The problem is that the lexer could inadvertently skip over the
end of a query being parsed if it encountered a malformed multibyte
character. A specially crated query string could cause the lexer
to jump up to six bytes past the end of the query buffer. Another
problem was that the laxer could use unfiltered user input as
a signed array index for the parser maps (having upper and lower
bounds 0 and 256 respectively).
The solution is to ensure that the lexer only skips over well-formed
multibyte characters and that the index value of the parser maps
is always a unsigned value.
mysql-test/r/ctype_recoding.result:
Update test case result: ending backtick is not skipped over anymore.
sql/sql_lex.cc:
Characters being analyzed must be unsigned as they can be
used as indexes for the parser maps. Only skip over if the
string is a valid multi-byte sequence.
tests/mysql_client_test.c:
Add test case for Bug#45010
All committed result differences have either been verified by me or copied from Oracle's provided
results (storage/innodb_plugin/mysql-test/*.result, storage/innodb_plugin/mysql-test/patches).
mysql-test/r/information_schema.result:
queries changed a bit
mysql-test/r/information_schema_db.result:
queries changed a bit
mysql-test/r/innodb-autoinc.result:
importing piece from storage/innodb_plugin/mysql-test/innodb-autoinc.result
mysql-test/r/innodb.result:
result close to storage/innodb_plugin/innodb.result, except 4095 pages instead of 8191, which makes
sense as Summit runs with a buffer pool of 64M, whereas the mentioned result file was made with
a buffer pool of 128M.
mysql-test/r/mysqlshow.result:
InnoDB I_S tables have arrived
mysql-test/suite/funcs_1/r/is_columns_is.result:
queries changed a bit
mysql-test/suite/funcs_1/r/is_columns_is_embedded.result:
queries changed a bit
mysql-test/suite/funcs_1/r/is_tables_is.result:
queries changed a bit
mysql-test/suite/funcs_1/t/is_columns_is.test:
making I_S queries ignore InnoDB I_S tables
mysql-test/suite/funcs_1/t/is_columns_is_embedded.test:
making I_S queries ignore InnoDB I_S tables
mysql-test/suite/funcs_1/t/is_tables_is.test:
making I_S queries ignore InnoDB I_S tables
mysql-test/suite/innodb/r/innodb-zip.result:
result update
mysql-test/suite/innodb/t/innodb-zip.test:
439, as size of prefix key, throws error with certain system zlib (ubuntu
"intrepid") but not with zlib bundled with MySQL, because zlib's
compressBound() are different (and used by InnoDB's page_zip_empty_size()).
mysql-test/suite/sys_vars/r/innodb_file_per_table_basic.result:
result update
mysql-test/suite/sys_vars/r/innodb_lock_wait_timeout_basic.result:
result update
mysql-test/suite/sys_vars/r/innodb_sync_spin_loops_basic_32.result:
result update (default value is 30 in the plugin, 20 in the builtin)
mysql-test/suite/sys_vars/r/innodb_sync_spin_loops_basic_64.result:
result update (default value is 30 in the plugin, 20 in the builtin)
mysql-test/suite/sys_vars/r/table_definition_cache_basic.result:
result update (default value is 400 in Summit)
mysql-test/suite/sys_vars/t/innodb_file_per_table_basic.test:
variable is writable in the plugin (patch from Oracle)
mysql-test/suite/sys_vars/t/innodb_lock_wait_timeout_basic.test:
variable is per-session in the plugin (patch from Oracle)
mysql-test/t/information_schema.test:
making I_S queries ignore InnoDB I_S tables
mysql-test/t/information_schema_db.test:
making I_S queries ignore InnoDB I_S tables
mysql-test/t/innodb-autoinc.test:
importing piece from storage/innodb_plugin/mysql-test/innodb-autoinc.test
mysql-test/t/innodb.test:
importing most of storage/innodb_plugin/t/innodb.test. Most replace_result were not needed (no path printed), some where (for --embedded).
mysql-test/t/mysqlshow.test:
output of test now depends on InnoDB or not InnoDB. As there is no way to make mysqlshow
produce a single output in those two cases (no way to make it exclude InnoDB I_S
tables), let the test depend on InnoDB, it isn't a very selective condition, and the
test is simple enough.
storage/innobase/CMakeLists.txt:
thanks Vlad for the noticing () vs {}
Problem 1:
When the 'Using index' optimization is used, the optimizer may still - after
cost-based optimization - decide to use another index in order to avoid using
a temporary table. But when this happens, the flag to the storage engine to
read index only (not table) was still set. Fixed by resetting the flag in the
storage engine and TABLE structure in the above scenario, unless the new index
allows for the same optimization.
Problem 2:
When a 'ref' access method was employed by cost-based optimizer, (when the column
is non-NULLable), it was assumed that it needed no initialization if 'quick' access
methods (since they are based on range scan). When ORDER BY optimization overrides
the decision, however, it expects to have this initialized and hence crashes.
Fixed in 5.1 (was fixed in 6.0 already) by initializing 'quick' even when there's
'ref' access.
mysql-test/r/order_by.result:
Bug#46454: Test result.
mysql-test/t/order_by.test:
Bug#46454: Test case.
sql/sql_select.cc:
Bug#46454:
Problem 1 fixed in make_join_select()
Problem 2 fixed in test_if_skip_sort_order()
sql/table.h:
Bug#46454: Added comment to field.
column on partitioned table
An assertion 'ASSERT_COULUMN_MARKED_FOR_READ' is failed if the query
is executed with index containing double column on partitioned table.
The problem is that assertion expects all the fields which are read,
to be in the read_set.
In this query only the field 'a' is in the readset as the tables in
the query are joined by the field 'a' and so the assertion fails
expecting other field 'b'.
Since the function cmp() is just comparison of two parameters passed,
the assertion is not required.
Fixed by removing the assertion in the double fields comparision
function and also fixed the index initialization to do ordered
index scan with RW LOCK which ensures all the fields from a key are in
the read_set.
Note: this bug is not reproducible with other datatypes because the
assertion doesn't exist in comparision function for other
datatypes.
mysql-test/r/partition.result:
Testcase for BUG#45816
mysql-test/t/partition.test:
Testcase for BUG#45816
sql/field.cc:
Removed the assertion ASSERT_COLUMN_MARED_FOR_READ in Field_double::cmp()
function
sql/ha_partition.cc:
Fixed index_int() method to make it initialize the read_set properly if
ordered index scan with RW lock is requested.
bzr branch mysql-5.1-performance-version mysql-trunk # Summit
cd mysql-trunk
bzr merge mysql-5.1-innodb_plugin # which is 5.1 + Innodb plugin
bzr rm innobase # remove the builtin
Next step: build, test fixes.
The test case fails sporadically on Windows while trying to overwrite an unused
binary log. The problem stems from the fact that MySQL on Windows does not
immediately unlock/release a file while the process that opened and closed it is
still running. In BUG 38603, this issue was circumvented by stopping the MySQL
process, copying the file and then restarting the MySQL process.
Unfortunately, such facilities are not available in the 5.0. Other approaches
such as stopping the slave and issuing change master do not work because the relay
log file and index are not closed when a slave is stopped. So to fix the problem,
we simply don't run on windows the part of the test that was failing.
engine to the partition_csv test. Also remove test case that was
duplicated. Fix connection procedure with the embedded server.
mysql-test/r/partition.result:
Update test case result.
mysql-test/r/partition_csv.result:
Update test case result.
mysql-test/t/partition.test:
Move test cases to the partition_csv test.
mysql-test/t/partition_csv.test:
Move tests from partition.test and remove duplicate.
Tweaky connection procedure to work with embedded.
when used with --tab
1) New syntax: added CHARACTER SET clause to the
SELECT ... INTO OUTFILE (to complement the same clause in
LOAD DATA INFILE).
mysqldump is updated to use this in --tab mode.
2) ESCAPED BY/ENCLOSED BY field parameters are documented as
accepting CHAR argument, however SELECT .. INTO OUTFILE
silently ignored rests of multisymbol arguments.
For the symmetrical behavior with LOAD DATA INFILE the
server has been modified to fail with the same error:
ERROR 42000: Field separator argument is not what is
expected; check the manual
3) Current LOAD DATA INFILE recognizes field/line separators
"as is" without converting from client charset to data
file charset. So, it is supposed, that input file of
LOAD DATA INFILE consists of data in one charset and
separators in other charset. For the compatibility with
that [buggy] behaviour SELECT INTO OUTFILE implementation
has been saved "as is" too, but the new warning message
has been added:
Non-ASCII separator arguments are not fully supported
This message warns on field/line separators that contain
non-ASCII symbols.
client/mysqldump.c:
mysqldump has been updated to call SELECT ... INTO OUTFILE
statement with a charset from the --default-charset command
line parameter.
mysql-test/r/mysqldump.result:
Added test case for bug #30946.
mysql-test/r/outfile_loaddata.result:
Added test case for bug #30946.
mysql-test/t/mysqldump.test:
Added test case for bug #30946.
mysql-test/t/outfile_loaddata.test:
Added test case for bug #30946.
sql/field.cc:
String conversion code has been moved from check_string_copy_error()
to convert_to_printable() for reuse.
sql/share/errmsg.txt:
New WARN_NON_ASCII_SEPARATOR_NOT_IMPLEMENTED message has been added.
sql/sql_class.cc:
The select_export::prepare() method has been modified to:
1) raise the ER_WRONG_FIELD_TERMINATORS error on multisymbol
ENCLOSED BY/ESCAPED BY field arguments like LOAD DATA INFILE;
2) warn with a new WARN_NON_ASCII_SEPARATOR_NOT_IMPLEMENTED
message on non-ASCII field or line separators.
The select_export::send_data() merhod has been modified to
convert item data to output charset (see new SELECT INTO OUTFILE
syntax). By default the BINARY charset is used for backward
compatibility.
sql/sql_class.h:
The select_export::write_cs field added to keep output
charset.
sql/sql_load.cc:
mysql_load has been modified to warn on non-ASCII field or
line separators with a new WARN_NON_ASCII_SEPARATOR_NOT_IMPLEMENTED
message.
sql/sql_string.cc:
New global function has been added: convert_to_printable()
(common code has been moved from check_string_copy_error()).
sql/sql_string.h:
New String::is_ascii() method and new global convert_to_printable()
function have been added.
sql/sql_yacc.yy:
New syntax: added CHARACTER SET clause to the
SELECT ... INTO OUTFILE (to complement the same clause in
LOAD DATA INFILE). By default the BINARY charset is used for
backward compatibility.
We disallow the partitioning of a log table. You could however
partition a table first, and then point logging to it. This is
not only against the docs, it also crashes the server.
We catch this case now.
mysql-test/r/partition.result:
results for 40281
mysql-test/t/partition.test:
test for 40281: show that trying to log to partitioned table fails rather
to crash the server
sql/ha_partition.cc:
Signal that we no longer support logging to partitioned tables,
as per the docs.
sql/sql_partition.cc:
Some commands like "USE ..." have no select, yet we may try
to parse partition info after their execution if user set a
partitioned table as log target. This shouldn't lead to a
NULL-deref/crash.
A REPLACE in the MERGE engine is actually a REPLACE
into one (FIRST or LAST) of the underlying MyISAM
tables. So in effect the server works on the meta
data of the MERGE table, while the real insert happens
in the MyISAM table.
The MERGE table has no index, while MyISAM has a
unique index. When a REPLACE into a MERGE table (
and the REPLACE conflicts with a duplicate in a
child table) is done, we try to access the duplicate
key information for the MERGE table. This information
actually does not exist, hence this results in a crash.
The problem can be resolved by modifying the MERGE
engine to provide us the duplicate key information
directly, instead of just returning the MyISAM index
number as the error key. Then the SQL layer (or "the
server") does not try to access the key_info of the
MERGE table, which does not exist.
The current patch modifies the MERGE engine to provide
the position for a record where a unique key violation
occurs.
include/myisammrg.h:
Bug#45800 crash when replacing into a merge table and there is a duplicate
Add a member to the st_mymerge_info structure that will
store the duplicate key offset in the MERGE table. This
offset will be the sum of the record offset of the MyISAM
table within the MERGE table and the offset of the record
within the MyISAM table.
mysql-test/r/merge.result:
Bug#45800 crash when replacing into a merge table and there is a duplicate
Result file for the test case.
mysql-test/t/merge.test:
Bug#45800 crash when replacing into a merge table and there is a duplicate
Added test case for both REPLACE and INSERT...ON DUPLICATE UPDATE.
storage/myisammrg/ha_myisammrg.cc:
Bug#45800 crash when replacing into a merge table and there is a duplicate
The info method now will process the HA_STATUS_ERRKEY flag
and will return the index and the offset of the duplicate
key.
storage/myisammrg/ha_myisammrg.h:
Bug#45800 crash when replacing into a merge table and there is a duplicate
Set the HA_DUPLICATE_POS flag to indicate that the duplicate
key information is now available in the MERGE storage engine.
storage/myisammrg/myrg_info.c:
Bug#45800 crash when replacing into a merge table and there is a duplicate
We modify the myrg_status function to return the position of the
duplicate key. The duplicate key position in the MERGE table will
be the MyISAM file_offset and the offset within the MyISAM table
of the start position of the records.
an assertion in a debug build.
The reason is that the C API doesn't support multiple result sets for prepared
statements and attempting to execute a stored routine which returns multiple result
sets sometimes lead to a network error. The network error sets the diagnostic area
prematurely which later leads to the assert when an attempt is made to set a second
server state.
This patch fixes the issue by changing the scope of the error code returned by
sp_instr_stmt::execute() to include any error which happened during the execution.
To assure that Diagnostic_area::is_sent really mean that the message was sent all
network related functions are checked for return status.
libmysqld/lib_sql.cc:
* Changed prototype to return success/failure status on net_send_error_packet(),
net_send_ok(), net_send_eof(), write_eof_packet().
mysql-test/r/sp_notembedded.result:
* Added test case for bug 44521
mysql-test/t/sp_notembedded.test:
* Added test case for bug 44521
sql/protocol.cc:
* Changed prototype to return success/failure status on net_send_error_packet(),
net_send_ok(), net_send_eof(), write_eof_packet().
sql/protocol.h:
* Changed prototype to return success/failure status on net_send_error_packet(),
net_send_ok(), net_send_eof(), write_eof_packet().
sql/sp_head.cc:
* Changed prototype to return success/failure status on net_send_error_packet(),
net_send_ok(), net_send_eof(), write_eof_packet().
those keywords do nothing in 5.1 (they are meant for future versions, for example featuring the Maria engine)
so they are here removed from the syntax. Adding those keywords to future versions when needed is:
- WL#5034 "Add TRANSACTIONA=0|1 and PAGE_CHECKSUM=0|1 clauses to CREATE TABLE"
- WL#5037 "New ROW_FORMAT value for CREATE TABLE: PAGE"
mysql-test/r/create.result:
test that syntax is not accepted
mysql-test/t/create.test:
test that syntax is not accepted
sql/handler.cc:
remove ROW_FORMAT=PAGE
sql/handler.h:
Mark unused objects, but I don't remove them by fear of breaking any plugin which includes this file
(see also table.h)
sql/lex.h:
removing syntax
sql/sql_show.cc:
removing output of noise keywords in SHOW CREATE TABLE and INFORMATION_SCHEMA.TABLES
sql/sql_table.cc:
removing TRANSACTIONAL
sql/sql_yacc.yy:
removing syntax
sql/table.cc:
removing TRANSACTIONAL, PAGE_CHECKSUM. Their place in the frm file is not reclaimed,
for compatibility with older 5.1.
sql/table.h:
Mark unused objects, but I don't remove them by fear of breaking any plugin which includes this file
(and there are several engines which use the content TABLE_SHARE and thus rely on a certain binary
layout of this structure).
One of the tests introduced for this bug was failing
because of path size restriction in windows.
Moved the test case to a new test which is disabled under windows.
mysql-test/r/partition_not_embedded.result:
updated test results after removing a test case.
mysql-test/r/partition_rename_longfilename.result:
Test result for partition_rename_longfilename
mysql-test/t/partition_not_embedded.test:
Removed the test case which tests renaming partition table such that
the file name is 255 char long.
mysql-test/t/partition_rename_longfilename.test:
Test case to test renaming partition table such that
the file name is 255 char long.
Moved from partition_no_embedded.
In create_myisam_from_heap() mark all errors as fatal except
HA_ERR_RECORD_FILE_FULL for a HEAP table.
Not doing so could lead to problems, e.g. in a case when a
temporary MyISAM table gets overrun due to its MAX_ROWS limit
while executing INSERT/REPLACE IGNORE ... SELECT.
The SELECT execution was aborted, but the error was
converted to a warning due to IGNORE clause, so neither 'ok'
nor 'error' packet could be sent back to the client. This
condition led to hanging client when using 5.0 server, or
assertion failure in 5.1.
mysql-test/r/insert_select.result:
Added a test case for bug #46075.
mysql-test/t/insert_select.test:
Added a test case for bug #46075.
sql/sql_select.cc:
In create_myisam_from_heap() mark all errors as fatal except
HA_ERR_RECORD_FILE_FULL for a HEAP table.
Problem was that a failing rename just left the partitions at the state
it was at the failure.
Solution was to try to revert the started rename if a failure occured.
mysql-test/r/partition_not_embedded.result:
Bug#30102: Rename table does corrupt tables with partition files on failure
New result file
mysql-test/t/partition_not_embedded.test:
Bug#30102: Rename table does corrupt tables with partition files on failure
New test file
(list_files does not report the files in embedded)
sql/ha_partition.cc:
Bug#30102: Rename table does corrupt tables with partition files on failure
Better error handling for rename partitions (reverting the started rename
operation)
Different order of files for delete.
sql/handler.cc:
Bug#30102: Rename table does corrupt tables with partition files on failure
Tries to remove as many table files as possible
if the first delete succeeds.
When during the optimization an item is moved to the upper select
the item's context left unchanged. This caused wrong result in the
PS/SP mode.
The Item_ident::remove_dependence_processor now sets the context
of the select to which the item is moved to.
mysql-test/r/subselect.result:
The test case for the bug#46051 is adjusted.
mysql-test/t/subselect.test:
The test case for the bug#46051 is adjusted.
sql/item.cc:
Bug#46051: Incorrectly market field caused wrong result.
The Item_ident::remove_dependence_processor now sets the context
of the select to which the item is moved to.
In a subselect all fields from outer selects are marked as dependent on
selects they are belong to. In some cases optimizer substitutes it for an
equivalent expression. For example "a_field IN (SELECT outer_field)" is
substituted with "a_field = outer_field". As we moved the outer_field to the
upper select it's not really outer anymore. But it was left marked as outer.
If exists an index over a_field optimizer choose wrong execution plan and thus
return wrong result.
Now the Item_in_subselect::single_value_transformer function removes dependent
marking from fields when a subselect is optimized away.
mysql-test/r/subselect.result:
Added a test case for the bug#46051.
mysql-test/t/subselect.test:
Added a test case for the bug#46051.
sql/item_subselect.cc:
Bug#46051: Incorrectly market field caused wrong result.
Now the Item_in_subselect::single_value_transformer function removes dependent
marking from fields when a subselect is optimized away.
table
The MERGE table storage engine does not support the HA_CAN_SQL_HANDLE feature
and any attempt to open the merge table will fail with ER_ILLEGAL_HA.
After an error occurred the tables that was opened must be closed again
or they will be left in an inconsistent state. However, the assumption
made in the code for closing and register handler tables was that only
one table will be opened, and this is not true for MERGE tables which
will cause multiple tables to open.
The next time a SELECT operation was issued on the merge table it
caused the system to freeze.
This patch fixes this issue by making sure that all tables which
are opened also are closed in the event of an error.
mysql-test/r/merge.result:
Added test case for bug 45781
mysql-test/t/merge.test:
Added test case for bug 45781
sql/sql_handler.cc:
* mysql_ha_open() was never ment to open more than one table. If we encounter more tables, we should
close all tables related to the current substatement and raise an exception.
match against.
Server crashes when executing prepared statement with duplicating
MATCH() function calls in SELECT and ORDER BY expressions, e.g.:
SELECT MATCH(a) AGAINST('test') FROM t1 ORDER BY MATCH(a) AGAINST('test')
This query gets optimized by the server, so the value returned
by MATCH() from the SELECT list is reused for ORDER BY purposes.
To make this optimization server is comparing items from
SELECT and ORDER BY lists. We were getting server crash because
comparision function for MATCH() item is not intended to be called
at this point of execution.
In 5.0 and 5.1 this problem is workarounded by resetting MATCH()
item to the state as it was during PREPARE.
In 6.0 correct comparision function will be implemented and
duplicating MATCH() items from the ORDER BY list will be
optimized.
mysql-test/r/fulltext.result:
Updated with the test case for Bug#37740
mysql-test/t/fulltext.test:
A test case for Bug#37740.
sql/item_func.h:
True initialization of 'table' happens in ::fix_fields(). As
Item_func_match::eq() may be called before ::fix_fields(), it is
expected that 'table' is initialized to 0 when it is reused.
This is mostly affecting prepared statements, when the same item
doesn't get destroyed, but rather cleaned up and reused.
without error
When using quick access methods for searching rows in UPDATE or
DELETE there was no check if a fatal error was not already sent
to the client while evaluating the quick condition.
As a result a false OK (following the error) was sent to the
client and the error was thus transformed into a warning.
Fixed by checking for errors sent to the client during
SQL_SELECT::check_quick() and treating them as real errors.
Fixed a wrong test case in group_min_max.test
Fixed a wrong return code in mysql_update() and mysql_delete()
mysql-test/r/bug40113.result:
Bug #40013: test case
mysql-test/r/group_min_max.result:
Bug #40013: fixed a wrong test case
mysql-test/t/bug40113-master.opt:
Bug #40013: test case
mysql-test/t/bug40113.test:
Bug #40013: test case
mysql-test/t/group_min_max.test:
Bug #40013: fixed a wrong test case
sql/sql_delete.cc:
Bug #40113: check for errors evaluating the quick select
sql/sql_update.cc:
Bug #40113: check for errors evaluating the quick select
