firebirds
71017c810b
nightly update
2013-01-26 08:38:50 +00:00
dimitr
a9358db6c7
Slightly optimized fetches for internal EXECUTE STATEMENT.
2013-01-25 12:37:45 +00:00
firebirds
94233be876
nightly update
2013-01-25 09:51:21 +00:00
dimitr
56924b0a2d
Slightly improved the condition for the Adriano's backward compatibility workaround, this optimizes METD_move() into a plain memcpy() for strings with an explicitly declared charset.
2013-01-24 16:47:03 +00:00
dimitr
ad604b8862
This trivial optimization saves a few percents of performance when reading/writing DBKEYs from/to compatible local variables.
2013-01-24 09:29:46 +00:00
dimitr
216ef35338
Backported dbkey/recver optimization.
2013-01-24 07:10:29 +00:00
firebirds
084005c7e0
nightly update
2013-01-19 03:55:29 +00:00
dimitr
38ae66edb0
A little bit more defensive approach to the dbb handle validation. It also avoids the second handle lookup.
2013-01-17 16:32:48 +00:00
firebirds
4851c402b1
nightly update
2013-01-17 03:56:03 +00:00
dimitr
e0068e7fcd
Simplest fix for CORE-4038: Broken optimization for the stored dbkeys.
2013-01-16 17:02:23 +00:00
firebirds
c68dc7ea6d
nightly update
2013-01-11 03:52:10 +00:00
firebirds
a1b7a1f584
nightly update
2013-01-09 04:24:05 +00:00
paul_reeves
d948dfffde
Fix for CORE-4020
...
Instantiate cpl applet methods with LPARAM instead of long.
2013-01-08 13:47:47 +00:00
firebirds
a24d30f926
nightly update
2012-12-30 03:35:20 +00:00
alexpeshkoff
e713fdb4c5
Backported fix for CORE-3932: can create but cant delete user with username containing double quotes with gsec
2012-12-29 11:16:34 +00:00
firebirds
016c34fbc4
nightly update
2012-12-29 03:35:35 +00:00
alexpeshkoff
9215878883
Use same context for ASTs and async API calls
2012-12-28 16:19:29 +00:00
alexpeshkoff
4d5faef00b
Once I've got NULL object here, could not reproduce, but fix is trivial
2012-12-28 16:16:34 +00:00
alexpeshkoff
198dd90a22
Avoid use of already dropped attachment in system transaction
2012-12-28 12:40:56 +00:00
alexpeshkoff
506021360e
Validating attachment by proving its existence in the attachment list makes no sense after validation in handle's array
2012-12-28 10:31:45 +00:00
alexpeshkoff
b8a04dae7e
Explicit NULL check makes no sense when validating against array of known handles (pointers)
2012-12-28 10:22:30 +00:00
firebirds
524948c027
nightly update
2012-12-28 03:37:44 +00:00
alexpeshkoff
c98ccc6f4c
Fixed CORE-3944: SuperClassic crashes when running script that delete attachments and move database to offline
2012-12-27 13:26:31 +00:00
firebirds
562620dd7d
nightly update
2012-12-27 04:05:51 +00:00
dimitr
95047dd88f
Easy fix for the crash reported by Alex privately. More generic solution is desirable.
2012-12-26 10:27:30 +00:00
firebirds
4a9dffec26
nightly update
2012-12-25 03:34:40 +00:00
dimitr
654b668434
Favor quick search over quick add/remove. This is more important in our code, especially in the debug build.
2012-12-24 14:26:02 +00:00
firebirds
72ede8a0cc
nightly update
2012-12-24 03:34:01 +00:00
dimitr
88635bfc4a
Corrected the check_database() logic.
2012-12-23 06:52:07 +00:00
firebirds
5f12ef1d9b
nightly update
2012-12-23 03:37:43 +00:00
dimitr
8ef43e01dc
Moved the statement cancellation logic from the transaction level to the attachment level. It reuses the already existing ATT_cancel_raise handling and simplifies the code.
...
Also, encapsulated the shutdown/cancellation signaling code.
2012-12-22 16:56:36 +00:00
dimitr
4d8273fe11
Fixed the missing dbb flags, especially DBB_security_db.
2012-12-22 15:02:18 +00:00
dimitr
e432c56660
We don't need the new flags for these tasks.
2012-12-22 12:55:34 +00:00
dimitr
916828a19e
Fixed the crash that appears after PublicHandle became an inherited class (thus containing vtable).
2012-12-22 12:30:13 +00:00
firebirds
ae650b596e
nightly update
2012-12-21 03:51:39 +00:00
alexpeshkoff
4c3978ea22
Backported fix for CORE-3995: Slight problem with GBAK parameter checking
2012-12-20 15:28:25 +00:00
firebirds
af103214d5
nightly update
2012-12-19 09:24:18 +00:00
dimitr
72fd24bf66
Slightly refactored the error handling.
2012-12-18 19:35:51 +00:00
dimitr
b2a6cadd84
Attempted to fix the user savepoint logic I've accidentally broken in a recent commit.
2012-12-18 13:30:16 +00:00
dimitr
1a70da9ff9
Fixed the Windows build.
2012-12-18 11:20:56 +00:00
firebirds
7435a2de81
nightly update
2012-12-15 04:00:52 +00:00
alexpeshkoff
74938ad13e
Fixed CORE-3935 and CORE-3993
2012-12-14 17:45:38 +00:00
firebirds
b37cf969a9
nightly update
2012-12-14 04:05:50 +00:00
dimitr
7a6b9f8966
1) Encapsulated the remaining external looper calls inside a SEH block. It allows to catch async exceptions coming while result set is being fetched. No performance penalty has been noticed.
...
2) Slightly refactored the looper call chain to avoid code duplication.
3) Fixed CORE-4010: Fetch cannot be subsequently interrupted via DELETE FROM MON$STATEMENTS.
2012-12-13 18:00:35 +00:00
firebirds
b422a1736e
nightly update
2012-12-13 04:08:58 +00:00
dimitr
79c9da1c05
Misc.
2012-12-12 08:31:00 +00:00
dimitr
9dbf5dce69
Yet another complementary bugfix for CORE-4004.
2012-12-12 07:46:09 +00:00
firebirds
a37e04ec4a
nightly update
2012-12-12 04:09:08 +00:00
dimitr
c848fc69a8
Avoid garbage to be written to the shadow's header page. It does not really hurt, but anyway.
2012-12-10 05:10:45 +00:00
firebirds
5a1ce71dc7
nightly update
2012-12-07 04:14:40 +00:00
dimitr
ae48c19a87
Another bunch of changes for CORE-4004: Sometimes long-running operations cannot be interrupted by asynchronous shutdown / cancellation requests. The looper nodes should never ignore req_unwind and transform it to something else.
...
Review and testing would be appreciated, especially in regard to EXE_assignment.
2012-12-06 11:43:48 +00:00
firebirds
5ae85cea08
nightly update
2012-12-05 04:06:14 +00:00
hvlad
543b50779f
Alternative fix for CORE-3034 (Bugcheсk 300 (can't find shared latch)), see also CORE-4004 (Sometimes long-running operations cannot be interrupted by asynchronous shutdown / cancellation requests)
2012-12-04 10:28:01 +00:00
firebirds
9062a43c89
nightly update
2012-12-04 04:47:17 +00:00
dimitr
65178c062a
Yet another complementary bugfix for CORE-4004 that avoids an index root page remaining locked after cooperative GC has been terminated by user.
2012-12-03 11:04:34 +00:00
dimitr
8d7cce247d
Complementary patch to the prior one. It avoids unnecessary operations (system table lookups) if we don't need index or constraint names to be reported in the error message.
2012-12-03 10:47:11 +00:00
dimitr
1c043b0e46
Fixed one part of CORE-4004: Sometimes long-running operations cannot be interrupted by asynchronous shutdown / cancellation requests. It's definitely not a good idea to ignore errors coming from inside BTR_key().
2012-12-03 10:44:39 +00:00
dimitr
f760ee4c58
Fixed the partially disabled interruptions, accidentally broken with CORE-3323 in v2.5.1.
2012-12-03 10:33:11 +00:00
firebirds
0499ea88d4
nightly update
2012-12-01 04:06:38 +00:00
hvlad
3e0920857c
Fixed bug CORE-4002 : Error message "index unexpectedly deleted" in database trigger on commit transaction
2012-11-30 08:12:43 +00:00
firebirds
d5decaf997
nightly update
2012-11-30 04:51:52 +00:00
alexpeshkoff
83ba186572
avoid const warnings
2012-11-29 14:37:33 +00:00
firebirds
faf9d908c8
nightly update
2012-11-28 05:08:49 +00:00
dimitr
bebc026b62
Backported CORE-1295: Bad optimization of queries with DB_KEY.
2012-11-27 17:37:14 +00:00
hvlad
8433046836
Backport change from trunk.
...
It fixes bug CORE-3998 : Parametrized execute statement fails
2012-11-27 11:12:50 +00:00
hvlad
12e97c4a07
Make BPB correctly
2012-11-27 11:10:37 +00:00
firebirds
0ef6458bd3
nightly update
2012-11-23 04:28:21 +00:00
dimitr
87c65731bc
Implemented CORE-3994: Improve the limbo transactions scan at the end of the sweep.
2012-11-22 14:05:37 +00:00
dimitr
5eaecbbfc0
Backported (from trunk) more reliable AST handling during the database cleanup.
2012-11-22 12:45:45 +00:00
dimitr
9f60a27a7f
Simplest possible fix for CORE-3989: Bad performance / slow response when many concurrent sorts are executed. It's mostly targeted at SC/CS architectures, as one can hardly have really concurrent sorts in v2.x SS. FB3 is likely to have a more complex solution.
2012-11-22 09:46:29 +00:00
firebirds
b66fb567fd
nightly update
2012-11-21 04:21:01 +00:00
hvlad
fd40f95751
Make sure relation's flags are loaded from disk before analyze them
2012-11-20 12:23:17 +00:00
dimitr
88bdcebaea
Additional fixes for CORE-3981 (Sub-optimal predicate checking while selecting from a view) that also resolve the regression CORE-3986 (GSEC couldn't change user password after fix for CORE-3981).
2012-11-20 10:15:35 +00:00
hvlad
18f91be9c6
FIxed bug CORE-3988 : Engine crashed when trace or audit is active
2012-11-20 09:32:55 +00:00
firebirds
89ccac15ae
nightly update
2012-11-20 04:17:55 +00:00
hvlad
c08c25a723
Make previous fix for CORE-3969 platform independent.
2012-11-19 08:41:50 +00:00
firebirds
4a59ad94ce
nightly update
2012-11-19 04:32:25 +00:00
dimitr
ef258b7712
Fixed CORE-3979: Server crashes while unwinding changes in an autonomous transaction.
2012-11-17 15:17:15 +00:00
firebirds
17b5568478
nightly update
2012-11-15 04:11:36 +00:00
dimitr
cf1212cc09
A more generic solution for CORE-1245: Incorrect column values with outer joins and views. It also fixes CORE-3981: Sub-optimal predicate checking while selecting from a view.
2012-11-14 08:14:35 +00:00
dimitr
ef51f56de4
Fixed CORE-3978: Invalid transaction counters may be reported in firebird.log during the sweep.
2012-11-13 09:38:37 +00:00
firebirds
05b31dcf16
nightly update
2012-11-12 04:39:35 +00:00
dimitr
3c9e272829
Fixed CORE-3977: DELETE FROM MON$STATEMENTS does not interrupt a longish fetch.
2012-11-11 15:26:39 +00:00
robocop
0ef0030c18
Misc.
2012-11-11 06:25:50 +00:00
firebirds
630441efd3
nightly update
2012-11-10 04:46:29 +00:00
firebirds
f10d591989
nightly update
2012-11-09 05:01:02 +00:00
firebirds
a19a912aeb
nightly update
2012-11-07 04:42:37 +00:00
firebirds
2bca9f8b4e
nightly update
2012-11-02 05:46:03 +00:00
hvlad
e1481ec095
Fixed bug CORE-3969 : Memory leak when doing a lot of iterations FBSVCMGR with action_trace_start, _list and _stop
2012-11-01 12:03:18 +00:00
alexpeshkoff
8ddf02f4d8
This should finally fix CORE-3786: Firebird 2.5.1 Hangs on MacOSX 10.7 (Lion) on DB create after reboot
2012-11-01 09:09:31 +00:00
firebirds
249e3d546a
nightly update
2012-11-01 04:38:59 +00:00
dimitr
183c3f9876
Avoid a sweep pass for temporary tables, as the sweeper always sees those tables as empty, but even setting them up as empty has a non-zero cost.
2012-10-31 14:16:00 +00:00
firebirds
7a162d89f2
nightly update
2012-10-15 03:40:58 +00:00
dimitr
e315f5a4a8
Fixed the backporting bug introduced while fixing CORE-3902.
2012-10-14 17:39:47 +00:00
firebirds
822c3262f7
nightly update
2012-10-13 03:45:37 +00:00
asfernandes
36dcd8e561
Backport fix for CORE-3946 - UNICODE collations does not work with ICU 49.
2012-10-12 21:20:30 +00:00
firebirds
c31eeef318
nightly update
2012-10-09 04:19:23 +00:00
hvlad
711ed0426d
Fix compilation
2012-10-08 17:57:47 +00:00
firebirds
1413f70a3c
nightly update
2012-09-30 03:42:38 +00:00
robocop
62dafac5f2
Misc.
2012-09-30 00:23:52 +00:00
