Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
100 check-ins occurring around 2aabe05e2e.
|
2024-12-16
| ||
| 13:32 | Sync with trunk. (check-in: eab619453a user: drh tags: wal2) | |
| 13:25 | Changes for better alignment with the wal2 branch. (check-in: 4fab85b3d0 user: drh tags: trunk) | |
| 13:05 | Synchronize the WASM code with trunk. (check-in: 3619baa19e user: drh tags: wal2) | |
| 12:41 | Merge all the latest trunk enhancements and fixes into the wal2 branch. (check-in: 011ce4ae69 user: drh tags: wal2) | |
| 12:19 | Resynchronize the begin-concurrent branch with trunk. (check-in: e4718a04be user: drh tags: begin-concurrent) | |
| 12:09 | Remove unnecessary divergence from trunk. (check-in: 019b043428 user: drh tags: begin-concurrent) | |
| 12:08 | Remove unnecessary end-of-line whitespace. (check-in: ae580443d2 user: drh tags: trunk) | |
| 11:19 | Merge the latest trunk enhancements into the begin-concurrent branch. (check-in: faef90eac3 user: drh tags: begin-concurrent) | |
| 07:03 | Tiny makefile/configure script doc tweaks. No functional changes. (check-in: ed5e912622 user: stephan tags: trunk) | |
|
2024-12-14
| ||
| 17:36 | Merge changes from wal2, including the fix to prevent non-PASSIVE checkpoints from ever taking the writer lock. (check-in: eb8449ea9a user: dan tags: bedrock) | |
| 17:19 | Update to walRestartLog() function to make it shorter easier to merge with branch "bedrock". (check-in: 88d7fb2aef user: dan tags: wal2) | |
| 16:50 | Fix a problem with non-PASSIVE wal2 checkpoints sometimes taking the WRITER lock. (check-in: 11d7865a56 user: dan tags: wal2) | |
| 16:41 | Do not attempt to switch wal files when writing if the current transaction read from both wal files. This fix is already on branch 'bedrock'. (check-in: 693cc8e8db user: dan tags: wal2) | |
|
2024-12-13
| ||
| 20:18 | Fix a problem with the output of "PRAGMA wal_checkpoint" introduced long ago by [63483e22c775183e]. (check-in: c5f7bba868 user: dan tags: wal2) | |
| 18:13 | Avoid a memset() when loading stat4 data. Add instrumentation to the same code. (check-in: fa87355f62 user: dan tags: cf8f1552-commit-instr) | |
| 17:01 | Merge all recent trunk enhancements and fixes into the wal2 branch. Addendum: This was part of an attempt to reduce divergence. I'm still trying to figure out the best way to do that. (Closed-Leaf check-in: be32067a2a user: drh tags: mistake) | |
| 16:37 | Try to reduce incidental divergence between trunk and wal2. (Leaf check-in: 01c8ebfb1f user: drh tags: wal2-conflict-reduction) | |
| 01:29 | Fix a harmless compiler warning that comes up when using SQLITE_DEBUG in separate compilation mode. (check-in: 52e0f8cab9 user: drh tags: trunk) | |
|
2024-12-12
| ||
| 20:59 | Fix a buffer overread in test code. Add test configurations for ENABLE_SETLK=1 and FOR_WINRT=1 builds to release testing. (check-in: b558db5db7 user: dan tags: win32-enable-setlk) | |
| 20:39 | Increase the maximum number of arguments on an SQL function to 1000 with the capability to increase it further up to 32767 using a compile-time option. (check-in: 3c25c69c93 user: drh tags: cf8f1552-commit-instr) | |
| 19:20 | Add extra logging for alternative stat4 code. (check-in: 9dde0365cb user: dan tags: cf8f1552-commit-instr) | |
| 19:15 | Experimental code to load stat4 data using a single pass. (check-in: e0bcd30581 user: dan tags: cf8f1552-commit-instr) | |
| 15:11 | Increase the maximum number of arguments on an SQL function to 1000 with the capability to increase it further up to 32767 using a compile-time option. (check-in: e8d7d68ba0 user: drh tags: trunk) | |
|
2024-12-11
| ||
| 20:42 | When a writer initializes a new *-shm file hash table, zero only the hash-table slots, not the array of page numbers. (check-in: ac5401c62b user: dan tags: cf8f1552-commit-instr) | |
| 15:30 | Ensure that all win32 API functions invocations in os_win.c are made via the function table. Fix asan error in test code. (check-in: 2ffdc799c6 user: dan tags: win32-enable-setlk) | |
|
2024-12-10
| ||
| 21:24 | Fix a problem with walsetlk.test on windows. (check-in: 704e6567c1 user: dan tags: win32-enable-setlk) | |
| 19:00 | Ensure windows may block on a SHARED (but not EXCLUSIVE) DMS lock. (check-in: 55e4e7bb00 user: dan tags: win32-enable-setlk) | |
| 18:12 | Fix walsetlk.test to work on unix. (check-in: 798dff041b user: dan tags: win32-enable-setlk) | |
| 17:36 | Merge latest trunk changes into this branch. (check-in: d67a42d625 user: dan tags: win32-enable-setlk) | |
| 14:56 | Fix a problem with the formatting of "slow schema" log messages. Also change logging version to v=17. (check-in: df4183ace9 user: dan tags: cf8f1552-commit-instr) | |
| 12:32 | Fix harmless typo in a comment describing the columnMallocFailure() function. (check-in: 0f3b484fd7 user: drh tags: trunk) | |
|
2024-12-09
| ||
| 21:26 | Fix a problem with non-PASSIVE wal2 checkpoints sometimes taking the WRITER lock. (check-in: 1a59cae3c3 user: dan tags: cf8f1552-commit-instr) | |
| 21:01 | Bring all the latest trunk fixes and enhancements into the commit-and-continue branch. (Leaf check-in: 19e47df37a user: drh tags: commit-and-continue) | |
| 20:46 | Make the TCL interface more rebust against very large strings coming out of TCL9. (check-in: e2bae4143a user: drh tags: trunk) | |
| 20:37 | Enhance the ".import" command of the CLI so that it is able to insert into a view that has an instead-of trigger. Forum post 3e03c73150f8b9f8. (check-in: 7dcc3731a9 user: drh tags: trunk) | |
| 20:36 | Enhance the ".import" command of the CLI so that it is able to insert into a view that has an instead-of trigger. (Originally checked into the wrong branch.) (check-in: c71acee1cf user: drh tags: commit-and-continue) | |
| 17:30 | Test the case where a db connection blocks for a while and then successfully obtains the lock. (check-in: 9cbc97ff84 user: dan tags: win32-enable-setlk) | |
| 16:01 | Enhance tests to ensure that blocking locks, and not calls to xSleep(), are being used by SQLITE_ENABLE_SETLK_TIMEOUT builds when they should be. (check-in: c892242bb6 user: dan tags: win32-enable-setlk) | |
| 13:02 | Add NEVER() around branches that seems to have been made obsolete by [a350ea7c6b89725b]. (check-in: cb5bad5c74 user: drh tags: trunk) | |
| 11:47 | Further improvements to the decimal-to-float conversion for values close to LARGEST_UINT64. (check-in: 453c949914 user: drh tags: trunk) | |
| 11:37 | Fix an obscure problem with multiple outer joins, ON clauses and query flattening. Forum thread 5c8a069d23. (check-in: a350ea7c6b user: dan tags: trunk) | |
| 11:12 | Fix an obscure problem with multiple outer joins, ON clauses and query flattening. Forum thread 5c8a069d23. (Closed-Leaf check-in: 289daf6cee user: dan tags: forum-5c8a069d23-fix) | |
| 10:57 | Add a note in Makefile.msc about EXTRA_SRC files possibly requiring manual editing. (check-in: a0a36bad45 user: drh tags: trunk) | |
| 10:52 | Resynchronize autoconf/Makefile.msc (Closed-Leaf check-in: 8f9c640818 user: drh tags: makefile-fix) | |
|
2024-12-08
| ||
| 18:38 | Add a note in Makefile.msc about EXTRA_SRC files possibly requiring manual editing, as discussed in forum thread 903f721f3e7c0d25. (check-in: 229c2f013c user: stephan tags: makefile-fix) | |
|
2024-12-07
| ||
| 23:29 | Merge the latest trunk fixes and enhancements into the bedrock branch. (check-in: 08cfa7e8b3 user: drh tags: bedrock) | |
| 23:24 | Update the wal2 branch with the latest trunk enhancements and fixes. (check-in: d3ce95f7e0 user: drh tags: wal2) | |
| 21:20 | Version 3.47.2 for the reuse-schema branch. (Leaf check-in: 96428a5137 user: drh tags: reuse-schema-3.47) | |
| 21:16 | Version 3.47.2 for wal2 (Leaf check-in: 11fd09dbb7 user: drh tags: wal2-3.47) | |
| 21:14 | Version 3.47.2 for begin-concurrent (Leaf check-in: e725fe5aec user: drh tags: begin-concurrent-3.47) | |
| 20:39 | Version 3.47.2 (Leaf check-in: 2aabe05e2e user: drh tags: release, branch-3.47, version-3.47.2) | |
| 20:01 | Yet another iteration of the solution to the floating-point conversion problem. This one avoids complaints about oversize double values from -fsanitize. (check-in: 07bd06b13c user: drh tags: branch-3.47) | |
| 19:57 | Yet another iteration of the solution to the floating-point conversion problem - this what avoids complaints about oversize double values from -fsanitize. (check-in: fc6904a508 user: drh tags: trunk) | |
| 19:12 | A cleaner and more robust solution to the floating-point conversion problem originally fixed earlier today. (check-in: 18b20494ce user: drh tags: branch-3.47) | |
| 19:06 | A cleaner and more robust solution to the floating-point conversion problem originally fixed by [81342fa6dd03fffb]. (check-in: 351de57f80 user: drh tags: trunk) | |
| 17:08 | Fix more harmless compiler warnings. (check-in: f5b8fd7763 user: drh tags: trunk) | |
| 16:53 | Fix harmless compiler warning caused by the previous check-in. (check-in: 462700aeb7 user: drh tags: trunk) | |
| 14:53 | Increment the version number to 3.47.2 (check-in: 42aa7157f8 user: drh tags: branch-3.47) | |
| 14:51 | On x64 hardware, round-trip uint64_t→double→uint64_t conversions fail for values greater than UINT64_MAX-2047. This caused the SQLite text-to-float converter routine to give incorrect results for values between '1.8446744073709550592eNNN' and '1.8446744073709551609eNNN' for any exponent NNN. Fixed by this check-in. (check-in: 17537a98cb user: drh tags: branch-3.47) | |
| 14:48 | On x64 hardware, round-trip uint64_t→double→uint64_t conversions fail for values greater than UINT64_MAX-2047. This caused the SQLite text-to-float converter routine to give incorrect results for values between '1.8446744073709550592eNNN' and '1.8446744073709551609eNNN' for any exponent NNN. This problem was introduced by check-in [761d8fd18b0ee868] and first appeared in version 3.47.0 and was reported by forum post 569a7209179a7f5e. Fixed by this check-in. (check-in: 81342fa6dd user: drh tags: trunk) | |
|
2024-12-06
| ||
| 18:35 | Add the SQLITE_PREPARE_DONT_LOG option for sqlite3_prepare_v3(), that prevents errors in the compilation of the SQL from being sent to sqlite3_log(). (check-in: 8704034254 user: drh tags: trunk) | |
| 17:52 | Update logging (now v=16) to more closely analyze all attempts to load the main db schema. (check-in: 65b753735b user: dan tags: cf8f1552-commit-instr) | |
| 09:36 | Guard the x64 assembler in wal.c with "ifdef __x86_64__". (check-in: 5fa1699e31 user: dan tags: cf8f1552-commit-instr) | |
| 00:12 | Minor doc update in auto.def. No functional changes. (check-in: d324be296d user: stephan tags: trunk) | |
| 00:09 | Ensure that the post-config checks for @UNEXPANDED_VARS@ pass even if --disable-tcl is used, as reported in forum post 74ef8059fc. (check-in: a38606bf44 user: stephan tags: trunk) | |
|
2024-12-05
| ||
| 23:53 | Fix the build of sqlite3_analyzer.exe on Windows that was broken by check-in [80f3bf8c2ee31ba1]. (check-in: 223f47b2db user: drh tags: trunk) | |
| 19:45 | Add logging to report on occasions when loading the ANALYZE data is slow. (check-in: 5f9f6764e9 user: dan tags: cf8f1552-commit-instr) | |
| 16:52 | Use an inline asm "ret stosp" instruction instead of a memset() when a writer is required to zero a hash table. (check-in: c10fa6af8b user: dan tags: cf8f1552-commit-instr) | |
|
2024-12-04
| ||
| 19:32 | Fix filectrl.test so that it works with winrt builds. (check-in: 70b0cfb049 user: dan tags: win32-enable-setlk) | |
| 18:27 | Assume an NT-based OS when running tests if sqlite_os_type is not set. (check-in: 22e4aeded0 user: dan tags: win32-enable-setlk) | |
| 16:01 | Minor doc correction in ext/misc/shathree.c, as reported in the forum. No functional changes. (check-in: 3b82d2c6b7 user: stephan tags: trunk) | |
| 11:01 | Temporarily enable zeroing of wal2-mode hash tables within PASSIVE, as well as non-PASSIVE, checkpoints. (check-in: a97a2b0da1 user: dan tags: cf8f1552-commit-instr) | |
|
2024-12-03
| ||
| 19:41 | Back out minor changes not required by this branch. (check-in: 68c34ffd23 user: dan tags: win32-enable-setlk) | |
| 19:30 | Merge latest trunk into this branch. (check-in: eb13b8b3ef user: dan tags: win32-enable-setlk) | |
| 19:09 | Further test file tweaks to run with SQLITE_OS_WINRT builds. (check-in: 54853421ba user: dan tags: win32-enable-setlk) | |
| 15:19 | Fix the lock_win32_file test command so that it works with winrt builds. (check-in: 5994656074 user: dan tags: win32-enable-setlk) | |
|
2024-12-02
| ||
| 20:48 | Fix a problem with SQLITE_OS_WINRT builds. (check-in: baf43e8c7d user: dan tags: win32-enable-setlk) | |
| 19:55 | Improve the output from the ".testctrl opt" command in the CLI. (check-in: de7064d118 user: drh tags: trunk) | |
| 19:32 | Ensure that the query flattener does not change an ON clause term to a WHERE clause term. (check-in: e487d3b3c0 user: drh tags: branch-3.47) | |
| 17:21 | Ensure that the query flattener does not change an ON clause term to a WHERE clause term. Forum post 3f676b1196. Follow-up to [f1eae192315335d7]. (check-in: bdd408a255 user: dan tags: trunk) | |
| 16:38 | Add the SVG version of the SQLite logo. (check-in: a2061fe09f user: drh tags: branch-3.47) | |
| 16:24 | Fix a comment typo on the sqlite3ExprIsSingleTableConstraint() routine. (check-in: 346a845bf1 user: drh tags: trunk) | |
| 16:07 | Fix a harmless typo in debugging output. (check-in: 578f76cfb2 user: drh tags: trunk) | |
| 14:14 | Reformulate [9e09d5d6] so that memsys5 shows up in the late-config summary as being enabled and does the right thing if both --debug and --memsys3 are used. (check-in: 447db1cd0a user: stephan tags: trunk) | |
| 13:47 | Enable MEMSYS5 with the --dev configure option. (check-in: 9e09d5d60a user: drh tags: trunk) | |
| 13:29 | Slighly less confusing output from treeview during debugging. (check-in: 7aef0b9305 user: drh tags: trunk) | |
|
2024-12-01
| ||
| 16:19 | Fix the percentile extension so that works as an independent extension. Forum post ab25469a350e0488 (check-in: 86256f2cd4 user: stephan tags: branch-3.47) | |
| 15:23 | Trim a bit more from the sqlite370.svg border to get it closer in scaled size to the logo currently on the docsrc site. (check-in: 81c7277fc5 user: stephan tags: trunk) | |
| 14:20 | Correct the fill color for the background of sqlite370.svg - it was slightly darker than it should have been. (check-in: 732132407b user: stephan tags: trunk) | |
| 14:14 | Re-export sqlite370.svg to retain the gradient part and clip the size using the cutting guides in sqlite370.eps. (check-in: 12017b01c8 user: stephan tags: trunk) | |
| 14:02 | Export a clipped copy of sqlite370.eps as sqlite370.svg, based on discussion in forum post 1bbd6d17ca. (check-in: dd3a13c120 user: stephan tags: trunk) | |
|
2024-11-30
| ||
| 20:00 | On windows, use a separate handle for each connection for xShmLock() locks. (check-in: 272d552f12 user: dan tags: win32-enable-setlk) | |
| 17:48 | At the end of the configure script ensure that none of the files which are filtered for @VARS@ contain any unresolved @VAR@ placeholders, failing fatally if any do. (check-in: 301df5c2be user: stephan tags: trunk) | |
| 14:13 | In sqlite_dbpage, cancel any pending truncate operation if there an error occurs later in the transaction. (check-in: 1abab10f85 user: drh tags: trunk) | |
| 12:00 | A NEVER() that was added by [eb5ac9e5b9a4f9c8] is violated by the ifnull() in-line function. This check-in fixes that problem. (check-in: 2220ccf4d6 user: drh tags: trunk) | |
|
2024-11-29
| ||
| 11:49 | Fix a NEVER() in the iif() logic that can be true if compiled with SQLITE_ENABLE_UNKNOWN_SQL_FUNCTIONS. Problem introduced by [eb5ac9e5b9a4f9c8]. (check-in: 3ec2df5a6c user: drh tags: trunk) | |
|
2024-11-28
| ||
| 23:55 | Remove some outdated docs from Makefile.in. (check-in: cc2c5fc98b user: stephan tags: trunk) | |
| 20:46 | Typo fixes and cleanups in autosetup/README.md. (check-in: 7f366565f4 user: stephan tags: trunk) | |
| 19:14 | Update logging version to "v=13". (check-in: 442167b86a user: dan tags: cf8f1552-commit-instr) | |
| 18:58 | In wal2 mode, attempt to zero old *-shm hash table pages within non-PASSIVE checkpoints. This saves writers from having to do it. (check-in: 40aafd4585 user: dan tags: cf8f1552-commit-instr) | |