#5860 "Multi-table UPDATE does not activate update triggers"
#6812 "Triggers are not activated for INSERT ... SELECT"
#8755 "Trigger is not activated by LOAD DATA".
This patch also implements proper handling of triggers for special forms
of insert like REPLACE or INSERT ... ON DUPLICATE KEY UPDATE.
Also now we don't call after trigger in case when we have failed to
inserted/update or delete row. Trigger failure should stop statement
execution.
I have not properly tested handling of errors which happen inside of
triggers in this patch, since it is simplier to do this once we will be
able to access tables from triggers.
Memory leak in locally evalutated expressions during SP execution fixed by
reusing allocated item slots when possible.
Note: No test case added, since the test is a stress test that tries to make
the machine to run out of memory.
Second attempt, now tested with debug build, valgrind build, max (optimized)
build, with and without --debug, --vagrind and --ps-protocol.
Errors in trigger and view test with --debug in debug build where present
before this patch, and likewise for valgrind warnings for view test in
valgrind build with --ps-protocol.
Memory leak in locally evalutated expressions during SP execution fixed by
reusing allocated item slots when possible.
Note: No test case added, since the test is a stress test that tries to make
the machine to run out of memory.
records if prepared statements is used".
This fix changes equality evaluation method of basic constants from
by-name to by-value, thus effectively enabling use of parameter markers
in some optimizations (constants propagation, evaluation of possible
keys for query).
Added a test case for bug #9338.
sql_select.cc:
Fixed bug #9338.
When an occurence of a field reference has to be replaced
by another field reference the whole Item_field must be
replaced.
item.cc:
Fixed bug #9338.
The method Item_field::replace_equal_field_processor was
replaced by Item_field::replace_equal_field. The new method
is used to replace the occurences of Item_field objects.
item.h:
Fixed bug #9338.
The virtual function replace_equal_field_processor was replaced
by replace_equal_field. The latter is supposed to be used as a
callback function in calls of the method transform.
fixing tests accordingly
item.cc:
Allow mixing non-binary collation and binary collation
even if coercibility is the same.
For easier 4.0 -> 4.1 migrating.
CAST() now produces warnings when casting a wrong INTEGER or CHAR values. This also applies to implicite string to number casts. (Bug #5912)
ALTER TABLE now fails in STRICT mode if it generates warnings.
Inserting a zero date in a DATE, DATETIME or TIMESTAMP column during TRADITIONAL mode now produces an error. (Bug #5933)
- Add function Item_param::fix_fields which will update any subselect they are part of and indicate that the subsleect is not const during prepare phase, and thus should not be executed during prepare.
Windows to call CreateFileMapping() with correct arguments, and
propogating the introduction of query_id_t to everywhere query ids are
passed around. (Bug #8826)
don't call escape_string_for_mysql() unnecesary
don't overwrite local buffer
escape_string_for_mysql():
take a length of the destination buffer as an argument
implementation of AVG(DISTINCT) which utilizes the approach with Fields.
The patch implemented in October is portede to the up-to-date tree
containing DECIMAL type.
Tests for AVG(DISTINCT) (although there is not much to test provided
that SUM(DISTINCT) works), cleanups for COUNT(DISTINCT) and GROUP_CONCAT()
will follow in another changeset.
The reported problems were due to two completely unrelated omissions.
1) The file sort procedure didn't correctly create the sort key in
make_sortkey when the sortkey was an unsigned integer.
2) The name resolution procedure for column references inside a HAVING
clause did not propagate the unsigned_flag of the resolved references.
This patch corrects both problems.
fixing test results accordingly.
func_system.test:
New test that illegal mix of collations does not happen anymore.
item_strfunc.h:
safe_charset_converter() was added for system constants.
item_strfunc.cc:
safe_charset_converter() was added for system constants.
item_func.cc, item.h, item.cc:
Bug#8291: Illegal collation mix with USER() function.
After discussion with PeterG and Serge, a new coercibility
level for "system constants" was introduced, between
COERRIBLE and IMPLICIT. Thus:
SELECT col1 = USER() FROM t1; - is done according to col1 collation.
SELECT 'string' = USER(); - is done according to USER() collation.
At the same time, "nagg" and "strong" members were removed as unused.
item_create.cc:
Version is a system constant too.
with view-merge algorithm and prepared statements: in case when some
Item_ref pointing to a view column was substituted with a reference
pointing to the view expression for that column
Item_ref::ref member of the original Item_ref was left pointing to
not_found_item (0x1).
As we currently perform expression substition part of the view-merge
algorithm per each execution of a prepared statement or stored procedure,
we need to preserve original Item_ref objects usable.
In Item_cache_decimal::store(item) the call item->val_decimal_result()
returns NULL if the passed item has an SQL null value. Don't try copying
NULL into Item_cache_decimal::val in this case.
Change string->float conversion to delay division as long as possible.
This gives us more exact integer->float conversion for numbers of type '123.45E+02' (Bug #7740)
Lots of small fixes to multi-precision-math path
Give Note for '123.4e'
Added helper functions type 'val_string_from_real()
Don't give warnings for end space for string2decimal()
Changed storage of values for SP so that we can detect length of argument without strlen()
Changed interface for str2dec() so that we must supple the pointer to the last character in the buffer
Ensure that references in HAVING, ORDER BY or GROUP BY are calculated after fields in SELECT.
This will ensure that any reference to these has a valid value.
Generalized the code for split_sum_func()
not default_charset_into. It fixes the
problem that in some cases numbers where
treated as CHAR(N), not as BINARY(N), e.g.
wrong 'charsetnr' when sent to the client side.
2. IFNULL didn't aggregate argument charsets
and collations, so IFNULL(1,'a') produced
a CHAR(N). Now produces a BINARY(N).
3. SELECT PROCEDURE ANALIZE now returns
BINARY columns, which is much better than it worked
previously: CHAR with the default character set.
But in the future it's worth to fix the fields
'Field_name' and 'Optimal_fieldtype' to use UTF8,
and 'Min_value' and 'Max_value' to inherit their charsets
from the original items. But it is not important,
and BINARY(N) is OK for now.
4. Tests were fixed accordingly. No new tests were
made, as the old onces cover everything.
Split TABLE to TABLE and TABLE_SHARE (TABLE_SHARE is still allocated as part of table, will be fixed soon)
Created Field::make_field() and made Field_num::make_field() to call this
Added 'TABLE_SHARE->db' that points to database name; Changed all usage of table_cache_key as database name to use this instead
Changed field->table_name to point to pointer to alias. This allows us to change alias for a table by just updating one pointer.
Renamed TABLE_SHARE->real_name to table_name
Renamed TABLE->table_name to alias
Renamed TABLE_LIST->real_name to table_name
* Added Item_ref::set_properties
* Item_ref::Item_ref now expects to get in *item either
NULL - then fix_fields() will be called later or
ptr to Item it will refer to - then an equivalent of fix_fields() call is performed
The problem in 4.1 was the same as in 4.0 - fix_fields() not called for created Item_ref.
The fix is similar too - initialize Item_refs in ctor (but don't interfere with cases when
Item_ref is used by subselects).
In Item_ref::Item_ref set maybe_null (and other fields fix_fields sets) to be the
same as in (*ref), because Item_ref::fix_fields() will not be called. Previously
maybe_null was 0 always and this produced a bogus state where
maybe_null==0 && is_null() == true
which broke evaluation for some upper-level Items, like AND and OR.
Renamed HA_VAR_LENGTH to HA_VAR_LENGTH_PART
Renamed in all files FIELD_TYPE_STRING and FIELD_TYPE_VAR_STRING to MYSQL_TYPE_STRING and MYSQL_TYPE_VAR_STRING to make it easy to catch all possible errors
Added support for VARCHAR KEYS to heap
Removed support for ISAM
Now only long VARCHAR columns are changed to TEXT on demand (not CHAR)
Internal temporary files can now use fixed length tables if the used VARCHAR columns are short
out of order". (final version)
Now instead of binding Item_trigger_field to TABLE objects during
trigger definition parsing at table open, we perform pass through
special list of all such objects in trigger. This allows easily check
all references to fields in old/new version of row in trigger during
execution of CREATE TRIGGER statement (this is more courtesy for users
since we can't check everything anyway).
We also report that such reference is bad by returning error from
Item_trigger_field::fix_fields() method (instead of setup_field())
This means that if trigger is broken we will bark during trigger
execution instead of trigger definition parsing at table open.
(i.e. now we allow to open tables with broken triggers).
In server we assume that datetime values stored in MYSQL_TIME struct
are normalized (and year is not greater than 9999), so we should
perform range checks in all places then we convert something to
MYSQL_TIME.
the result takes its charset/collation
attributes from the character string,
e.g. SELECT func(NULL, _latin2'string')
now returns a latin2 result. This is
done by introducing a new derivation
(aka coercibility) level DERIVATION_IGNORABLE,
which is used with Item_null.
2. 'Pure' NULL is now BINARY(0), not CHAR(0).
I.e. NULL is now more typeless.
of client equals to character set of connection, possibly required
conversion to character set of column is not performed
(prepared statements, data is supplied using placeholders).
FOUND is not a reserved keyword anymore
Added Item_field::set_no_const_sub() to be able to mark fields that can't be substituted
Added 'simple_select' method to be able to quickly determinate if a select_result is a normal SELECT
Note that the 5.0 tree is not yet up to date: Sanja will have to fix multi-update-locks for this merge to be complete
- Changed name resolution for GROUP BY so that derived columns do not shadow table columns
from the FROM clause. As a result GROUP BY now is handled as a true ANSI extentsion.
- Issue a warning when HAVING is resolved into ambiguous columns, and prefer the columns from
the GROUP BY clause over SELECT columns.
constant with a column. The string is converted into the column
character set. It conversion doesn't lose data, then operation
is possible. Otherwise, give an error, as it was earlier.
Simple optimzations and cleanups
Removed compiler warnings and fixed portability issues
Added client functions 'mysql_embedded()' to allow client to check if we are using embedded server
Fixes for purify
Added the code processing on expressions for applying
multiple equalities.
sql_select.cc:
Post-merge fixes for Item_equal patch.
Added the code processing on expressions for applying
multiple equalities.
Many files:
Post-merge fixes for Item_equal patch.
item_cmpfunc.cc:
Post-merge fixes for Item_equal patch.
Fixed a problem when an equality field=const cannot be applied to
the predicate P(field,c) for constant propagation as a conversion
of field is needed.
item.h, item.cc:
Fixed a problem when an equality field=const cannot be applied to
the predicate P(field,c) for constant propagation as a conversion
of field is needed.
statements and negative time/date values".
The bug was in wrong sprintf format used in the client library.
The fix moves TIME -> string conversion functions to sql-common and
utilized them in the client library.
crashes server (prepared statements)": the bug was that all boolean
items always recovered its original arguments at statement cleanup
stage.
This collided with Item_subselect::select_transformer, which tries to
permanently change the item tree to use a transformed subselect instead of
original one.
So we had this call sequence for prepare:
mysql_stmt_prepare -> JOIN::prepare ->
Item_subselect::fix_fields -> the item tree gets transformed ->
Item_bool_rowready_func2::cleanup, item tree is recovered to original
state, while it shouldn't have been;
mysql_stmt_execute -> attempts to execute a broken tree -> crash.
Now instead of bluntly recovering all arguments of bool functions in
Item_bool_rowready_func2::cleanup, we recover only those
which were changed, and do it in one place.
There still would exist a possibility for a collision with subselect
tranformation, if permanent and temporary changes were performed at the
same stage.
But fortunately subselect transformation is always done first, so it
doesn't conflict with the optimization done by propogate_cond_constants.
Now we have:
mysql_stmt_prepare -> JOIN::prepare -> subselect transformation
permanently changes the tree -> cleanup doesn't recover anything,
because nothing was registered for recovery.
mysql_stmt_execute -> JOIN::prepare (the tree is already transformed,
so it doesn't change), JOIN::optimize ->
propogate_cond_constants -> temporary changes the item tree
with constants -> JOIN::execute -> cleanup ->
the changes done by propogate_cond_constants are recovered, as
they were registered for recovery.
crashes mysqld": implementation for a generic item tree modifications
registry. Every item tree modification which should be rolled back for
subsequent execution of a prepared statement or stored procedure should
be saved in the registry. All such modifications are rolled back at once
during cleanup stage of PS.
Actual fix for the bug just adds a call to register modifications to
convert_constant_item.
Post review fixes implemented.
More tests.
Better error messages.
Fixed bug when checking if we updated all needed columns for INSERT.
Give an error if we encounter a wrong float value during parsing.
Don't print DEFAULT for columns without a default value in SHOW CREATE/SHOW FIELDS.
Fixed UPDATE IGNORE when using STRICT mode.
identical to another in result"
According to SQL standard queries like
"select t1.a as col from t1, t2 order by a" should return an error if
both tables contain field a.
Fixed (together with Guilhem) bugs in mysqlbinlog regarding --offset
Prefix addresses with 0x for easier comparisons of debug logs
Fixed problem where MySQL choosed index-read even if there would be a much better range on the same index
This fix changed some 'index' queries to 'range' queries in the test suite
Don't create 'dummy' WHERE clause for trivial WHERE clauses where we can remove the WHERE clause.
This fix removed of a lot of 'Using where' notes in the test suite.
Give NOTE instead of WARNING if table/function doesn't exists when using DROP IF EXISTS
Give NOTE instead of WARNING for safe field-type conversions
Mostly needed for Monty for him getting notion what needed for triggers
from new .FRM format.
Things to be done:
- Right placement of trigger's invocations
- Right handling of errors in triggers (including transaction rollback)
- Support for priviliges
- Right handling of DROP/RENAME table (hope that it will be handled automatically
with merging of .TRG into .FRM file)
- Saving/restoring some information critical for trigger creation and replication
with their definitions (e.g. sql_mode, creator, ...)
- Replication
Already has some known bugs so probably not for general review.
work (prepared statements)" and after-review fixes:
- str_to_TIME renamed to str_to_datetime to pair with str_to_time
- functions str_to_time and str_to_TIME moved to sql-common
- send_data_str now supports MYSQL_TYPE_TIME, MYSQL_TIME_DATE,
MYSQL_TIME_DATETIME types of user input buffers.
- few more comments in the client library
- a test case added.
Added basic per-thread time zone functionality (based on public
domain elsie-code). Now user can select current time zone
(from the list of time zones described in system tables).
All NOW-like functions honor this time zone, values of TIMESTAMP
type are interpreted as values in this time zone, so now
our TIMESTAMP type behaves similar to Oracle's TIMESTAMP WITH
LOCAL TIME ZONE (or proper PostgresSQL type).
WL#1266 "CONVERT_TZ() - basic time with time zone conversion
function".
Fixed problems described in Bug #2336 (Different number of warnings
when inserting bad datetime as string or as number). This required
reworking of datetime realted warning hadling (they now generated
at Field object level not in conversion functions).
Optimization: Now Field class descendants use table->in_use member
instead of current_thd macro.
character set introducer": add new item type to be returned before from
Item_param until it's value is set.
This way items like Item_bool_func2 and udf_handler won't treat this item
as constant literal when statement is prepared.
after Monty's review.
- Item_param was rewritten.
- it turns out that we can't convert string data to character set of
connection on the fly, because they first should be written to the binary
log.
To support efficient conversion we need to rewrite prepared statements
binlogging code first.
