couple of usage of not initialized warnings (unlikely cases) Suppress compiler warnings from bdb and (for the moment) warnings from ndb
2007-02-23 12:13:55 +01:00
fn_format(path_buff, file.str, dir.str, "", MY_UNPACK_FILENAME);
2004-07-16 00:15:55 +02:00
path.length= strlen(path_buff);
if (!access(path.str, F_OK))
{
if (mode == VIEW_CREATE_NEW)
{
my_error(ER_TABLE_EXISTS_ERROR, MYF(0), view->alias);
error= -1;
goto err;
2004-07-16 00:15:55 +02:00
}
if (!(parser= sql_parse_prepare(&path, thd->mem_root, 0)))
{
error= 1;
goto err;
}
if (!parser->ok() || !is_equal(&view_type, parser->type()))
2004-07-16 00:15:55 +02:00
{
A fix and a test case for Bug#19022 "Memory bug when switching db during trigger execution" Bug#17199 "Problem when view calls function from another database." Bug#18444 "Fully qualified stored function names don't work correctly in SELECT statements" Documentation note: this patch introduces a change in behaviour of prepared statements. This patch adds a few new invariants with regard to how THD::db should be used. These invariants should be preserved in future: - one should never refer to THD::db by pointer and always make a deep copy (strmake, strdup) - one should never compare two databases by pointer, but use strncmp or my_strncasecmp - TABLE_LIST object table->db should be always initialized in the parser or by creator of the object. For prepared statements it means that if the current database is changed after a statement is prepared, the database that was current at prepare remains active. This also means that you can not prepare a statement that implicitly refers to the current database if the latter is not set. This is not documented, and therefore needs documentation. This is NOT a change in behavior for almost all SQL statements except: - ALTER TABLE t1 RENAME t2 - OPTIMIZE TABLE t1 - ANALYZE TABLE t1 - TRUNCATE TABLE t1 -- until this patch t1 or t2 could be evaluated at the first execution of prepared statement. CURRENT_DATABASE() still works OK and is evaluated at every execution of prepared statement. Note, that in stored routines this is not an issue as the default database is the database of the stored procedure and "use" statement is prohibited in stored routines. This patch makes obsolete the use of check_db_used (it was never used in the old code too) and all other places that check for table->db and assign it from THD::db if it's NULL, except the parser. How this patch was created: THD::{db,db_length} were replaced with a LEX_STRING, THD::db. All the places that refer to THD::{db,db_length} were manually checked and: - if the place uses thd->db by pointer, it was fixed to make a deep copy - if a place compared two db pointers, it was fixed to compare them by value (via strcmp/my_strcasecmp, whatever was approproate) Then this intermediate patch was used to write a smaller patch that does the same thing but without a rename. TODO in 5.1: - remove check_db_used - deploy THD::set_db in mysql_change_db See also comments to individual files.
2006-06-26 22:47:52 +02:00
my_error(ER_WRONG_OBJECT, MYF(0), view->db, view->table_name, "VIEW");
error= -1;
goto err;
2004-07-16 00:15:55 +02:00
}
/*
read revision number
TODO: read dependence list, too, to process cascade/restrict
TODO: special cascade/restrict procedure for alter?
*/
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
if (parser->parse((uchar*)view, thd->mem_root,
view_parameters + revision_number_position, 1,
&file_parser_dummy_hook))
2004-07-16 00:15:55 +02:00
{
error= thd->net.report_error? -1 : 0;
goto err;
2004-07-16 00:15:55 +02:00
}
}
else
2007-06-02 02:58:46 +02:00
{
2004-07-16 00:15:55 +02:00
if (mode == VIEW_ALTER)
{
my_error(ER_NO_SUCH_TABLE, MYF(0), view->db, view->alias);
error= -1;
goto err;
2004-07-16 00:15:55 +02:00
}
}
}
2007-06-02 02:58:46 +02:00
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
/* Initialize view creation context from the environment. */
view->view_creation_ctx= View_creation_ctx::create(thd);
/*
Set LEX_STRING attributes in view-structure for parser to create
frm-file.
*/
lex_string_set(&view->view_client_cs_name,
view->view_creation_ctx->get_client_cs()->csname);
lex_string_set(&view->view_connection_cl_name,
view->view_creation_ctx->get_connection_cl()->name);
view->view_body_utf8= lex->view_body_utf8;
/*
Check that table of main select do not used in subqueries.
This test can catch only very simple cases of such non-updateable views,
all other will be detected before updating commands execution.
(it is more optimisation then real check)
NOTE: this skip cases of using table via VIEWs, joined VIEWs, VIEWs with
UNION
*/
if (view->updatable_view &&
!lex->select_lex.master_unit()->is_union() &&
!((TABLE_LIST*)lex->select_lex.table_list.first)->next_local &&
find_table_in_global_list(lex->query_tables->next_global,
lex->query_tables->db,
lex->query_tables->table_name))
{
view->updatable_view= 0;
}
2004-09-03 14:18:40 +02:00
if (view->with_check != VIEW_CHECK_NONE &&
!view->updatable_view)
{
my_error(ER_VIEW_NONUPD_CHECK, MYF(0), view->db, view->table_name);
error= -1;
goto err;
2004-09-03 14:18:40 +02:00
}
2004-07-16 00:15:55 +02:00
if (sql_create_definition_file(&dir, &file, view_file_type,
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
(uchar*)view, view_parameters, num_view_backups))
2004-07-16 00:15:55 +02:00
{
error= thd->net.report_error? -1 : 1;
goto err;
2004-07-16 00:15:55 +02:00
}
DBUG_RETURN(0);
err:
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
view->select_stmt.str= NULL;
view->select_stmt.length= 0;
view->md5.str= NULL;
view->md5.length= 0;
DBUG_RETURN(error);
2004-07-16 00:15:55 +02:00
}
2004-07-16 00:15:55 +02:00
/*
read VIEW .frm and create structures
SYNOPSIS
mysql_make_view()
thd Thread handler
parser parser object
table TABLE_LIST structure for filling
flags flags
RETURN
0 ok
1 error
2004-07-16 00:15:55 +02:00
*/
bool mysql_make_view(THD *thd, File_parser *parser, TABLE_LIST *table,
uint flags)
2004-07-16 00:15:55 +02:00
{
SELECT_LEX *end, *view_select;
LEX *old_lex, *lex;
Query_arena *arena, backup;
TABLE_LIST *top_view= table->top_table();
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
bool parse_status;
bool result, view_is_mergeable;
TABLE_LIST *view_main_select_tables;
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
2004-07-16 00:15:55 +02:00
DBUG_ENTER("mysql_make_view");
DBUG_PRINT("info", ("table: 0x%lx (%s)", (ulong) table, table->table_name));
2004-07-16 00:15:55 +02:00
if (table->view)
{
/*
It's an execution of a PS/SP and the view has already been unfolded
into a list of used tables. Now we only need to update the information
about granted privileges in the view tables with the actual data
stored in MySQL privilege system. We don't need to restore the
required privileges (by calling register_want_access) because they has
not changed since PREPARE or the previous execution: the only case
when this information is changed is execution of UPDATE on a view, but
the original want_access is restored in its end.
*/
if (!table->prelocking_placeholder && table->prepare_security(thd))
{
DBUG_RETURN(1);
}
2004-07-16 00:15:55 +02:00
DBUG_PRINT("info",
("VIEW %s.%s is already processed on previous PS/SP execution",
2004-07-16 00:15:55 +02:00
table->view_db.str, table->view_name.str));
DBUG_RETURN(0);
}
/* check loop via view definition */
for (TABLE_LIST *precedent= table->referencing_view;
precedent;
precedent= precedent->referencing_view)
{
if (precedent->view_name.length == table->table_name_length &&
precedent->view_db.length == table->db_length &&
my_strcasecmp(system_charset_info,
precedent->view_name.str, table->table_name) == 0 &&
my_strcasecmp(system_charset_info,
precedent->view_db.str, table->db) == 0)
{
my_error(ER_VIEW_RECURSIVE, MYF(0),
top_view->view_db.str, top_view->view_name.str);
DBUG_RETURN(TRUE);
}
}
2004-07-16 00:15:55 +02:00
/*
For now we assume that tables will not be changed during PS life (it
will be TRUE as far as we make new table cache).
*/
old_lex= thd->lex;
arena= thd->stmt_arena;
2004-09-10 01:22:44 +02:00
if (arena->is_conventional())
arena= 0;
else
thd->set_n_backup_active_arena(arena, &backup);
2004-07-16 00:15:55 +02:00
/* init timestamp */
if (!table->timestamp.str)
2004-07-16 00:15:55 +02:00
table->timestamp.str= table->timestamp_buffer;
/* prepare default values for old format */
table->view_suid= TRUE;
table->definer.user.str= table->definer.host.str= 0;
table->definer.user.length= table->definer.host.length= 0;
2004-07-16 00:15:55 +02:00
/*
TODO: when VIEWs will be stored in cache, table mem_root should
be used here
*/
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
if (parser->parse((uchar*)table, thd->mem_root, view_parameters,
required_view_parameters, &file_parser_dummy_hook))
2004-07-16 00:15:55 +02:00
goto err;
/*
check old format view .frm
*/
if (!table->definer.user.str)
{
DBUG_ASSERT(!table->definer.host.str &&
!table->definer.user.length &&
!table->definer.host.length);
push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
ER_VIEW_FRM_NO_USER, ER(ER_VIEW_FRM_NO_USER),
table->db, table->table_name);
get_default_definer(thd, &table->definer);
}
if (flags & OPEN_VIEW_NO_PARSE)
{
DBUG_RETURN(FALSE);
}
2004-07-16 00:15:55 +02:00
/*
Save VIEW parameters, which will be wiped out by derived table
processing
*/
table->view_db.str= table->db;
table->view_db.length= table->db_length;
table->view_name.str= table->table_name;
table->view_name.length= table->table_name_length;
2004-07-16 00:15:55 +02:00
2004-07-16 14:20:51 +02:00
/*TODO: md5 test here and warning if it is differ */
2004-07-16 00:15:55 +02:00
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
/*
Initialize view definition context by character set names loaded from
the view definition file. Use UTF8 character set if view definition
file is of old version and does not contain the character set names.
*/
table->view_creation_ctx= View_creation_ctx::create(thd, table);
2004-07-16 14:20:51 +02:00
/*
TODO: TABLE mem root should be used here when VIEW will be stored in
TABLE cache
now Lex placed in statement memory
*/
table->view= lex= thd->lex= (LEX*) new(thd->mem_root) st_lex_local;
2004-07-16 00:15:55 +02:00
{
char old_db_buf[NAME_LEN+1];
LEX_STRING old_db= { old_db_buf, sizeof(old_db_buf) };
bool dbchanged;
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
Lex_input_stream lip(thd,
table->select_stmt.str,
table->select_stmt.length);
/*
Use view db name as thread default database, in order to ensure
that the view is parsed and prepared correctly.
*/
if ((result= sp_use_new_db(thd, table->view_db, &old_db, 1, &dbchanged)))
goto end;
lex_start(thd);
view_select= &lex->select_lex;
view_select->select_number= ++thd->select_number;
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
ulong saved_mode= thd->variables.sql_mode;
2004-07-16 00:15:55 +02:00
/* switch off modes which can prevent normal parsing of VIEW
- MODE_REAL_AS_FLOAT affect only CREATE TABLE parsing
+ MODE_PIPES_AS_CONCAT affect expression parsing
+ MODE_ANSI_QUOTES affect expression parsing
+ MODE_IGNORE_SPACE affect expression parsing
- MODE_NOT_USED not used :)
* MODE_ONLY_FULL_GROUP_BY affect execution
* MODE_NO_UNSIGNED_SUBTRACTION affect execution
- MODE_NO_DIR_IN_CREATE affect table creation only
- MODE_POSTGRESQL compounded from other modes
- MODE_ORACLE compounded from other modes
- MODE_MSSQL compounded from other modes
- MODE_DB2 compounded from other modes
- MODE_MAXDB affect only CREATE TABLE parsing
- MODE_NO_KEY_OPTIONS affect only SHOW
- MODE_NO_TABLE_OPTIONS affect only SHOW
- MODE_NO_FIELD_OPTIONS affect only SHOW
- MODE_MYSQL323 affect only SHOW
- MODE_MYSQL40 affect only SHOW
- MODE_ANSI compounded from other modes
(+ transaction mode)
? MODE_NO_AUTO_VALUE_ON_ZERO affect UPDATEs
+ MODE_NO_BACKSLASH_ESCAPES affect expression parsing
*/
thd->variables.sql_mode&= ~(MODE_PIPES_AS_CONCAT | MODE_ANSI_QUOTES |
MODE_IGNORE_SPACE | MODE_NO_BACKSLASH_ESCAPES);
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
/* Parse the query. */
parse_status= parse_sql(thd, &lip, table->view_creation_ctx);
/* Restore environment. */
if ((old_lex->sql_command == SQLCOM_SHOW_FIELDS) ||
(old_lex->sql_command == SQLCOM_SHOW_CREATE))
lex->sql_command= old_lex->sql_command;
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
thd->variables.sql_mode= saved_mode;
if (dbchanged && mysql_change_db(thd, &old_db, TRUE))
goto err;
2004-07-16 00:15:55 +02:00
}
Patch for the following bugs: - BUG#11986: Stored routines and triggers can fail if the code has a non-ascii symbol - BUG#16291: mysqldump corrupts string-constants with non-ascii-chars - BUG#19443: INFORMATION_SCHEMA does not support charsets properly - BUG#21249: Character set of SP-var can be ignored - BUG#25212: Character set of string constant is ignored (stored routines) - BUG#25221: Character set of string constant is ignored (triggers) There were a few general problems that caused these bugs: 1. Character set information of the original (definition) query for views, triggers, stored routines and events was lost. 2. mysqldump output query in client character set, which can be inappropriate to encode definition-query. 3. INFORMATION_SCHEMA used strings with mixed encodings to display object definition; 1. No query-definition-character set. In order to compile query into execution code, some extra data (such as environment variables or the database character set) is used. The problem here was that this context was not preserved. So, on the next load it can differ from the original one, thus the result will be different. The context contains the following data: - client character set; - connection collation (character set and collation); - collation of the owner database; The fix is to store this context and use it each time we parse (compile) and execute the object (stored routine, trigger, ...). 2. Wrong mysqldump-output. The original query can contain several encodings (by means of character set introducers). The problem here was that we tried to convert original query to the mysqldump-client character set. Moreover, we stored queries in different character sets for different objects (views, for one, used UTF8, triggers used original character set). The solution is - to store definition queries in the original character set; - to change SHOW CREATE statement to output definition query in the binary character set (i.e. without any conversion); - introduce SHOW CREATE TRIGGER statement; - to dump special statements to switch the context to the original one before dumping and restore it afterwards. Note, in order to preserve the database collation at the creation time, additional ALTER DATABASE might be used (to temporary switch the database collation back to the original value). In this case, ALTER DATABASE privilege will be required. This is a backward-incompatible change. 3. INFORMATION_SCHEMA showed non-UTF8 strings The fix is to generate UTF8-query during the parsing, store it in the object and show it in the INFORMATION_SCHEMA. Basically, the idea is to create a copy of the original query convert it to UTF8. Character set introducers are removed and all text literals are converted to UTF8. This UTF8 query is intended to provide user-readable output. It must not be used to recreate the object. Specialized SHOW CREATE statements should be used for this. The reason for this limitation is the following: the original query can contain symbols from several character sets (by means of character set introducers). Example: - original query: CREATE VIEW v1 AS SELECT _cp1251 'Hello' AS c1; - UTF8 query (for INFORMATION_SCHEMA): CREATE VIEW v1 AS SELECT 'Hello' AS c1;
2007-06-28 19:34:54 +02:00
if (!parse_status)
2004-07-16 00:15:55 +02:00
{
TABLE_LIST *view_tables= lex->query_tables;
TABLE_LIST *view_tables_tail= 0;
TABLE_LIST *tbl;
2004-07-16 14:20:51 +02:00
2004-07-16 00:15:55 +02:00
/*
Check rights to run commands (EXPLAIN SELECT & SHOW CREATE) which show
underlying tables.
Skip this step if we are opening view for prelocking only.
2004-07-16 00:15:55 +02:00
*/
if (!table->prelocking_placeholder &&
(old_lex->sql_command == SQLCOM_SELECT && old_lex->describe))
2004-07-16 00:15:55 +02:00
{
if (check_table_access(thd, SELECT_ACL, view_tables, 1) &&
check_table_access(thd, SHOW_VIEW_ACL, table, 1))
2004-07-16 00:15:55 +02:00
{
2004-11-12 13:34:00 +01:00
my_message(ER_VIEW_NO_EXPLAIN, ER(ER_VIEW_NO_EXPLAIN), MYF(0));
2004-07-16 00:15:55 +02:00
goto err;
}
}
else if (!table->prelocking_placeholder &&
(old_lex->sql_command == SQLCOM_SHOW_CREATE) &&
!table->belong_to_view)
{
if (check_table_access(thd, SHOW_VIEW_ACL, table, 0))
goto err;
}
2004-07-16 00:15:55 +02:00
if (!(table->view_tables=
(List<TABLE_LIST>*) new(thd->mem_root) List<TABLE_LIST>))
goto err;
/*
mark to avoid temporary table using and put view reference and find
last view table
*/
for (tbl= view_tables;
tbl;
tbl= (view_tables_tail= tbl)->next_global)
{
tbl->skip_temporary= 1;
tbl->belong_to_view= top_view;
tbl->referencing_view= table;
tbl->prelocking_placeholder= table->prelocking_placeholder;
/*
First we fill want_privilege with SELECT_ACL (this is needed for the
tables which belongs to view subqueries and temporary table views,
then for the merged view underlying tables we will set wanted
privileges of top_view
*/
tbl->grant.want_privilege= SELECT_ACL;
/*
After unfolding the view we lose the list of tables referenced in it
(we will have only a list of underlying tables in case of MERGE
algorithm, which does not include the tables referenced from
subqueries used in view definition).
Let's build a list of all tables referenced in the view.
*/
table->view_tables->push_back(tbl);
}
/*
Put tables of VIEW after VIEW TABLE_LIST
NOTE: It is important for UPDATE/INSERT/DELETE checks to have this
tables just after VIEW instead of tail of list, to be able check that
table is unique. Also we store old next table for the same purpose.
*/
if (view_tables)
{
if (table->next_global)
{
view_tables_tail->next_global= table->next_global;
table->next_global->prev_global= &view_tables_tail->next_global;
}
else
{
old_lex->query_tables_last= &view_tables_tail->next_global;
}
view_tables->prev_global= &table->next_global;
table->next_global= view_tables;
}
* Mixed replication mode * : 1) Fix for BUG#19630 "stored function inserting into two auto_increment breaks statement-based binlog": a stored function inserting into two such tables may fail to replicate (inserting wrong data in the slave's copy of the second table) if the slave's second table had an internal auto_increment counter different from master's. Because the auto_increment value autogenerated by master for the 2nd table does not go into binlog, only the first does, so the slave lacks information. To fix this, if running in mixed binlogging mode, if the stored function or trigger plans to update two different tables both having auto_increment columns, we switch to row-based for the whole function. We don't have a simple solution for statement-based binlogging mode, there the bug remains and will be documented as a known problem. Re-enabling rpl_switch_stm_row_mixed. 2) Fix for BUG#20630 "Mixed binlogging mode does not work with stored functions, triggers, views", which was a documented limitation (in mixed mode, we didn't detect that a stored function's execution needed row-based binlogging (due to some UUID() call for example); same for triggers, same for views (a view created from a SELECT UUID(), and doing INSERT INTO sometable SELECT theview; would not replicate row-based). This is implemented by, after parsing a routine's body, remembering in sp_head that this routine needs row-based binlogging. Then when this routine is used, the caller is marked to require row-based binlogging too. Same for views: when we parse a view and detect that its SELECT needs row-based binary logging, we mark the calling LEX as such. 3) Fix for BUG#20499 "mixed mode with temporary table breaks binlog": a temporary table containing e.g. UUID has its changes not binlogged, so any query updating a permanent table with data from the temporary table will run wrongly on slave. Solution: in mixed mode we don't switch back from row-based to statement-based when there exists temporary tables. 4) Attempt to test mysqlbinlog on a binlog generated by mysqlbinlog; impossible due to BUG#11312 and BUG#20329, but test is in place for when they are fixed.
2006-07-09 17:00:47 +02:00
/*
If the view's body needs row-based binlogging (e.g. the VIEW is created
from SELECT UUID()), the top statement also needs it.
*/
if (lex->is_stmt_unsafe())
old_lex->set_stmt_unsafe();
view_is_mergeable= (table->algorithm != VIEW_ALGORITHM_TMPTABLE &&
lex->can_be_merged());
LINT_INIT(view_main_select_tables);
if (view_is_mergeable)
{
/*
Currently 'view_main_select_tables' differs from 'view_tables'
only then view has CONVERT_TZ() function in its select list.
This may change in future, for example if we enable merging of
views with subqueries in select list.
*/
view_main_select_tables=
(TABLE_LIST*)lex->select_lex.table_list.first;
/*
Let us set proper lock type for tables of the view's main
select since we may want to perform update or insert on
view. This won't work for view containing union. But this is
ok since we don't allow insert and update on such views
anyway.
*/
for (tbl= view_main_select_tables; tbl; tbl= tbl->next_local)
tbl->lock_type= table->lock_type;
/*
If the view is mergeable, we might want to
INSERT/UPDATE/DELETE into tables of this view. Preserve the
original sql command and 'duplicates' of the outer lex.
This is used later in set_trg_event_type_for_command.
*/
lex->sql_command= old_lex->sql_command;
lex->duplicates= old_lex->duplicates;
}
/*
This method has a dependency on the proper lock type being set,
so in case of views should be called here.
*/
lex->set_trg_event_type_for_tables();
* Mixed replication mode * : 1) Fix for BUG#19630 "stored function inserting into two auto_increment breaks statement-based binlog": a stored function inserting into two such tables may fail to replicate (inserting wrong data in the slave's copy of the second table) if the slave's second table had an internal auto_increment counter different from master's. Because the auto_increment value autogenerated by master for the 2nd table does not go into binlog, only the first does, so the slave lacks information. To fix this, if running in mixed binlogging mode, if the stored function or trigger plans to update two different tables both having auto_increment columns, we switch to row-based for the whole function. We don't have a simple solution for statement-based binlogging mode, there the bug remains and will be documented as a known problem. Re-enabling rpl_switch_stm_row_mixed. 2) Fix for BUG#20630 "Mixed binlogging mode does not work with stored functions, triggers, views", which was a documented limitation (in mixed mode, we didn't detect that a stored function's execution needed row-based binlogging (due to some UUID() call for example); same for triggers, same for views (a view created from a SELECT UUID(), and doing INSERT INTO sometable SELECT theview; would not replicate row-based). This is implemented by, after parsing a routine's body, remembering in sp_head that this routine needs row-based binlogging. Then when this routine is used, the caller is marked to require row-based binlogging too. Same for views: when we parse a view and detect that its SELECT needs row-based binary logging, we mark the calling LEX as such. 3) Fix for BUG#20499 "mixed mode with temporary table breaks binlog": a temporary table containing e.g. UUID has its changes not binlogged, so any query updating a permanent table with data from the temporary table will run wrongly on slave. Solution: in mixed mode we don't switch back from row-based to statement-based when there exists temporary tables. 4) Attempt to test mysqlbinlog on a binlog generated by mysqlbinlog; impossible due to BUG#11312 and BUG#20329, but test is in place for when they are fixed.
2006-07-09 17:00:47 +02:00
/*
If we are opening this view as part of implicit LOCK TABLES, then
this view serves as simple placeholder and we should not continue
further processing.
*/
if (table->prelocking_placeholder)
goto ok2;
old_lex->derived_tables|= (DERIVED_VIEW | lex->derived_tables);
/* move SQL_NO_CACHE & Co to whole query */
old_lex->safe_to_cache_query= (old_lex->safe_to_cache_query &&
lex->safe_to_cache_query);
/* move SQL_CACHE to whole query */
if (view_select->options & OPTION_TO_QUERY_CACHE)
old_lex->select_lex.options|= OPTION_TO_QUERY_CACHE;
if (table->view_suid)
{
/*
Prepare a security context to check underlying objects of the view
*/
if (!(table->view_sctx= (Security_context *)
thd->stmt_arena->alloc(sizeof(Security_context))))
goto err;
/* Assign the context to the tables referenced in the view */
if (view_tables)
{
DBUG_ASSERT(view_tables_tail);
for (tbl= view_tables; tbl != view_tables_tail->next_global;
tbl= tbl->next_global)
tbl->security_ctx= table->view_sctx;
}
/* assign security context to SELECT name resolution contexts of view */
for(SELECT_LEX *sl= lex->all_selects_list;
sl;
sl= sl->next_select_in_list())
sl->context.security_ctx= table->view_sctx;
}
/*
Setup an error processor to hide error messages issued by stored
routines referenced in the view
*/
for (SELECT_LEX *sl= lex->all_selects_list;
sl;
sl= sl->next_select_in_list())
{
sl->context.error_processor= &view_error_processor;
sl->context.error_processor_data= (void *)table;
}
2004-07-16 00:15:55 +02:00
/*
check MERGE algorithm ability
- algorithm is not explicit TEMPORARY TABLE
- VIEW SELECT allow merging
2004-07-16 00:15:55 +02:00
- VIEW used in subquery or command support MERGE algorithm
*/
if (view_is_mergeable &&
2004-07-16 00:15:55 +02:00
(table->select_lex->master_unit() != &old_lex->unit ||
old_lex->can_use_merged()) &&
!old_lex->can_not_use_merged())
2004-07-16 00:15:55 +02:00
{
/* lex should contain at least one table */
DBUG_ASSERT(view_main_select_tables != 0);
2004-07-16 00:15:55 +02:00
List_iterator_fast<TABLE_LIST> ti(view_select->top_join_list);
2004-07-16 00:15:55 +02:00
table->effective_algorithm= VIEW_ALGORITHM_MERGE;
DBUG_PRINT("info", ("algorithm: MERGE"));
table->updatable= (table->updatable_view != 0);
table->effective_with_check=
old_lex->get_effective_with_check(table);
table->merge_underlying_list= view_main_select_tables;
/* Fill correct wanted privileges. */
for (tbl= view_main_select_tables; tbl; tbl= tbl->next_local)
tbl->grant.want_privilege= top_view->grant.orig_want_privilege;
/* prepare view context */
lex->select_lex.context.resolve_in_table_list_only(view_main_select_tables);
lex->select_lex.context.outer_context= 0;
lex->select_lex.context.select_lex= table->select_lex;
lex->select_lex.select_n_having_items+=
table->select_lex->select_n_having_items;
2004-07-16 00:15:55 +02:00
/*
Tables of the main select of the view should be marked as belonging
to the same select as original view (again we can use LEX::select_lex
for this purprose because we don't support MERGE algorithm for views
with unions).
2004-07-16 00:15:55 +02:00
*/
for (tbl= lex->select_lex.get_table_list(); tbl; tbl= tbl->next_local)
tbl->select_lex= table->select_lex;
{
if (view_main_select_tables->next_local)
{
table->multitable_view= TRUE;
if (table->belong_to_view)
table->belong_to_view->multitable_view= TRUE;
}
/* make nested join structure for view tables */
NESTED_JOIN *nested_join;
if (!(nested_join= table->nested_join=
(NESTED_JOIN *) thd->calloc(sizeof(NESTED_JOIN))))
goto err;
nested_join->join_list= view_select->top_join_list;
/* re-nest tables of VIEW */
ti.rewind();
while ((tbl= ti++))
{
tbl->join_list= &nested_join->join_list;
tbl->embedding= table;
}
}
2004-07-16 00:15:55 +02:00
/* Store WHERE clause for post-processing in setup_underlying */
table->where= view_select->where;
/*
Add subqueries units to SELECT into which we merging current view.
unit(->next)* chain starts with subqueries that are used by this
view and continues with subqueries that are used by other views.
We must not add any subquery twice (otherwise we'll form a loop),
to do this we remember in end_unit the first subquery that has
been already added.
NOTE: we do not support UNION here, so we take only one select
*/
SELECT_LEX_NODE *end_unit= table->select_lex->slave;
SELECT_LEX_UNIT *next_unit;
for (SELECT_LEX_UNIT *unit= lex->select_lex.first_inner_unit();
unit;
unit= next_unit)
{
if (unit == end_unit)
break;
SELECT_LEX_NODE *save_slave= unit->slave;
next_unit= unit->next_unit();
unit->include_down(table->select_lex);
unit->slave= save_slave; // fix include_down initialisation
}
/*
We can safely ignore the VIEW's ORDER BY if we merge into union
branch, as order is not important there.
*/
if (!table->select_lex->master_unit()->is_union())
table->select_lex->order_list.push_back(&lex->select_lex.order_list);
2004-07-16 00:15:55 +02:00
/*
This SELECT_LEX will be linked in global SELECT_LEX list
to make it processed by mysql_handle_derived(),
but it will not be included to SELECT_LEX tree, because it
will not be executed
*/
2004-07-16 00:15:55 +02:00
goto ok;
}
table->effective_algorithm= VIEW_ALGORITHM_TMPTABLE;
DBUG_PRINT("info", ("algorithm: TEMPORARY TABLE"));
view_select->linkage= DERIVED_TABLE_TYPE;
table->updatable= 0;
table->effective_with_check= VIEW_CHECK_NONE;
old_lex->subqueries= TRUE;
2004-07-16 00:15:55 +02:00
/* SELECT tree link */
lex->unit.include_down(table->select_lex);
lex->unit.slave= view_select; // fix include_down initialisation
2004-07-16 00:15:55 +02:00
table->derived= &lex->unit;
}
else
goto err;
ok:
/* global SELECT list linking */
end= view_select; // primary SELECT_LEX is always last
2004-07-16 00:15:55 +02:00
end->link_next= old_lex->all_selects_list;
old_lex->all_selects_list->link_prev= &end->link_next;
old_lex->all_selects_list= lex->all_selects_list;
lex->all_selects_list->link_prev=
(st_select_lex_node**)&old_lex->all_selects_list;
ok2:
DBUG_ASSERT(lex == thd->lex);
thd->lex= old_lex; // Needed for prepare_security
result= !table->prelocking_placeholder && table->prepare_security(thd);
2004-07-16 00:15:55 +02:00
lex_end(lex);
end:
2004-07-16 00:15:55 +02:00
if (arena)
thd->restore_active_arena(arena, &backup);
thd->lex= old_lex;
DBUG_RETURN(result);
err:
DBUG_ASSERT(thd->lex == table->view);
lex_end(thd->lex);
delete table->view;
2004-07-16 00:15:55 +02:00
table->view= 0; // now it is not VIEW placeholder
result= 1;
goto end;
2004-07-16 00:15:55 +02:00
}
/*
drop view
SYNOPSIS
mysql_drop_view()
thd - thread handler
views - views to delete
drop_mode - cascade/check
RETURN VALUE
FALSE OK
TRUE Error
2004-07-16 00:15:55 +02:00
*/
bool mysql_drop_view(THD *thd, TABLE_LIST *views, enum_drop_mode drop_mode)
2004-07-16 00:15:55 +02:00
{
char path[FN_REFLEN];
TABLE_LIST *view;
String non_existant_views;
char *wrong_object_db= NULL, *wrong_object_name= NULL;
bool error= FALSE;
enum legacy_db_type not_used;
DBUG_ENTER("mysql_drop_view");
2004-07-16 00:15:55 +02:00
VOID(pthread_mutex_lock(&LOCK_open));
2004-07-16 00:15:55 +02:00
for (view= views; view; view= view->next_local)
{
TABLE_SHARE *share;
2006-08-01 22:38:38 +02:00
frm_type_enum type= FRMTYPE_ERROR;
build_table_filename(path, sizeof(path),
Bug#18775 - Temporary table from alter table visible to other threads Continued implementation of WL#1324 (table name to filename encoding) The intermediate (not temporary) files of the new table during ALTER TABLE was visible for SHOW TABLES. These intermediate files are copies of the original table with the changes done by ALTER TABLE. After all the data is copied over from the original table, these files are renamed to the original tables file names. So they are not temporary files. They persist after ALTER TABLE, but just with another name. In 5.0 the intermediate files are invisible for SHOW TABLES because all file names beginning with "#sql" were suppressed. This failed since 5.1.6 because even temporary table names were converted when making file names from them. The prefix became converted to "@0023sql". Converting the prefix during SHOW TABLES would suppress the listing of user tables that start with "#sql". The solution of the problem is to continue the implementation of the table name to file name conversion feature. One requirement is to suppress the conversion for temporary table names. This change is straightforward for real temporary tables as there is a function that creates temporary file names. But the generated path names are located in TMPDIR and have no relation to the internal table name. This cannot be used for ALTER TABLE. Its intermediate files need to be in the same directory as the old table files. And it is necessary to be able to deduce the same path from the same table name repeatedly. Consequently the intermediate table files must be handled like normal tables. Their internal names shall start with tmp_file_prefix (#sql) and they shall not be converted like normal table names. I added a flags parameter to all relevant functions that are called from ALTER TABLE. It is used to suppress the conversion for the intermediate table files. The outcome is that the suppression of #sql in SHOW TABLES works again. It does not suppress user tables as these are converted to @0023sql on file level. This patch does also fix ALTER TABLE ... RENAME, which could not rename a table with non-ASCII characters in its name. It does also fix the problem that a user could create a table like `#sql-xxxx-yyyy`, where xxxx is mysqld's pid and yyyy is the thread ID of some other thread, which prevented this thread from running ALTER TABLE. Some of the above problems are mentioned in Bug 1405, which can be closed with this patch. This patch does also contain some minor fixes for other forgotten conversions. Still known problems are reported as bugs 21370, 21373, and 21387.
2006-08-02 17:57:06 +02:00
view->db, view->table_name, reg_ext, 0);
2006-08-01 22:38:38 +02:00
if (access(path, F_OK) ||
FRMTYPE_VIEW != (type= mysql_frm_type(thd, path, &not_used)))
2004-07-16 00:15:55 +02:00
{
char name[FN_REFLEN];
my_snprintf(name, sizeof(name), "%s.%s", view->db, view->table_name);
2004-07-16 00:15:55 +02:00
if (thd->lex->drop_if_exists)
{
push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_NOTE,
ER_BAD_TABLE_ERROR, ER(ER_BAD_TABLE_ERROR),
name);
continue;
}
if (type == FRMTYPE_TABLE)
{
if (!wrong_object_name)
{
wrong_object_db= view->db;
wrong_object_name= view->table_name;
}
}
2004-07-16 00:15:55 +02:00
else
{
if (non_existant_views.length())
non_existant_views.append(',');
non_existant_views.append(String(view->table_name,system_charset_info));
}
continue;
2004-07-16 00:15:55 +02:00
}
if (my_delete(path, MYF(MY_WME)))
error= TRUE;
/*
For a view, there is only one table_share object which should never
be used outside of LOCK_open
*/
if ((share= get_cached_table_share(view->db, view->table_name)))
{
DBUG_ASSERT(share->ref_count == 0);
pthread_mutex_lock(&share->mutex);
share->ref_count++;
share->version= 0;
pthread_mutex_unlock(&share->mutex);
release_table_share(share, RELEASE_WAIT_FOR_DROP);
}
query_cache_invalidate3(thd, view, 0);
sp_cache_invalidate();
2004-07-16 00:15:55 +02:00
}
if (error)
{
2006-10-20 23:44:03 +02:00
VOID(pthread_mutex_unlock(&LOCK_open));
DBUG_RETURN(TRUE);
}
if (wrong_object_name)
{
2006-10-20 23:44:03 +02:00
VOID(pthread_mutex_unlock(&LOCK_open));
my_error(ER_WRONG_OBJECT, MYF(0), wrong_object_db, wrong_object_name,
"VIEW");
DBUG_RETURN(TRUE);
}
if (non_existant_views.length())
{
2006-10-20 23:44:03 +02:00
VOID(pthread_mutex_unlock(&LOCK_open));
my_error(ER_BAD_TABLE_ERROR, MYF(0), non_existant_views.c_ptr());
DBUG_RETURN(TRUE);
}
2006-10-20 23:44:03 +02:00
write_bin_log(thd, TRUE, thd->query, thd->query_length);
2006-10-20 23:44:03 +02:00
2004-07-16 00:15:55 +02:00
send_ok(thd);
2006-10-20 23:44:03 +02:00
VOID(pthread_mutex_unlock(&LOCK_open));
DBUG_RETURN(FALSE);
2004-07-16 00:15:55 +02:00
}
/*
Check type of .frm if we are not going to parse it
SYNOPSIS
mysql_frm_type()
path path to file
RETURN
FRMTYPE_ERROR error
FRMTYPE_TABLE table
FRMTYPE_VIEW view
*/
frm_type_enum mysql_frm_type(THD *thd, char *path, enum legacy_db_type *dbt)
2004-07-16 00:15:55 +02:00
{
File file;
uchar header[10]; //"TYPE=VIEW\n" it is 10 characters
int error;
2004-07-16 00:15:55 +02:00
DBUG_ENTER("mysql_frm_type");
*dbt= DB_TYPE_UNKNOWN;
if ((file= my_open(path, O_RDONLY | O_SHARE, MYF(0))) < 0)
2004-07-16 00:15:55 +02:00
DBUG_RETURN(FRMTYPE_ERROR);
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
error= my_read(file, (uchar*) header, sizeof(header), MYF(MY_WME | MY_NABP));
2004-07-16 00:15:55 +02:00
my_close(file, MYF(MY_WME));
if (error)
DBUG_RETURN(FRMTYPE_ERROR);
if (!strncmp((char*) header, "TYPE=VIEW\n", sizeof(header)))
DBUG_RETURN(FRMTYPE_VIEW);
/*
This is just a check for DB_TYPE. We'll return default unknown type
if the following test is true (arg #3). This should not have effect
on return value from this function (default FRMTYPE_TABLE)
*/
if (header[0] != (uchar) 254 || header[1] != 1 ||
(header[2] != FRM_VER && header[2] != FRM_VER+1 &&
(header[2] < FRM_VER+3 || header[2] > FRM_VER+4)))
DBUG_RETURN(FRMTYPE_TABLE);
*dbt= (enum legacy_db_type) (uint) *(header + 3);
DBUG_RETURN(FRMTYPE_TABLE); // Is probably a .frm table
2004-07-16 00:15:55 +02:00
}
/*
check of key (primary or unique) presence in updatable view
SYNOPSIS
check_key_in_view()
thd thread handler
view view for check with opened table
DESCRIPTION
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 it is VIEW and query have LIMIT clause then check that underlying
table of view contain one of following:
1) primary key of underlying table
2) unique key underlying table with fields for which NULL value is
impossible
3) all fields of underlying table
2004-07-16 00:15:55 +02:00
RETURN
FALSE OK
TRUE view do not contain key or all fields
*/
bool check_key_in_view(THD *thd, TABLE_LIST *view)
{
TABLE *table;
Field_translator *trans, *end_of_trans;
KEY *key_info, *key_info_end;
2004-07-16 00:15:55 +02:00
DBUG_ENTER("check_key_in_view");
/*
2004-11-11 20:18:10 +01:00
we do not support updatable UNIONs in VIEW, so we can check just limit of
LEX::select_lex
*/
if ((!view->view && !view->belong_to_view) ||
thd->lex->sql_command == SQLCOM_INSERT ||
thd->lex->select_lex.select_limit == 0)
DBUG_RETURN(FALSE); /* it is normal table or query without LIMIT */
table= view->table;
view= view->top_table();
trans= view->field_translation;
key_info_end= (key_info= table->key_info)+ table->s->keys;
2004-07-16 00:15:55 +02:00
end_of_trans= view->field_translation_end;
DBUG_ASSERT(table != 0 && view->field_translation != 0);
2004-07-16 00:15:55 +02:00
{
/*
We should be sure that all fields are ready to get keys from them, but
this operation should not have influence on Field::query_id, to avoid
marking as used fields which are not 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
enum_mark_columns save_mark_used_columns= thd->mark_used_columns;
thd->mark_used_columns= MARK_COLUMNS_NONE;
DBUG_PRINT("info", ("thd->mark_used_columns: %d", thd->mark_used_columns));
for (Field_translator *fld= trans; fld < end_of_trans; fld++)
{
if (!fld->item->fixed && fld->item->fix_fields(thd, &fld->item))
2005-08-12 11:17:27 +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
thd->mark_used_columns= save_mark_used_columns;
return TRUE;
2005-08-12 11:17:27 +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
thd->mark_used_columns= save_mark_used_columns;
DBUG_PRINT("info", ("thd->mark_used_columns: %d", thd->mark_used_columns));
}
/* Loop over all keys to see if a unique-not-null key is used */
for (;key_info != key_info_end ; key_info++)
2004-07-16 00:15:55 +02:00
{
if ((key_info->flags & (HA_NOSAME | HA_NULL_PART_KEY)) == HA_NOSAME)
2004-07-16 00:15:55 +02:00
{
KEY_PART_INFO *key_part= key_info->key_part;
KEY_PART_INFO *key_part_end= key_part + key_info->key_parts;
/* check that all key parts are used */
for (;;)
2004-07-16 00:15:55 +02:00
{
Field_translator *k;
for (k= trans; k < end_of_trans; k++)
2004-07-16 00:15:55 +02:00
{
Item_field *field;
if ((field= k->item->filed_for_view_update()) &&
field->field == key_part->field)
2004-07-16 00:15:55 +02:00
break;
}
if (k == end_of_trans)
break; // Key is not possible
if (++key_part == key_part_end)
DBUG_RETURN(FALSE); // Found usable key
2004-07-16 00:15:55 +02:00
}
}
}
DBUG_PRINT("info", ("checking if all fields of table are used"));
2004-07-16 00:15:55 +02:00
/* check all fields presence */
{
Field **field_ptr;
Field_translator *fld;
for (field_ptr= table->field; *field_ptr; field_ptr++)
2004-07-16 00:15:55 +02:00
{
for (fld= trans; fld < end_of_trans; fld++)
2004-07-16 00:15:55 +02:00
{
Item_field *field;
if ((field= fld->item->filed_for_view_update()) &&
field->field == *field_ptr)
2004-07-16 00:15:55 +02:00
break;
}
if (fld == end_of_trans) // If field didn't exists
2004-07-16 00:15:55 +02:00
{
/*
Keys or all fields of underlying tables are not found => we have
to check variable updatable_views_with_limit to decide should we
issue an error or just a warning
*/
if (thd->variables.updatable_views_with_limit)
2004-07-16 00:15:55 +02:00
{
/* update allowed, but issue warning */
push_warning(thd, MYSQL_ERROR::WARN_LEVEL_NOTE,
ER_WARN_VIEW_WITHOUT_KEY, ER(ER_WARN_VIEW_WITHOUT_KEY));
DBUG_RETURN(FALSE);
2004-07-16 00:15:55 +02:00
}
/* prohibit update */
DBUG_RETURN(TRUE);
2004-07-16 00:15:55 +02:00
}
}
}
DBUG_RETURN(FALSE);
}
/*
insert fields from VIEW (MERGE algorithm) into given list
SYNOPSIS
insert_view_fields()
thd thread handler
2004-07-16 00:15:55 +02:00
list list for insertion
view view for processing
RETURN
2004-11-25 01:23:13 +01:00
FALSE OK
TRUE error (is not sent to cliet)
2004-07-16 00:15:55 +02:00
*/
bool insert_view_fields(THD *thd, List<Item> *list, TABLE_LIST *view)
2004-07-16 00:15:55 +02:00
{
Field_translator *trans_end;
Field_translator *trans;
2004-07-16 00:15:55 +02:00
DBUG_ENTER("insert_view_fields");
if (!(trans= view->field_translation))
2004-11-25 01:23:13 +01:00
DBUG_RETURN(FALSE);
trans_end= view->field_translation_end;
2004-07-16 00:15:55 +02:00
for (Field_translator *entry= trans; entry < trans_end; entry++)
2004-07-16 00:15:55 +02:00
{
Item_field *fld;
if ((fld= entry->item->filed_for_view_update()))
list->push_back(fld);
else
{
my_error(ER_NON_INSERTABLE_TABLE, MYF(0), view->alias, "INSERT");
2004-11-25 01:23:13 +01:00
DBUG_RETURN(TRUE);
}
2004-07-16 00:15:55 +02:00
}
2004-11-25 01:23:13 +01:00
DBUG_RETURN(FALSE);
2004-07-16 00:15:55 +02:00
}
/*
checking view md5 check suum
SINOPSYS
view_checksum()
thd threar handler
view view for check
RETUIRN
HA_ADMIN_OK OK
HA_ADMIN_NOT_IMPLEMENTED it is not VIEW
HA_ADMIN_WRONG_CHECKSUM check sum is wrong
*/
int view_checksum(THD *thd, TABLE_LIST *view)
{
char md5[MD5_BUFF_LENGTH];
if (!view->view || view->md5.length != 32)
return HA_ADMIN_NOT_IMPLEMENTED;
view->calc_md5(md5);
return (strncmp(md5, view->md5.str, 32) ?
HA_ADMIN_WRONG_CHECKSUM :
HA_ADMIN_OK);
}
2005-09-16 17:13:21 +02:00
/*
rename view
2005-09-16 17:13:21 +02:00
Synopsis:
renames a view
2005-09-16 17:13:21 +02:00
Parameters:
thd thread handler
new_name new name of view
view view
2005-09-16 17:13:21 +02:00
Return values:
FALSE Ok
TRUE Error
*/
2005-09-16 17:13:21 +02:00
bool
mysql_rename_view(THD *thd,
const char *new_name,
TABLE_LIST *view)
2005-09-16 17:13:21 +02:00
{
LEX_STRING pathstr;
2005-09-16 17:13:21 +02:00
File_parser *parser;
char path_buff[FN_REFLEN];
bool error= TRUE;
2005-09-16 17:13:21 +02:00
DBUG_ENTER("mysql_rename_view");
pathstr.str= (char *) path_buff;
pathstr.length= build_table_filename(path_buff, sizeof(path_buff) - 1,
view->db, view->table_name,
reg_ext, 0);
2005-09-16 17:13:21 +02:00
if ((parser= sql_parse_prepare(&pathstr, thd->mem_root, 1)) &&
is_equal(&view_type, parser->type()))
{
TABLE_LIST view_def;
char dir_buff[FN_REFLEN];
LEX_STRING dir, file;
2005-09-16 17:13:21 +02:00
/*
To be PS-friendly we should either to restore state of
TABLE_LIST object pointed by 'view' after using it for
view definition parsing or use temporary 'view_def'
object for it.
*/
bzero(&view_def, sizeof(view_def));
view_def.timestamp.str= view_def.timestamp_buffer;
view_def.view_suid= TRUE;
2005-09-16 17:13:21 +02:00
/* get view definition and source */
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
if (parser->parse((uchar*)&view_def, thd->mem_root, view_parameters,
array_elements(view_parameters)-1,
&file_parser_dummy_hook))
goto err;
2005-09-16 17:13:21 +02:00
/* rename view and it's backups */
if (rename_in_schema_file(view->db, view->table_name, new_name,
view_def.revision - 1, num_view_backups))
goto err;
2005-09-16 17:13:21 +02:00
dir.str= dir_buff;
dir.length= build_table_filename(dir_buff, sizeof(dir_buff) - 1,
view->db, "", "", 0);
2005-09-16 17:13:21 +02:00
pathstr.str= path_buff;
pathstr.length= build_table_filename(path_buff, sizeof(path_buff) - 1,
view->db, new_name, reg_ext, 0);
2005-09-16 17:13:21 +02:00
file.str= pathstr.str + dir.length;
file.length= pathstr.length - dir.length;
2005-09-16 17:13:21 +02:00
if (sql_create_definition_file(&dir, &file, view_file_type,
WL#3817: Simplify string / memory area types and make things more consistent (first part) The following type conversions was done: - Changed byte to uchar - Changed gptr to uchar* - Change my_string to char * - Change my_size_t to size_t - Change size_s to size_t Removed declaration of byte, gptr, my_string, my_size_t and size_s. Following function parameter changes was done: - All string functions in mysys/strings was changed to use size_t instead of uint for string lengths. - All read()/write() functions changed to use size_t (including vio). - All protocoll functions changed to use size_t instead of uint - Functions that used a pointer to a string length was changed to use size_t* - Changed malloc(), free() and related functions from using gptr to use void * as this requires fewer casts in the code and is more in line with how the standard functions work. - Added extra length argument to dirname_part() to return the length of the created string. - Changed (at least) following functions to take uchar* as argument: - db_dump() - my_net_write() - net_write_command() - net_store_data() - DBUG_DUMP() - decimal2bin() & bin2decimal() - Changed my_compress() and my_uncompress() to use size_t. Changed one argument to my_uncompress() from a pointer to a value as we only return one value (makes function easier to use). - Changed type of 'pack_data' argument to packfrm() to avoid casts. - Changed in readfrm() and writefrom(), ha_discover and handler::discover() the type for argument 'frmdata' to uchar** to avoid casts. - Changed most Field functions to use uchar* instead of char* (reduced a lot of casts). - Changed field->val_xxx(xxx, new_ptr) to take const pointers. Other changes: - Removed a lot of not needed casts - Added a few new cast required by other changes - Added some cast to my_multi_malloc() arguments for safety (as string lengths needs to be uint, not size_t). - Fixed all calls to hash-get-key functions to use size_t*. (Needed to be done explicitely as this conflict was often hided by casting the function to hash_get_key). - Changed some buffers to memory regions to uchar* to avoid casts. - Changed some string lengths from uint to size_t. - Changed field->ptr to be uchar* instead of char*. This allowed us to get rid of a lot of casts. - Some changes from true -> TRUE, false -> FALSE, unsigned char -> uchar - Include zlib.h in some files as we needed declaration of crc32() - Changed MY_FILE_ERROR to be (size_t) -1. - Changed many variables to hold the result of my_read() / my_write() to be size_t. This was needed to properly detect errors (which are returned as (size_t) -1). - Removed some very old VMS code - Changed packfrm()/unpackfrm() to not be depending on uint size (portability fix) - Removed windows specific code to restore cursor position as this causes slowdown on windows and we should not mix read() and pread() calls anyway as this is not thread safe. Updated function comment to reflect this. Changed function that depended on original behavior of my_pwrite() to itself restore the cursor position (one such case). - Added some missing checking of return value of malloc(). - Changed definition of MOD_PAD_CHAR_TO_FULL_LENGTH to avoid 'long' overflow. - Changed type of table_def::m_size from my_size_t to ulong to reflect that m_size is the number of elements in the array, not a string/memory length. - Moved THD::max_row_length() to table.cc (as it's not depending on THD). Inlined max_row_length_blob() into this function. - More function comments - Fixed some compiler warnings when compiled without partitions. - Removed setting of LEX_STRING() arguments in declaration (portability fix). - Some trivial indentation/variable name changes. - Some trivial code simplifications: - Replaced some calls to alloc_root + memcpy to use strmake_root()/strdup_root(). - Changed some calls from memdup() to strmake() (Safety fix) - Simpler loops in client-simple.c
2007-05-10 11:59:39 +02:00
(uchar*)&view_def, view_parameters,
num_view_backups))
{
2005-09-16 17:13:21 +02:00
/* restore renamed view in case of error */
rename_in_schema_file(view->db, new_name, view->table_name,
view_def.revision - 1, num_view_backups);
goto err;
2005-09-16 17:13:21 +02:00
}
} else
DBUG_RETURN(1);
/* remove cache entries */
query_cache_invalidate3(thd, view, 0);
sp_cache_invalidate();
error= FALSE;
err:
DBUG_RETURN(error);
2005-09-16 17:13:21 +02:00
}