SQLite

Timeline
Login

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

396 check-ins using file tool/speed-check.sh version e8d20cc2

2024-05-25
23:17
Ensure that sqlite3ViewGetColumnNames() returns non-zero on any error. (Leaf check-in: 01ead0a2 user: drh tags: branch-3.46)
23:13
Ensure that sqlite3ViewGetColumnNames() returns non-zero on any error. dbsqlfuzz d9de2a508f693e5a228d7a01d5341e3c64326fbb. (Leaf check-in: 57aeb3a2 user: drh tags: trunk)
2024-05-24
23:44
Ensure that queries like "SELECT func(a) FROM tab GROUP BY 1" only call the func() function once per row. Fix for the issue described in forum thread bf8f43aa522c2299. (check-in: 663f5dd3 user: drh tags: trunk)
21:38
Fix another fts5 problem related to unknown tokenizer modules. (check-in: e78b7783 user: dan tags: trunk)
21:23
Fix a crash in fts5 that could occur within "PRAGMA integrity_check" if the specified tokenizer was unknown. (check-in: 061d7b69 user: dan tags: trunk)
20:18
When omitting result columns for the vector-IN optimization, make sure that any result-set column references in both the ORDER BY clause and the GROUP BY clause are updated appropriately. (Closed-Leaf check-in: f890812d user: drh tags: group-by-consistency)
18:31
Ensure that queries like "SELECT indeterministic(a) FROM tbl GROUP BY 1" invoke the indeterministic function only once for each row of tbl. (check-in: 4555d665 user: dan tags: group-by-consistency)
14:16
Add the stmtrand() extension function for use in testing. (check-in: 5c97a5b9 user: drh tags: trunk)
13:33
Fix an "ifcapable" expression in values.test. (check-in: b8442d2a user: dan tags: trunk)
2024-05-23
23:34
Increase the version number to 3.46.1 (check-in: 170e3a91 user: drh tags: branch-3.46)
23:30
Fix the window-function group_concat() so that it returns an empty string if it has one or more empty string inputs. (check-in: 7fe11274 user: drh tags: branch-3.46)
23:26
Fix the window-function group_concat() so that it returns an empty string if it has one or more empty string inputs. This fixes a bug introduced by [c6da39115d3e2b0f] on 2019-03-26 (version 3.28.0) and reported by [forum/forumpost/bf8f43aa522c2299|forum post bf8f43aa522c2299]. (check-in: cec6bb3f user: drh tags: trunk)
15:05
Merge various enhancements that were being held on a branch until after the 3.46.0 release. (1) Allow arbitrary expressions in the second argument to the RAISE() function. (2) Performance optimization to the OP_OpenRead and OP_OpenWrite bytecode operators. (3) Allow FTS5 tables to be dropped even if the associated tokenizer is not available. (4) Performance optimizations in FTS5. (5) Generate better bytecode to improve performance of ONEROW scans. (check-in: 1e5385ff user: drh tags: trunk)
14:58
Increase the version number to 3.47.0 to begin the next development cycle. (check-in: 20e228a2 user: drh tags: trunk)
14:09
Version 3.46.0 for the reuse-schema branch (Leaf check-in: 8f6b8594 user: drh tags: reuse-schema)
14:09
Version 3.46.0 for the bedrock branch (Leaf check-in: 483d112a user: drh tags: bedrock)
14:05
Version 3.46.0 for the wal2 branch. (Leaf check-in: fdc0e148 user: drh tags: wal2)
14:04
Version 3.46.0 for the begin-concurrent branch (Leaf check-in: e3f8c70e user: drh tags: begin-concurrent)
13:25
Version 3.46.0 (check-in: 96c92aba user: drh tags: trunk, release, major-release, version-3.46.0)
2024-05-21
17:37
Ensure an sqlite_dbdata cursor is properly reset before being used again, even if it has already encountered database corruption. (check-in: 3210e1ca user: dan tags: trunk)
15:24
Merge the latest trunk enhancements into the bedrock branch. (check-in: c65a82f1 user: drh tags: bedrock)
15:20
Merge the lates trunk enhancements into the reuse-schema branch. (check-in: 97e0af88 user: drh tags: reuse-schema)
15:20
Merge the latest trunk enhancements into the wal2 branch. (check-in: d4217d9f user: drh tags: wal2)
15:17
Merge the latest trunk enhancements into the begin-concurrent branch. (check-in: 5c1c8651 user: drh tags: begin-concurrent)
14:54
Do not attempt to run mmapcorrupt.test with builds that do not support mmap(). (check-in: 54a49f18 user: dan tags: trunk)
13:34
Fix harmless compiler warnings. (check-in: aab59f9f user: drh tags: trunk)
11:11
Fix the -> and ->> operators so that when the RHS is a string that looks like a number, it is still treated as a string, because that is what PG does. Forum post 9e52cdfe15c3926e. (check-in: de8182cf user: drh tags: trunk)
2024-05-20
19:59
Omit a capi3 test from 2007 that depends upon undefined behavior (UAF). (check-in: a60a0b75 user: drh tags: trunk)
19:51
Do not run the tests in recovercorrupt4.test with the inmemory_journal permutation. (check-in: c13205d6 user: dan tags: trunk)
17:52
Improvements to the documentation of sqlite3_vtab_distinct(). Do not allow sqlite3_vtab_distinct() to return 2 or 3 if the rowid column of the virtual table is used, as doing so might lead to an incorrect answer. (check-in: 08058d66 user: drh tags: trunk)
16:00
Improvements to the sqlite3_vtab_distinct() documentation. (Closed-Leaf check-in: 6ee041d3 user: drh tags: vtab-distinct-fix)
2024-05-18
20:00
Fix the definition of sqlite3_vtab_distinct() such that return codes 2 and 3 mean that all rows must be distinct over "colUsed" which is a superset of "aOrderBy". Also, disallow return codes 2 and 3 if the rowid of the virtual table is accessed. (check-in: 922731ce user: drh tags: vtab-distinct-fix)
15:21
JS error message and doc typos reported in the forum. No code changes. (check-in: 1a073f9a user: stephan tags: trunk)
2024-05-17
22:51
Improvements to a VDBE comment. No changes to anything that affects queries. (check-in: c2188da8 user: drh tags: trunk)
14:26
Merge branches "fix-onerow-opt", "faster-openread", "fts5-delay-tokenizer" and "enhanced-raise", each containing minor enhancements prepared for 3.46, into this branch. (Closed-Leaf check-in: 6dc64721 user: dan tags: pending-3.46)
13:56
Fix a problem in where.c preventing SQLite from identifying some ONEROW scans, leading to (slightly) sub-optimal byte-code. (Closed-Leaf check-in: b41e7120 user: dan tags: fix-onerow-opt)
2024-05-16
16:22
Add extra tests for the error messages generated by sqlite3_declare_vtab() when passed something other than a well-formed CREATE TABLE statement. (check-in: 5fa0c2a0 user: dan tags: trunk)
15:57
The sqlite3_declare_vtab() interface should return SQLITE_ERROR, not SQLITE_MISUSE when a nonsense SQL input is provided, for legacy compatibility. Forum post 4b875d7b2a. (check-in: 1ec2961a user: drh tags: trunk)
00:20
Fix a case where an error in the ORDER BY clause of an aggregate function that is part of a trigger body might cause an assert() to fail in an ALTER TABLE command. dbsqlfuzz 899d9255555a07f1fe11a0bb51a2fe197b3150cd (check-in: 953271f2 user: drh tags: trunk)
2024-05-15
23:36
Back out the assert() fixes from [a9443dbfbe25e588] and [3afaeac56dff58db] (but keep the corresponding test cases) because they are both made obsolete by the previous fix on this branch. (Closed-Leaf check-in: 3120c7c8 user: drh tags: alter-table-fix)
21:38
Fix a case where an error in the ORDER BY clause of an aggregate function that is part of a trigger body might cause an assert() to fail in an ALTER TABLE command. (check-in: b79a8be3 user: dan tags: alter-table-fix)
15:06
If there is no data available for a field in a corrupted record, but the type is available, have the recovery extension substitute 0, 0.0, '' or X'' instead of a NULL value. (check-in: 8eb326c2 user: dan tags: trunk)
2024-05-14
19:41
Optimize queries against fts5vocab(row) tables that do use the "cnt" column. (Closed-Leaf check-in: 96a591c2 user: dan tags: fts5-delay-tokenizer)
19:09
Avoid calculating the value of the "cnt" column for fts5vocab(row) queries that do not use it. (check-in: c0c91e30 user: dan tags: fts5-delay-tokenizer)
17:16
Have fts5 tables delay initializing the tokenizer until it is first used in all cases where the tokenizer is not "trigram". (check-in: ca4fdcb8 user: dan tags: fts5-delay-tokenizer)
2024-05-13
20:06
Allow existing fts5 tables to be dropped even if the associated tokenizer is not available. (check-in: 69ef47ee user: dan tags: fts5-delay-tokenizer)
2024-05-12
00:26
Optimization in the implementation of OP_OpenRead and OP_OpenWrite (Closed-Leaf check-in: 3a256a16 user: drh tags: faster-openread)
2024-05-11
16:44
Add tests to bestindexC.test. No changes to code. (check-in: 58d938c0 user: dan tags: trunk)
2024-05-10
18:24
Fix aggregate function processing to correctly deal with OOMs inside of sqlite3ParserAddCleanup(). (dbsqlfuzz b2d11ca70e55ee8bde48ae0b53fa3e9355812f95). Also add improved testing support by causing sqlite3FaultSim(300) to simulate an OOM inside of sqlite3ParserAddCleanup() and by adding improved fault-sim support to the CLI. (check-in: c6fd70b3 user: drh tags: trunk)
18:15
Add a test for the fixes on this branch. (Closed-Leaf check-in: 716088d6 user: dan tags: cleanup-testing)
18:10
The sqlite3FaultSim(300) error from [1e8863909ac369e5] must be treated as an OOM, since it simulates an OOM. Also fix deferred deletion of Expr objects from [a53bdd311c4154fd] so that it does not corrupt the AST if an OOM occurs. (check-in: c623d9a5 user: drh tags: cleanup-testing)
17:05
Improvements to the ".testctrl fault_install" command in the CLI. (check-in: 48b34ded user: drh tags: cleanup-testing)
16:28
Add an sqlite3FaultSim(300) call to the sqlite3ParserAddCleanup() routine for more precise simulation of OOM errors in that routine. (check-in: 1e886390 user: drh tags: cleanup-testing)
09:26
Squelch two OMIT-flag-related warnings reported in forum post 388243d9882067a9. No functional changes. (check-in: 1d671605 user: stephan tags: trunk)
2024-05-09
18:39
Add a maintenance-reminder comment explaining [dc58ed65492857]. (check-in: 857f5d7f user: stephan tags: trunk)
18:00
Reformulate a for-in loop to a for loop in wasm.xWrap() to work around downstream frameworks modifying the JS Array prototype, as reported in forum post b549992634b55104. (check-in: dc58ed65 user: stephan tags: trunk)
2024-05-08
17:57
Rename a shadowed local var, as reported in forum post 2ba65af31. No functional changes. (check-in: d030c87c user: stephan tags: trunk)
17:42
Allow arbitrary expressions as the second argument to RAISE(). (Closed-Leaf check-in: 003e1c8c user: drh tags: enhanced-raise)
11:51
Fix a hyperlink typo in session documentation. (check-in: 42d67c6f user: drh tags: trunk)
2024-05-07
19:45
Fix bug in test script sessionconflict.test. (check-in: e29decb8 user: dan tags: trunk)
17:58
Ensure successive scans of pragma virtual tables return consistent rowids. (check-in: 6838bf3a user: dan tags: trunk)
13:20
Fix a compilation error in test/lemon-test01.y, reported in forum post f0ad095705. (check-in: 576b68c8 user: stephan tags: trunk)
2024-05-06
20:21
When extracting links from sqlar archives, clobber any existing file or link, and do not call utimes() to set the timestamp - it looks through the link and operates on the target. (check-in: 2bf8c3f9 user: dan tags: trunk)
20:18
Omit redundant unary + operators from the AST. (check-in: f81cc149 user: drh tags: trunk)
19:04
Avoid unnecessary recursion in sqlite3ExprDeleteNN(). This complicates the code, but it is needed to prevent nuisance "stack overflow" reports from OSSFuzz while it is running the latest ASAN. (check-in: 70abc144 user: drh tags: trunk)
17:55
Add new sessions API sqlite3changegroup_add_change(). (check-in: 5eaab43c user: dan tags: trunk)
17:32
Edit comments in sqlite3session.h. (Closed-Leaf check-in: 49f84878 user: dan tags: changegroup-add-change)
17:21
Add tests for sqlite3changegroup_add_change(). (check-in: 9d67bedc user: dan tags: changegroup-add-change)
11:27
Fix problem in sqlar_uncompress() when being used to extract symbolic links. (check-in: 4d90c3f1 user: dan tags: trunk)
2024-05-05
10:00
Add comment to the test/lemon-test01.y file indicating that it does not work and is for historical reference only. (check-in: 9bdf73a8 user: drh tags: trunk)
2024-05-04
21:10
Add new sessions API sqlite3changegroup_add_change(). (check-in: 73647db1 user: dan tags: changegroup-add-change)
16:50
Add test cases to test/in7.test. No code changes. (check-in: 8c3086f9 user: dan tags: trunk)
15:10
Assume that an SQL function is able to return a subtype if any of its arguments are SQL functions that are able to return subtypes. This closes a corner-case hole in the patch at [ba789a7804ab96d8]. (check-in: b36d499e user: drh tags: trunk)
15:04
Assume that a function is able to return a subtype if either (1) it is itself marked with SQLITE_RESULT_SUBTYPE, or (2) one of its arguments is a function that is able to return a subtype. This check-in backs out the code changes from the previous two on this same branch, but keeps the test cases from the previous two. (Closed-Leaf check-in: f16b200f user: drh tags: func-rw-subtype)
11:31
Functions that pass through the sqlite3_value of one of their arguments must also be marked as SQLITE_RESULT_SUBTYPE, in case one of their arguments has a subtype. (check-in: 2f9fba93 user: drh tags: func-rw-subtype)
2024-05-03
19:35
The coalesce(), ifnull(), and iif() functions pass through subtype values from their arguments, and hence need to have the SQLITE_RESULT_SUBTYPE flag set. (check-in: cdd1610c user: drh tags: func-rw-subtype)
2024-05-02
19:22
Fix a problem with queries that use "IN(...) ORDER BY ... NULLS LAST" or similar introduced by [d7648e21605a0b38]. (check-in: 26b2d939 user: dan tags: trunk)
18:16
Avoid an assert failure accessible only when internal testing features are enabled. dbsqlfuzz crash-66abacda6bca6bd531e25595e8c8068e2c710326.txt. (check-in: fbc446da user: dan tags: trunk)
17:46
Fix a case where the recovery extension might write uninitialized data to the recovered database. (check-in: 8519b4d0 user: dan tags: trunk)
14:48
Avoid an OP_Next in cases where an IN(...) query against a UNIQUE index may return at most 1 row. (check-in: d7648e21 user: dan tags: trunk)
12:14
Fix the CREATE TABLE performance issue described by forum thread 4cf69794d9dfff7c in two different ways: (1) Omit the call to PRAGMA integrity_check('X') that was being done after CREATE TABLE "X" because the result was being ignored and the integrity_check was not doing anything other than burning CPU cycles. (2) Do not interpret the argument to PRAGMA integrity_check as a number if it is in fact a string that looks like a number. (check-in: 71f08b91 user: drh tags: trunk)
12:00
Add a test case to the fix to PRAGMA integrity_check in the previous check-in. (Closed-Leaf check-in: 39a57b59 user: drh tags: faster-create)
11:52
Do not accept a string that looks like a number used as an argument to PRAGMA integrity_check as a number. Treat it as a table name that just happens to look like a number. (check-in: b04e7a23 user: drh tags: faster-create)
11:51
Omit the OP_SqlExec to "PRAGMA integrity_check" added by [348fa7aaf7958b3f] because it is a no-op. Even if the integrity_check failes, the CREATE TABLE is stull successful. The OP_SqlExec just burns CPU cycles for no reason. (check-in: 532795ac user: drh tags: faster-create)
2024-05-01
19:48
Avoid an OP_Next in cases where an IN(...) query against a UNIQUE index may return at most 1 row. (Closed-Leaf check-in: 560f6415 user: dan tags: unique-in-opt)
16:25
Fix another problem in the recovery extension where a corrupt sqlite_schema table could lead to excessive memory consumption. (check-in: 1c7e33a8 user: dan tags: trunk)
14:42
Fix a problem causing the recovery extension to use excessive memory and CPU time in some cases. (check-in: 8de85170 user: dan tags: trunk)
2024-04-30
19:34
Avoid 32-bit overflow when calculating ncycle for ".scanstats vm". (check-in: 2858efa0 user: dan tags: trunk)
2024-04-26
19:10
Fix issues in LIMIT/OFFSET support for virtual tables. The first problem was reported by forum post c243b8f856. That report prompted an enhancement to the generate_series() (also included in this merge) which in turn identified other similar issues. (check-in: 5f6c079d user: drh tags: trunk)
18:42
Internal JS doc touchups. No functional changes. (check-in: 626b997b user: stephan tags: trunk)
18:24
Additional internal docs in the wasm utilities. (check-in: 5a2245a9 user: stephan tags: trunk)
18:13
Allow virtual table implementations to handle OFFSET but not LIMIT, but not LIMIT but not OFFSET. (Closed-Leaf check-in: 90e5c822 user: dan tags: vtab-limit-fix)
17:19
Have where.c ignore any plan from a virtual table that tries to use LIMIT/OFFSET without also using all WHERE constraints. (check-in: 7d305964 user: dan tags: vtab-limit-fix)
17:09
Further fixes and improvements to the generate_series() enhancements on this branch. (check-in: b7d9bd7e user: drh tags: vtab-limit-fix)
14:36
Use hex flag masks rather than decimals in ext/misc/series.c. (check-in: a94e2cd0 user: drh tags: vtab-limit-fix)
14:32
Fix a problem allowing a LIMIT constraint to be passed to a virtual table in cases where there exist WHERE terms that cannot also be passed. (check-in: 72c8ed96 user: dan tags: vtab-limit-fix)
13:38
Fix handling of LIMIT and OFFSET in virtual tables that are part of a compound SELECT. (check-in: 40421c1c user: drh tags: vtab-limit-fix)
13:30
Enhance the generated_series() table-valued-function to respond to LIMIT and OFFSET. Use this to add new test cases for LIMIT and OFFSET on virtual tables in a compound SELECT. (check-in: 408d47ec user: drh tags: vtab-limit-fix)
12:01
Add test demonstrating the problem at forum post c243b8f856. No fix yet. (check-in: 1685495c user: dan tags: vtab-limit-fix)
11:32
Fix the TreeView display of a LIMIT clause on a compound query. Debugging code only - not part of production builds. (check-in: 38f1e5ce user: drh tags: trunk)
2024-04-25
23:26
Improvement to the way that affinity is determined for columns of a compound subquery. The affinity is the affinity of the left-most arm of the compound subquery that has an affinity other than NONE, adjusted to accommodate the data types coming out of the other arms. (check-in: e6df846f user: drh tags: trunk)
17:52
Further improvements to the computation of affinity for compound subqueries: make sure that the selected affinity is compatible with a literal values in arms to the left of the arm that is used to determine affinity. (Closed-Leaf check-in: bbdf22e3 user: drh tags: compound-subquery-affinity)
16:55
Back out the previous change on this branch. In its place: Determine the affinity of a subquery by the left-most arm of the subquery that has an affinity other than NONE. In other words, scan from left to right looking for an arm of the compound subquery with an affinity of BLOB, TEXT, INTEGER, or REAL and pick the first one found. Or stay with NONE if no arm has a defined affinity. Test cases added. (check-in: b8ec8511 user: drh tags: compound-subquery-affinity)
06:52
wasm: add a makefile comment about the static sqlite3.h/c version info injected into the JS files possibly differing from the runtime-emited version info when a user provides their own sqlite3.c. (check-in: d99a01a0 user: stephan tags: trunk)
2024-04-24
19:49
An experimental minor tweak to the way affinities are computed for compound subqueries, when the actual affinity is ambiguous. (check-in: 779723ad user: drh tags: compound-subquery-affinity)
16:36
If a RETURNING clause contains a subquery that references the table that is being updated in the statement that includes the RETURNING clause, then mark the subquery as correlated so that it is recomputed for each result and not just computed once and reused. See forum post 2c83569ce8945d39. (check-in: 9ea6bcc8 user: drh tags: trunk)
14:18
Fix a case where a corrupt stat4 record could go unrecognized due to integer overflow. (check-in: 240a4a48 user: dan tags: trunk)
11:51
Add comments linking the assert() added in [cef4d9e3ba586735] to the places where the precondition that the assert() tests are actually required. (check-in: 6f0e7e19 user: drh tags: trunk)
11:40
Avoid slowdown when exprNodeIsConstant() is called on deeply nested SQL functions. (check-in: 1c0b0345 user: dan tags: trunk)
11:21
Ensure that temporary SrcItem objects created by trigger processing have either SrcItem.zName or SrcItem.pSelect defined. Every SrcItem should have one or the other. (check-in: cef4d9e3 user: drh tags: trunk)
06:43
Remove an undocumented feature from the JS Worker1.open API because it relies on broken behavior which has no VFS-agnostic workaround. (check-in: 3a640038 user: stephan tags: trunk)
2024-04-23
12:02
Fix a problem with vector IN operators used with an index where the affinities and collations for the various vector terms are different. (check-in: 86e8c782 user: drh tags: trunk)
11:56
Fix an uninitialized variable problem in the new code on this branch. (Closed-Leaf check-in: 85625b38 user: dan tags: vector-in-fix)
06:49
Remove some dead WASM-side code. (check-in: 0a07ee27 user: stephan tags: trunk)
06:36
Minor touchups in JS docs and exception messages. (check-in: 9e7fc937 user: stephan tags: trunk)
05:38
When running the 'dist' target in ext/wasm for an SEE-capable build, ensure that the resulting zip file and directory name include '-see'. (check-in: 04c552b1 user: stephan tags: trunk)
2024-04-22
20:09
Add extra tests for aborting conflicts in the sessions module. (check-in: b4a6d326 user: dan tags: trunk)
17:03
Minor cleanups to [8fbda563d2f5]. (check-in: 5ee2594b user: stephan tags: trunk)
16:46
Extend the JS/WASM SEE build support by (A) filtering SEE-related bits out of the JS when not building with SEE and (B) accepting an optional key/textkey/hexkey option to the sqlite3.oo1.DB and subclass constructors to create/open SEE-encrypted databases with. Demonstrate SEE in the test app using the kvvfs. This obviates the changes made in [5c505ee8a7]. (check-in: 8fbda563 user: stephan tags: trunk)
13:31
Extra robustness in the code that causes cursors to return NULL when they are participating in an OUTER JOIN. (check-in: 672c2869 user: drh tags: trunk)
11:48
For sqlite3.oo1.DB JavaScript classes, bypass execution of any on-open() SQL in SEE-capable builds because it would necessarily run before the client has an opportunity to provide their decryption key, which would leave the db handle in an unusable state and cause the ctor to throw. This currently affects only the OPFS VFSes. We may want to consider extending the ctor options object to optionally accept an SEE key and apply it when opening the db. (check-in: 5c505ee8 user: stephan tags: trunk)
00:42
Continuation of the fix at [8c0f69e0e4ae0a44]: If a viaCoroutine FROM clause term is participating in a RIGHT or FULL JOIN, we have to create an always-NULL pseudo-cursor for that term when processing the RIGHT join. dbsqlfuzz 6fd1ff3a64bef4a6c092e8d757548e95698b0df5. (check-in: e1040e51 user: drh tags: trunk)
2024-04-21
23:35
Fix an indentation problem in debugging routine sqlite3ShowSrcList(). No changes to production code. (check-in: 5560df27 user: drh tags: trunk)
2024-04-20
19:19
Slight performance improvement for the new code on this branch. (check-in: 500c67f1 user: dan tags: vector-in-fix)
15:14
Fix a problem where an expression like (a, b) IN (SELECT ...) might not use an index on (a, b) if the affinities and collation sequences of "a" and "b" are not identical. (check-in: 4d870fd8 user: dan tags: vector-in-fix)
2024-04-18
16:11
Correct handling of OUTER JOIN when on or the other operand is a subquery implemented using the VALUES-as-coroutine optimization. dbsqlfuzz bde3bf80aedf25afa56e2997a0545a314765d3f8. (check-in: 8c0f69e0 user: drh tags: trunk)
2024-04-17
19:48
Prevent tokenize-blobs from being used with non-contentless tables. Fix some other issues with the new code on this branch. (Leaf check-in: 6a640ea4 user: dan tags: fts5-tokenize-blob)
2024-04-16
14:23
Add some tests for the fts5 tokenize-blob functionality on this branch. (check-in: c2f9d125 user: dan tags: fts5-tokenize-blob)
2024-04-15
20:43
The first assert() added in [0ebc65481f4a3e79] is not necessarily true in a corrupt database file. So add a term to make it true. (check-in: 6b21cccd user: drh tags: trunk)
20:24
Add experimental way to specify an alternative tokenizer when writing to or querying an fts5 table. (check-in: 6c51c9c6 user: dan tags: fts5-tokenize-blob)
14:45
Merge all of the latest trunk enhancements into the bedrock branch. (check-in: 16dac180 user: drh tags: bedrock)
14:39
Merge the latest trunk enhancements into the wal2 branch. (check-in: 4a72acbc user: drh tags: wal2)
14:36
Merge the latest trunk enhancements into the begin-concurrent branch. (check-in: eedf6bed user: drh tags: begin-concurrent)
14:29
Merge from wrong branch. Bedrock should be updated from wal2, not from trunk. (Closed-Leaf check-in: 4c23cad9 user: drh tags: mistake)
2024-04-12
18:46
If a build fails in testrunner.tcl, do not attempt to run the jobs that depend on that build. Instead, report those jobs as having been skipped. (check-in: b40580be user: drh tags: trunk)
15:02
Fixes and new tests logic to ensure that the btree overflow page cache is only used when it is consistent. This resolves the malfunction observed in forum post 284955a3cd454a15. (check-in: 5dede50d user: drh tags: trunk)
13:05
Add assert() statements to help verify the overflow page cache. (Closed-Leaf check-in: 0ebc6548 user: drh tags: btree-ovfl-cache)
12:32
New assert() statements to verify the correctness of the BTCF_AtLast flag on btree cursors. (check-in: 4efecd61 user: drh tags: btree-ovfl-cache)
11:28
Test case for the fix in the previous check-in. (check-in: 0cf4d835 user: drh tags: btree-ovfl-cache)
02:57
Mark the BTree cell overflow cache as invalid whenever the rowid goes invalid. (check-in: 74c9e19c user: drh tags: btree-ovfl-cache)
2024-04-11
19:08
Fix handling of the SQLITE_OMIT_WAL check in one place. Resolves forum post 87cc13302de160eb. (check-in: 5dae6e6d user: stephan tags: trunk)
2024-04-09
19:23
Fix a "shift exponent is too large" usan error that could occur in the fts3 snippet() function when a query phrase contains more than 64 tokens. (check-in: 6d2c7ac4 user: dan tags: trunk)
15:09
Avoid some performance problems in the recover extension when recovering strategically corrupted databases. (check-in: e66a834b user: dan tags: trunk)
13:57
The read-only CHECK-constraint optimization of [34ddf02d3d21151b] inhibits the xfer optimization for tables with CHECK constraints. However, the xfer optimization is required for correct operation of VACUUM INTO on tables that contain generated columns. Fix this by ignoring CHECK constraints when qualifying the xfer optimization while running VACUUM. Problem reported by forum post 3ec177d68fe7fa2c. (check-in: a6e26e77 user: drh tags: trunk)
2024-04-08
17:55
Update tests in returning1.test to account for [c7896e88]. (check-in: c02956cd user: dan tags: trunk)
12:02
Merge the latest trunk enhancements, and especially the enhancement to the WHERE-clause push-down optimization, into the bedrock branch. (check-in: 0f93b52c user: drh tags: bedrock)
11:57
Merge the latest trunk enhancements into the wal2 branch. (check-in: 59ccea7d user: drh tags: wal2)
11:50
Enhance the WHERE-clause push-down optimization so that it is able to push down WHERE clause terms that contain uncorrelated subqueries. (check-in: 69ec714b user: drh tags: trunk)
11:45
Fix a harmless compiler warning in the ICU extension due to [acddbc489d5231da]. (check-in: 3d8750e8 user: drh tags: trunk)
06:37
Enhancements to the fuzzer invariant checker to turn optimizations on and off. (Closed-Leaf check-in: 67594481 user: drh tags: pushdown-subquery)
2024-04-07
18:55
Merge trunk testing enhancements into the pushdown-subquery branch. (check-in: 287ff24b user: drh tags: pushdown-subquery)
18:36
Omit all rowid-in-view restrictions from the fuzzinvariant.c test module as they are no longer necessary, as of the previous check-in. (check-in: 6431538f user: drh tags: trunk)
18:23
If SQLITE_ALLOW_ROWID_IN_VIEW is set to 2, then all rowids for views return a value of NULL. (check-in: 0a53dde2 user: drh tags: trunk)
10:30
Merge trunk enhancements into the pushdown-subquery branch. (check-in: 27865e31 user: drh tags: pushdown-subquery)
10:27
Add comments to note the name abiguity between the MySQL push-down optimization and the WHERE-clause push-down optimization. (check-in: 3d5fb1ec user: drh tags: trunk)
2024-04-06
23:09
Add comments to note the name abiguity between the MySQL push-down optimization and the WHERE-clause push-down optimization. (Leaf check-in: 681dfe2d user: drh tags: wrong-branch)
18:30
Do not allow changes to sqlite3ExprIsTableConstant() that support pushdown of subqueries interfere with the hash-join logic. (check-in: 8682931f user: drh tags: pushdown-subquery)
17:42
Merge testing logic fixes on trunk into the pushdown-subquery branch. (check-in: a4e1f03d user: drh tags: pushdown-subquery)
17:37
Improvements to the accurancy of the fuzzinvariants.c testing module when SQLITE_ALLOW_ROWID_IN_VIEW is defined and the test query involves rowids. (check-in: c6e873d4 user: drh tags: trunk)
12:19
Generalize pushdown to allow any uncorrelated subquery to be pushed down. (check-in: 87c45fb0 user: drh tags: pushdown-subquery)
11:59
Generalize the pushdown of IN right-hand sides to cover more SELECTs. (Closed-Leaf check-in: 4120bc74 user: drh tags: pushdown-IN-table)
2024-04-05
20:01
Experimental enhancement in which expressions of the form "expr IN table" can be pushed down into subexpressions. (check-in: 2cbd7838 user: drh tags: pushdown-IN-table)
15:04
Merge all recent trunk enhancements into the reuse-schema branch. (check-in: e556f3d3 user: drh tags: reuse-schema)
14:54
Merge the latest trunk enhancements into the bedrock branch. (check-in: 7a2a64b6 user: drh tags: bedrock)
14:50
Merge the latest trunk enhancements into the wal2 branch. (check-in: a8a8a2db user: drh tags: wal2)
14:46
Merge the latest trunk enhancements into the begin-concurrent branch. (check-in: 4ff83342 user: drh tags: begin-concurrent)
13:56
Check-in [a9657c87c53c1922] is wrong: the IndexedExpr.bMaybeNullRow flag is required for virtual columns if they are part of an outer join. Add a test case (derived from dbsqlfuzz b9e65e2f110df998f1306571fae7af6c01e4d92b) to prove it. (check-in: 4484ec6d user: drh tags: trunk)
11:23
When compiling with SQLITE_ALLOW_ROWID_IN_VIEW, if the RETURNING clause of an UPDATE of a view specifies a rowid, then return NULL for the value of that rowid. dbsqlfuzz 7863696e9e5ec10b29bcf5ab2681cd6c82a78a4a. (check-in: c7896e88 user: drh tags: trunk)
2024-04-04
22:53
Optimize sqlite3.oo1.DB.exec() for the rowMode='object' case to avoid converting the object property keys (column names) from native code to JS for each row. This speeds up large data sets considerably and addresses the report in forum post 3632183d2470617d. (check-in: 8b41ef86 user: stephan tags: trunk)
14:26
Add the "interstage-heuristic" that attempts to avoid wildly inefficient queries that use table scans instead of index lookups because the output row estimates are inaccurate. (check-in: 7bf49e2c user: drh tags: trunk)
13:36
Fix an ASAN problem in part of the test harness. No changes to SQLite itself. (check-in: 797cda7d user: drh tags: trunk)
13:07
Fix test cases due to an error message change. (check-in: c49ec239 user: drh tags: trunk)
12:36
Fix exception propagation when the opfs-sahpool VFS fails to initialize to address report in forum post be0141c639. (check-in: cdb3a900 user: stephan tags: trunk)
2024-04-02
21:55
Add an optional 3rd argument to the icu_load_collation() function that specifies the "strength" of the comparison. (check-in: acddbc48 user: drh tags: trunk)
14:12
Fix typos in comments. Provided ".wheretrace" debugging output for the interstage heuristic module. Do omit automatic index loops in the interstage heuristic. (Closed-Leaf check-in: 186dcae1 user: drh tags: interstage-heuristic)
11:44
Add a heuristic in between the two solver() passes of the query planner that tries to prevent a very slow query plan in cases where the output row count estimate is imprecise. (check-in: 8018417b user: drh tags: interstage-heuristic)
2024-04-01
15:38
Improved comments in the query planner logic that computes the cost for a particular step in a query plan. No code changes. (check-in: 0b2ac2cd user: drh tags: trunk)
2024-03-30
14:11
Make explicit that sqlite3_keyword_name()'s index is 0-based, per forum request. Doc changes only. (check-in: 090943dc user: stephan tags: trunk)
2024-03-29
18:19
Fix handling of an OOM case broken by [11015b4e]. (check-in: 1f3a805d user: dan tags: bedrock)
18:16
Fix handling of an OOM case broken by [49263c91]. (check-in: 9e6b64de user: dan tags: begin-concurrent)
18:07
Fix a problem in SQLITE_DIRECT_OVERFLOW_READ builds that could allow a concurrent transaction to be committed even if it read from an overflow page that was modified concurrently, in cases where the overflow page was written without also writing the b-tree page to which it is linked. (check-in: 11015b4e user: dan tags: bedrock)
18:04
Improve the log message emitted when a BEGIN CONCURRENT transaction cannot be committed due to conflicts so that it identifies the conflicting table in a few more cases. (check-in: 19ff2d4e user: dan tags: bedrock)
17:58
Fix a problem in SQLITE_DIRECT_OVERFLOW_READ builds that could allow a concurrent transaction to be committed even if it read from an overflow page that was modified concurrently, in cases where the overflow page was written without also writing the b-tree page to which it is linked. (check-in: 49263c91 user: dan tags: begin-concurrent)
17:32
Improve the log message emitted when a BEGIN CONCURRENT transaction cannot be committed due to conflicts so that it identifies the conflicting table in a few more cases. (check-in: 5d30e362 user: dan tags: begin-concurrent)
2024-03-28
10:58
Document that the order of an update hook call is unspecied vis-a-vis the final result of the operation which triggers that hook. Doc changes only. (check-in: 3d4b1f07 user: stephan tags: trunk)
2024-03-27
20:34
Update comments in fts5.h. (check-in: 862945d5 user: dan tags: trunk)
2024-03-26
18:07
Add extra tests to vtabL.test. (check-in: fe209099 user: dan tags: trunk)
11:14
Avoid expanding integer values in columns with real affinity to the full 8-byte representation when editing records as part of a DROP COLUMN command. (check-in: a49296de user: dan tags: trunk)
10:48
Merge all recent trunk enhancements into the reuse-schema branch. (check-in: e469b022 user: drh tags: reuse-schema)
10:45
Merge all recent trunk enhancements into the bedrock branch. (check-in: 33d3453f user: drh tags: bedrock)
10:42
Merge all recent trunk enhancements into the wal2 branch. (check-in: c2139d82 user: drh tags: wal2)
10:42
Merge all recent trunk enhancements into the begin-concurrent branch. (check-in: 9b854e54 user: drh tags: begin-concurrent)
2024-03-25
20:35
The RAISE() operator is not a constant expression and cannot participate in the VALUE-as-coroutine optimization. dbsqlfuzz 74cf7c9904360322a6c917e4934b127543d1cd51 (check-in: 6a06dc73 user: drh tags: trunk)
20:00
Add further tests for sqlite3_declare_vtab(). (check-in: f3c4433f user: dan tags: trunk)
19:31
Simplifications to the sqlite3_declare_vtab() implementation. Changes some conditionals into assert() statements, for coverage. (check-in: ff7b898a user: drh tags: trunk)
18:30
Fix a theoretical OOB memory access in sqlite3_stmt_scanstatus_v2(). (check-in: 87be9580 user: dan tags: trunk)
18:24
Revert the previous change. Instead, do a pre-check of the CREATE TABLE statement that is the second argument to sqlite3_declare_vtab() and if the first two keywords are not "CREATE" and "TABLE", then raise an SQLITE_MISUSE error. (check-in: 6a2ff835 user: drh tags: trunk)
17:43
Remove an ALWAYS() from a condition that can be true. (check-in: 715fcf03 user: drh tags: trunk)
14:44
Fix an inaccuracy in automatic indexes that was exposed by the fix at [80c4223098c1827f]. (check-in: fdc9406f user: drh tags: trunk)
13:18
Fix the pragma4.test script so that it works with SQLITE_OMIT_VIRTUALTABLE. (check-in: f8fe6fb6 user: drh tags: trunk)
12:28
Fix recent test case changes so that they also work on Windows. (check-in: 38578b5a user: drh tags: trunk)
11:34
Use the SQLITE_CONSTRAINT return value from xBestIndex to prohibit bad query plans in the pragma virtual table. (check-in: b1259d44 user: drh tags: trunk)
10:55
Improvements to the vtablog.c extension: Eliminate memory leaks. More diagnostic output for xBestIndex. (check-in: 92e9a71b user: drh tags: trunk)
10:28
Flag sqlite3_trace() and sqlite3_profile() as deprecated so that the doc generator for funclist.html sees them as such, and add 'Deprecated' to their page's title for consistency with other deprecated APIs. Comment changes only. Addresses forum post 0901025836. (check-in: 87c54f93 user: stephan tags: trunk)
00:38
Improvements to the vtablog.c extension, for better logging of virtual table interactions. (check-in: e253bb36 user: drh tags: trunk)
2024-03-24
21:10
Fix the xBestIndex method of the pragma virtual table so that it correctly gives a higher cost to plans where the schema hidden parameter is unconstrained. Fix for the problem reported by forum post 85b6a8b6705fb77a. (check-in: bc516ff5 user: drh tags: trunk)
20:26
Improved "wheretrace" debugging output for calls to xBestIndex. No changes to release code. (check-in: 33966bb0 user: drh tags: trunk)
16:33
Do not automatically assume that ROWID is NOT NULL when compiled with SQLITE_ALLOW_ROWID_IN_VIEW. dbsqlfuzz 31b38eeb63a4e1562de665078f52b7b47a7543cf (check-in: 80c42230 user: drh tags: trunk)
2024-03-23
15:17
Fix an adverse interaction between CREATE TABLE AS and the new VALUES-as-coroutine optimization. dbsqlfuzz c2c5e7e08b7e489d270a26d895077a03f678c33b (check-in: 84b6fdea user: drh tags: trunk)
2024-03-22
10:32
Fix incorrect boundary assert()s on the new OP_IfSizeBetween opcode. (check-in: 8eda4797 user: drh tags: trunk)
2024-03-21
22:18
Test case for the assert() problem fixed by [9ba3cff9d48fe8fc]. (check-in: 865f0de7 user: drh tags: trunk)
15:52
Add comments (only, no code) to the JS sqlite3_config() bindings reminding us why SQLITE_CONFIG_ROWID_IN_VIEW is specifically not included. JNI does not need this treatment because it explicitly supports only a very small subset of config options. (check-in: 4fea3962 user: stephan tags: trunk)
11:46
Enhancements to testrunner.tcl: (1) Add the "--config CONFIGS" option, to that only configuration named on the comma-separated list CONFIGS are run. (2) Add the "--omit CONFIGS" to omit configuration on the CONFIGS list (3) Add the Android configuration to "release" (check-in: 4ccb3729 user: drh tags: trunk)
11:37
Have os_unix.c reuse cached file-descriptors in the case when a read-write fd is requested on a read-only file and a read-only fd returned. (check-in: a678e854 user: dan tags: trunk)
10:35
Fix all test cases so that they work with SQLITE_ALLOW_ROWID_IN_VIEW. (check-in: 66c69e2f user: drh tags: trunk)
2024-03-20
16:56
Change EXPLAIN QUERY PLAN output to say "USE TEMP B-TREE FOR LAST TERM OF ORDER BY", or "LAST N TERMS OF ORDER BY", instead of "RIGHT PART OF ORDER BY". (check-in: 7e13a2c7 user: dan tags: trunk)
16:27
Change EXPLAIN QUERY PLAN output to say "USE TEMP B-TREE FOR LAST TERM OF ORDER BY", or "LAST N TERMS OF ORDER BY", instead of "RIGHT PART OF ORDER BY". (Closed-Leaf check-in: cd547c50 user: dan tags: eqp-orderby)
10:45
Fix the SQLITE_ALLOW_ROWID_IN_VIEW bug-compatibility option so that it works again. Add the SQLITE_CONFIG_ROWID_IN_VIEW sqlite3_config() value to optionally disable the bug-compatibility mode at start-time. Enable testing of SQLITE_ALLOW_ROWID_IN_VIEW. (check-in: b6802565 user: drh tags: trunk)
10:40
Fix a typo in a comment. (Closed-Leaf check-in: 54680d0f user: drh tags: rowid-in-view)
09:45
Fix typos in the SQLITE_CONFIG_ROWID_IN_VIEW documentation. (check-in: 1ad1ca49 user: drh tags: rowid-in-view)
2024-03-19
23:01
Change the SQLITE_CONFIG_NO_ROWID_IN_VIEW configuration option to be just SQLITE_CONFIG_ROWID_IN_VIEW (without the "_NO_" in the middle) and give it the ability to turn the option on and off. Otherwise, it is difficult to test. (check-in: 5d412edc user: drh tags: rowid-in-view)
19:56
Return the error "ambiguous column name: rowid", instead of "no such column: rowid", when a "rowid" reference is ambiguous. (check-in: 0615bdae user: drh tags: rowid-in-view)
19:48
Add the --no-rowid-in-view option to the --help output of the CLI. (check-in: 134da9c3 user: drh tags: rowid-in-view)
19:42
Make sure the new u1.nRow value is copied when making a copy of a SrcItem object. This fixes a problem in [ac6f095e13e43d66] from yesterady. (check-in: 262f8f9d user: drh tags: trunk)
16:51
Update test scripts to account for [8fcea4cd]. (check-in: 0c543c51 user: dan tags: rowid-in-view)
16:34
In the name resolver when SQLITE_ALLOW_ROWID_IN_INDEX is enabled, if there are multiple views that might resolve to the "rowid" but only one real table, then use that one real table and ignore the views. (check-in: 8fcea4cd user: drh tags: rowid-in-view)
15:27
Update tests to work with SQLITE_ALLOW_ROWID_IN_VIEW. (check-in: 495b8c7b user: dan tags: rowid-in-view)
14:55
Return the error "ambiguous column name: rowid", instead of "no such column: rowid", when a "rowid" reference is ambiguous. (Closed-Leaf check-in: 021f34fc user: dan tags: ambiguous-rowid-error)
14:38
Fix the fuzzinvariants.c test module so that it does not generate invariants that depend on rowid when SQLITE_ALLOW_ROWID_IN_VIEW is used, since rowid values can be unreliable in that context. (check-in: 0d711aaa user: drh tags: rowid-in-view)
14:21
Improvements to the first check-in on this branch, causing less collateral damage. (check-in: 7bc88289 user: drh tags: rowid-in-view)
13:55
On second thought, change SQLITE_TESTCTRL_ROWID_IN_VIEW into a start-time option SQLITE_CONFIG_NO_ROWID_IN_VIEW. (check-in: b8e045c9 user: drh tags: rowid-in-view)
13:31
When compiled with SQLITE_ALLOW_ROWID_IN_VIEW, rowid-in-view is on by default but can now be turned off using SQLITE_TESTCTRL_ROWID_IN_VIEW. Without the compile-time option, rowid-in-view is always off. (check-in: 8a6196ab user: drh tags: rowid-in-view)
10:41
Adjustments to the test cases for the previous check-in to make them cross-platform Linux/Mac/Windows. Test changes only. No change to core code. (check-in: 0dddadbd user: drh tags: trunk)
10:16
Fix another problem from [c63e26e705f5e967] involving infinities and the Kahan-Babushka-Neumaier summation algorithm. Problem reported by forum post 23b8688ef4. (check-in: 45d272ef user: drh tags: trunk)
02:30
First steps toward getting -DSQLITE_ALLOW_ROWID_IN_VIEW to work again. That compile-time option is untested, undocumented, and unsupported. But it was mentioned in the release notes for version 3.36.0, so I think that means we need to support it forever. (check-in: 7c46ff64 user: drh tags: rowid-in-view)
2024-03-18
21:08
Fix asan problems in sqlite3rbu.c. Enable RBU for sanitizer release tests. (check-in: ad2ae771 user: dan tags: trunk)
20:52
Avoid spurious SQLITE_IOERR_DATA errors when reading partial pages from a cksumvfs database using a SQLITE_DIRECT_OVERFLOW_READ build. (check-in: 0f485dbe user: dan tags: trunk)
18:56
Merge the latest trunk enhancements into the reuse-schema branch. (check-in: c2d4adab user: drh tags: reuse-schema)
18:55
Merge the latest trunk enhancements into the bedrock branch. (check-in: 10ab5e85 user: drh tags: bedrock)
18:51
Merge the latest trunk enhancements into the wal2 branch. (check-in: c0af1242 user: drh tags: wal2)
18:47
Merge the latest trunk enhancements into the begin-concurrent branch. (check-in: 50b43b02 user: drh tags: begin-concurrent)
18:03
Remove unnecessary blank lines from build commands in the default Makefile.in. (check-in: 76fb3a90 user: drh tags: trunk)
18:00
Add the "VALUES-as-coroutine" optimization. Large VALUES clauses on an INSERT, for example, prepare and run in about half the time and with half the memory. This check-in also includes enhancements to the internal sqlite3ExprIsConstant() routine to recognize pure SQL functions as constant if they have constant arguments. (check-in: a120c923 user: drh tags: trunk)
17:13
Fix harmless compiler (scan-build) warnings. (Closed-Leaf check-in: c86f9f2a user: drh tags: exp-values-clause2)
16:30
Allow the VALUES-as-coroutine optimization to be applied to later rows of a VALUES clause even if earlier rows do not qualify. (check-in: 9a47ea7f user: drh tags: exp-values-clause2)
15:10
Change a test case to use a window function rather than random() to disable the VALUES clause optimization. (check-in: c589149a user: drh tags: exp-values-clause2)
13:31
Add assert() statements to validate access to the SrcItem.u1.nRow union member. (check-in: 21f616d9 user: drh tags: exp-values-clause2)
13:10
Fix the patch at [10ee6fcba08ab128] so that it works on reduced-size Expr nodes. (check-in: 260bd764 user: drh tags: exp-values-clause2)
12:49
Improvements to EXPLAIN QUERY PLAN output for multi-row VALUES claues. (check-in: ac6f095e user: drh tags: exp-values-clause2)
11:12
Ensure the database schema has been loaded and the database encoding gleaned before beginning to code a multi-row VALUES clause. (check-in: 2ff476eb user: dan tags: exp-values-clause2)
10:54
Fix a problem caused by a non-aggregate function with an OVER clause in a multi-row VALUES clause. (check-in: 10ee6fcb user: dan tags: exp-values-clause2)
2024-03-17
16:01
Omit the "noop(X)" test SQL function. Accomplish the same thing using the idiom: "coalesce(X,random())". (check-in: 0eb2fbb8 user: drh tags: exp-values-clause2)
00:13
Add an ALWAYS on an unreachable branch. (check-in: 0dce6211 user: drh tags: exp-values-clause2)
2024-03-16
21:12
Recognize when an SQL function is constant and can occur inside an optimized VALUES clause. (check-in: c0e5ceb8 user: drh tags: exp-values-clause2)
21:04
Fix a assert() that is incorrect, though harmless. (Closed-Leaf check-in: 9ba3cff9 user: drh tags: enhanced-expr-is-const)
20:15
Further expand the scope of usage for the enhacement to expr-is-constant. (check-in: d7eadcf7 user: drh tags: enhanced-expr-is-const)
19:48
Assert that the eCode is always non-zero upon entry into exprNodeIsConstant(). (check-in: f71b28f6 user: drh tags: enhanced-expr-is-const)
19:36
Activate the enhanced constant-expression for the [constant propagation optimization](/info/f4229707ac08d66c). (check-in: 92afea38 user: drh tags: enhanced-expr-is-const)
17:04
Extend the scope expr-is-constant for function to cover the IN-to-EQ optimization. (check-in: dae66924 user: drh tags: enhanced-expr-is-const)
14:27
Fix exprNodeIsConstantFunction() so that it returns WRC_Continue, not WRC_Abort, if the function really is constant. (check-in: d85dd4de user: dan tags: enhanced-expr-is-const)
13:18
Enhance the sqlite3ExprIsConstant() function so that it recognizes constant functions. So far the enhancement only applies to the multi-row VALUES clause, but it could possibly be applied in many other places. (check-in: c9e0488c user: drh tags: enhanced-expr-is-const)
2024-03-15
17:57
New options for testrunner.tcl: --stop-on-error and --stop-on-coredump. (check-in: 82035b9c user: drh tags: trunk)
17:04
Fix (totally harmless) memory leaks in Lemon to avoid warnings during ASAN builds. (check-in: ce009205 user: drh tags: trunk)
13:16
Tweaks to the func4.test test module (which tests the rarely used totype.c extension) so that it works on i586 compiled using gcc-13 with -O0. Details at forum post 0a7553b0f734c033. (check-in: d0fbe779 user: drh tags: trunk)
12:59
Attempt to use the Dekker high-precision floating point algorithm in the ext/misc/totype.c extension. This check-in should have been derived from trunk, not the exp-values-clause2 branch. But that's ok, because both trunk and the branch are the same in all aspects that relate to this check-in and this check-in is just parking the proposed enhancement for future consideration - it is not a serious proposal at this point. (Leaf check-in: dd41db3b user: drh tags: totype-dekker)
2024-03-14
20:39
Fix testcase to account for new behaviors with this branch. (check-in: 823e5793 user: drh tags: exp-values-clause2)
19:31
Fix a problem with ALTER TABLE and correlated multi-row VALUES clauses. (check-in: d543c829 user: dan tags: exp-values-clause2)
19:01
Fix a problem handling "INSERT INTO ... SELECT ... UNION VALUES(...), (...)" and similar statements. (check-in: e8a2a819 user: dan tags: exp-values-clause2)
17:04
Avoid some OP_SCopy instructions in "INSERT INTO .. VALUES" statements that insert more than one row in cases where the VALUES clause contains a value for all columns of the table. (check-in: 988f0ea7 user: dan tags: exp-values-clause2)
13:18
JS: remove an unused function argument and some debug-related comments. No functional changes. (check-in: f1570943 user: stephan tags: trunk)
11:45
Add further tests for VALUES clauses in various contexts. (check-in: 973d197d user: dan tags: exp-values-clause2)
2024-03-13
20:04
Change the way parse.y handles multi-row VALUES clauses to save a few cycles. (check-in: 88d5bc91 user: dan tags: exp-values-clause2)
18:41
Remove unreachable code from this branch. (check-in: 657c7b4f user: dan tags: exp-values-clause2)
18:00
Merge the latest trunk enhancements into the bedrock branch. (check-in: 5c8e80f4 user: drh tags: bedrock)
17:55
Merge the latest trunk enhancements into the wal2 branch. (check-in: 2b12da72 user: drh tags: wal2)
17:51
Merge the latest trunk enhancements into the begin-concurrent branch. (check-in: 238fa84a user: drh tags: begin-concurrent)
17:33
Fix some OOM handling problems on this branch. (check-in: 2085c7f1 user: dan tags: exp-values-clause2)
16:49
Do not automatically activate SQLITE_ENABLE_SHARED_SCHEMA in the Makefile. Instead, enhance the mdevtest and sdevtest targets of testrunner.tcl to set that option sometimes and sometimes not. (check-in: 693b8f93 user: drh tags: reuse-schema)
16:35
Merge the testrunner.tcl enhancement from trunk. (check-in: da471515 user: drh tags: reuse-schema)
16:32
Minor change to the "mdevtest" and "sdevtest" in testrunner.tcl, to make it easier to add configurations in branches. (check-in: 9349d94e user: drh tags: trunk)
15:59
Merge the latest trunk enhancements into the reuse-schema branch. Fix the reuse-schema build of the CLI so that it works again. (check-in: 70ef3784 user: drh tags: reuse-schema)
15:47
Merge trunk into this branch. (check-in: 38635651 user: dan tags: exp-values-clause2)
15:44
Ensure that if a multi-row VALUES cannot be coded while being parsed, the SF_MultiValue flag is correctly set. (check-in: 94791824 user: dan tags: exp-values-clause2)
15:34
Avoid creating a co-routine incrementally for a VALUES clause that has affinities other than NONE. (check-in: 4229b12b user: dan tags: exp-values-clause2)
13:43
If the NJOB environment variable is a positive integer, it causes testrunner.tcl to run that many jobs. (check-in: 924281b9 user: drh tags: trunk)
13:27
Document the testrunner.tcl --explain option in doc/testrunner.md. (check-in: 5a7b5fc9 user: drh tags: trunk)
13:09
Add the --explain option to testrunner.tcl. (check-in: f0c9eb03 user: drh tags: trunk)
11:46
Improve one of the error messages from testrunner.tcl. (check-in: 233155c1 user: drh tags: trunk)
11:29
Minor tweaks to the testrunner documentation. (check-in: 3128b98d user: drh tags: trunk)
10:55
Improvements to the "help" message generated by testrunner.tcl. Add the special "list" permutation that does nothing but list all allowed values for the PERMUTATION argument to testrunner.tcl. (check-in: 1f5bfc6a user: drh tags: trunk)
10:54
Add pointer to documentation to testrunner.tcl help message. (check-in: eff85510 user: dan tags: trunk)
10:47
Have testrunner.tcl set SQLITE_TMPDIR on windows as well. (check-in: 64f4dad1 user: dan tags: trunk)
00:37
On testrunner.tcl: (1) Add the "help" command. (2) Add the "script" command to the help message. (3) Improve the error message generated by "script" when an incorrect CONFIG option is provided. (check-in: 1b7f0be4 user: drh tags: trunk)
2024-03-12
20:10
Correctly initialize the SrcItem.iCursor field to -1 when creating a co-routine from a multi-values VALUES clause. (check-in: c32953ff user: dan tags: exp-values-clause2)
18:33
Have testrunner.tcl have each test store its temp files in its working directory. To avoid unlikely, but possible, collisions. (check-in: 18842d0d user: dan tags: trunk)
18:04
Have testrunner.tcl have each test store its temp files in its working directory. To avoid unlikely, but possible, collisions. (Closed-Leaf check-in: 775a10d1 user: dan tags: testrunner-tmpdir-fix)
2024-03-11
20:34
Fix another testrunner.tcl problem on windows. (check-in: 7805844b user: dan tags: trunk)
19:43
Fix a missing brace problem in testrunner.tcl introduced by the previous commit. (check-in: 1f5e334d user: dan tags: trunk)
19:33
Update mdevtest and sdevtest to do shell builds in parallel. And only if one or more of the tcl scripts run requires it. (check-in: 6eae4547 user: dan tags: trunk)
18:21
Run shell*.test modules with "make mdevtest". (check-in: 76c47106 user: drh tags: trunk)
17:27
Attempt to reduce the memory used by VALUES clauses in as many statements as possible, not just INSERT. This branch still has problems. (check-in: 17d1f7cf user: dan tags: exp-values-clause2)
11:24
Fix the ".import" command in the CLI so that it works correctly with tables that contain computed columns. forum post ca014d7358. (check-in: 95a9c88b user: drh tags: trunk)
09:39
Add more docs for the OPFS delete-before-open feature. (check-in: cb8d9c26 user: stephan tags: trunk)
09:34
Add delete-before-open=1 URI flag to the 'opfs' VFS to tell it to xDelete the db file before opening it, primarily to enable users to work around a corrupt db without having to reach into OPFS-specific APIs to remove the db file. (check-in: e87cf0d7 user: stephan tags: trunk)
09:32
Add delete-before-open=1 URI flag to the 'opfs' VFS to tell it to xDelete the db file before opening it, primarily to enable users to work around a corrupt db without having to reach into OPFS-specific APIs to remove the db file. (Closed-Leaf check-in: e83f9788 user: stephan tags: opfs-delete-before-open)
2024-03-09
18:41
Fix the .expert command in the shell so that it does not leak memory if not followed by SQL and so that it works with reverse_unordered_selects. (check-in: 7ead022e user: drh tags: trunk)
13:31
Fuzzcheck is run automatically by testrunner.tcl, so there is no need to include fuzzcheck as a separate dependency for the releasetest target in the MSVC makefile. (check-in: 76629b2b user: drh tags: trunk)
2024-03-08
21:37
The NOT NULL strength reduction optimization from [de9c86c9e4cdb34f] should be applied to the WHERE clause only. Otherwise, the operand of the IS NULL or IS NOT NULL operator might be a reference to a bare column of an aggregate table, and we can't tell if it is NULL or not based only on its NOT NULL attribute. Forum post 440f2a2f17. (check-in: 51704fea user: drh tags: trunk)
19:03
One of the assert()s added by the previous check-in was not quite correct. This commit fixes it. (check-in: d4013583 user: drh tags: trunk)
15:39
Add new assert() statements, which if they had existed three years ago, would have detected the code generator problem fixed by the previous check-in. (check-in: 166d1e5d user: drh tags: trunk)
14:01
Silently ignore redundant ON CONFLICT clauses in an UPSERT. Only the first ON CONFLICT for each index is active. Do not issue an error, since that might break legacy queries. But ignore the redundant ON CONFLICT clauses to prevent problems such as described in forum post 919c6579c8. (check-in: d0ea6b6b user: drh tags: trunk)
13:49
Omit some extra tests for the intck extension in OMIT_VIRTUAL_TABLE or OMIT_PRAGMA builds. (check-in: 29d9eb7d user: dan tags: trunk)
11:13
Fix pragma6.test so that it works with ENABLE_OVERSIZE_CELL_CHECK builds. (check-in: a02551de user: dan tags: trunk)
11:08
Do not run tests for the intck extension in OMIT_VIRTUAL_TABLE or OMIT_PRAGMA builds. (check-in: de8130c0 user: dan tags: trunk)
03:24
Must use sqlite3IntFloatCompare() for accurate comparisons between very large integer and floating point values in RTREE. Otherwise the comparison does not work on all platforms. Further fix to [027e5336acc26f57]. (check-in: 820f106a user: drh tags: trunk)
2024-03-07
19:29
Introducing JS worker1 promiser v2, which initializes via Promise (instead of a callback function) and can be loaded as an ESM module. (check-in: 2fbaf2f5 user: stephan tags: trunk)
18:53
Resolve duplicate loading of promiser v2 demo code (a side effect of having done precisely what it was told to do). Consolidate demo-worker1-promiser(-esm).html/(m)js variants into central copies processed with c-pp. (Closed-Leaf check-in: 9347d9b9 user: stephan tags: wasm-promiser1-v2)
18:47
Improved error handling. (Leaf check-in: cf2d1ad0 user: drh tags: exp-values-clause)
17:56
wasm promiser.v2 is essentially working but the demo code is double-loading the module for as-yet-undetermined reasons. (check-in: aa877ce0 user: stephan tags: wasm-promiser1-v2)
17:11
Simplify the parser changes on this branch. (check-in: 89a6c7b5 user: dan tags: exp-values-clause)
16:25
Merge all the latest trunk enhancements into the exp-values-clause branch. (check-in: c4923f26 user: drh tags: exp-values-clause)
16:04
Initial work at getting sqlite3Worker1Promiser.v2() to return a Promise instead of using an onready() callback, and also creating an ESM build for promiser1 per user request. It seems to work but requires more testing. (check-in: 0e272123 user: stephan tags: wasm-promiser1-v2)
15:58
Fix harmless compiler warnings in test code for the intck extension. (check-in: 7fbdc1a8 user: drh tags: trunk)
15:44
Add further tests for the code on this branch. (check-in: 5d8065e8 user: dan tags: exp-values-clause)
12:34
Do not allow the query planner to be tricked into thinking that an index on a constant expression might be useful for something. Problem reported on forum post ecdfc02339. This is a follow-up to the fixes at [44200596aa943963] and [2d2b91cc0f6fed8c]. (check-in: 720ce06d user: drh tags: trunk)
2024-03-06
20:49
Add the json_pretty() SQL function. (check-in: ceb51c1c user: drh tags: trunk)
20:38
Add test cases for json_pretty(). (Closed-Leaf check-in: 6448b907 user: drh tags: json-pretty)
20:34
Avoid making too many calls to sqlite3Select/sqlite3WhereBegin() when processing large VALUES clauses. (check-in: 356158c1 user: dan tags: exp-values-clause)
20:24
Work around obscure floating point issue seen with older versions of MSVC. (check-in: 42d39f91 user: mistachkin tags: trunk)
14:55
Handle VALUES clauses having different number of values in different rows. Also the case where SQLITE_LIMIT_COMPOUND_SELECT is set to 0. (check-in: f5074b29 user: dan tags: exp-values-clause)
14:42
Add a couple of json_pretty() examples to /fiddle. (check-in: d5e1687b user: stephan tags: json-pretty)
14:30
Add the json_pretty(J) function for pretty-printing of JSON. An optional 2nd argument is text used for indentation, with a default value being four spaces. (check-in: 39552bd3 user: drh tags: json-pretty)
12:28
Correction to the previous check-in. (check-in: 483fa296 user: drh tags: trunk)
11:35
Fix handling of "id=?" corner cases in rtree when the value on the RHS is a real value. Problem reported by forum post 1bb055be17. (check-in: 027e5336 user: dan tags: trunk)
08:45
Wasm build tweaks to attempt to get the node.js-for-node.js build to use Emscripten's nodefs filesystem driver for persistent storage. This is completely untested - validating it requires a nodeist. (Leaf check-in: 0bcbde7c user: stephan tags: wasm-nodefs)
2024-03-05
20:33
Add comments for the new code on this branch. (check-in: 551caba9 user: dan tags: exp-values-clause)
18:41
Remove code that added a P4 parameter to the OP_Variable opcode. This is no longer required. (check-in: dd5977c9 user: dan tags: trunk)
18:03
Merge trunk changes into this branch. (check-in: a003fffa user: dan tags: exp-values-clause)
17:48
Fix a problem with large values clauses and named variables. (check-in: 4784a92e user: dan tags: exp-values-clause)
17:33
Reformulate [34439fe3aeea7cbb] slightly to resolve a false-positive OOM reported in forum post 2eadfe94e3. (check-in: 82906467 user: stephan tags: trunk)
17:23
Simplify changes on this branch by avoiding any special handling until SQLITE_LIMIT_COMPOUND_SELECT rows are reached. (check-in: 9cbceaac user: dan tags: exp-values-clause)
16:47
The value returned by the json_each.json field needs to survive longer than the json_each virtual table itself, in the event that the value is used in an aggregate expression. dbsqlfuzz 5120bd0b3bb0b73eebbcb79ac37c6b0663fccad6 (check-in: 952ed71b user: drh tags: trunk)
12:48
/fiddle: reorder/replace example entries to make them more coherent when tried in order and clear the example selection list when the input area is cleared, both based on forum feedback. (check-in: b4ce270b user: stephan tags: trunk)
08:48
/fiddle 'Mandlebrot' typo fix reported in the forum. (check-in: ceefe6ba user: stephan tags: trunk)
07:55
Update fiddle.make to account for Makefile changes in [178b7d46f9]. (check-in: 7a5d8105 user: stephan tags: trunk)
06:31
/fiddle: before resetting a db, roll back any transactions (resolves problem reported in forum post 0b41a25d65) and remove an obsolete/broken reference to a long-gone API which could cause initialization to fail prematurely. (check-in: ee164ca7 user: stephan tags: trunk)
2024-03-04
18:50
Fix some problems with "VALUES(...) UNION ..." and similar on this branch. (check-in: 5d679771 user: dan tags: exp-values-clause)
18:22
Fix a compiler warning in date.c. Update makefiles to include all necessary dependencies for building shell.c. (check-in: 178b7d46 user: drh tags: trunk)
18:02
The ability to DROP multiple tables (or views or indexes or triggers) at once, as implemented by this branch, appears to work. However, this branch adds over 1000 bytes to the code footprint. And the changes is significant, adding risk. The benefit of being able to drop multiple tables in a single statement is not seen to be enough to overcome those downsides, and so there are no immediate plans to merge this branch. (Leaf check-in: 1a49788a user: drh tags: multi-drop)
16:44
Change test results in corruptC.test to account for the added detail in the output of "PRAGMA integrity_check" from [b736519d3d2e93c7]. (check-in: 0d19eeaa user: drh tags: trunk)
16:18
Fix harmless compiler warnings in the increment integrity_check extension. (check-in: 596668e6 user: drh tags: trunk)
13:58
Adjust date/time functions so that they do a better job of keeping track of whether the current time is UTC or localtime, and no-op the 'utc' and 'localtime' modifiers accordingly. See forum post e7a939e074. Also add the datedebug() function, available only under -DSQLITE_DEBUG, for improved visibility of the DateTime object during debugging and testing. (check-in: dc569683 user: drh tags: trunk)
11:12
Fix assert() statements in date/time computations: The month and day numbers can be zero if an error has been seen. (check-in: fc773f6c user: drh tags: trunk)
09:56
Document why the wasm build does not use Emscripten's -sSTRICT=1 flag and account for a behaviour change in emsdk 3.1.55 regarding the HEAPxyz symbols. (check-in: f6e88720 user: stephan tags: trunk)
09:23
When warning about multiple invocations of JS's sqlite3ApiBootstrap(), use the warning logger installed by the first invocation, rather than console.warn. (check-in: d162af2d user: stephan tags: trunk)
06:54
After wasm bootstrapping has read the optional global-scope config objects, delete those objects because any further changes would not have any useful effect. Expand the docs regarding the sqlite3ApiBootstrap() config object and explain the necessary evil of a separate global-scope object for that configuration. (check-in: e6b14f73 user: stephan tags: trunk)
2024-03-03
20:15
Back out the previous change. Replace it with new date modifiers "ceiling" and "floor". (check-in: f0831cce user: drh tags: trunk)
2024-03-02
21:02
New date/time modifiers "mnth" and "yr" work like "month" and "year" but resolve day-of-month overflow by truncating rather than rolling over into the next month. Forum thread 232d1abb5d (check-in: 5d392c16 user: drh tags: trunk)
20:50
Change the magic time-interval names that do truncate-to-same-month to be "mnth" and "yr" - "month" and "year" without the vowels. (Closed-Leaf check-in: 296b46c5 user: drh tags: month-truncate)
20:39
Attempt to use less memory when handling a large VALUES clause attached to an INSERT statement. This branch is buggy. (check-in: 6d4f1ae2 user: dan tags: exp-values-clause)
13:38
Proof-of-concept for new time-interval operator "pg-month" and "pg-year" that use the truncate-to-month algorithm for month overflow instead of the wrap-to-next-month algorithm that is used by SQLite by default. (check-in: b606c096 user: drh tags: month-truncate)
12:17
Remove an unused line of code. (check-in: 26272590 user: drh tags: multi-drop)
11:02
Simplifications to multi-DROP logic. (check-in: e70d6be7 user: drh tags: multi-drop)
00:29
Improved rebustness against OOM and corruption. (check-in: 3992f450 user: drh tags: multi-drop)
2024-03-01
22:42
When doing DROP TABLE or DROP INDEX, use a heap to ensure that the various btrees are dropped in the right order for autovacuum. (check-in: e5bf9556 user: drh tags: multi-drop)
19:08
RTREE uses multi-DROP to erase shadow tables. (check-in: cf8a58d6 user: drh tags: multi-drop)
18:05
Omit unreachable branches in the parser. (check-in: 90de1f73 user: drh tags: multi-drop)
15:56
Fix harmless compiler warning. (check-in: 33d3bb84 user: drh tags: multi-drop)
14:47
Improved handling of foreign-key constraints on a multi-drop of tables. (check-in: 144e332f user: drh tags: multi-drop)
12:21
When doing a multi-drop, delete btrees beginning with the largest root page first and working toward smaller root pages, to avoid problems with root page renumbering due to autovacuum. (check-in: 37d4c5e2 user: drh tags: multi-drop)
2024-02-29
19:40
Attempt to fix multi-object DROP. Working better, but still not perfect. (check-in: 02f9fc73 user: drh tags: multi-drop)
13:44
Add the ability to DROP one or more objects of the same class in a single statement by listing the objects as multiple arguments to the DROP command. (check-in: 2266086c user: drh tags: multi-drop)
10:55
Fix two -Werror=lto-type-mismatch warnings reported in forum post ef62b57bd5. (check-in: 803481f2 user: stephan tags: trunk)
03:45
Fix two -Werror=lto-type-mismatch warnings reported in forum post ef62b57bd5. (Closed-Leaf check-in: 29f94610 user: stephan tags: lto-type-mismatch)
2024-02-28
19:44
Put NEVER() on a branch in STAT4 that appears to be unreachable. (Closed-Leaf check-in: 0ca35e92 user: drh tags: mistake)
15:32
Add in a VdbeCoverage() macro needed by STAT4 that should have been part of check-in [63ef234e88857a65]. (check-in: d51c699a user: drh tags: trunk)
01:12
Always convert 32-bit integer literals into EP_IntValue notation, even if they contain "_" separators. (check-in: 2dfc427f user: drh tags: trunk)
2024-02-27
20:19
Fix an eputf() macro without its argument in the CLI. (check-in: c2f5e257 user: drh tags: trunk)
19:57
Add an assert() to help scan-build recognize that a variable is initialized. (check-in: 10d26f1e user: drh tags: trunk)
16:36
Adjust a VdbeCoverage() macro in PRAGMA integrity_check to account for the fact that arguments to the OP_Eq opcode used are never NULL. (check-in: 759ccea3 user: drh tags: trunk)
16:25
Adjust a VdbeCoverage() macro in PRAGMA integrity_check to account for the fact that arguments to the OP_Eq opcode used are never NULL. (Closed-Leaf check-in: 8c788fc4 user: drh tags: wrong-branch)
15:56
Fix typo in the help text of the CLI - a typo that prevented the display of help for the ".imposter" command. Turns out the comma is deliberate. The ".imposter" command is deliberately undocumented to avoid confusing non-devs. (Closed-Leaf check-in: 0d9a2596 user: drh tags: mistake)
15:33
Remove unreachable branches from the PRAGMA integrity_check enhancement of [b736519d3d2e93c7]. (check-in: 6d02fb23 user: drh tags: trunk)
14:31
Ensure memory cells are properly freed before they are overwritten by OP_IntegrityCk. (check-in: e50acaf9 user: dan tags: trunk)
11:03
Have "PRAGMA quick_check" compare the number of entries in tables and indexes. (check-in: b736519d user: dan tags: trunk)
11:02
Extend sqlite3.c makefile rule to support EXTRA_SRC=list-of-c-files to append to the generated sqlite3.c, as discussed in/around forum post ccda88cf6f1754c5. (check-in: 61676f1e user: stephan tags: trunk)
10:52
Allow "_" characters to appear between any two digits in an integer, real or hexadecimal SQL literal. (check-in: 0e6700f4 user: dan tags: trunk)
2024-01-23
11:20
Add extra checks for the validity of a numeric literal to sqlite3DequoteNumber(). (Closed-Leaf check-in: d57407ef user: dan tags: digit-separators)
2024-01-22
19:42
Fix a problem in the previous commit with hex literals that start with "0X" instead of "0x". (check-in: c063c89b user: dan tags: digit-separators)
19:38
Allow underscores to occur in hex literals. (check-in: 81a56229 user: dan tags: digit-separators)
19:00
Add extra tests for the code on this branch. (check-in: b15bbb20 user: dan tags: digit-separators)
17:18
Do not allow digit separators that are not surrounded on both sides by digits. (check-in: 5bbc378a user: drh tags: digit-separators)
11:12
Fix test script literal.test so that it works with SQLITE_OMIT_ALTER_TABLE builds. (check-in: b57327be user: dan tags: digit-separators)
2024-01-20
18:45
Merge trunk changes into this branch. (check-in: 03ade4a8 user: dan tags: digit-separators)
16:46
Add test cases for the new code on this branch. (check-in: 49f29a7b user: dan tags: digit-separators)
16:18
Allow "_" characters to appear following any digit in an integer or real SQL literal. (check-in: 401650aa user: dan tags: digit-separators)