/ Timeline
Login

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

100 most recent check-ins

2021-01-17
00:13
Minor simplification of the EXISTS-to-IN logic. (Leaf check-in: cac90a9f user: drh tags: exists-to-in)
2021-01-16
20:22
Improved handling of vector equalities in the EXISTS-to-IN translator. (check-in: ef49ee4a user: drh tags: exists-to-in)
18:55
Give the EXISTS-to-IN optimization the ability to handle some cases that involve vector comparisons, instead of throwing a mysterious error in those cases. (check-in: 87e78a19 user: drh tags: exists-to-in)
18:22
Add debugging output about the EXISTS-to-IN optimization when the ".wheretrace" flag has the 0x20 bit set. (check-in: 0dad5ce3 user: drh tags: exists-to-in)
12:15
Fix a hyperlink in the Lemon documentation. (Leaf check-in: 2ffb2ffa user: drh tags: trunk)
2021-01-15
17:51
Add OOM injection tests for new code on this branch. (check-in: 9a181dba user: dan tags: exists-to-in)
16:37
Update header comments for routines added by this branch. (check-in: 950030d6 user: dan tags: exists-to-in)
15:32
Ensure the EXISTS->IN transformation preserves the collation sequence of the comparison operation. (check-in: a373baae user: dan tags: exists-to-in)
15:21
Fix a potential NULL pointer dereference following OOM. (check-in: 8ce3cb90 user: drh tags: exists-to-in)
15:17
Add a new optimizer disabling bit to close off the exists-to-in optimization, for testing purposes. (check-in: a80c9a07 user: drh tags: exists-to-in)
14:25
Small performance improvement in the EXISTS-to-IN translator for the common case where the EXISTS operator is not found in the WHERE clause. (check-in: dcb7772d user: drh tags: exists-to-in)
14:15
Merge the latest trunk enhancements into the exists-to-in branch. (check-in: 13c4c908 user: drh tags: exists-to-in)
11:39
Add simple tests (and a fix) for the change on this branch. (check-in: 897f3f40 user: dan tags: exists-to-in)
2021-01-14
20:57
Improvements to the min/max optimization. Fix for a performance regression introduced at [b8ba2f17f938c035] reported by forum post 4050026ab8 (check-in: 249a71cc user: drh tags: trunk)
20:50
Allow the planner to convert an EXISTS(SELECT...) expression in a WHERE clause to the equivalent IN(...) expression in situations where this is possible and advantageous. (check-in: 9f90a882 user: dan tags: exists-to-in)
00:53
The early-out of the inner loop on the min/max optimization was overly aggressive for the cases where there is a join and outer loops contain IN operators. Fix this. Test case in TH3. (Closed-Leaf check-in: ccd3bae1 user: drh tags: minmax-opt-exp)
2021-01-13
21:05
Fix a harmless compiler warning. (check-in: 83ec01e3 user: drh tags: minmax-opt-exp)
19:28
Expand the number of optimization-disable bits from 16 to 32. Use one of the new bits to disable the min/max optimization, so that we can more easily verify that we get the same answer both with and within that optimization. (check-in: fd0c9a12 user: drh tags: minmax-opt-exp)
15:23
Further enhancements to the min/max optimization of check-in b8ba2f17f938c035 to fix the performance regression identified by forum post 4050026ab8. (check-in: 188772a1 user: drh tags: minmax-opt-exp)
12:59
cli: Omit surplus whitespace at the end of lines in .explain output. (check-in: 11e4eb09 user: drh tags: trunk)
11:44
In the .selecttrace output, show the ORDER BY clause added by the Min/Max optimization. (check-in: db0ecfe6 user: drh tags: trunk)
2021-01-12
20:16
Lexer and grammar rules for a RETURNING clause on DELETE/INSERT/UPDATE. Actually making this work, though, will involve a lot more code which will likely slow down processing for the common case where there is no RETURNING clause. Furthermore, RETURNING seems to be of limited usefulness and it is not standard SQL. So we abandon it here. These experimental changes are parked in a branch as an historical reference. If circumstances changes, we might take up the cause again some day. (Leaf check-in: abf8da81 user: drh tags: returning)
16:26
Merge the ParseCleanup enhancement to trunk. (check-in: 35824c1b user: drh tags: trunk)
15:30
Fix a potential use-after-free following an OOM in sqlite3ParserAddCleanup() and add a mechanism to detect situations where this might occur in the future. (Closed-Leaf check-in: 38ef8ab9 user: drh tags: parse-cleanup)
14:23
Rerun autoconf. (check-in: 8f3ab5da user: drh tags: trunk)
14:19
Fix for the top-level configure script so that it works with tcl 8.7. (check-in: 4810f814 user: dan tags: trunk)
2021-01-11
20:37
Add a linked list of ParseCleanup objects to the end of a Parse object and use that list as a place to put other sub-objects that need to be deallocated. Have a single such list for infrequently used sub-objects is more efficient than doing an a separate check for each kind of sub-object. Edit: Now possible to have use-after-free following an OOM. (check-in: affa2b7b user: drh tags: parse-cleanup)
2021-01-09
19:10
New CLI command: ".stats vmstep" enables the display of the virtual-machine step count only, after each command. Useful for optimization problems. (check-in: 49dfce46 user: drh tags: trunk)
18:24
More detailed compile-time testing before attempting to use atomic load intrinsics. See forum post fc0237a39b30ac0a. (check-in: 5204c2c4 user: drh tags: trunk)
2021-01-08
19:53
Fix an issue with sha3_query() when the first argument contains blank SQL statements. (check-in: 24baab9a user: drh tags: trunk)
2021-01-07
16:59
Update cksumvfs to check that the xCurrentTimeGetInt64 method of the underlying VFS is not NULL before invoking it. (check-in: c71f6cad user: dan tags: trunk)
16:29
Fix problems with some "crashsql" tests. (check-in: 0c8e2ede user: dan tags: trunk)
16:10
Fix harmless typos in comments per forum post 7849e58dd5 (check-in: d1e22e2f user: drh tags: trunk)
2021-01-04
18:28
Add extra test for handling of embedded nul characters in the fts4 unicode61 tokenizer. (check-in: c2c2c7e9 user: dan tags: trunk)
2021-01-02
23:56
In Lemon, factor the parser stack overflow detection logic out of the yy_reduce() subroutine and into the main parser routine, so that when overflow is detected, it can exit immediately. This saves a single conditional in the main loop of the parser. (check-in: 203c049c user: drh tags: trunk)
2021-01-01
22:06
Avoid allocating space to hold the prepared statements for CREATE statements when parsing the schema of an existing database, since those prepared statements are never used.. This helps to make startup faster, (check-in: d01e9f2d user: drh tags: trunk)
21:02
Faster and smaller test to ensure that the sqlite_schema.sql field is always a CREATE statement of some kind. (check-in: 76de2bb0 user: drh tags: trunk)
20:04
Change the unions of the Table.addColOffset field from characters to bytes. This makes the query that implements ALTER TABLE ADD COLUMN more complex and slightly slower, but also makes CREATE TABLE statement parsing faster by avoiding a call to sqlite3UtfCharLen(). Since, CREATE TABLE parsing is far more common than ALTER TABLE, this is a net win for performance. (check-in: 6f25f252 user: drh tags: trunk)
19:17
Use the column name hash to improve performance of column name collision detection while parsing CREATE TABLE statements. (check-in: d02820f0 user: drh tags: trunk)
18:32
Modify the makefile rule for "startup" so that it always builds with -Os and -DSQLITE_THREADSAFE=0 and no other options, for consistency of performance. (check-in: 5ac939e0 user: drh tags: trunk)
18:23
Size reduction and performance increase in sqlite3Prepare(). (check-in: 41f45c8e user: drh tags: trunk)
17:01
Small size reduction and performance improvement in sqlite3VdbeMakeReady() by linking the new prepared statement into the prepared statement list sooner rather than later. (check-in: 2996e800 user: drh tags: trunk)
16:43
Streamline processing of the authenticator callback for the common case when there is no callback. (check-in: d3196685 user: drh tags: trunk)
15:13
Add the "startup" test program designed to measure startup performance, and in particular schema parsing time. (check-in: 7b3b31ef user: drh tags: trunk)
01:44
Do not attempt to take a pointer to the ceil() and floor() functions as those routines are intrinsics on some versions of MSVC. (check-in: e5d7209e user: drh tags: trunk)
2020-12-30
13:20
New test case for the HAVING fix of check-in [f62f983b56623f0e]. (check-in: 45f46317 user: drh tags: trunk)
13:10
New test cases for cursor renumbering in the UNION ALL query flattener. (check-in: 270babf2 user: drh tags: trunk)
2020-12-29
16:48
Add an ALWAYS on an always-true conditional in the cursor renumbering logic of the UNION ALL flattener. Edit: turns out this change is incorrect and the conditional is reachable after all. See the nearby trunk check-in for test cases. (Closed-Leaf check-in: 5774318e user: drh tags: mistake)
15:06
Do not set the P3 parameter on OP_RowCell when copying an index btree, as P3 is not used in that case. (check-in: eef070a4 user: drh tags: trunk)
2020-12-28
21:42
Fix missing comma in ctime.c that would cause the ENABLE_MATH_FUNCTIONS output rw to merge with whatever row followed. Problem reported in forum post aacac97680. (check-in: 328bc4a0 user: drh tags: trunk)
2020-12-23
16:46
Add the sqlite3session_memory_used() API to the sessions module. For querying the amount of heap memory currently being used by a session object. (check-in: 823f75c2 user: dan tags: trunk)
2020-12-22
20:35
Fix SQLITE_OMIT_WINDOWFUNC builds by moving declaration of sqlite3ExpandSubquery out of "ifndef SQLITE_OMIT_WINDOWFUNC" block. (check-in: 9587fa8b user: dan tags: trunk)
19:57
Fix a couple spelling typos in comments. (check-in: 907ddf86 user: mistachkin tags: trunk)
16:23
Fix a problem handling sub-queries with both a correlated WHERE clause and a "HAVING 0" clause where the parent query is itself an aggregate. The problem was apparently introduced by check-in [6e6b3729e0549de0] (check-in: f62f983b user: dan tags: trunk)
14:54
Simplification to the aggregate-function analysis error detection logic at the end of sqlite3Select(). (check-in: 82884438 user: drh tags: trunk)
2020-12-21
19:50
Fix problems with joining UNION ALL sub-queries against other sub-queries that contain LEFT JOIN. (check-in: d554f710 user: dan tags: trunk)
18:39
Fix a problem when flattening joins between a UNION ALL sub-query and another sub-query that uses more than one window function. (check-in: ef9733fe user: dan tags: trunk)
14:51
Enhance documentation to show that "ro" is the correct way to say "readonly" in the mode= query parameter. (check-in: 788b9685 user: drh tags: trunk)
12:14
Add the --timer option to fuzzcheck. Get the --timeout option working in fuzzcheck when running dbsql tests. (check-in: 3b0c9b41 user: drh tags: trunk)
2020-12-20
14:51
Always declare the sqlite3WhereTrace variable, even for non-debug builds. (check-in: 88d93ee3 user: drh tags: trunk)
2020-12-19
15:39
Fix a broken assert() in fts5 that could be triggered by corrupt database records. (check-in: b79f59f9 user: dan tags: trunk)
13:58
Allow UNION ALL sub-queriesto be flattened even if the parent query is a join. (check-in: df1d6482 user: drh tags: trunk)
2020-12-18
18:04
Fix for the previous fix in the case where a UNION ALL sub-query is joined against some other compound query. (Closed-Leaf check-in: 63c5cfb9 user: dan tags: union-all-flattener)
16:13
When flattening UNION ALL subqueries into a join query, ensure that separate cursor numbers are used for each segment of the newly flattened query. (check-in: c510377b user: dan tags: union-all-flattener)
2020-12-17
17:17
Fix a part of the header comment for flattenSubquery(). (check-in: dc0937ce user: dan tags: union-all-flattener)
16:48
Add test cases and minor fixes to this branch. (check-in: 5d6dc29d user: dan tags: union-all-flattener)
15:17
In the CLI, add the ".filectrl data_version" command. And put the various ".filectrl" subcommands in alphabetical order. (check-in: 34344521 user: drh tags: trunk)
11:30
Merge latest trunk changes into this branch. (Leaf check-in: 9f41f604 user: dan tags: reuse-schema)
11:24
Fix a problem with sqlite3_expanded_sql() that could occur with statements that use both numbered (e.g. "?1") and unnumbered (i.e. "?") parameters. (check-in: 2a6cd683 user: dan tags: trunk)
2020-12-16
21:09
Enhance the sqlite3BtreeTransferRow() routine so that it does more careful checks for corrupt database pages. (check-in: 85952e71 user: drh tags: trunk)
20:00
Allow sub-queries that use UNION ALL to be flattened, even if the parent query is a join. Still some problems on this branch. (check-in: 00e4bf74 user: dan tags: union-all-flattener)
14:20
Remove an unnecessary and incorrect #ifdef. Fix harmless compiler warnings. (check-in: 31cd1bbf user: drh tags: trunk)
13:20
Fix a typo in the sqlite3_free_filename() documentation. (Leaf check-in: 7316ee19 user: drh tags: branch-3.34)
13:17
Fix a typo in the sqlite3_free_filename() documentation. (check-in: 48301edc user: drh tags: trunk)
2020-12-15
19:27
Fix another integer overflow triggered by a corrupt database in recently modified vacuum code. (check-in: 4e2dd2a5 user: dan tags: trunk)
16:28
When the -statstep option is passed to the "rbu" executable, print out memory stats right before exiting, as well as every -statstep steps. (check-in: 94f81b51 user: dan tags: trunk)
13:55
Change an fts5 assert() that can be triggered by a corrupt database to an if() condition. (check-in: ea0a7f10 user: dan tags: trunk)
2020-12-14
16:50
Merge the latest trunk enhancements into the begin-concurrent-pnu branch. (Leaf check-in: daf04650 user: drh tags: begin-concurrent-pnu)
16:26
Merge recent trunk enhancements into the begin-concurrent branch. (Leaf check-in: a1708e84 user: drh tags: begin-concurrent)
15:39
Enhance UPSERT so that it allows multiple ON CONFLICT clauses and does not require a conflict target for DO UPDATE. (check-in: 6b01a24d user: drh tags: trunk)
15:25
Fix an integer overflow problem in new VACUUM code. (check-in: 59b4367f user: dan tags: trunk)
13:52
Minor changes for test coverage. (Closed-Leaf check-in: e5a8fa50 user: drh tags: generalized-upsert)
2020-12-12
00:43
More test cases. No new problems discovered. (check-in: f34dd67e user: drh tags: generalized-upsert)
00:28
New test cases with corresponding bug fixes. (check-in: f22c21a9 user: drh tags: generalized-upsert)
2020-12-11
19:36
Begin adding test cases. Fix one bug found so far. More are pending. (check-in: aadd67dd user: drh tags: generalized-upsert)
19:01
Fix an assert() broken by recent changes to vacuum. (check-in: dd058da8 user: dan tags: trunk)
17:11
Small performance tweaks. (check-in: 5321d60c user: drh tags: generalized-upsert)
16:49
Bug fixes so that legacy tests pass. New tests for new functionality have not yet been added. (check-in: aa76790e user: drh tags: generalized-upsert)
16:03
Add an "#ifndef SQLITE_AMALGAMATION" block around the typedef for u64 in shathree.c. (check-in: b411f29e user: dan tags: trunk)
14:34
Fix the configure script on the autoconf distribution so that the --disable-threadsafe option works. See FreeBSD bug 244459. (check-in: e64ff2cc user: drh tags: trunk)
14:22
Fix the expert extension so that it handles generated columns correctly. (check-in: c2ae7ba6 user: dan tags: trunk)
01:17
Logic is in place to handle multiple ON CONFLICT clauses, but it does not work. Any use of ON CONFLICT will likely lead to memory faults. This is an incremental check-in to save my place. (check-in: 15514231 user: drh tags: generalized-upsert)
2020-12-10
20:31
Use an iterator for the index loop in sqlite3GenerateConstraintChecks(). The idea is that this iterator can be enhanced to traverse the indexes in any order, as required by multi-index UPSERT. (check-in: 64a4a91e user: drh tags: generalized-upsert)
19:54
Remove a redundant branch added by [56a54258560]. (check-in: b4d6f6d7 user: dan tags: trunk)
19:51
Changes to avoid loading large records into memory within VACUUM. (check-in: 56a54258 user: dan tags: trunk)
19:20
Fix minor issues with new code on this branch. (Closed-Leaf check-in: f7fa76d0 user: dan tags: vacuum-lomem)
18:07
Merge trunk changes into this branch. (check-in: 7337eed6 user: dan tags: vacuum-lomem)
18:06
Better integrate the changes on this branch with OP_Insert and OP_IdxInsert. (check-in: 101cef14 user: dan tags: vacuum-lomem)
13:49
Add an optional function to the chsumvfs extension. When activated by the SQLITE_CKSUMVFS_INIT_FUNCNAME macro, this function will invoke the file-control that sets the number of reserved bytes to 8. This can be used to initialize a cksumvfs database file by programming languages that do not have access to the sqlite3_file_control() interface. (check-in: 01841fb4 user: drh tags: trunk)
12:49
The DO UPDATE code generator searches for the correct ON CONFLICT clause to use. (check-in: a47e35ee user: drh tags: generalized-upsert)