Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
100 check-ins occurring around 2d3a40c05c.
2023-09-15
| ||
20:28 | Simplifications and performance optimizations for the RTree extension. (check-in: 04a333f5 user: drh tags: trunk) | |
20:04 | Drop support for the view-scan optimization as it was causing multiple performance regressions. In its place, reduce the estimated row count for DISTINCT subsqueries by a factor of 8. (check-in: 796a65fa user: drh tags: branch-3.28) | |
19:51 | Drop support for the view-scan optimization (check-in [609fbb94b8f01d67]) as it was causing multiple performance regressions. In its place, reduce the estimated row count for DISTINCT subsqueries by a factor of 8. (check-in: f911f1c4 user: drh tags: trunk) | |
19:27 | Minor simplification to the DISTINCT output row count change. (Closed-Leaf check-in: 0738386d user: drh tags: rethink-viewscan) | |
19:00 | Do not reduce subquery output row count estimates due to DISTINCT until after the decision of whether or not to use an index for ORDER BY has been made. (check-in: 27390051 user: drh tags: rethink-viewscan) | |
18:36 | Allow expressions like "<tbl>.rowid" to refer to implicit rowid columns of tables in nested FROM clauses. (check-in: 59a1bbc6 user: dan tags: nested-from-rowid-expansion) | |
16:15 | Experimental: Assume that a DISTINCT clause on a subquery reduces the number of rows returned by that subquery by a factor of 8. (check-in: 5a940e47 user: drh tags: rethink-viewscan) | |
14:06 | The view-scan optimization was added to enhance the performance of one specific query, but it causes performance regressions on a host of others. Disable it (at least temporarily) in order to try to find an alternative way of fixing the one specific query that it was created to fix - an alternative way that does not cause performance problems for other unrelated queries. (check-in: d9625a9e user: drh tags: rethink-viewscan) | |
10:27 | Do not try to convert a double into an unsigned 64-bit integer, as that does not work on all platforms. A double can only be converted into a signed 64-bit integer. (check-in: 4ce5a03e user: drh tags: branch-3.43) | |
10:24 | Do not try to convert a double into an unsigned 64-bit integer, as that does not work on all platforms. A double can only be converted into a signed 64-bit integer. This is a fix for the problem reported in forum post 9f6db917e1c05d40. (check-in: ce339046 user: drh tags: trunk) | |
2023-09-14
| ||
20:56 | Tests for java versions of Fts5ExtensionApi.xSetAuxdata() and xGetAuxdata(). (check-in: 858fa236 user: dan tags: trunk) | |
20:02 | JNI: move XTokenizeCallback interface out of the Fts5 object - that level of indirection is a holdover from when those classes were in the core package. (check-in: 385ec085 user: stephan tags: trunk) | |
20:01 | Add a few more tests for Fts5ExtensionApi methods. (check-in: 4ce2a7e8 user: dan tags: trunk) | |
17:42 | Add a test for the java version of Fts5ExtensionApi.xRowid(). (check-in: 227facf8 user: dan tags: trunk) | |
16:45 | According to forum post 9f6db917e1, older MSVC compilers are unable to convert a double directly into an unsigned long long int, but must first go through a signed long long int. Work around this by restricting the range of doubles that are converted into long long integers so that only the lower 63 bits are used. (Closed-Leaf check-in: 0ab05871 user: drh tags: legacy-msvc-workaround) | |
16:02 | Document the use of OPFLAG_ISNOOP as the P2 argument on OP_Delete. No functional code changes. (check-in: d88f41b2 user: drh tags: trunk) | |
14:05 | Avoid a segfault in test program showdb.c that could occur when analyzing a corrupt db. (check-in: 6bf3e906 user: dan tags: trunk) | |
13:10 | Improve the sqlite3IntFloatCompare() routine so that it is testable. (check-in: bb221381 user: drh tags: trunk) | |
10:46 | Do not set the global bUseLongDouble flag based purely on sizeof(long double) being more than 8. That is a necessary but not sufficient condition to use the long double routines. Instead, do a run-time test to verify that long double really does provide extra precision. (check-in: aa999d49 user: drh tags: trunk) | |
01:46 | Reduce the number of calls to sqlite3_mprintf() made by RTree. (Closed-Leaf check-in: f158b7d4 user: drh tags: rtree-performance) | |
2023-09-13
| ||
21:32 | JNI: enable the fts5-customization bits by default but elide them from the docs until it's clearer where they're headed. Unrelated minor doc tweaks. (check-in: 21adb1ae user: stephan tags: trunk) | |
20:35 | Improved comments on the hasHighPrecisionDouble() routine. No changes to the underlying code. (Closed-Leaf check-in: 810c635c user: drh tags: runtime-longdouble-test) | |
20:06 | Determine at start time whether or not the underlying hardware supports high-precision long double computations. (check-in: 9a854b91 user: drh tags: runtime-longdouble-test) | |
17:30 | Omit the Reinsert algorithm from RTree. This causes most benchmarks to run faster, at the expense of having a slightly less dense and hence larger index (example: 33 entries/node versus 34 entries/node). (check-in: b3049a1d user: drh tags: rtree-performance) | |
17:16 | Doc corrections for the previous checkin. (check-in: 9ea0a9f3 user: stephan tags: trunk) | |
17:11 | Add JNI sqlite3_prepare_multi(), based on feedback. (check-in: fa1c1534 user: stephan tags: trunk) | |
14:07 | Performance optimization to cellContains() in RTREE. (check-in: 43cde22b user: drh tags: rtree-performance) | |
13:12 | In the ChooseLeaf algorithm of RTREE, do an initial pass through the cells of a node looking for solutions that involve no cell enlargement. Only look for the minimum cell enlargement if the enlargement is non-zero for all cells. This results in a performance improvement by reducing the number of calls to cellUnion(). (check-in: 59f0e239 user: drh tags: rtree-performance) | |
12:06 | Reduce the number of calls to cellArea() in the ChooseLeaf() algorithm of rtree. (check-in: c633fe6b user: drh tags: rtree-performance) | |
11:36 | Remove unused variable from the ChooseLeaf algorithm in r-tree. (check-in: fb2e4a80 user: drh tags: rtree-performance) | |
11:24 | Replace an if() condition in fts5 that is always true with an assert(). (check-in: 2170312c user: dan tags: trunk) | |
2023-09-12
| ||
23:21 | CLI .import to accept EOF in lieu of record terminator on last field of CSV (with multiple field records), per RFC 4180. forum post 5b21c25bdfa (check-in: 6f7842f5 user: larrybr tags: trunk) | |
23:11 | Fix a use-after-free error in fts5 that could occur when querying the "rank" column immediately after another connection changes its definition. (check-in: dfc42411 user: drh tags: branch-3.43) | |
18:36 | Fix a use-after-free error in fts5 that could occur when querying the "rank" column immediately after another connection changes its definition. forum post a2dd636330. (check-in: cb54c2da user: dan tags: trunk) | |
17:49 | Clarify some metrics output in Tester1.java. (check-in: 4ae7199a user: stephan tags: trunk) | |
17:06 | Re-add the fts5 decls to sqlite3-jni.h, removed in a rebuild without fts5 enabled. Need a better solution for this problem. (check-in: 6f269790 user: stephan tags: trunk) | |
17:03 | Rename JNI sqlite3_errmsg() to sqlite3_errmsg16() for overall constency with the text()/text16() family of functions. (check-in: c4a8d8a5 user: stephan tags: trunk) | |
16:33 | JNI readme fix. (check-in: 7dde000e user: stephan tags: trunk) | |
14:36 | 64-bit builds on Windows default to using UTF-8 output (as if the -utf8 command-line option had been specified.) 32-bit builds continue to use Windows code pages. There is also a new -no-utf8 command-line option to force the use of legacy code pages. (check-in: d8c7560e user: drh tags: trunk) | |
14:09 | Show the "(utf8)" tag on the CLI banner in Windows if in UTF8 mode. (Closed-Leaf check-in: 822699a4 user: drh tags: cli-utf8) | |
2023-09-11
| ||
20:09 | Add the -no-utf8 option to the Windows CLI to cause UTF-8 mode to be omitted. (check-in: 4cb799c6 user: drh tags: cli-utf8) | |
20:02 | Make the -utf8 option the default behavior in the CLI on Windows for 64-bit builds. (check-in: 8b3c306d user: drh tags: cli-utf8) | |
15:35 | Remove a harmlessly unused variable from FTS5. (check-in: 958d62c7 user: drh tags: branch-3.43) | |
15:28 | Refuse to build amalgamation-tarball and snapshot-tarball if the TEA version number disagrees with ./VERSION. (check-in: ad5cc375 user: drh tags: branch-3.43) | |
15:27 | Refuse to build amalgamation-tarball and snapshot-tarball if the TEA version number disagrees with ./VERSION. (check-in: 3308fdda user: drh tags: trunk) | |
15:20 | Merge in JNI's significant divergence since it adopted sqlite3_client_data(). (check-in: e8e07dbb user: stephan tags: trunk) | |
15:04 | Fix the TEA version number to be 3.43.1 (check-in: e00f1fed user: drh tags: branch-3.43) | |
14:55 | Add support for the sqlite3_get_clientdata() and sqlite3_set_clientdata() interfaces, to better support JNI. (check-in: 9806c0dd user: drh tags: trunk) | |
13:06 | Merge version 3.43.1 into the reuse-schema-3.43 subbranch. (check-in: d96586b5 user: drh tags: reuse-schema-3.43) | |
12:01 | Version 3.43.1 (check-in: 2d3a40c0 user: drh tags: release, branch-3.43, version-3.43.1) | |
2023-09-10
| ||
10:56 | Roll back [84e38341aeab] because a direct pointer comparison is ill-advised when the client-data string is dynamically allocated. (Closed-Leaf check-in: 7b884832 user: stephan tags: db-client-data) | |
2023-09-09
| ||
17:53 | Remove out-of-date comment regarding use of Parse.pConstExpr. (check-in: 71548f72 user: dan tags: trunk) | |
15:30 | Merge post-3.43.0 wal2 fixes into the bedrock-3.43 branch. (check-in: 48ecf390 user: drh tags: bedrock-3.43) | |
15:22 | Merge post-3.43.0 wal2 fixes into the wal2-3.43 sub-branch. (Leaf check-in: f82b1a0b user: drh tags: wal2-3.43) | |
11:20 | In the JNI build, emit a reminder to not check in the javac-generated sqlite3-jni.h when FTS5 is disabled because changes in that feature flag result in unnecessary and large diffs in checked-in generated code. (Closed-Leaf check-in: caceb52f user: stephan tags: jni-client-data) | |
07:52 | Disable FTS5 in the JNI build by default until it can be thoroughly tested. This is mainly so that the public-facing javadoc does not include it. (check-in: 6f25f02b user: stephan tags: jni-client-data) | |
2023-09-08
| ||
20:24 | Merge candidate patches for version 3.43.1 into the reuse-schema-3.43 branch. (check-in: c94be460 user: drh tags: reuse-schema-3.43) | |
20:19 | Merge the latest candidate patches for 3.43.1 into the bedrock-3.43 branch. (check-in: 4cefd16e user: drh tags: bedrock-3.43) | |
20:14 | Merge the latest 3.43.1 candidate patches into the wal2-3.43 branch. (check-in: b4491c9f user: drh tags: wal2-3.43) | |
20:09 | Merge patches queued for version 3.43.1 into the begin-concurrent-3.43 branch. (check-in: c7867985 user: drh tags: begin-concurrent-3.43) | |
19:35 | Fix a problem with fts5 secure-delete mode causing fts5 to corrupt its records. (check-in: fb24ca17 user: drh tags: branch-3.43) | |
19:32 | Fix a problem with fts5 secure-delete mode causing fts5 to corrupt its records. (check-in: cc0f82a4 user: dan tags: branch-3.42) | |
19:18 | Add extra tests for [4b60a1c3]. (check-in: 6357491b user: dan tags: trunk) | |
18:34 | Fix a problem with fts5 secure-delete mode causing fts5 to corrupt its records. (check-in: 4b60a1c3 user: dan tags: trunk) | |
16:49 | Fix a sanitizer runtime error in wal2 code (that (1 << 31) cannot be represented as an int). (check-in: 79a920c0 user: dan tags: wal2) | |
16:07 | Fix an apparently harmless buffer-overrun that could occur on this branch when committing multi-file transactions in wal2 mode. (check-in: ffe043ba user: dan tags: bedrock) | |
16:05 | Fix an apparently harmless buffer-overrun that could occur on this branch when committing multi-file transactions in wal2 mode. (check-in: dde9fcbd user: dan tags: wal2) | |
2023-09-07
| ||
17:54 | Add -DSQLITE_OMIT_LOOKASIDE=1 to the "Sanitize" configuration in testrunner_data.tcl. (check-in: a917ebf1 user: dan tags: trunk) | |
16:53 | Enhance the ./configure script and its associated Makefile.in so that the --with-linenoise=DIR argument cause the linenoise command-line editing library located in directory DIR to be linked with the sqlite3 CLI. (check-in: 6c7822b5 user: drh tags: trunk) | |
16:36 | Enhance the ./configure script and its associated Makefile.in so that the --with-linenoise=DIR argument cause the linenoise command-line editing library located in directory DIR to be linked with the sqlite3 CLI. (Closed-Leaf check-in: 9c4e4928 user: drh tags: wrong-branch) | |
14:10 | Fix a bad assert() in FTS3. (check-in: 61d2074c user: drh tags: branch-3.43) | |
14:04 | Fix a buffer overread in the sessions extension that could occur when processing a corrupt changeset. (check-in: 6009c871 user: drh tags: branch-3.43) | |
13:53 | Fix a buffer overread in the sessions extension that could occur when processing a corrupt changeset. (check-in: 0e4e7a05 user: dan tags: trunk) | |
13:48 | Fix a harmless compiler warning in the sqldiff.c utility. (check-in: e6390a65 user: drh tags: trunk) | |
13:43 | Add the "tool-zip" makefile target with the intention of using it to build the "sqlite-tool-*.zip" deliverables. deliverables. (check-in: 06e4af49 user: drh tags: trunk) | |
11:48 | Version number to 3.43.1. (check-in: fdfe4d60 user: drh tags: branch-3.43) | |
02:13 | Fix a problem with the new xIntegrity method for virtual tables, and also fix a bad assert() in FTS3 that was found by the new xIntegrity method. (check-in: 52bbf44f user: drh tags: trunk) | |
2023-09-06
| ||
17:07 | Add documentation to sqlite3_get/set_clientdata() to make it clear that these are security-sensitive interfaces that should not be exposed to potential attackers. (check-in: 2837061d user: drh tags: db-client-data) | |
16:51 | Add the xIntegrity method to the sqlite3_module object, thus enabling PRAGMA integrity_check to operate on virtual tables. Make use of this new method in the FTS3/4, FTS5, and RTREE virtual tables. (check-in: 2f08e7d2 user: drh tags: trunk) | |
16:22 | Fix unreachable branches. (Closed-Leaf check-in: f00fc163 user: drh tags: vtab-integrity-check) | |
14:00 | Change the xIntegrity virtual table method signature so that it returns an integer error code and writes the error message into a parameter. (check-in: f1d4024a user: drh tags: vtab-integrity-check) | |
12:52 | Add the xIntegrity method to the sqlite3_module object. Implement this method in RTREE, FTS3/4, and FTS5 so that "PRAGMA integrity_check" also verifies the correctness of shadow tables associated with those virtual tables. (check-in: 17bede8c user: drh tags: vtab-integrity-check) | |
07:39 | Move Java-side FTS5 bits into the fts5 subpackage to (A) make it easy to optionally bundle it (or not) and (B) set a precedent for puting extension APIs in their own package. (check-in: 4c773e1f user: stephan tags: jni-client-data) | |
06:54 | Java FTS5 API renaming, for consistency. (check-in: 5e4cc9f3 user: stephan tags: jni-client-data) | |
2023-09-05
| ||
15:43 | Add debugging routines to check each page for corruption after it it is manipulated by the fts5 secure-delete code, and to log said page via sqlite3_log() if it is found to be corrupt. (Leaf check-in: da4b7385 user: dan tags: fts5-secure-delete-debug) | |
15:03 | Fix a few SQLITE_MISUSE returns so that they call sqlite3MisuseError(). (check-in: 93f74490 user: drh tags: trunk) | |
2023-09-04
| ||
16:48 | Add tests for, and source code comments to, fts5. No changes to code. (check-in: b12afff4 user: dan tags: trunk) | |
12:50 | Do not make assumptions about the byteorder of PowerPC processors. (check-in: 4a2498fe user: drh tags: trunk) | |
06:15 | In JNI routines which set an OutputPointer.value, always do so, even on error, for consistency. (check-in: d0df62f0 user: stephan tags: jni-client-data) | |
05:58 | Remove several unnececessary functions. (check-in: 84433c72 user: stephan tags: jni-client-data) | |
05:34 | Move an internal utility class out of the public Java API. (check-in: 8e1ae319 user: stephan tags: jni-client-data) | |
04:23 | Expose sqlite3_bind_parameter_name() to JNI. Extend the definition of null for the @NotNull annotation, to consider closed/finalized Java-side handles wrapping C-side resources as null for its purposes. (check-in: fc34e2bf user: stephan tags: jni-client-data) | |
2023-09-03
| ||
15:01 | In mksqlite3c.tcl, pass EXTRA_SRC files through verbatim instead of applying the post-processing needed for sqlite's own source files. (check-in: 44f3d7d0 user: stephan tags: extra-src) | |
14:41 | Add the missing sqlite3_blob_read() to [9a9945c405cf]. Elaborate on the definition of the @NotNull annotation. (check-in: 85952b4a user: stephan tags: jni-client-data) | |
14:13 | Expose the sqlite3_blob_...() APIs to JNI. (check-in: 9a9945c4 user: stephan tags: jni-client-data) | |
12:46 | Expose sqlite3_system_errno() to JNI. (check-in: bac54da2 user: stephan tags: jni-client-data) | |
12:32 | Expose sqlite3_stmt_status() to JNI. (check-in: d266acc2 user: stephan tags: jni-client-data) | |
12:17 | Expose sqlite3_randomness() to JNI. (check-in: de969224 user: stephan tags: jni-client-data) | |
11:58 | Expose sqlite3_get/set_auxdata() to JNI. (check-in: 0de32717 user: stephan tags: jni-client-data) | |
11:16 | Expose sqlite3_get_autocommit() to JNI. (check-in: 67391c65 user: stephan tags: jni-client-data) | |