/ Timeline
Login

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

50 check-ins occurring around 84ae41fd2f50cb7c.

2019-09-02
00:58
When computing an expression value for an index-on-expression or a CHECK constraint and the expressions uses a REAL table column, but the value of that column is an integer (in other words, when it is using the store-real-as-integer optimization) be sure to promote the value to real before evaluating the expression. Ticket [57af00b6642ecd68]. (check-in: 0658c16e31 user: drh tags: trunk)
2019-09-01
23:36
Remove an obsolete paragraph from the OP_Column documentation. No code changes. (check-in: f6d8956cf8 user: drh tags: trunk)
2019-08-31
20:29
Fix a harmless compiler warning. (check-in: 63c67a54b4 user: drh tags: trunk)
20:26
Mark new VDBE branches never taken. (check-in: 83c2adffbf user: drh tags: trunk)
20:13
Improvements to the algorithm that determines which SELECT in a sequence of nested SELECT statements that an aggregate function belongs to. This resolves an issue identified by dbsqlfuzz. (check-in: d768007473 user: drh tags: trunk)
17:14
If a DELETE trigger fired by an UPDATE OR REPLACE statement deletes the row being updated, do not attempt to proceed with the original UPDATE operation. Fix for [d6a0fbc1]. (check-in: 4145b3e050 user: dan tags: trunk)
01:33
Remove some affinity tests that became unreachable due to the prior change. (check-in: e2db1123fa user: drh tags: trunk)
2019-08-30
23:56
When the affinity of a table column is INT or REAL, make the affinity of corresponding index columns NUMERIC. This increases the precision of index lookups for large numbers so that it matches the precision of ordinary comparison operators. Ticket [40812aea1fde9594] (check-in: e0d909c740 user: drh tags: trunk)
23:15
Make sure OP_RealAffinity has been applied to all columns of type REAL in the excluded.* pseudo-table of an UPSERT. Ticket [5a3dba8104421320] (check-in: 67381daded user: drh tags: trunk)
21:52
Improvements to VDBE tracing. No changes to code in normal deliverables. (check-in: 54553bf16f user: drh tags: trunk)
19:45
Add test case to window8.test. Also fix an error in a comment in window.c. (check-in: 2925bfa597 user: dan tags: trunk)
18:02
Back out the change at [47cd634c98b502d4] which was incorrect. Add a test case so that we don't accidently back out that change again. (check-in: 596ac2a4ea user: drh tags: trunk)
17:28
Add further comments to window.c. (check-in: 900464567b user: dan tags: trunk)
16:46
New test cases for window functions with RANGE BETWEEN and DESC NULLS FIRST. (check-in: f7002f86c7 user: drh tags: trunk)
16:14
Fix a bug in RANGE window functions that use "ORDER BY <expr> DESC NULLS FIRST" as the window-frame ORDER BY clause. (check-in: 39b4cad4a5 user: dan tags: trunk)
16:00
The expression "(X IS FALSE) IN (FALSE)" does not imply that X is NOT NULL. Ticket [f8f472cbc77ba9c9] (check-in: dd66134817 user: drh tags: trunk)
15:11
The expression "(X IS FALSE) BETWEEN FALSE AND TRUE" does not implie that X is not NULL. Ticket [fba33c8b1df6a915] (check-in: 057fb8b180 user: drh tags: trunk)
13:07
The expression "(x IS FALSE) IS FALSE" does not imply that X is not NULL. Ticket [a6408d42b9f44462] (check-in: 45ff2b1f26 user: drh tags: trunk)
2019-08-29
23:24
Remove a faulty testcase() macro. (check-in: 80124ae552 user: drh tags: trunk)
21:16
Improve vdbe branch coverage of NULLS LAST code. (check-in: e8e9f77d52 user: dan tags: trunk)
19:34
Fix other problems similar to ticket [c0390363]. (check-in: 96ff2ba9c4 user: dan tags: trunk)
16:48
Add a missing VdbeCoverage() macro. (check-in: 33da6092d3 user: drh tags: trunk)
16:17
Remove an unreachable branch in the NULLS LAST logic of RANGE window functions. (check-in: 47cd634c98 user: drh tags: trunk)
15:50
Fix another case where SQLite assumes that if "~(? AND FALSE)" is true, "?" must be non-null. (check-in: 616f5663b3 user: dan tags: trunk)
15:06
Avoid assuming that for "~ (? AND FALSE)" to be true, "?" must not be NULL, just as we do for "NOT (? AND FALSE)". Fix for ticket [c0390363]. (check-in: 84ae41fd2f user: dan tags: trunk)
14:25
Fix a potential buffer overrun in fts5 caused by corrupted database records. (check-in: c465d0eb47 user: dan tags: trunk)
00:27
Improve Lemon so that it enlarges some of its tables slightly in order to avoid having to index range checks on table lookups for a performance increase. (check-in: 4be6a23a18 user: drh tags: trunk)
2019-08-28
11:49
Fix an unreachable branch in sqlite3ParserFallback() (Closed-Leaf check-in: e059178b47 user: drh tags: lemon-optimization)
11:31
Further improvements to parser speed by enlarging lookup tables to eliminate the need to do range checking on the index prior to lookup. (check-in: 47d3e091ae user: drh tags: lemon-optimization)
02:09
Increase the size of the yy_lookahead table so that it is never necessary to down bounds checking on the index. (check-in: bafd872398 user: drh tags: lemon-optimization)
2019-08-27
19:59
Add support for "ORDER BY ... NULLS FIRST" and "ORDER BY ... NULLS LAST". Use this to fix ticket [f8a7060e]. (check-in: 94085fb3e7 user: dan tags: trunk)
17:59
The ALWAYS() added by the previous check-in was incorrect. Take it back out. (check-in: 336235db2b user: drh tags: trunk)
17:28
Add ALWAYS() to an always true conditional that results from the previous check-in. Add a test case for ticket [dbaf8a6820be1ece] to supplement those already checked into TH3. (check-in: aff2098047 user: drh tags: trunk)
17:01
Omit the "x IN (y)" to "x==y" optimization of check-in [e68b427afbc82e20] (and ticket [e39d032577df6942]) as it causes difficult affinity problems as demonstrated by ticket [dbaf8a6820be1ece] and the original assertion fault is no longer a factor due to countless other changes of the previous 5 years. (check-in: 7f5168a76a user: drh tags: trunk)
10:05
If a TEMP TRIGGER references an auxiliary schema, and that auxiliary schema is detached, move the trigger to reference the TEMP schema before completing the detach, so that the trigger does not hold a dangling schema pointer. Ticket [ac8dd4a32ba4322f] (check-in: 069c2f4c61 user: drh tags: trunk)
2019-08-26
20:41
Merge in recent fixes from trunk. (Closed-Leaf check-in: ad816d01d4 user: drh tags: nulls-last)
14:57
Tweak the shell tool ".recover" command so that it can recover rows that consist of a rowid and no fields. (check-in: 279ac7fdec user: dan tags: trunk)
14:18
Improved detection of number of column mismatch for vector assignment in UPDATE statements. Ticket [78acc9d40f0786e8] (check-in: bd4bda73df user: drh tags: trunk)
13:45
Fix typo in a comment. No code changes. (check-in: d0cc06d8a3 user: drh tags: trunk)
12:50
Enforce 80-character line discipline in the CLI. Minor improvements to the CLI built-in help. (check-in: 9690013a00 user: drh tags: trunk)
2019-08-24
21:02
Minor performance improvement for balance_nonroot(). (check-in: d7434cae3e user: dan tags: trunk)
17:11
Fix a potential buffer overrun in fts5 caused by corrupt database records. (check-in: 156d612800 user: dan tags: trunk)
2019-08-23
23:05
Fix the built-in edit() SQL function in the CLI so that it works with zero-length blobs. (check-in: e324901286 user: drh tags: trunk)
21:11
Attempt to parse the NULLS LAST clause more efficiently. A few cycles were saved, but at the expense of more code. And there is a bug was introduced somewhere. Not worth continuing down this line. (Closed-Leaf check-in: 12d2cf8871 user: drh tags: dead-end)
20:33
Move some things in parse.y so that TK_COLUMN and TK_AGG_COLUMN are assigned the same values as they are on trunk for a very small speedup. (check-in: d26fdfa3bc user: dan tags: nulls-last)
17:09
Revise the VDBE comments for NULL-scan so that they also work make sense when reading a NULLS FIRST plan. (check-in: bfe793780f user: drh tags: nulls-last)
17:00
Additional simplifications of the WHERE loop code generator logic for NULLS LAST. (check-in: 1383680d92 user: drh tags: nulls-last)
16:12
Simplification of the WHERE loop code generator for NULLS LAST saves a few CPU cycles and about a hundred bytes of code space. (check-in: e3ed2f496f user: drh tags: nulls-last)
13:32
Invert the meaning of the regBignull flag so that it is 1 when doing the normal scan and 1 when scanning nulls. This enables the re-do jump at the bottom of the loop to be coded with a single OP_IfNotZero opcode, rather than a sequence of OP_If, OP_Integer, OP_Goto. (check-in: bf875e1a25 user: drh tags: nulls-last)
13:08
Extra VDBE comments in the NULLS LAST logic provide landmarks to help understand the EXPLAIN output. (check-in: 649b08ead5 user: drh tags: nulls-last)