firebirds
0fb3676c67
nightly update
2011-02-21 03:21:28 +00:00
dimitr
dbd1775fc1
Fixed CORE-3176: View with "subselect" column join table and not use index.
2011-02-20 13:25:14 +00:00
firebirds
68f88c546b
nightly update
2011-02-19 03:25:38 +00:00
dimitr
4b03ba59be
Restored the pre-v2.5 logic by the customer request.
2011-02-18 06:52:44 +00:00
dimitr
3447e70433
Fixed the main reason for CORE-3233: LIKE, STARTING and CONTAINING fail if second operand >= 32K. Thanks to Adriano.
2011-02-18 06:26:47 +00:00
robocop
b3d34b25c2
Make sure at least one operand is promoted to FB_UINT64 before doing the multiplication.
2011-02-18 04:37:12 +00:00
robocop
0c3d7f3c5d
Misc.
2011-02-18 04:35:36 +00:00
robocop
8ac0024ba1
Misc.
2011-02-18 04:33:25 +00:00
robocop
66546ce4c4
Misc.
2011-02-18 04:32:11 +00:00
robocop
d54b4a731f
Simplification.
2011-02-18 04:30:45 +00:00
firebirds
c916dc94e9
nightly update
2011-02-18 03:21:36 +00:00
dimitr
55f5ce3def
Fixed inability to execute LIKE/STARTING/etc if the second argument is a small (<32KB) blob and the first one is a blob of charset NONE. This is partially related to CORE-3233, see my comment there.
2011-02-17 17:25:47 +00:00
dimitr
f0075041fc
Fixed CORE-3353: Predicate (blob_field LIKE ?) describes the parameter as VARCHAR(30) rather than as BLOB.
...
Also, let's never transfer the CS_dynamic pseudo-charset to the client side.
2011-02-17 07:38:34 +00:00
dimitr
b3450346ca
Fixed CORE-3058: New generators are created with wrong value when more than 32K generators was previously created.
2011-02-17 06:00:50 +00:00
firebirds
17b5cc059b
nightly update
2011-02-16 03:21:56 +00:00
hvlad
0b3cac18ba
Fixed bug CORE-3351 : Windows client could put 10054 error messages into firebird.log at connection time
2011-02-15 10:07:18 +00:00
dimitr
e642bc0f42
Reworked my yesterday's fix to simplify it.
2011-02-15 08:38:32 +00:00
firebirds
35008d189e
nightly update
2011-02-15 03:22:57 +00:00
dimitr
1974ff619e
Refactored the shared counter to support multiple number spaces.
...
Changed att_id and tra_id to be generated sequentially (without caching), as agreed with Vlad privately.
Fixed CORE-2305: Make mon$statement_id value constant among monitoring snapshots.
2011-02-14 17:05:23 +00:00
hvlad
f1ddbd3b10
Fixed bug CORE-3349 : Constants isc_lock_conflict / isc_eds_statement not recognized in PSQL
2011-02-14 15:20:10 +00:00
dimitr
f2c890e1fc
Fixed CORE-3138: Internal error or crash occurs when accessing any MON$ table after altering its structure.
2011-02-14 11:10:58 +00:00
firebirds
3bb7d22126
nightly update
2011-02-14 03:21:51 +00:00
hvlad
eb5b4a564b
Correction
2011-02-13 19:14:13 +00:00
firebirds
06cb2d009f
nightly update
2011-02-10 03:27:16 +00:00
dimitr
1933d37b9f
Simplified fix for CORE-3340 (Error in autonomous transaction with empty exception handler: can insert duplicate values into PK/UK column) and its cousin CORE-3341 (Events might not be posted inside the autonomous transaction). It's actually the same bug with different visible effects.
2011-02-09 10:47:42 +00:00
firebirds
08fbcc4c74
nightly update
2011-02-08 03:22:55 +00:00
asfernandes
a5b72a367e
Backport simplified fix for CORE-3255 - The server could crash using views with GROUP BY.
2011-02-07 14:32:21 +00:00
firebirds
c4c77f3a42
nightly update
2011-02-05 03:21:27 +00:00
dimitr
3652f11f36
Fixed CORE-3335: Wrong results (internal wrapping occured) for the multi-byte blob SUBSTRING function and its boundary arguments.
2011-02-04 08:41:02 +00:00
firebirds
6106bc6e72
nightly update
2011-02-04 03:21:56 +00:00
dimitr
263d9532e2
Removed the supposedly wrong checkout (it was added in v2.0), as theoretically it can lead to race conditions in the page cache.
2011-02-03 08:00:48 +00:00
hvlad
6722a282e5
Fixed bug CORE-3327 : Thread pool in network server could create more threads than necessary
2011-02-02 14:55:15 +00:00
hvlad
4d0a3d3d67
Fixed bug CORE-3326 : Fast mutex could be left in locked state by dead process
2011-02-02 12:10:12 +00:00
hvlad
241df76bb8
Fixed bug CORE-3325 : At high load it is possible that new process could fail to map shared memory
2011-02-02 11:50:27 +00:00
hvlad
914a21e766
Improvement CORE-3324 : ftracemgr should flush its output from time to time
2011-02-02 11:36:42 +00:00
alexpeshkoff
1633598aa3
Backported fix for CORE-3329:Windows administrator gets RDB$ADMIN role when this is not expected
2011-02-02 11:31:38 +00:00
hvlad
b2b911fe6c
Improvement CORE-3323 : Ability to cancel waiting in lock manager
2011-02-02 10:27:48 +00:00
dimitr
cb34e0c6c9
Fixed CORE-3245: SUBSTRING on long blobs truncates result to 32767 if third argument not present.
2011-02-02 10:12:40 +00:00
dimitr
9755df067e
Backported Adriano's correction.
2011-02-02 08:15:27 +00:00
alexpeshkoff
d1df5354f6
Fixed deadlock caused by my commit for CORE-3194
2011-02-02 07:54:14 +00:00
firebirds
615665b646
nightly update
2011-02-02 03:24:52 +00:00
hvlad
3db1dd93c8
Warnings
2011-02-01 15:17:49 +00:00
hvlad
4b638d6e28
Misc
2011-02-01 15:05:30 +00:00
hvlad
b877fa6fab
Move VALIDATE_LOCK_TABLE from lock.h into lock.cpp : at allows to rebuild just lock.cpp if some need to disable validation of lock table in DEBUG build or enable validation in RELEASE build.
2011-02-01 15:01:10 +00:00
firebirds
65d73e0655
nightly update
2011-02-01 03:26:59 +00:00
hvlad
bbd043ffa3
Fixed bug CORE-3095 : Client receive event's with count equal to 1 despite of how many times EVENT was POSTed in same transaction
2011-01-31 10:32:17 +00:00
asfernandes
b07b7e6996
Backported fix for CORE-3320 - Some MERGE syntax can crash the server.
2011-01-30 14:31:10 +00:00
dimitr
071a98a3aa
Attempted to fix CORE-3314: Dependencies are not removed after dropping the procedure and the table it depends on in the same transaction.
...
The main goal of MET_clear_cache() is to release the cached requests, not to cleanup the metadata at all (and it's not done for triggers, BTW). So setting PRC_obsolete and keeping the object cached seems safe, it will be rescanned upon the next reference.
The fix is to be field tested by the customer.
2011-01-29 12:15:21 +00:00
dimitr
e0648e8078
These checks don't seem required for the new ODS 11 optimizer logic. But they may cause bad effects, see the second example in CORE-3312: Sub-optimal join plan when the slave table depends on the master one via the OR predicate.
...
Let's remove them and see whether any regressions arise.
2011-01-29 11:30:22 +00:00
alexpeshkoff
8af62682b2
Fixed CORE-3194: Number of connections to linux superclassic has limit equal 508
2011-01-28 16:52:06 +00:00
