SQLite

Timeline
Login

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

100 ancestors of 5727562b75edf251

2016-01-12
19:37
Add the ".vfslist" command to the command-line shell. (check-in: 5727562b user: drh tags: trunk)
14:10
Fix a harmless use of an uninitialized variable following system errors in the multi-threaded sorter. (check-in: 4bb2c1df user: drh tags: trunk)
04:08
Remove an unused field from the MemPage object. (check-in: 5fc0a4c1 user: drh tags: trunk)
02:28
Make sure all bytes of the WAL file are initialized even if the checksums are not being computed until commit due to WAL overwrite. (check-in: 1f44a9bd user: drh tags: trunk)
02:00
Fix an uninitialized field in the Select object when SELECTTRACE is enabled. (check-in: fc628516 user: drh tags: trunk)
00:37
Various #ifdef enhancements for improved VxWorks support. (check-in: 75cd41ff user: drh tags: trunk)
2016-01-11
22:58
Performance optimization in the sqlite3VdbeChangeP4() routine of the code generator. (check-in: 28bd8d5f user: drh tags: trunk)
22:43
Fix duplicate test number in 'waloverwrite'. (check-in: f5fb8268 user: mistachkin tags: trunk)
18:05
Performance optimization on the Column opcode. (check-in: ecc98bef user: drh tags: trunk)
14:19
Performance optimization in sqlite3WalkExpr(). (check-in: 2f8bd5fa user: drh tags: trunk)
12:52
Increase the version number to 3.11.0 due to the WAL overwrite enhancement. (check-in: 8e807bfa user: drh tags: trunk)
12:13
If a single page is written to the wal file more than once, instead of appending the second and subsequent copy to the wal file, overwrite the first. Update: See the important bug fix at [f694e60a]! (check-in: d493d4f1 user: dan tags: trunk)
08:12
Add a test case to waloverwrite.test to check that savepoint rollback does not cause a problem. (Closed-Leaf check-in: 87dae56c user: dan tags: wal-overwrite-frames)
03:48
Slightly smaller and faster LIKE/GLOB comparison implementation. (check-in: 5dddcc78 user: drh tags: trunk)
00:52
Improved testability. (check-in: e83d3a2a user: drh tags: wal-overwrite-frames)
2016-01-09
23:55
All WAL frame overwrites even if there are active savepoints. This is safe because a ROLLBACK TO will cause all reverted pages to be rewritten to the WAL file prior to COMMIT. (check-in: 99b31a6b user: drh tags: wal-overwrite-frames)
18:57
If it is known that checksums will be recalculated on transaction commit, skip calculating checksums when appending frames to the wal file. When recalculating checksums, recalculate them starting with the first overwritten frame - not the first frame in the transaction. (check-in: 16b34f25 user: dan tags: wal-overwrite-frames)
16:39
If a single page is written to the wal file more than once, have each subsequent copy overwrite the original frame. (check-in: 5d113aef user: dan tags: wal-overwrite-frames)
2016-01-08
22:31
Typo fix in the previous commit. (check-in: 52c16603 user: drh tags: trunk)
22:18
More #ifdef fixes for the unix VFS. (check-in: 122c111e user: drh tags: trunk)
19:34
Fix the build for SQLITE_OMIT_SHARED_CACHE, which was broken, possibly by check-in [2d96aeba]. (check-in: 3392f8fa user: drh tags: trunk)
19:17
Fix a problem with #ifdefs on the system calls in the unix VFS that causes problems when compiling with SQLITE_OMIT_WAL. (check-in: ceceea4c user: drh tags: trunk)
17:28
Remove an obsolete #ifdef related to iOS support. (check-in: e9a51d2a user: drh tags: trunk)
12:46
Reduce the number of calls to memset() in sqlite3DbMallocZero(). (check-in: cede2bf8 user: drh tags: trunk)
08:35
Fix test file vtabH.test so that it works with SQLITE_ENABLE_ICU builds. (check-in: 19c1f448 user: dan tags: trunk)
2016-01-07
21:12
Omit performance tracking from the memsys5 memory allocator if neither SQLITE_DEBUG nor SQLITE_TEST are defined. (check-in: af5c7714 user: drh tags: trunk)
17:19
Fix to cause BtShared.db to be set correctly on shared-cache connections in SQLITE_THREADSAFE=0 builds. Added assert()s to verify the correct setting of BtShared.db. (check-in: 2d96aeba user: drh tags: trunk)
17:09
Small performance and size optimization spotted while working on the shared-cache problem. (check-in: 828958ff user: drh tags: trunk)
16:43
Add some assert() statements to verify, where possible, that BtShared.db is set correctly. (Closed-Leaf check-in: 359277e0 user: dan tags: shared-cache-fix)
16:06
Have the vdbe layer call sqlite3BtreeEnter() on all b-trees in use from within sqlite3VdbeExec() even in SQLITE_THREADSAFE=0 builds. This ensures that BtShared.db is set correctly. (check-in: d0214602 user: dan tags: shared-cache-fix)
02:06
Add the "sqlite3 -sourceid" command in the TCL interface. Use that command and others to output additional information about the test configuration at the conclusion of a TCL test. (check-in: 4f80440b user: drh tags: trunk)
2016-01-06
15:54
Disable the --limit-mem option on fuzzcheck unless the SQLITE_ENABLE_MEMSYS5 or SQLITE_ENABLE_MEMSYS3 compile-time options are used. (check-in: 55a11fd6 user: drh tags: trunk)
11:01
Version 3.10.0 (check-in: fd0a50f0 user: drh tags: trunk, release, version-3.10.0)
2016-01-05
03:39
Always use -DSQLITE_ENABLE_MEMSYS5 for fuzzcheck so that the --limit-mem option will work. (check-in: a8556977 user: drh tags: trunk)
03:39
Another attempt to get reuse of excess opcode array space working correctly on all architectures and platforms. (check-in: 2f858374 user: drh tags: trunk)
01:48
Permit the 'test_fs' test module to be compiled and used on MinGW. (check-in: ac27f38e user: mistachkin tags: trunk)
2016-01-04
23:43
Fix a corner case in the opcode-array reuse logic where the number of bytes of reusable space might be computed to be a negative number, due to unusual system alignment restrictions and rounding error. (check-in: 1aa53014 user: drh tags: trunk)
13:06
Fix the conflict2.test module which was broken by check-in [e30062e9f6c]. (check-in: b779ca8a user: drh tags: trunk)
12:07
A new approach to very large sqlite3_mprintf() strings: Back out the [d655a665] check-in and instead make the size fields in StrAccum unsigned. Strings generated by sqlite3_mprintf() can now be as large as 2^31-1 bytes. (check-in: 7adb789f user: drh tags: trunk)
04:44
Improved comment on the vtabH-3.1 test. (check-in: 22abe080 user: drh tags: trunk)
04:41
Adjust the vtabH.test module to ignore files in the root directory of Windows machines that contain the "$" character. (check-in: 73e33550 user: drh tags: trunk)
04:28
Limit the length of sqlite3_mprintf() output to 2^30 bytes, even if SQLITE_MAX_LENGTH is set larger at compile-time. (check-in: d655a665 user: drh tags: trunk)
03:48
Fix a harmless compiler warning. (check-in: 4b404a99 user: drh tags: trunk)
01:08
Add releasetest.tcl cases for SQLITE_LIKE_DOESNT_MATCH_BLOB and for SQLITE_ENABLE_CURSOR_HINTS. (check-in: f300c35e user: drh tags: trunk)
2016-01-03
18:07
Bug fix: the conflict resolution behavior for the implied NOT NULL constraint on the PRIMARY KEY of a WITHOUT ROWID table should be ABORT. (check-in: e30062e9 user: drh tags: trunk)
11:27
Fix a premature reference to utf8_printf in shell.c. (check-in: 76f5efa6 user: drh tags: trunk)
11:17
Add the number of cores used and the name of the host computer to the summary at the end of releasetest.tcl. (check-in: cd3ae0f1 user: drh tags: trunk)
2016-01-02
05:00
Fix a (harmless) typo in an #ifdef for SQLITE_ENABLE_CURSOR_HINTS (check-in: 2047d2de user: drh tags: trunk)
2016-01-01
16:42
Remember the size of the Vdbe.aOp[] array in bytes, to avoid unnecessary calls to sqlite3_msize(). (check-in: 3e852804 user: drh tags: trunk)
16:26
Avoid misaligned memory allocations on Sparc in sqlite3VdbeMakeReady(). This fixes a problem introduced by check-in [a9e819082b]. (check-in: a304e346 user: drh tags: trunk)
03:37
Fix an assert() that might be false for a corrupt database. (check-in: 68360cd2 user: drh tags: trunk)
00:15
Remove unreachable branches from the new JSON aggregate functions. (check-in: 6e30f8c6 user: drh tags: trunk)
2015-12-31
22:29
Fix the ieee754.test module so that it works both on systems that report infinity as "inf" and as "Inf". (check-in: d41d4d73 user: drh tags: trunk)
19:06
In the shell, use utf8_printf() when any string formatting is present and raw_printf() otherwise. (check-in: 8cf01fff user: mistachkin tags: trunk)
19:04
Fix harmless compiler warning seen with MSVC. (check-in: 98ec778d user: mistachkin tags: trunk)
15:34
Fix harmless compiler warnings associated with SQLITE_ENABLE_HIDDEN_COLUMNS. (check-in: 9c392c10 user: drh tags: trunk)
04:34
Small size and performance optimization to the VDBE comparison opcodes. (check-in: 7a0b9413 user: drh tags: trunk)
2015-12-30
18:18
Reduce the size of the CellInfo object from 32 to 24 bytes on 64-bit machines, for a small performance increase and library size reduction. (check-in: 6a4cfc7a user: drh tags: trunk)
17:03
Enhance the command-line shell so that it can handle MBCS characters on input and output. (check-in: 3d81dfe3 user: drh tags: trunk)
16:51
Changes to the way that the default BINARY collating sequence is recorded result in a slightly smaller and slightly faster executable. More work could be done to make this cleaner. (check-in: 2081d757 user: drh tags: trunk)
15:18
Simplification to the xfer-optimization logic. (check-in: f35ba018 user: drh tags: trunk)
14:06
Remove unnecessary tests from the LIKE pattern matcher. Slightly faster and slightly smaller, and it should also now works with EBCDIC. (check-in: 0a99a8c4 user: drh tags: trunk)
13:36
Enhance the command-line shell to handle MBCS characters on input and output. (Closed-Leaf check-in: a0a08b8c user: drh tags: mbcs-shell)
01:07
Add the json_group_array() and json_group_object() aggregate functions to the JSON1 extension. (check-in: 7f386a93 user: drh tags: trunk)
2015-12-24
14:53
Avoid a harmless compiler warning on systems where the byteorder cannot be determined at compile-time. (check-in: 7c7b7f26 user: drh tags: trunk)
2015-12-23
16:42
Fix some harmless gcc compiler warnings. Mostly in fts5, but also two in the core code. (check-in: 5d44d4a6 user: dan tags: trunk)
10:54
Enhance the 'utc' modifier on date/time functions so that if the LHS is already known to be in UTC, the modifier becomes a no-op. This is not an incompatibility because the behavior is documented as "undefined" in that scenario. (check-in: b910a3d5 user: drh tags: trunk)
2015-12-21
15:22
Ensure that the Expr objects that describe indexed expressions are not modified by code generation. Fix for an assert() problem found by Jon Metzman using AFL. (check-in: 34073ce8 user: drh tags: trunk)
2015-12-18
16:29
Micro-optimizations and comment fixes on the mem5.c memory allocator module. (check-in: 8bf5e056 user: drh tags: trunk)
03:59
Reduce the size of the CellInfo object from 32 to 24 bytes on 64-bit machines. (Closed-Leaf check-in: 78507154 user: drh tags: optimize-cellinfo)
2015-12-17
14:18
Fix the spellfix1_scriptcode() function to ignore whitespace and punctuation, and to recognize hebrew and arabic scripts. (check-in: 7adfa4a5 user: drh tags: trunk)
13:28
Fixes for harmless compiler warnings. (check-in: 85ebd46c user: drh tags: trunk)
2015-12-16
23:30
Fix even more harmless compiler warnings. (Closed-Leaf check-in: 1d0e6aa1 user: mistachkin tags: msvcWarn)
22:06
Fix some more harmless compiler warnings. (check-in: 18f531e1 user: mistachkin tags: msvcWarn)
21:09
Fix compiler warning seen with MSVC. (check-in: 18a4c054 user: mistachkin tags: msvcWarn)
13:07
Make greater use of the SQLITE_WITHIN macro. (check-in: f2fc3aed user: drh tags: trunk)
2015-12-15
19:32
Update the mkautoconfamal.sh script to make the building of the sqlite3-autoconf-VERSION.tar.gz package more orthodox. (check-in: 1928f894 user: dan tags: trunk)
13:50
Add the SQLITE_OMIT_PARSER_TRACE compile-time option. (check-in: 2fda43e6 user: drh tags: trunk)
2015-12-14
19:42
Add a missing memAboutToChange() macro to vdbe.c, the lack of which was causing an assert() to fail incorrectly. Problem found by libfuzzer. (check-in: 1a247911 user: dan tags: trunk)
2015-12-11
14:59
Add a new assert() statement to the snapshot-specific part of sqlite3WalBeginReadTrans(). (check-in: d8a12023 user: dan tags: trunk)
13:51
Simplified alignment constraints in the memory reuse logic of sqlite3VdbeMakeReady(). (check-in: e998513e user: drh tags: trunk)
12:44
Add the experimental snapshot interface. Because it is experimental, it is subject to change or removal at a later date. (check-in: 9b124a5a user: drh tags: trunk)
04:11
Fix a memory allocation bug introduced last week by check-in [a9e819082ba]. The bug only appears on systems where the size of a structure is not always a multiple of 8 - which in practice means only on 32-bit windows systems. (check-in: 96d3e99f user: drh tags: trunk)
03:27
Mention that the snapshot interfaces are only available if SQLite is compiled with SQLITE_ENABLE_SNAPSHOT. (Closed-Leaf check-in: 843c15a5 user: drh tags: snapshot-get)
03:20
Disable the snapshot test scripts if not compiled with SQLITE_ENABLE_SNAPSHOT. (check-in: 5fd008f0 user: drh tags: snapshot-get)
03:16
Fix a variable type mismatch problem in the snapshot logic. (check-in: 93fb8010 user: drh tags: snapshot-get)
01:22
Fix a bad assert related to snapshots. (check-in: 767ee30e user: drh tags: snapshot-get)
2015-12-10
22:48
Snapshot documentation updates. Comment changes only - no changes to code. (check-in: ef51a398 user: drh tags: snapshot-get)
20:03
Update the sqlite3_snapshot_get() API so that if the caller does not have an open read transaction on the named database, one is automatically opened. (check-in: b9c90f10 user: dan tags: snapshot-get)
19:44
Add tests to snapshot.test. (check-in: f3b74362 user: dan tags: snapshot-get)
19:11
Return SQLITE_BUSY (not SQLITE_BUSY_SNAPSHOT) if sqlite3_snapshot_open() fails to obtain the shared checkpointer lock. (check-in: 5343060b user: dan tags: snapshot-get)
18:06
Add tests to ensure that an sqlite3_snapshot_open() client cannot be tricked into reading a corrupt snapshot even if another process fails mid-checkpoint. (check-in: b908048b user: dan tags: snapshot-get)
17:59
Move pointer range comparisons into a macro, where they can be dealt with in a more portable way. (check-in: 05bc4f92 user: drh tags: trunk)
15:45
Have sqlite3_snapshot_open() avoid a race condition by taking a shared CHECKPOINTER lock while checking pInfo->nBackfillAttempted. (check-in: 8084eae0 user: dan tags: snapshot-get)
15:09
Move pointer range comparisons into a macro, where they can be dealt with in a more portable way. (Closed-Leaf check-in: ad3124c8 user: drh tags: stdint.h)
03:16
Fix spacing typo in comment. No changes to code. (check-in: 3a18526f user: mistachkin tags: snapshot-get)
02:15
Add the nBackfillAttempted field in formerly unused space in WalCkptInfo and use that field to close the race condition on opening a snapshot. (check-in: cb68e9d0 user: drh tags: snapshot-get)
2015-12-09
20:05
Update sqlite3_snapshot_open() to reduce the chances of reading a corrupt snapshot created by a checkpointer process exiting unexpectedly. (check-in: 7315f7cb user: dan tags: snapshot-get)
17:23
Further simplifications to the VDBE code generation logic that flow out of the previous check-in. (check-in: 6a5dfe85 user: drh tags: trunk)
16:26
Simplification of the DROP TRIGGER logic using sqlite3NestedParse() instead of hand-coded VDBE code. This is a manual cherry-pick of the key change from check-in [c80bbf14b365d]. (check-in: 8021b4c8 user: drh tags: trunk)
16:04
Merge unrelated fixes from trunk. (check-in: 362615b4 user: drh tags: snapshot-get)