This actually, fix for the patch for bug-27354. The problem with
the patch was that Item_func_sp::used_tables() was updated, but
Item_func_sp::const_item() was not. So, for Item_func_sp, we had
the following inconsistency:
- used_tables() returned RAND_TABLE, which means that the item
can produce "random" results;
- but const_item() returned TRUE, which means that the item is
a constant one.
The fix is to change Item_func_sp::const_item() behaviour: it must
return TRUE (an item is a constant one) only if a stored function
is deterministic and each of its arguments (if any) is a constant
item.
mysql-test/r/sp.result:
Update result file.
mysql-test/t/sp.test:
Add test cases for BUG-31035, BUG-31191, BUG-31226
sql/item_func.cc:
Make Item::const_item() be consistent with Item::used_tables().
Adding support for correct handling of pthread_mutex_trylock() on Win32
systems as well as when using the safe mutexes.
include/my_pthread.h:
Adding win_pthread_mutex_trylock() function as wrapper function for
Windows implementation of pthread_mutex_trylock().
Adding flag to safe_mutex_lock() that is shall not abort if the mutex
is already locked and instead return EBUSY since this is the POSIX
behaviour.
mysys/my_winthread.c:
Added Win32 wrappper function to handle pthread_mutex_trylock().
mysys/thr_mutex.c:
Added parameter 'try_lock' to safe_mutex_lock() to allow it to do double-
duty as a pthread_mutex_trylock() in addition to working as
pthread_mutex_lock().
The code needs to return EBUSY instead of throwing an error in the event
that the mutex is re-locked (we do not have recursive mutexes), but it
also requires some special handling to avoid race conditions when
calling the real pthread_mutex_{lock,trylock}().
The problem was that aborted_threads variable was updated
twice when a client connection had been aborted.
The fix is to refactor a code to have aborted_threads updated
only in one place.
sql/mysql_priv.h:
Make do_command() a private function.
sql/sql_parse.cc:
1. Make do_command() a private function;
2. Update aborted_threads in the only one place.
into weblab.(none):/home/marcsql/TREE/mysql-5.0-rt-merge
mysql-test/r/sp.result:
Auto merged
mysql-test/t/mysql.test:
Auto merged
mysql-test/t/sp.test:
Auto merged
Invaldating a subset of a sufficiently large query cache can take a long time.
During this time the server is efficiently frozen and no other operation can
be executed. This patch addresses this problem by setting a time limit on
how long time a dictionary access request can take before giving up on the
attempt. This patch does not work for query cache invalidations issued by
DROP, ALTER or RENAME TABLE operations.
sql/sql_cache.cc:
Changed mutex locking to a timed spinn lock. If access to query cache dictionary
takes "a long time" (currently more than 0.1 seconds) the system fall backs on
ordinary statement execution.
The problem is that a SELECT on one thread is blocked by INSERT ... ON
DUPLICATE KEY UPDATE on another thread even when low_priority_updates is
activated.
The solution is to possibly downgrade the lock type to the setting of
low_priority_updates if the INSERT cannot be concurrent.
sql/sql_insert.cc:
Possibly downgrade lock type to the the setting of low_priority_updates if
if the INSERT cannot be concurrent.
Problem:
In cases when a client-side macro appears inside a server-side comment, the add_line() function in mysql.cc discarded all characters until the next delimiter to remove macro arguments from the query string. This resulted in broken queries being sent to the server when the next delimiter character appeared past the comment's boundaries, because the comment closing sequence ('*/') was discarded.
Fix:
If a client-side macro appears inside a server-side comment, discard all characters in the comment after the macro (that is, until the end of the comment rather than the next delimiter).
This is a minimal fix to allow only simple cases used by the mysqlbinlog utility. Limitations that are worth documenting:
- Nested server-side and/or client-side comments are not supported by mysql.cc
- Using client-side macros in multi-line server-side comments is not supported
- All characters after a client-side macro in a server-side comment will be omitted from the query string (and thus, will not be sent to server).
client/mysql.cc:
If a client-side macro appears inside a server-side comment, discard all characters in the comment after the macro.
mysql-test/r/mysql.result:
Added a test case for bug #30164.
mysql-test/t/mysql.test:
Added a test case for bug #30164.
comments)
Before this fix, the server would accept queries that contained comments,
even when the comments were not properly closed with a '*' '/' marker.
For example,
select 1 /* + 2 <EOF>
would be accepted as
select 1 /* + 2 */ <EOF>
and executed as
select 1
With this fix, the server now rejects queries with unclosed comments
as syntax errors.
Both regular comments ('/' '*') and special comments ('/' '*' '!') must be
closed with '*' '/' to be parsed correctly.
mysql-test/r/comments.result:
Unbalanced comments are a syntax error.
mysql-test/t/comments.test:
Unbalanced comments are a syntax error.
sql/sql_lex.cc:
Unbalanced comments are a syntax error.
seems to be converted as varbinary.
The bug has been already fixed. This CS just adds a test case for it.
mysql-test/r/sp.result:
Update result file.
mysql-test/t/sp.test:
Test case for BUG#13675.
mysql-test/lib/mtr_misc.pl:
Add function 'mtr_rmtree' it will try 'rmtree' and if that fails (most likely
due to permission problems we will fun File::find to chmod all files and dirs
to 0777 and then delete.
mysql-test/mysql-test-run.pl:
Use 'mtr_rmtree' in favour of 'rmtree'
This is a performance bug, affecting in particular the bison generated code
for the parser.
Prior to this fix, the grammar used a long chain of reduces to parse an
expression, like:
bit_expr -> bit_term
bit_term -> bit_factor
bit_factor -> value_expr
value_expr -> term
term -> factor
etc
This chain of reduces cause the internal state automaton in the generated
parser to execute more state transitions and more reduces, so that the
generated MySQLParse() function would spend a lot of time looping to execute
all the grammar reductions.
With this patch, the grammar has been reorganized so that rules are more
"flat", limiting the depth of reduces needed to parse <expr>.
Tests have been written to enforce that relative priorities and properties
of operators have not changed while changing the grammar.
See the bug report for performance data.
mysql-test/r/parser_precedence.result:
Improved test coverage for operator precedence
mysql-test/t/parser_precedence.test:
Improved test coverage for operator precedence
sql/sql_yacc.yy:
Simplified the grammar to improve performances
Fix typo, "next_con" -> "con"
client/mysqltest.c:
Reuse "find_connection_by_name" both from "do_close_connection" and "do_send_quit"
Adjust alignment of comment
Fix typo in "do_close_connection", it used the global variable "next_con" instead
of local variable "con"
to perform this analyzis)
client/mysqltest.c:
Add function 'show_query' and use it to output some debug queries when
"sync_with_master" has failed.
mysql-test/mysql-test-run.pl:
Move "analyze_testcase_failure" to mysqltest
- Merge sslaccept and sslconnect.
- Atomically "reset" vio to VIO_TYPE_SSL when the SSL connection has
succeeded, this avoids having to revert anything and thus protects
against "close_active_vio" in the middle.
- Add some variance to the testcase
mysql-test/t/rpl_ssl.test:
Add some variance by running two selects before stopping the slave
Check that number of records in t1 are equal on master and slave
vio/viossl.c:
Rewrite sslconnect and sslaccept to automically "reset" the vio
to VIO_TYPE_SSL. Also use the fd from 'SSL_get_fd' to avoid
setting vio->sd to -1, that previously occured when "close_active_vio"
was called during connect/accept.
Merge the two function since they were exactly the same except for one line.
Update the DBUG printouts to be generic(i.e use peer instead of client/server).
If, after the tables are locked, one of the conditions to read from a
HANDLER table is not met, the handler code wrongly jumps to a error path
that won't unlock the tables.
The user-visible effect is that after a error in a handler read command,
all subsequent handler operations on the same table will hang.
The fix is simply to correct the code to jump to the (same) error path that
unlocks the tables.
mysql-test/r/handler.result:
Bug#30632 test case result
mysql-test/t/handler.test:
Bug#30632 test case
sql/sql_handler.cc:
Always unlock the internal and external table level locks if any of the conditions
(including errors) to read from a HANDLER table are not met.
The problem from a user's perspective: user creates table A, and then tries
to CREATE TABLE a SELECT from A - and this causes a deadlock error, a hang,
or fails with a debug assert, but only if the storage engine is InnoDB.
The origin of the problem: InnoDB uses case-insensitive collation
(system_charset_info) when looking up the internal table share, thus returning
the same share for 'a' and 'A'.
Cause of the user-visible behavior: since the same share is returned to SQL
locking subsystem, it assumes that the same table is first locked (within the
same session) for WRITE, and then for READ, and returns a deadlock error.
However, the code is wrong in not properly cleaning up upon an error, leaving
external locks in place, which leads to assertion failures and hangs.
Fix that has been implemented: the SQL layer should properly propagate the
deadlock error, cleaning up and freeing all resources.
Further work towards a more complete solution: InnoDB should not use case
insensitive collation for table share hash if table names on disk honor the case.
mysql-test/r/innodb-deadlock.result:
Bug#25164 test case result
mysql-test/t/innodb-deadlock.test:
Bug#25164 test case. The CREATE TABLE may fail depending on the character set
of the system and filesystem, but it should never hang.
sql/lock.cc:
Unlock the storage engine "external" table level locks, if the MySQL thr_lock
locking subsystem detects a deadlock error.
- "mysql" and "mysqlcheck" should not read defaults file
client/mysql_upgrade.c:
Instruct "mysql" and "mysqlcheck" that is invoked by "mysql_upgrade" not
to read defaults file, they should get all the parameters they need from
mysql_upgrade(that read the default file)
- Chop off .libs/ part of path if running in non installed builddir
using libtool
client/mysql_upgrade.c:
Chop off .libs part of path to avoid executing "non relinked" binaries
that would use the system installed dynamic libraries instead of the
newly built ones.
1) Ensure "init_db.sql" and "test_db-sql" really get built.
2) Ensure the "*.def" files with NetWare linker options get distributed to the proper directories.
netware/BUILD/compile-netware-END:
Ensure the "*.def" files are built for NetWare.
This is a backport of a 5.1 fix which may not be needed in 5.0 but cannot do any harm:
the general "link_sources" step might fall victim to a cleanup which would be fatal
just for NetWare, because of problems in the ordering of SUBDIR entries.
netware/Makefile.am:
1) The scripts "init_db.sql" and "test_db.sql" must be built in the NetWare phase.
2) Use "basename", not sed.
