Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
100 descendants and 94 ancestors of 61853bc171e7c3af
2017-07-10
| ||
13:24 | For sqlite3TreeView() debugging output, show the Expr.flags field on scalar subqueries. (check-in: dc857a96b0 user: drh tags: trunk) | |
12:07 | Update error message text for standard error codes to better describe the latest usage of those error codes. Modify sqlite3_open_v2() so that it does return a valid sqlite3 object in the event of SQLITE_MISUSE due to bad open flags, so that sqlite3_errmsg() does not report "out of memory" in that case. (check-in: f27b637040 user: drh tags: trunk) | |
11:17 | Remove the error message text from disused error codes such as SQLITE_EMPTY and SQLITE_FORMAT. (check-in: 871752f292 user: drh tags: trunk) | |
2017-07-09
| ||
18:55 | Always make "column%d" column-names 1-based, never 0-based. (check-in: 70096c505d user: drh tags: trunk) | |
00:30 | Make sure the columns of a VALUES() clause are named "column%d" even if the VALUES() clause is a subquery in the FROM clause of an outer query. (check-in: acf3b9cc9c user: drh tags: trunk) | |
2017-07-08
| ||
22:30 | Add the --icc, --gcc7, and --orm options to the speed-check.sh test script. (check-in: 70dae08a9a user: drh tags: trunk) | |
20:46 | Support read-only MVCC transactions in server-mode. Started using "BEGIN READONLY". (check-in: 5a043aa8dd user: dan tags: server-process-edition) | |
18:27 | Disable compiler intrinsics when using the Intel C Compiler. (Leaf check-in: dbb59f6156 user: drh tags: icc-fix) | |
2017-07-07
| ||
22:58 | Merge in all the latest trunk enhancements. (check-in: 7f48f6ab2a user: drh tags: apple-osx) | |
22:47 | Basic test cases for PRAGMA secure_delete=FAST. (check-in: 0c246017b4 user: drh tags: trunk) | |
21:20 | Fix comment in the Win32 VFS for lsm1. (check-in: 12a421bcac user: mistachkin tags: trunk) | |
21:15 | For lsmtest, use a more portable means of setting open() files to binary. (check-in: aea6e0ffd3 user: mistachkin tags: trunk) | |
20:35 | Adjust copy_file() lsmtest function so it works properly for locked database files on Win32. (check-in: 54a3855583 user: mistachkin tags: trunk) | |
20:06 | Add the "PRAGMA secure_delete=FAST" option, which overwrites most deleted content without increasing the amount of I/O. Deleted content might persist on the free page list, however. And extra CPU cycles are used for zeroing, of course. (check-in: 38978ce65b user: drh tags: trunk) | |
19:43 | Change the error message text for SQLITE_ERROR to omit the part about "missing database" as that meaning is now obsolete (since approx SQLite 2.0). (check-in: 732f90d632 user: drh tags: trunk) | |
19:22 | Fix typos in Win32 VFS for lsm1. (check-in: 0961ec873b user: mistachkin tags: trunk) | |
19:12 | Fix the memcpy calls in the lsmWin32OsShmMap function for lsm1. (check-in: aa4890b9f9 user: mistachkin tags: trunk) | |
19:02 | Enhance the MSVC makefile 'clean' target to handle lsm1. (check-in: 2d6faebc31 user: mistachkin tags: trunk) | |
18:31 | Add support for LSM_DEBUG to the MSVC makefile for lsm1. (check-in: 277a7fedaf user: mistachkin tags: trunk) | |
18:06 | In the command-line shell, when running ".schema" give a sensible error when the database file is locked. (check-in: cc329eb8d6 user: drh tags: trunk) | |
17:57 | Fix the usleep() macro in the Win32 test code for lsm1. (check-in: 63599fa524 user: mistachkin tags: trunk) | |
17:43 | Add new PRAGMAs: "function_list", "module_list", and "pragma_list". All are enclosed within #ifdef SQLITE_INTROSPECTION_PRAGMAS. The compile-time option is not on ctime.c yet, since these are still experimental and one can always test "PRAGMA pragma_list" and see whether or not it returns an empty set. (check-in: e0b6ae92ad user: drh tags: trunk) | |
17:33 | Add the "Pragma_list" pragma. Put all three pragmas created on this branch inside of #ifdef SQLITE_INTROSPECTION_PRAGMAS. (Closed-Leaf check-in: 2f3c3781ef user: drh tags: list-pragmas) | |
16:46 | Improve clarity in the Win32 VFS for lsm1. (check-in: 9c7a9e805d user: mistachkin tags: trunk) | |
16:40 | Merge latest trunk changes with this branch. (check-in: 216c757f92 user: dan tags: server-process-edition) | |
16:29 | Add support for LSM_DEBUG_MEM to the MSVC makefile for lsm1. (check-in: 3340de83df user: mistachkin tags: trunk) | |
16:27 | Remove a block from the Win32 VFS for lsm1 that is now superfluous. (check-in: 5f726f2830 user: dan tags: trunk) | |
16:09 | In lsmtest, use an empty string instead of ":memory:" when requesting a temporary database from SQLite. This prevents other test code from trying to unlink() ":memory:", which fails on win32. (check-in: 281ad505d7 user: dan tags: trunk) | |
16:07 | Remove a block from the Win32 VFS for lsm1 that is now superfluous. (check-in: db4a41408e user: mistachkin tags: trunk) | |
16:06 | Memory reallocation fix for the Win32 VFS for lsm1. (check-in: 5a3140e4de user: mistachkin tags: trunk) | |
16:00 | Enhance the sqlite3VdbeMultiLoad() interface to automatically generate the OP_ResultRow opcode on PRAGMA implementations, for a small reduction in the library footprint. (check-in: c46f0f076c user: drh tags: trunk) | |
15:43 | Add new pragmas: "function_list" and "module_list" (check-in: e5f01d7fde user: drh tags: list-pragmas) | |
14:54 | Fix an obsolete comment on the sqlite3_namelist() function implementation. (Closed-Leaf check-in: f4229857a1 user: drh tags: experimental-namelist) | |
14:26 | Merge recent enhancements from trunk. (check-in: 73d0fc027d user: drh tags: bind-pointer) | |
13:59 | Exploit the fact that Expr.pRight and Expr.x are never used at the same time for a small performance gain. (check-in: aacbb9a461 user: drh tags: trunk) | |
12:58 | More efficient and compact implementation of walkExpr(). (check-in: 115d4b8339 user: drh tags: trunk) | |
12:43 | More aggressive use of EP_Leaf on expression nodes, to help prune searches. (check-in: c1a1d68c82 user: drh tags: trunk) | |
11:49 | Very slightly smaller and faster sqlite3WalkSelect(). (check-in: 9bc65635b2 user: drh tags: trunk) | |
2017-07-06
| ||
22:43 | Small adjustment to main.mk that facilitates giving non-standard compile-time options to the shell. (check-in: 7c7d53a9bb user: drh tags: trunk) | |
22:40 | Always load the schema before starting tab-completion. (check-in: 907fd3aab6 user: drh tags: experimental-namelist) | |
20:08 | Tab-completion now also works using readline/editline. (check-in: c906739f0c user: drh tags: experimental-namelist) | |
19:26 | Use the sqlite3_namelist() interface to implement simple tab-completion using linenoise. (check-in: 5cc7b0e2ca user: drh tags: experimental-namelist) | |
18:52 | Change the sqlite3_namelist() interface to return a pointer to an array of pointers to strings, and to avoid duplicates. (check-in: 7029111044 user: drh tags: experimental-namelist) | |
18:25 | The sqlite3_namelist() routine now works for all name types. (check-in: e41d62175b user: drh tags: experimental-namelist) | |
17:36 | Initial implementation of a highly experimental interface for listing all keywords and symbolic names for an SQLite database connection. (check-in: 04ef6783a5 user: drh tags: experimental-namelist) | |
16:33 | Change the (machine-generated) keywordhash.h file to increase the scope of the tables used for keyword matching, so that the tables are accessible to functions other then keywordCode(). (check-in: c5ed5ebdf6 user: drh tags: trunk) | |
13:51 | More compact implementation of the typeof() SQL function. (check-in: efb4aab0ca user: drh tags: trunk) | |
13:23 | Avoid unnecessary upper-to-lower case conversion for function names when registering the built-in functions. (check-in: 0626925764 user: drh tags: trunk) | |
03:06 | Small size reduction in findCollSeqEntry(). (check-in: 55ecd30304 user: drh tags: trunk) | |
02:49 | Avoid unnecessary calls to sqlite3GetCollSeq() for a small performance gain. (check-in: 503ba172e4 user: drh tags: trunk) | |
01:28 | Slightly more compact implementation of the byte-code generator for the COMMIT and ROLLBACK commands. (check-in: 4da663d986 user: drh tags: trunk) | |
01:02 | Small performance increase in sqlite3SrcListAppend(). (check-in: 7b2623f199 user: drh tags: trunk) | |
2017-07-05
| ||
23:33 | Make the hash table implementation a little smaller and faster. (check-in: f762f1effe user: drh tags: trunk) | |
18:48 | Have fts3 virtual table cursors free internal resources when they reach EOF, instead of waiting until the xClose method is called. (check-in: b6b14ab6c8 user: dan tags: trunk) | |
14:29 | Fix a problem in lsmtest causing one test to fail for multi-threaded LSM connections. (check-in: 2bce64c886 user: dan tags: trunk) | |
2017-07-04
| ||
19:34 | Add the count-of-view optimization when compiled using SQLITE_COUNTOFVIEW_OPTIMIZATION. (check-in: d1ba200234 user: drh tags: trunk) | |
19:23 | Fix further OOM handling cases in LSM. (check-in: ca757c8665 user: dan tags: trunk) | |
17:25 | Fix a couple of problems in handling OOM conditions within LSM. (check-in: 989a5c1745 user: dan tags: trunk) | |
12:50 | Fix a problem in the lempar.c Lemon template for YYSTACKDEPTH<=0 introduced by check-in [36e54cd8b1fb37] (check-in: 268a40f44c user: drh tags: trunk) | |
2017-07-03
| ||
21:09 | Fix compiler warnings in LSM1, especially in the test logic. (check-in: 9bd3be92b8 user: drh tags: trunk) | |
20:19 | Fix a bug in lsm queries on levels undergoing incremental an merge that contain range-delete markers. (check-in: b2c2487928 user: dan tags: trunk) | |
17:37 | Attempt to improve documentation on sqlite3_column_ and sqlite3_value_ interfaces. (check-in: 9111ac69bf user: drh tags: trunk) | |
09:00 | Add more tests for LSM log file recovery. Fix a problem in recovering log files that contain range deletes. (check-in: e34eafd4c5 user: dan tags: trunk) | |
2017-07-01
| ||
20:59 | Fix a memory management problem in lsm log recovery code. (check-in: dd55af30b4 user: dan tags: trunk) | |
15:21 | Enhance the RTree module to detect node truncation early and report an error. (check-in: 66de6f4a95 user: drh tags: trunk) | |
2017-06-30
| ||
23:46 | Update the carray() and remember() extension functions so that they user the new sqlite3_value_pointer() interface. (check-in: a99fa94db7 user: drh tags: bind-pointer) | |
23:09 | Add APIs for binding pointers that can be used by app-defined functions. (check-in: d9f4a831ba user: drh tags: bind-pointer) | |
20:11 | Improved documentation for sqlite3_value_type(). (check-in: 0db20efe20 user: drh tags: trunk) | |
19:22 | Add some assert() statements in the Win32 interface for lsm1. (check-in: d076d58ff1 user: mistachkin tags: trunk) | |
18:12 | Fix some minor typos in lsm1. (check-in: 0ef777d747 user: mistachkin tags: trunk) | |
11:44 | Avoid constantly freeing and reallocing small internal buffers associated with LSM cursors. Instead, allow them to persist for the lifetime of the cursor if they are LSM_SEGMENTPTR_FREE_THRESHOLD (default 1024) bytes or smaller. This change is based on research by Martijn Blaauw. (check-in: bacfe8cb3e user: dan tags: trunk) | |
2017-06-29
| ||
21:33 | In the command-line shell, add the -quote option to start up in quote mode. Enhance the ".mode" command so that it reports the current output mode if given no arguments. (check-in: 5e3f9ea5c4 user: drh tags: trunk) | |
21:11 | Remove the ".explain" command from the ".help" output of the command-line shell, though keep the implementation around for backwards compatibility. (check-in: 7782c04e9b user: drh tags: trunk) | |
20:31 | Some simple tests for the STMT virtual table. (check-in: a26e17ee56 user: drh tags: trunk) | |
20:23 | Avoid reading or writing the 32 locking bytes at the end of the first meta-page of an LSM database. (check-in: 3ed6877f89 user: dan tags: trunk) | |
20:13 | Avoid reading or writing the 32 locking bytes at the end of the first meta-page of an LSM database. (Closed-Leaf check-in: 2b5df3e8a8 user: dan tags: lsm-metapage-fix) | |
19:08 | Fix issues in the POSIX and Win32 interfaces for lsm1. (check-in: 38ec414166 user: mistachkin tags: trunk) | |
17:27 | Edit comments in sqlite.h.in used for generating documentation, to improve the description of the new sqlite3_prepare_v3() interfaces, and other miscellaneous cleanup. No changes to executable code. (check-in: 284707a7b3 user: drh tags: trunk) | |
16:51 | More updates to the Win32 interface for lsm1. (check-in: cae6476967 user: mistachkin tags: trunk) | |
15:57 | Correct typo in the Win32 interface for lsm1. (check-in: 1b4facb1ad user: mistachkin tags: trunk) | |
15:24 | Add interfaces sqlite3_prepare_v3() and sqlite3_prepare16_v3() with the extra prepFlags argument. Add the SQLITE_PREPARE_PERSISTENT option as one bit in that argument. Use the new option in FTS3, FTS5, and RTREE. (check-in: 03977248e6 user: drh tags: trunk) | |
15:13 | Further corrections to the Win32 interface for lsm1. (check-in: 4df6e2476c user: mistachkin tags: trunk) | |
14:33 | Rename the "stmts" virtual table to just "stmt" without the final "s". (check-in: adfdb80105 user: drh tags: trunk) | |
14:17 | A couple fixes for the Win32 interface for lsm1. (check-in: ebbd98e941 user: mistachkin tags: trunk) | |
13:41 | Add the stmts virtual table to testfixture builds. Add new compile-time options SQLITE_ENABLE_QPSG and SQLITE_ENABLE_STMTSVTAB to ctime.c (check-in: 60c628293a user: drh tags: trunk) | |
13:35 | Compilation fix for lsm1 using MSVC. (check-in: b7794cc5d5 user: mistachkin tags: trunk) | |
13:19 | Add the LSM1 extension. (check-in: 824e83274c user: drh tags: trunk) | |
12:54 | Implement xUnlink, xShmMap, and xShmUnmap for lsm1 on Win32. (Closed-Leaf check-in: 680cc064c9 user: mistachkin tags: lsm-vtab) | |
00:20 | Implement xRemap for lsm1 on Win32. Also, zero file handle when closing it. (check-in: 93c9aa7d9a user: mistachkin tags: lsm-vtab) | |
2017-06-28
| ||
21:36 | Implement xLock and xTestLock for lsm1 on Win32. (check-in: 9112117dad user: mistachkin tags: lsm-vtab) | |
2017-06-27
| ||
22:33 | Improve a local variable name. (check-in: d0f6973d93 user: mistachkin tags: lsm-vtab) | |
22:27 | Initial work on porting lsmtest to Win32. (check-in: 7e669d9bfa user: mistachkin tags: lsm-vtab) | |
18:15 | Implement xFullpath for lsm1 on Win32. (check-in: dbe9c8aa8d user: mistachkin tags: lsm-vtab) | |
06:28 | Minor corrections to the previous check-in. (check-in: e1cf8a78a0 user: mistachkin tags: lsm-vtab) | |
05:59 | Work in progress porting lsm1 to Win32. (check-in: 2017636e93 user: mistachkin tags: lsm-vtab) | |
2017-06-26
| ||
11:46 | Remove an invalid assert() from lsm test code. (check-in: ca8a7e995c user: dan tags: lsm-vtab) | |
08:24 | Fix another problem with multi-threaded mode in the LSM test code. (check-in: 461ced77d9 user: dan tags: lsm-vtab) | |
07:31 | Update a test case to reflect the fact that the lsm blocksize is configured in KiB. And defaults to 1024, not 2048. (check-in: d7015475f1 user: dan tags: lsm-vtab) | |
06:53 | Fix some compiler warnings in lsm_file.c. (check-in: 23f16382ed user: dan tags: lsm-vtab) | |
2017-06-15
| ||
15:06 | Merge the latest changes from trunk. (check-in: 60105c2253 user: drh tags: lsm-vtab) | |
2017-06-07
| ||
19:46 | Fix another lsmtest build problem. Add the "-trans BOOLEAN" option to "lsmtest speed2". (check-in: 61853bc171 user: dan tags: lsm-vtab) | |
2017-06-06
| ||
16:14 | Fix a threading problem in lsm test code. (check-in: 501238b9fb user: dan tags: lsm-vtab) | |
2017-06-02
| ||
23:44 | Merge all recent trunk enhancements. (check-in: 1d23294d88 user: drh tags: lsm-vtab) | |
23:32 | Add the --fsync flag to kvtest, and document the --nosync flag. (check-in: 7fdc78a672 user: drh tags: trunk) | |
19:31 | Work toward enhancing kvtest to measure write performance. (check-in: fc73e7d2f1 user: drh tags: trunk) | |
15:44 | Change the name of the OP_Seek opcode into OP_DeferredSeek for better clarity of function. No functional code changes. (check-in: ab33d299c7 user: drh tags: trunk) | |
13:16 | Fix a typo in the header comment of the "series.c" virtual table extension. (check-in: d637feb4e3 user: drh tags: trunk) | |
2017-06-01
| ||
16:13 | Add test code for LSM to the ext/lsm1/lsm-test directory. (check-in: bb7436e84a user: dan tags: lsm-vtab) | |
01:53 | Add a comment to the Lemon documentation regarding the security of the lemon.exe command-line tool. (check-in: 4c2458c190 user: drh tags: trunk) | |
2017-05-31
| ||
17:30 | Add the SQLITE_STMTSTATUS_MEMUSED opcode to sqlite3_stmt_status() for finding the heap memory usage by a single prepared statement. (check-in: c26cf978ee user: drh tags: trunk) | |
16:21 | Add the SQLITE_STMTSTATUS_MEMUSED option for sqlite3_stmt_status() that reports the amount of heap memory used for a single prepared statement. (Closed-Leaf check-in: b57d510465 user: drh tags: stmtstatus-memused) | |
16:09 | Add the experimental "stmts" virtual table for introspection of prepared statements. (check-in: cb4c5c66ab user: drh tags: trunk) | |
13:45 | Avoid allocating excess memory to the KeyInfo objects. (check-in: df78594351 user: drh tags: trunk) | |
13:27 | Remove a completely unnecessary memset() from the INSERT code generator. (check-in: 7d58836b9d user: drh tags: trunk) | |
11:24 | Very small performance increase in sqlite3VdbeSetNumCols(). (check-in: 8885b44610 user: drh tags: trunk) | |
03:20 | Reorder fields in the PgHdr object for a performance increase. (check-in: 326e63d71e user: drh tags: trunk) | |
02:58 | Avoid unnecessary memory zeroing during expression list allocation. (check-in: de28e6514a user: drh tags: trunk) | |
00:49 | Size and performance optimizations to sqlite3ResolveExprNames(). (check-in: af8c0fed93 user: drh tags: trunk) | |
2017-05-30
| ||
18:34 | Change sqlite3BtreeNext() and sqlite3BtreePrevious() so that they return SQLITE_DONE if they have already reached the end (or beginning) of the table. This gives a performance increase and size reduction. (check-in: e972a38608 user: drh tags: trunk) | |
18:14 | Omit a test of codepoint 0x202F (non-break narrow space) from the fts3 ICU tests. Different versions of ICU apparently handle this obscure codepoint slightly differently. (check-in: 69ae688982 user: dan tags: trunk) | |
2017-05-29
| ||
22:44 | More efficient error handling and reset in the binding mechanism. (check-in: 9a3e3b34ba user: drh tags: trunk) | |
17:59 | Omit unnecessary fstat() calls for determining the database file size, resulting in smaller and faster code. (check-in: dd61c73677 user: drh tags: trunk) | |
17:43 | Fix an issue with OPT_FEATURE_FLAGS in configure.ac. (check-in: 43ce3bd3a4 user: drh tags: trunk) | |
16:45 | Avoid unnecessary fstat() calls to determine the database size when transitioning into a read transaction on a rollback-mode database. (Closed-Leaf check-in: 6c3476b513 user: drh tags: io-opt) | |
14:57 | Avoid unnecessary calls to fstat() to determine the database file size. (check-in: c733a1dea3 user: drh tags: io-opt) | |
14:30 | Fix a typo in a comment. No code changes. (check-in: cb29b83888 user: drh tags: trunk) | |
14:26 | Optimizations to the Walker object and its methods to make the code a little smaller and to help it run a little faster. (check-in: 6854a34ed7 user: drh tags: trunk) | |
13:09 | Correctly initialize the iSelectId of FROM clause terms that are a self join of a reused materialized subquery. Without this, the EXPLAIN QUERY PLAN output for the query will identify the subquery using the uninitialized (and arbitrary) iSelectId. (check-in: 43c9ae371f user: drh tags: trunk) | |
03:48 | Improve shell help text for the '.open' command. (check-in: 7cc940a97e user: mistachkin tags: trunk) | |
2017-05-27
| ||
22:42 | Smaller and faster vdbeSorterCompareText(). (check-in: 542dc4c5eb user: drh tags: trunk) | |
2017-05-25
| ||
21:35 | Small performance increase and size decrease in the btreeInitPage() routine. (check-in: 6f415833e0 user: drh tags: trunk) | |
17:27 | Merge the LEFT JOIN query flattener fixes from 3.19.2. (check-in: 6513e4a121 user: drh tags: trunk) | |
16:50 | Version 3.19.2 (check-in: edb4e819b0 user: drh tags: release, version-3.19.2, branch-3.19) | |
11:39 | The SQLITE_EXTRA_IFNULLROW compile-time option causes OP_IfNullRow opcodes to be issued for references to the right-hand side table of *any* flattened join, not just LEFT JOINs. This puts extra stress on the OP_IfNUllRow opcodes for testing purposes. (check-in: 1a074c8a2b user: drh tags: branch-3.19) | |
00:28 | The TK_IF_NULL_ROW expression node must be treated as a variable that references the table Expr.iTable. Proposed fix for ticket [7fde638e94287d2c]. (check-in: b30a364a12 user: drh tags: branch-3.19) | |
00:12 | Increase the version number to 3.19.2 since ticket [7fde638e94287d2] is going to necessitate another patch release. (check-in: c315727acd user: drh tags: branch-3.19) | |
00:08 | The TK_IF_NULL_ROW expression node must be treated as a variable that references the table Expr.iTable. Proposed fix for ticket [7fde638e94287d2c]. (check-in: 77fc23013c user: drh tags: trunk) | |
2017-05-24
| ||
15:32 | Fix harmless compiler warnings in FTS5. (check-in: 57a60e959c user: mistachkin tags: trunk) | |
13:08 | Version 3.19.1 (check-in: f6d7b988f4 user: drh tags: release, version-3.19.1, branch-3.19) | |
04:18 | Fix a problem in STAT4 equality estimation for multi-column indexes introduced by check-in [3e0590dee0e68cc1599]. (check-in: cfb0d9e020 user: drh tags: trunk) | |
2017-05-23
| ||
20:00 | Remove the msvc.h header file from the autoconf tarball. (check-in: 84996aef68 user: drh tags: trunk) | |
19:35 | Do not use strcpy() in the (obsolete) amatch extension. In releasetest.tcl, run Apple tests using -Os instead of -O1 (check-in: 4440e42031 user: drh tags: trunk) | |
15:33 | Disable the LEFT JOIN flattening optimization for aggregate queries, as it does not currently work. Further fix for ticket [cad1ab4cb7b0fc344]. (check-in: 05ada74155 user: drh tags: branch-3.19) | |
15:21 | Disable the LEFT JOIN flattening optimization for aggregate queries, as it does not currently work. Further fix for ticket [cad1ab4cb7b0fc344]. (check-in: 44b21e35c9 user: drh tags: trunk) | |
12:44 | Ensure that the expression rewriter inside the query flattener decends into the substructure of the TK_IF_NULL_ROW operator. This is a continuation of the fix for ticket [cad1ab4cb7b0fc344]. (check-in: 28d2902d8f user: drh tags: branch-3.19) | |
12:36 | Ensure that the expression rewriter inside the query flattener decends into the substructure of the TK_IF_NULL_ROW operator. This is a continuation of the fix for ticket [cad1ab4cb7b0fc344]. (check-in: 941d8142b7 user: drh tags: trunk) | |
01:35 | Change the version number to 3.19.1. (check-in: 16656b8ff5 user: drh tags: branch-3.19) | |
01:30 | Remove a surplus semicolon, which was harmless on gcc and clang but caused compile-time errors on msvc. This only comes up using the encryption extension. (check-in: 4510d55d1e user: drh tags: branch-3.19) | |
01:29 | When flattening a query, make sure iTable attribute of TK_IF_NULL_ROW operators (that result from a prior flattening of a LEFT JOIN) are updated correctly. Fix for ticket [cad1ab4cb7b0fc344]. (check-in: 919d36e6ea user: drh tags: branch-3.19) | |
01:21 | When flattening a query, make sure iTable attribute of TK_IF_NULL_ROW operators (that result from a prior flattening of a LEFT JOIN) are updated correctly. Fix for ticket [cad1ab4cb7b0fc344]. (check-in: 92c178507d user: drh tags: trunk) | |
00:32 | Remove a surplus semicolon, which was harmless on gcc and clang but caused compile-time errors on msvc. This only comes up using the encryption extension. (check-in: 987a2b5537 user: drh tags: trunk) | |
2017-05-22
| ||
18:09 | Fix incompatibilities between the "sqldiff --changeset" command and the sessions module. Specifically, allow sessions to process changesets containing tables with zero operations on them and have sqldiff output the expected output for tables with multi-column primary keys. (check-in: 0bb23c4806 user: dan tags: trunk) | |
18:00 | Add the ".cd" command to the command-line shell. (check-in: 5fe28e15b1 user: drh tags: trunk) | |
17:39 | Increase the version number to 3.20 for the next release cycle. (check-in: ab471f61ef user: drh tags: trunk) | |
14:04 | Merge the last-minute 3.19.0 changes into trunk. (check-in: e6ba2a93a8 user: drh tags: trunk) | |
13:58 | Version 3.19.0 (check-in: 28a94eb282 user: drh tags: release, version-3.19.0, branch-3.19) | |
13:06 | Include the "msvc.h" header file in the amalgamation tarball. (check-in: 3dd6fe53c7 user: drh tags: branch-3.19) | |
08:04 | Fix a case where NULL was being passed to memcmp() following an OOM. This is probably not a real problem, as the number-of-bytes parameter was passed 0 in this case, but it was causing a santizer complaint. (check-in: 3ea2bad27e user: dan tags: trunk) | |
00:45 | When planning a query using sorting, resolve ties in the solver by selecting loop plans with the smaller unsorted cost. (check-in: f261678c90 user: drh tags: trunk) | |
00:27 | When planning a query using sorting, resolve ties in the solver by selecting loop plans with the smaller unsorted cost. (check-in: 962531e7c1 user: drh tags: branch-3.8.9) | |
2017-05-19
| ||
23:04 | Prevent a possible NULL pointer dereference in the OP_Found opcode that can follow an OOM error. Problem found by OSS-Fuzz. (check-in: 50ad60ded5 user: drh tags: branch-3.19) | |
22:51 | Prevent a possible NULL pointer dereference in the OP_Found opcode that can follow an OOM error. Problem found by OSS-Fuzz. (check-in: c2de178fe7 user: drh tags: trunk) | |
20:55 | Improved comments on one routine in the query planner. Improved diagnostic output for ".wheretrace". No production code changes. (check-in: 946b87a528 user: drh tags: trunk) | |
20:47 | Fix a documentation typo. No changes to code. (check-in: 68942a4fee user: drh tags: trunk) | |
20:46 | Branch for the 3.19 release. (check-in: f2b829ec52 user: drh tags: branch-3.19) | |
12:32 | Fix some problems in fts5 code detected by -fsanitize=undefined. (check-in: 35f721045d user: dan tags: trunk) | |
2017-05-18
| ||
18:17 | Fix a documentation typo. No changes to code. (Closed-Leaf check-in: cfa4aa2036 user: drh tags: doc-type) | |
2017-05-16
| ||
09:49 | Update the tool/warnings.sh script to automatically use the right options on OpenBSD. (check-in: 7940bff32a user: dan tags: trunk) | |
2017-05-15
| ||
15:12 | Fix the build so that it works again with SQLITE_OMIT_SUBQUERY. (check-in: bb0d928158 user: drh tags: trunk) | |
15:05 | Improvements to the omittest.tcl script. (check-in: 0f05d7e93a user: drh tags: trunk) | |
2017-05-12
| ||
14:05 | Make sure the 'sqlite3changegroup_*' functions are included in the SQLITE_APICALL handling. (check-in: a123cb9330 user: mistachkin tags: trunk) | |
2017-05-11
| ||
19:09 | Add another test case for the new authorizer functionality. (check-in: 339df63f40 user: dan tags: trunk) | |
18:42 | Enhance the json_extract() function to reuse parses of the same JSON when the function appears multiple times in the same query. (check-in: 3ba9e7ab79 user: drh tags: trunk) | |
18:14 | Do not save the state of an fts5 merge operation mid-way through an input term, even if no previous entry for that term has caused any output. Doing so may corrupt the FTS index. (check-in: 9a2de4f05f user: dan tags: trunk) | |
16:49 | Cache the JSON parse used by json_extract(). (Closed-Leaf check-in: 44ca6c2c46 user: drh tags: auxdata-cache) | |
15:20 | Negative N values in sqlite3_get_auxdata() and sqlite3_set_auxdata() can be used to access an auxiliary data cache over all functions in a single prepared statement. (check-in: ff5306752e user: drh tags: auxdata-cache) | |
13:43 | New requirements marks and documentation for the authorizer. (check-in: 3980ea0911 user: drh tags: trunk) | |
12:27 | Improvements to the sqlite3_set_authorizer() documentation. (check-in: 47629b1911 user: drh tags: trunk) | |
12:05 | Change the SQLITE_READ authorization call for unreferenced tables to use an empty string for the column name, as this is less likely to impact legacy authorization callbacks that assume column names are always non-NULL. (check-in: 4139953ab5 user: drh tags: trunk) | |
2017-05-10
| ||
19:42 | Rename fields of the internal AuxData object to make them unique and easier to search for. (check-in: 2be9850cef user: drh tags: trunk) | |
16:33 | Improved documentation for the SQLITE_READ authorizer callback. No code changes. (check-in: 92c5ea7047 user: drh tags: trunk) | |
16:12 | Invoke the SQLITE_READ authorizer callback with a NULL column name for any table referenced by a query but from when no columns are extracted. (check-in: 92ab1f7257 user: drh tags: trunk) | |
13:36 | Fix a couple of test scripts so that they work with -DSQLITE_DISABLE_FTS4_DEFERRED builds. (check-in: 30018d3106 user: dan tags: trunk) | |
12:58 | Avoid unnecessary codec operations on in-memory subjournals. (check-in: 199b2a8499 user: drh tags: trunk) | |
12:49 | Do not invoke codec macros for in-memory subjournals. (Closed-Leaf check-in: d2bb0066f7 user: drh tags: codecless-inmemory-subjournal) | |
2017-05-08
| ||
18:29 | Do not invoke codec macros when reading or writing an in-memory sub-journal. (check-in: 2c145ee6c9 user: dan tags: codecless-inmemory-subjournal) | |
2017-05-06
| ||
18:09 | Fix an obscure assertion fault that can follow an OOM. The problem was introduced by check-in [a1cf44763277b6c7]. (check-in: 04e7e5650e user: drh tags: trunk) | |
17:12 | Fix requirements marks and harmless compiler warnings. (check-in: 198ff4c01d user: drh tags: trunk) | |
2017-05-04
| ||
11:13 | Fix a collision of the "B0" identifier name between the termios.h header file and the SHA3 implementation in the shell. (check-in: b9a58daca8 user: drh tags: trunk) | |
2017-05-03
| ||
19:36 | Remove the unused "sqlite3_stack_used" TCL command from the test harness. (check-in: e24b73820c user: drh tags: trunk) | |
17:44 | Fix the sqlite3GetInt32() function so that it correctly returns 0 on a zero-length input string. (check-in: 05eba9e3a5 user: drh tags: trunk) | |
15:54 | Always enable URI filenames when compiling with SQLITE_HAS_CODEC. Also allow plaintext keys using the key= query parameter. (check-in: 31a51b4d16 user: drh tags: trunk) | |
2017-05-02
| ||
18:00 | Fix the fts3EvalAverageDocsize() routine so that it returns errors from sqlite3_reset() rather than always returning SQLITE_OK. (check-in: 430f539cbb user: drh tags: trunk) | |
17:54 | Reuse the same materialization of a view when that view appears in a query more than once, such as in a self-join. (check-in: 9e35c89dbe user: drh tags: trunk) | |
16:55 | Move terms of the HAVING clause that reference only columns in the GROUP BY clause over to the WHERE clause, resulting in a faster query plan. (check-in: 47cbb471d0 user: drh tags: trunk) | |