Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
200 most recent check-ins using file test/fts3expr3.test version 9e91b8ed
|
2018-12-14
| ||
| 20:28 | Add a missing OP_ColumnsUsed opcode to code for expressions like "? IN (SELECT ...)" in cases where expression can use an index that may contain NULL values. (Backported fix from version 3.14.0). (Leaf check-in: 11368e67 user: drh tags: branch-3.8.9) | |
| 20:20 | Add the OP_ColumnsUsed opcode (when compiled with SQLITE_ENABLE_COLUMN_USED_MASK) as a hint to the b-tree layer as to which columns of a btree cursor will be used. (Backport from 3.8.11) (check-in: b29e02f8 user: drh tags: branch-3.8.9) | |
|
2017-11-14
| ||
| 18:26 | Fix the SQLITE_ENABLE_UPDATE_DELETE_LIMIT functionality so that it works with views and WITHOUT ROWID tables. This is a modified cherrypick of [dae4a97a]. (check-in: b2679d3b user: dan tags: branch-3.8.9) | |
|
2017-11-09
| ||
| 04:13 | Disallow ORDER BY and LIMIT on UPDATE and DELETE of views and WITHOUT ROWID tables. This is a temporary fix for ticket [d4beea1633f1b88f] until a better solution can be found. (check-in: 30aa941f user: drh tags: branch-3.8.9) | |
|
2017-07-27
| ||
| 22:16 | Enhance the like optimization so that it works with an ESCAPE clause. (check-in: 2495acf7 user: drh tags: branch-3.8.9) | |
|
2017-07-21
| ||
| 04:08 | Add new interfaces sqlite3_result_pointer(), and sqlite3_value_pointer() and use them to transfer the eponymous FTS3 column pointer to the snippet() and offsets() routines. This is a cherry-pick of [f0f49224] which was in turn inspired by [72de49f2]. (Leaf check-in: 5a3022e0 user: drh tags: branch-3.8.10) | |
| 03:23 | Add new interfaces sqlite3_result_pointer(), and sqlite3_value_pointer() and use them to transfer the eponymous FTS3 column pointer to the snippet() and offsets() routines. This is a cherry-pick of [f0f492245e95], which is in turn inspired by check-in [72de49f2]. (Leaf check-in: a66a5b39 user: drh tags: branch-3.8.6) | |
|
2017-07-17
| ||
| 19:37 | Try to push relevant WHERE clause terms of the outer query down into the subquery in order to help the subquery run faster and/or use less memory. This is a cherry-pick of [6df18e949d36] with bug fixes. (check-in: cd6ac078 user: drh tags: branch-3.8.9) | |
| 19:25 | When checking for the WHERE-clause push-down optimization, verify that all terms of the compound inner SELECT are non-aggregate, not just the last term. Fix for ticket [f7f8c97e97597]. (Closed-Leaf check-in: adc082c1 user: drh tags: push-down-backport) | |
| 19:14 | Do not apply the WHERE-clause pushdown optimization to terms that originate in the ON or USING clause of a LEFT JOIN. Fix for ticket [c2a19d81652f40568c]. (check-in: 52674f94 user: drh tags: push-down-backport) | |
| 19:07 | For FROM-clause subqueries that cannot be flattened, try to push relevant WHERE clause terms of the outer query down into the subquery in order to help the subquery run faster and/or use less memory. Cherry-pick from [6df18e949d36]. Still need to backport bug fixes associated with that check-in. (check-in: 043d6ce8 user: drh tags: push-down-backport) | |
|
2017-07-08
| ||
| 01:01 | Fix the covering index OR optimization (check-in [fcbd6abd]) so that it works with SQLITE_MAX_ATTACHED>30. Broken by a bad cherry-pick merge. (check-in: d227de8a user: drh tags: branch-3.8.9) | |
|
2017-07-05
| ||
| 16:20 | Make use of covering indexes in the OR optimization. (check-in: fcbd6abd user: drh tags: branch-3.8.9) | |
| 14:54 | Add the count-of-view optimization when compiled using SQLITE_COUNTOFVIEW_OPTIMIZATION. (check-in: b7ae4b87 user: drh tags: branch-3.8.9) | |
|
2017-06-06
| ||
| 18:22 | Add the SQLITE_DEFAULT_ROWEST compile-time option for changing the estimated number of rows in tables that lack sqlite_stat1 entries. (check-in: 802b82f3 user: drh tags: branch-3.8.9) | |
|
2017-05-24
| ||
| 04:32 | Fix a problem in STAT4 equality estimation for multi-column indexes introduced by check-in [3e0590dee0e68cc1599]. (check-in: 19dad0a7 user: drh tags: branch-3.8.9) | |
|
2017-05-22
| ||
| 00:27 | When planning a query using sorting, resolve ties in the solver by selecting loop plans with the smaller unsorted cost. (check-in: 962531e7 user: drh tags: branch-3.8.9) | |
|
2017-01-11
| ||
| 14:21 | In the STAT4 computations, ensure that the aAvgEq values do not go negative. (check-in: 4f83f680 user: drh tags: branch-3.8.9) | |
|
2016-11-02
| ||
| 16:29 | When the block sorting optimization is used in a scalar subquery, be sure to exit the loop as soon as the first valid output row is received. Fix for ticket [cb3aa0641d9a4] backported to the 3.8.9 branch. (check-in: 8e4ba115 user: drh tags: branch-3.8.9) | |
|
2016-09-23
| ||
| 18:06 | Fix the ORDER BY LIMIT optimization backport so that it works when the ORDER BY uses the DESC direction. (check-in: 0c3cafb7 user: drh tags: branch-3.8.9) | |
|
2016-09-14
| ||
| 01:43 | Backport the ORDER BY LIMIT optimization to version 3.8.9. (check-in: db361482 user: drh tags: branch-3.8.9) | |
|
2016-08-10
| ||
| 03:35 | Try to make pcache1TruncateUnsafe() run faster for the case where iLimit is very close to iMaxKey. (Leaf check-in: 9ab53605 user: drh tags: debug) | |
| 02:54 | Remove all timers and other debugging logs except for the one timer on pcache1TruncateUnsafe(). (check-in: 5980e625 user: drh tags: debug) | |
|
2016-08-09
| ||
| 11:44 | Add another timer on pcache1TruncateUnsafe(). (check-in: 42ce53f6 user: drh tags: debug) | |
| 11:23 | Add a debug timer to pcache1Truncate(). Change the formatting of some other messages to make them easier to read. (check-in: 7d0af4b5 user: dan tags: debug) | |
| 01:21 | Refocus the timer on the pager_end_transaction() routine. (check-in: 9d4eb6e4 user: drh tags: debug) | |
|
2016-08-04
| ||
| 14:08 | Add extra logging calls to this branch. (check-in: 491f1ef3 user: dan tags: debug) | |
|
2016-07-30
| ||
| 03:33 | A new version of the slow mutex log that uses gettimeofday() instead of trying to access the hardware timer. (check-in: 92b9fead user: drh tags: debug) | |
|
2016-07-29
| ||
| 16:32 | Turn memory status off by default. (check-in: ea3c7162 user: drh tags: debug) | |
| 14:31 | Enhance a debugging log message to better identify a mutex. (check-in: e60cb6d8 user: dan tags: debug) | |
| 14:23 | Use __sync_synchronous(), when available, instead of a mutex when a memory barrier is needed. (check-in: 7d4562e9 user: drh tags: debug) | |
| 14:10 | Add calls to sqlite3_log() when sqlite3_mutex_enter() takes more than 100,000 CPU cycles. (check-in: 33c7ef7a user: drh tags: debug) | |
|
2016-07-16
| ||
| 11:47 | Temporarily add extra sqlite3_log() calls to this version to help with debugging a performance problem. (check-in: 613c1cea user: dan tags: debug) | |
|
2016-07-14
| ||
| 20:19 | Experimental patch restore the slight bias in btree-balancing immediately after the change to reduces the amount of memcpy() work done by balancing. (Leaf check-in: 46bd9533 user: drh tags: btree-balance-bias) | |
|
2015-10-05
| ||
| 11:57 | Add fts5txt2db.tcl, a tool for creating sample fts4/5 databases from text files. (check-in: 44f1ce30 user: dan tags: trunk) | |
|
2015-10-03
| ||
| 15:38 | Update fts5 to support the table function syntax. "... FROM fts5_tbl WHERE fts5_tbl MATCH ?1 AND rank MATCH ?1" can now be written "FROM fts5_tbl(?1, ?2)". (check-in: 41d17d9e user: dan tags: trunk) | |
| 12:23 | Add tests for the rtree module to verify that attempts to insert non-integer primary key values or non-numeric dimensions into an rtree table are handled correctly. (check-in: f653fce9 user: dan tags: trunk) | |
|
2015-10-02
| ||
| 20:04 | Update fts5 to avoid using a statement journal for UPDATE and DELETE operations that affect at most a single row. (check-in: 5c83b9db user: dan tags: trunk) | |
|
2015-10-01
| ||
| 18:31 | Fix an fts3 bug causing NEAR queries on uncommitted data to malfunction. (check-in: 6f90839e user: dan tags: trunk) | |
| 16:35 | Changes to the sesssions module ensure that tables appear within changesets and patchsets in the same order that they were attached to the sessions object. (check-in: 76958902 user: dan tags: sessions) | |
|
2015-09-30
| ||
| 14:50 | Merge recent enhancements from trunk, and especially the fix for ticket [1b266395d6bc10]. (check-in: b2face9a user: drh tags: sessions) | |
| 14:30 | Add a corrupt database to test/fuzzdata3.db to validate the previous check-in. (check-in: e796c0ef user: drh tags: trunk) | |
| 14:19 | Fix an assert() in btree.c that is only true for non-corrupt databases. (check-in: 29f7227f user: drh tags: trunk) | |
| 12:59 | Clear the BTCF_ValidNKey flag when putting a cursor into REQUIRESEEK state. Fix for [1b266395]. (check-in: a6d5e4e8 user: dan tags: trunk) | |
| 11:19 | Improve error handling in shell command ".tables". (check-in: 31a91ee7 user: dan tags: trunk) | |
|
2015-09-29
| ||
| 17:51 | This condition was not always true after all..... Change an always-true condition in the virtual table transaction interface into an assert(). (Closed-Leaf check-in: 3e15dea5 user: drh tags: mistake) | |
| 17:20 | Use symbolic names XN_ROWID and XN_EXPR in place of the (-1) and (-2) magic numbers in Index.aiColumn[]. Add asserts to help verify that Index.aiColumn[] is always used correctly. Fix one place in FK processing where Index.aiColumn[] was not being used correctly. (check-in: 7d272aa6 user: drh tags: trunk) | |
| 16:47 | Add the sqlite3_index_info.idxFlags field, allowing xBestIndex() implementations to specify to SQLite that a strategy may visit at most one row. Add support for this to fts3/4. Omit the statement journal from virtual table UPDATE and DELETE operations that are guaranteed not to affect more than one row. (check-in: a1d08fd3 user: dan tags: trunk) | |
| 16:41 | Ensure that the xSavepoint() virtual table method is correctly invoked if there are already open savepoints (or statement transactions) the first time a virtual table is written within a transaction. (check-in: 77948b5e user: dan tags: trunk) | |
| 15:50 | Remove dead code, replacing with assert() statements that make sure the code really was dead. (Closed-Leaf check-in: 0e317dda user: drh tags: vtab-onepass) | |
| 13:25 | Create the sqlite3IsToplevel(Parse*) interface to check to see if a top-level VDBE is being coded (versus a trigger) and use that interface. (check-in: 59662cd2 user: drh tags: vtab-onepass) | |
| 12:32 | Fix compiler warnings. (check-in: d1a07838 user: drh tags: vtab-onepass) | |
| 12:19 | Fix an off-by-one error in test function fts5_decode(). (check-in: 3a9f0762 user: dan tags: trunk) | |
| 11:59 | Merge latest trunk change into this branch. (check-in: b519c0d6 user: dan tags: vtab-onepass) | |
| 11:57 | Change the name of the new sqlite3_index_info.flags field to "idxFlags". Add documentation for the same to sqlite.h.in. (check-in: f61203bc user: dan tags: vtab-onepass) | |
| 10:11 | Add tests for fts3 and onepass update/delete operations. Also fix a problem with onepass updates that do not affect any rows. (check-in: 820c8044 user: dan tags: vtab-onepass) | |
|
2015-09-28
| ||
| 23:45 | Avoid unnecessary cursors and seeking when running a DELETE against a WITHOUT ROWID table. (Leaf check-in: 70ec88b2 user: drh tags: delete-without-rowid-opt) | |
| 20:03 | Also allow UPDATE on virtual tables to use the onepass strategy. (check-in: 1aa27d70 user: dan tags: vtab-onepass) | |
| 17:05 | Extra information provided by .wheretrace on input flags to the query planner and on the result of sqlite3WhereOkOnePass(). (check-in: c5566bb3 user: drh tags: trunk) | |
| 15:23 | Update fts3 to use the onepass strategy for delete operations. (check-in: fffab4f7 user: dan tags: vtab-onepass) | |
| 15:20 | Changes to allow DELETE operations on virtual tables to use the onepass strategy under some circumstances. (check-in: e73f919f user: dan tags: vtab-onepass) | |
| 15:08 | Add test cases to the ONEPASS optimization corruption problem fixed by the previous check-in. (check-in: 5c14d447 user: drh tags: trunk) | |
| 14:40 | Fix a database corruption bug caused by the ONEPASS optimization added in check-in [8b93cc5937000535]. Bug detected (prior to release) by sqllogictest. Test cases to follow. (check-in: 9d057f52 user: drh tags: trunk) | |
|
2015-09-26
| ||
| 17:44 | Enable adding JSON1 by appending the json1.c source file to the amalgamation and compiling with -DSQLITE_ENABLE_JSON1 (check-in: 33404b20 user: drh tags: trunk) | |
| 11:15 | Fix a memory leak that can occur following a syntax error in CREATE VIEW. (check-in: f4704035 user: drh tags: trunk) | |
| 03:31 | Fix the title of the documentation on sqlite3_value_subtype(). This is a comment change only. (check-in: b7902354 user: drh tags: trunk) | |
| 03:23 | Expand the header comment to clarify the purpose for the sqlite3MemoryBarrier() function. (check-in: b89495ae user: drh tags: trunk) | |
| 01:28 | Remove an assert in the windows worker-thread logic that can fail in a race condition. (check-in: d7bfb960 user: drh tags: trunk) | |
|
2015-09-25
| ||
| 23:40 | Fix threads.c so that setting sqlite3FaultSim(200) using the SQLITE_TESTCTRL_FAULT_INSTALL callback causes both unix and windows worker threads to be sequential and deterministic, so that they can be tested reliably. (check-in: 2d867c05 user: drh tags: trunk) | |
| 20:49 | Fix PRAGMA integrity_check so that it works with a UNIQUE index over expressions. (check-in: 113181ce user: drh tags: trunk) | |
| 20:18 | Fix a faulty assert() on the benign-malloc controller callback in test_malloc.c. (Leaf check-in: 64f2cb8d user: drh tags: malloc-testing) | |
| 16:29 | Disable the SQLITE_TESTCTRL_BENIGN_MALLOC_HOOKS mechanism and replace it with SQLITE_TESTCTRL_BENIGN_MALLOC_CTRL, which gives better control. (check-in: 2c57b2f3 user: drh tags: malloc-testing) | |
| 13:42 | Report an error if the number of named columns in a VIEW is different from the number of result columns in the SELECT statement that implements the VIEW. (check-in: 6e4ac0be user: drh tags: trunk) | |
| 01:09 | Fix a problem computing affinities for indexes during skip-scan code generation when SQLITE_ENABLE_STAT4 is on. Bug introduced by check-in [1ee089a72d789002]. (check-in: 1350dd42 user: drh tags: trunk) | |
|
2015-09-24
| ||
| 18:47 | Enhance the query planner so that it is able to use indexed expressions to help fulfill an ORDER BY clause. (check-in: 668fc1eb user: drh tags: trunk) | |
| 17:38 | Fix over-length source code lines in where.c. No logic changes. (check-in: 1c8c5380 user: drh tags: trunk) | |
| 15:17 | Merge all recent trunk enhancements and fixes into the begin-concurrent branch. (check-in: c63c1e15 user: drh tags: begin-concurrent) | |
| 15:06 | Merge trunk changes into the cursor-hints branch. (check-in: fbe63762 user: drh tags: cursor-hints) | |
| 14:43 | Merge recent trunk enhancements into the apple-osx branch. (check-in: 4dd06d8b user: drh tags: apple-osx) | |
| 14:26 | Merge all the latest trunk enhancements into the sessions branch. (check-in: c91065f8 user: drh tags: sessions) | |
| 12:40 | Strengthen the implementations of xShmMemoryBarrier on both the unix and windows VFSes, so that they likely work even if SQLITE_THREADSAFE=0 is used. (check-in: c6ab807b user: drh tags: trunk) | |
| 12:19 | Make sure joins work correctly when both sides of the join are connected using indexed expressions. (check-in: c2fcb032 user: drh tags: trunk) | |
| 11:26 | Correctly handle the case of a WHERE-clause term with indexed expressions on both sides of the == sign. (check-in: d9b716a6 user: drh tags: trunk) | |
| 11:06 | Fix a JSON1 test case so that it works on builds that omit virtual tables. (check-in: a4444c0f user: drh tags: trunk) | |
| 01:40 | Another (smaller) performance optimization for the JSON parser. (check-in: c43daa8c user: drh tags: trunk) | |
| 01:06 | Performance optimizations on the JSON parser. (check-in: 7dd4b07a user: drh tags: trunk) | |
|
2015-09-23
| ||
| 19:17 | Take care that the number of reserved bits per page is consistent between the source and destination databases when doing the back-copy on a VACUUM. (check-in: 5b61b72f user: drh tags: trunk) | |
| 16:35 | Merge updates from trunk. (check-in: 0a75ebd2 user: mistachkin tags: mutexInitIsInitReCheck) | |
| 16:33 | Correct superfluous whitespace difference. No functional changes. (check-in: 8d69983d user: mistachkin tags: mutexInitSimpleCmpSwap) | |
| 16:24 | Simplify thread-safety of mutex initialization. (check-in: da0587c5 user: mistachkin tags: mutexInitSimpleCmpSwap) | |
| 15:54 | Merge updates from trunk. (Closed-Leaf check-in: 6b85f8cd user: mistachkin tags: mutexInitCmpSwap) | |
| 11:59 | Capture AFL-generated fuzz tests for json1.c into the test/fuzzdata4.db file. (check-in: 10a214fd user: drh tags: trunk) | |
| 01:10 | Do not allow a comma at the end of a JSON array or object. (check-in: 7c7a3f3e user: drh tags: trunk) | |
|
2015-09-22
| ||
| 18:51 | Add the --export-sql and --export-db options to the fuzzcheck utility program. (check-in: 760af445 user: drh tags: trunk) | |
| 17:46 | Fix a typo in the --help screen of the fuzzcheck utility. (check-in: b6ae61fe user: drh tags: trunk) | |
| 01:15 | Futher simplifications to json1.c. Also an obscure bug-fix in the initial output of json_tree() when using a path to an object contained within an array. (check-in: fcb1e327 user: drh tags: trunk) | |
| 00:21 | Fix json_set() so that it can overwrite a value that was previously overwritten during the same call. (check-in: 0f160416 user: drh tags: trunk) | |
|
2015-09-21
| ||
| 23:53 | For MSVC, have the 'sqlite3.c' target depend on 'sqlite3ext.h' as well as other targets may depend on this behavior (e.g. extensions). (check-in: 737ac3fa user: mistachkin tags: trunk) | |
| 23:49 | Fix compilation under older versions of MSVC. (check-in: 17150ada user: mistachkin tags: trunk) | |
| 22:53 | Simplifications to the json1.c logic. (check-in: 1646a2bd user: drh tags: trunk) | |
|
2015-09-20
| ||
| 22:57 | Fix an overly-strict assert() in the btree logic. (check-in: 825ce320 user: drh tags: trunk) | |
|
2015-09-19
| ||
| 20:27 | Do not allow table-valued function syntax to be used on a view. (check-in: 97cfe346 user: drh tags: trunk) | |
| 19:36 | Add the --disable-lookaside option to the fuzzershell utility program. (check-in: 0b043744 user: drh tags: trunk) | |
| 18:54 | Fix a memory leak in json1.c that could occur after misuse of json_object(). (check-in: 394b81b1 user: drh tags: trunk) | |
| 18:11 | Add an assert() statement to delete.c. (check-in: 40fce67e user: dan tags: trunk) | |
| 18:08 | Fix compiler warnings. (check-in: 5c31ee62 user: drh tags: trunk) | |
| 16:51 | Add json1.c to the fuzzer test shells. (check-in: 809d67b6 user: drh tags: trunk) | |
| 16:45 | Adjust Makefiles to use SQLITE_CORE when statically linking json1.c into test shells. (Closed-Leaf check-in: 91682ac3 user: drh tags: json-fuzz) | |
| 14:57 | Adjust fuzzershell so that it adds the json1 extension separately for each no connection. (check-in: ca2d1e90 user: drh tags: json-fuzz) | |
| 14:32 | Include json1.c in fuzzcheck and fuzzershell. (check-in: c4b68eff user: drh tags: json-fuzz) | |
| 12:04 | Fix a documentation typo in sqlite3_bind_parameter_index(). No code changes. (check-in: 650605a8 user: drh tags: trunk) | |
| 11:57 | Fix an off-by-one error (really off-by-2 in this case) in the buffer resize logic of json1. (check-in: d2a02737 user: drh tags: trunk) | |
| 03:07 | Do not allow rowid in a UNIQUE or PRIMARY KEY constraint. (This restores the same behavior exhibited by all prior releases.) (check-in: b1278301 user: drh tags: trunk) | |
|
2015-09-18
| ||
| 18:09 | Fix a crash that can occur with a skip-scan on an index with expressions when SQLITE_ENABLE_EXPLAIN_COMMENTS is defined. (check-in: 25f34cb9 user: drh tags: trunk) | |
| 15:38 | Adjustments to the spellfix2.test module so that it works reliably on windows. (check-in: 09a313fb user: drh tags: trunk) | |
| 15:35 | Adjust two test cases so that they conform to the current code. (check-in: 228e3d7c user: drh tags: trunk) | |
| 15:13 | Fix sqlilimits.test module so that it knows that the SELECT in a CREATE VIEW is not checked until the view is actually used. (check-in: acf5d87f user: drh tags: trunk) | |
| 14:45 | Ensure that "PRAGMA integrity_check" reports an error if the free-list count header field contains a value smaller than the actual number of pages on the database free-list. (check-in: 26f64986 user: dan tags: trunk) | |
| 14:42 | Fix the orderby9.test case so that it works with 32-bit versions of TCL (check-in: 4b6af774 user: drh tags: trunk) | |
| 14:22 | Remove a test made obsolete by the ONEPASS DELETE optimization. (check-in: c88b62c2 user: drh tags: trunk) | |
|
2015-09-17
| ||
| 17:21 | Avoid passing (signed char) values directly to isspace(), isalnum() or isdigit() in json1.c. Cast the value to (unsigned char) first. (check-in: 6713e35b user: dan tags: trunk) | |
| 09:48 | Fix a crash in FTS5 caused by specifying NULL as the rhs of a MATCH operator. (check-in: e8ed62f8 user: dan tags: trunk) | |
|
2015-09-16
| ||
| 17:46 | Add a "flush" to force incremental output to releasetest-out.txt while running the test/releasetest.tcl. script. (check-in: 0f3de2d3 user: drh tags: trunk) | |
| 12:31 | Fix typo in fts5.h. (check-in: 07721c6c user: dan tags: trunk) | |
|
2015-09-15
| ||
| 19:16 | Merge enhancements from trunk. (check-in: fc4d1de8 user: drh tags: begin-concurrent) | |
| 17:31 | Merge enhancements from trunk. (check-in: 66fe0683 user: drh tags: cursor-hints) | |
| 17:20 | Merge trunk enhancements into the apple-osx branch. (check-in: f12b8a0f user: drh tags: apple-osx) | |
| 15:55 | Merge the latest trunk enhancements with this branch. (check-in: b7469c44 user: dan tags: sessions) | |
| 14:39 | Add test cases to cover fts5 integrity-check code. (check-in: 1d018c35 user: dan tags: trunk) | |
| 13:42 | Reformat some code to make it easier to merge with sessions. No logic changes. (check-in: eade355f user: drh tags: trunk) | |
| 11:58 | Fix a problem with fts5 "ORDER BY rowid DESC" queries and large terms. (check-in: b26d8f79 user: dan tags: trunk) | |
|
2015-09-14
| ||
| 22:54 | Remove a comment that is no longer accurate. (check-in: cf114c18 user: mistachkin tags: mutexInitIsInitReCheck) | |
| 22:53 | Merge updates from trunk. (check-in: 25c15771 user: mistachkin tags: mutexInitIsInitReCheck) | |
| 22:52 | Re-check sqlite3GlobalConfig.isInit after the mutex subsystem has been initialized. (check-in: fea8c0b2 user: mistachkin tags: mutexInitIsInitReCheck) | |
| 19:51 | Improve test coverage of fts5_index.c. (check-in: c1f76686 user: dan tags: trunk) | |
| 19:26 | Testability improvements for the ONEPASS_MULTI enhancement. (check-in: d2df93f2 user: drh tags: trunk) | |
| 14:49 | Fix a typo in an SQLITE_USER_AUTHENTICATION macro. (check-in: 379455af user: drh tags: trunk) | |
| 14:45 | Use symbolic names ONEPASS_OFF, ONEPASS_SINGLE, and ONEPASS_MULTI for the various modes of the one-pass optimization. (check-in: 16e56bda user: drh tags: trunk) | |
| 14:08 | Fix a compiler warning and providing missing VdbeCoverage() calls. (check-in: 2edd2e5e user: drh tags: trunk) | |
| 11:09 | Use a single-pass approach for DELETE statements on non-virtual tables that do not require trigger or foreign key processing. (check-in: 8b93cc59 user: dan tags: trunk) | |
| 10:47 | Merge the latest trunk enhancements. (check-in: 22ce9218 user: drh tags: sessions) | |
| 09:23 | Avoid calling sqlite3VdbeCursorMoveto() from more than one point in vdbe.c. Doing so prevents it from being inlined as part of OP_Column. (Closed-Leaf check-in: 166d5af8 user: dan tags: onepass-delete) | |
|
2015-09-13
| ||
| 20:15 | Add static VFS mutexes to the primary header file comments. (check-in: 9a867d9f user: mistachkin tags: trunk) | |
| 18:45 | Experimental changes to avoid recusrively calling xMutexInit. (check-in: c9400ff1 user: mistachkin tags: mutexInitCmpSwap) | |
| 18:43 | Merge updates from trunk. (check-in: b42c2e20 user: mistachkin tags: mutexInitCmpSwap) | |
|
2015-09-12
| ||
| 19:50 | Fix compiler warnings in delete.c. (check-in: 0a4d285e user: dan tags: onepass-delete) | |
| 19:27 | Fix API documentation typos. Emphasize that the sqlite3_config() routine is not threadsafe. (check-in: 786333e0 user: drh tags: trunk) | |
| 19:26 | Experimental change to use a single-pass approach for DELETE statements on non-virtual tables that do not fire triggers or require foriegn-key processing. (check-in: eaeb2b80 user: dan tags: onepass-delete) | |
| 18:58 | Merge updates from trunk. (check-in: d8051f61 user: mistachkin tags: mutexInitCmpSwap) | |
| 18:57 | Import common changes from the mutex initialization branch. (check-in: 334720c0 user: mistachkin tags: trunk) | |
| 16:59 | Merge updates from trunk. (check-in: 48597789 user: mistachkin tags: mutexInitCmpSwap) | |
| 04:22 | Fix comments. (check-in: de3fa51e user: mistachkin tags: mutexInitCmpSwap) | |
| 04:19 | Specifying an invalid mutex implementation (via SQLITE_CONFIG_MUTEX) should cause the default one to be used instead. (check-in: 1a97bc81 user: mistachkin tags: mutexInitCmpSwap) | |
| 03:40 | Clarify the new mutex implementation reset logic in sqlite3_shutdown(). (check-in: f9a03483 user: mistachkin tags: mutexInitCmpSwap) | |
| 03:35 | Add more asserts to the mutex subsystem. (check-in: 7562f1fb user: mistachkin tags: mutexInitCmpSwap) | |
| 01:17 | Fix typo in the Win32 implementation of sqlite3CompareAndSwap. (check-in: 31a26a1d user: mistachkin tags: mutexInitCmpSwap) | |
|
2015-09-11
| ||
| 23:24 | Make sure that the mutex implementation can be altered after calling sqlite3_shutdown(). (check-in: dc2cf897 user: mistachkin tags: mutexInitCmpSwap) | |
| 20:54 | Update evidence marks due to wording changes in requirements text. No changes to code. (check-in: 86781093 user: drh tags: trunk) | |
| 18:05 | Fix harmless compiler warnings. (check-in: bfc7b84b user: mistachkin tags: trunk) | |
| 15:32 | More test cases in test/json102.test corresponding to new examples in the json1 documentation. (check-in: f599a42e user: drh tags: trunk) | |
| 14:15 | Fix some compiler warnings in fts5 code. (check-in: 0dc43611 user: dan tags: trunk) | |
| 05:06 | Enhance mutex initialization to prevent possible race conditions between sqlite3_initialize() and sqlite3_config(). Also, re-check sqlite3GlobalConfig.isInit after the mutex subsystem has been initialized. (check-in: f6a8f577 user: mistachkin tags: mutexInitCmpSwap) | |
| 01:22 | Updates to the sqlite3_value_subtype() and sqlite3_result_subtype() documentation and to test cases for json1 dealing with those interfaces. (check-in: d6cadbe9 user: drh tags: trunk) | |
| 00:26 | Add new interfaces sqlite3_value_subtype() and sqlite3_result_subtype(). Update the json1.c extension to take advantages of those interfaces to avoid the goofy '$$' path syntax and to allow nested calls to json_array() and json_object() that work as expected. (check-in: db4152ae user: drh tags: trunk) | |
| 00:06 | Take out the goofy '$$' path syntax. Instead, use subtypes to communicate when a string is JSON. Add the json() function that validates and minifies the JSON and sets the appropriate subtype. (Closed-Leaf check-in: 8a80d645 user: drh tags: subtypes) | |
|
2015-09-10
| ||
| 20:40 | Make the sqlite3_value_subtype() and sqlite3_result_subtype() interfaces available to loadable extensions. (check-in: c6fca0be user: drh tags: subtypes) | |
| 20:34 | Experimental implementation of sqlite3_result_subtype() and sqlite3_value_subtype() interfaces. (check-in: 7b5be299 user: drh tags: subtypes) | |
| 19:22 | Fix a potential NULL pointer deref in the testing logic of pcache1. NB: The -DSQLITE_TEST compile-time option is needed to hit the problem. (check-in: f5580f08 user: drh tags: trunk) | |
| 17:23 | Modify the fts5 leaf page format to permit faster seek operations. This is a file-format change. Any existing databases can be upgraded by running the fts5 'rebuild' command. (check-in: 0c0c4ae9 user: dan tags: trunk) | |
| 17:20 | Create separate "path" and "root" columns in the json_each() and json_tree() virtual tables. "Root" is the 2nd parameter and is fixed. "Path" varies as json_tree() walks the hierarchy. (check-in: 127cce3e user: drh tags: trunk) | |
| 16:39 | Increment the fts5 version value to indicate that the on-disk format has changed. (Closed-Leaf check-in: 99de5e36 user: dan tags: fts5-incompatible) | |
| 16:19 | Fix a segfault in fts5 that could occur if the database contents were corrupt. (check-in: 4931e37d user: dan tags: fts5-incompatible) | |
| 15:52 | Merge latest changes from trunk. Including fts5_expr.c fixes. (check-in: 716e7e74 user: dan tags: fts5-incompatible) | |
| 15:49 | Update description of on-disk format in fts5_index.c. (check-in: 85aac7b8 user: dan tags: fts5-incompatible) | |
| 15:24 | Make the sqlite3ext.h header file responsive to -DSQLITE_OMIT_LOAD_EXTENSION. (check-in: 47a46a9f user: drh tags: trunk) | |
| 15:22 | Disable tests for json_each() and json_tree() on builds where virtual tables are not supported (check-in: bb8ee3b1 user: drh tags: trunk) | |
| 15:20 | Fix the --help option on test/releasetest.tcl. Also fix the final error count so that it includes a count of subtest crashes. (check-in: 9ecf684d user: drh tags: trunk) | |
| 10:40 | Revert an accidentally committed makefile change. (check-in: 402704b1 user: dan tags: fts5-incompatible) | |
| 10:01 | Fix an fts5 problem that could occur if a term and the first associated rowid are on different leaf pages. (check-in: ffe2796a user: dan tags: fts5-incompatible) | |
| 05:40 | Change the array of 16-bit offsets at the end of each page to an array of varints. (check-in: fab245be user: dan tags: fts5-incompatible) | |
| 04:17 | Attempt to declare sqlite3MemoryBarrier() correctly for all possible build configurations. (check-in: da8a288f user: drh tags: trunk) | |
| 03:29 | Fix the json_tree() scan for the case when a path is supplied. Add new json1 test cases. (check-in: 6adc7de7 user: drh tags: trunk) | |
| 01:22 | No-op the sqlite3_memory_alarm() interface in a different way, that does not break legacy memory behavior. This is a re-do of check-in [5d3f5df4da9f40d5]. (check-in: 8250e2a4 user: drh tags: trunk) | |
|
2015-09-09
| ||
| 23:54 | Fix harmless compiler warning in FTS5. (check-in: 86146a73 user: mistachkin tags: trunk) | |
| 19:44 | Fix an assert() enabled by SQLITE_ENABLE_EXPENSIVE_ASSERT in wal.c. (check-in: 8d2ed150 user: dan tags: trunk) | |
| 19:27 | Fix a possible NULL pointer deref when using SQLITE_ENABLE_MEMORY_MANAGEMENT. (check-in: 89bfdbfe user: drh tags: trunk) | |
| 17:23 | Fix harmless compiler warning. (check-in: 280fd3a4 user: mistachkin tags: trunk) | |
| 17:17 | Fix harmless compiler warnings in FTS5. (check-in: 2cdb1877 user: mistachkin tags: trunk) | |
| 13:28 | When running a CREATE TABLE AS, make the initial temporary sqlite_master entry for the new table a real record rather than a NULL, in case the query after the AS actually tries to read the sqlite_master table. Fix for ticket [acd12990885d9276]. (check-in: 4a18d8bd user: drh tags: trunk) | |
| 08:15 | Fix a bug in preprocessor macros within fts5_main.c. (check-in: 0eb2b952 user: dan tags: fts5-incompatible) | |
|
2015-09-08
| ||
| 21:16 | Remove an unused local variable from Lemon. (check-in: fe9ffe5e user: drh tags: trunk) | |
| 21:12 | Enhance the DBSTAT virtual table with a new hidden table "schema" that if set will cause the table to report on the specified schema rather than on "main". Also: Fix a faulty assert in sqlite3_context_db_handle(). (check-in: 6beb512c user: drh tags: trunk) | |
| 20:26 | Eponymous virtual tables exist in the "main" schema only. Enforce this rule. (check-in: 06f90bb2 user: drh tags: trunk) | |
| 19:55 | Remove the 0x00 terminators from the end of fts5 doclists stored on disk. (check-in: 00d99006 user: dan tags: fts5-incompatible) | |
| 17:31 | Fix the help message that sqlite3_analyzer.exe generates for invalid arguments. (check-in: 33a14e7b user: drh tags: trunk) | |
|
2015-09-07
| ||
| 23:40 | Minor tweaks to Lemon. (check-in: 98667722 user: drh tags: trunk) | |
| 20:22 | Merge parser enhancements and other improvements and bug fixes from trunk. (check-in: 9cf3e51b user: drh tags: begin-concurrent) | |
| 20:11 | Enhance the Lemon parser generator to add SHIFTREDUCE states that reduce the sizes of some of the parser tables. (check-in: 99b992fa user: drh tags: trunk) | |
| 20:02 | Fix an unreachable branch in the new parse automaton. (Closed-Leaf check-in: e9d604b4 user: drh tags: lemon-update) | |
| 19:52 | Change the parser engine so that it (once again) waits for a lookahead token before reducing, even in a SHIFTREDUCE action. (check-in: 2c17a135 user: drh tags: lemon-update) | |