Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
50 ancestors of ef94ea061d2ec8ee
|
2017-10-04
| ||
| 14:13 | New test cases for ticket [b899b6042f97f5] derived from a bug report on the mailing list from Wout Mertens. (check-in: ef94ea061d user: drh tags: trunk) | |
| 12:08 | Amend recent changes to busy.test to avoid failing with SQLITE_ENABLE_STAT4 builds. (check-in: 9ff4944194 user: dan tags: trunk) | |
| 12:06 | Turn restriction 20 on the query flattener into an assert since the situation restricted can no longer occur because of the more aggressive use of co-routines. (check-in: 4464f40ccd user: drh tags: trunk) | |
| 10:39 | Add tests to verify that the busy-handler is invoked correctly when processing "PRAGMA optimize" and ANALYZE commands. (check-in: fb83c3d8df user: dan tags: trunk) | |
| 05:59 | Remove a redundant restriction from the query flattener. (check-in: 66629b2a09 user: drh tags: trunk) | |
| 02:30 | Fix a minor typo in the query flattener header comment. No code changes. (check-in: d050dc605c user: drh tags: trunk) | |
|
2017-10-03
| ||
| 19:53 | Simplify the computation of types on columns of a view. (check-in: 772b0db146 user: drh tags: trunk) | |
| 18:35 | Do not enable SELECT tracing with -DSQLITE_DEBUG. Require the -DSQLITE_ENABLE_SELECTTRACE compile-time option to enable SELECT tracing. (check-in: f5c395834c user: drh tags: trunk) | |
| 17:29 | Add some OK_IF_ALWAYS_TRUE() marks on optimization branches in select.c. (check-in: 1dd828088d user: drh tags: trunk) | |
| 17:17 | Remove an unused variable from sqlite3Insert() and fix harmless compiler warnings associated with -DSQLITE_MUTATION_TEST. (check-in: 7be760e907 user: drh tags: trunk) | |
| 16:57 | Fix the sqlite3_column_decltype() interface to return the correct datatype when the value comes through a CTE. (check-in: 966438bd25 user: drh tags: trunk) | |
| 14:24 | The simplification to sqlite3SelectExpand() in last night's [3a4ffb21] check-in was not completely correct. This adjustment is needed for correct OOM handling in some configurations. (check-in: c5ad5e1675 user: drh tags: trunk) | |
| 03:01 | Do not compute column and row size estimates for the transient Table objects associated with the result set of a SELECT statement, since those estimates are never used for anything constructive. (check-in: e4342fd401 user: drh tags: trunk) | |
|
2017-10-02
| ||
| 23:50 | Mark a single branch within sqlite3SelectExpand() as an optimization. (check-in: 3a4ffb21d0 user: drh tags: trunk) | |
| 23:30 | Simplification to sqlite3SelectPrep() for improved testability. (check-in: 5bf2ccb908 user: drh tags: trunk) | |
| 21:29 | Add the OK_IF_ALWAYS_TRUE() and OK_IF_ALWAYS_FALSE() macros for marking conditionals that improve performance but do not change the outcome. (check-in: 6035c9b272 user: drh tags: trunk) | |
| 13:20 | Avoid using lookaside memory for persistent virtual table structures. (check-in: d815f255df user: drh tags: trunk) | |
| 02:52 | The query planner now always prefers to implement a FROM-clause subquery using a co-routine rather than flattening the subquery into the outer query. (check-in: c9104b59c7 user: drh tags: trunk) | |
| 02:32 | Improvements to a comment in lempar.c. No code changes. (check-in: 8000d23017 user: drh tags: trunk) | |
|
2017-09-30
| ||
| 11:47 | More details in the header comment of pushDownWhereTerms(). No changes to code. (Closed-Leaf check-in: 928486231f user: drh tags: prefer-coroutine-sort-subquery) | |
| 10:50 | Make sure the SQLITE_Stat34 optimization switch is always 0x800, a value which is hard-coded in the TH3 test suite. (check-in: 6aed4ea34c user: drh tags: prefer-coroutine-sort-subquery) | |
| 01:25 | Fix unreachable conditionals and revise a testcase that was made obsolete by the changes on this branch. (check-in: 71f0adf7ca user: drh tags: prefer-coroutine-sort-subquery) | |
|
2017-09-29
| ||
| 22:13 | Always render a subquery that is not part of a join as a co-routine. (check-in: 6b1651d711 user: drh tags: prefer-coroutine-sort-subquery) | |
| 16:08 | Merge the query flattener comment improvements from trunk. (check-in: f62cd4d940 user: drh tags: prefer-coroutine-sort-subquery) | |
| 16:07 | Clean up the comments on the query flattener to more clearly and accurately express the conditions under which the flattener is able to run. (check-in: 0840f9f824 user: drh tags: trunk) | |
| 14:31 | Merge changes from trunk. (check-in: 06f432fb7c user: drh tags: prefer-coroutine-sort-subquery) | |
| 12:44 | Make sure the 6th parameter to the authorizer callback for view subqueries has the correct view name. (check-in: 2a45bbc9fd user: drh tags: trunk) | |
| 12:12 | Better names for subqueries in EXPLAIN comments. (check-in: 04ef40a8fe user: drh tags: trunk) | |
|
2017-09-28
| ||
| 20:06 | Do not flatten subqueries that contain an ORDER BY or GROUP BY clause and can be implemented using a co-routine. (check-in: 042d655dd9 user: drh tags: prefer-coroutine-sort-subquery) | |
| 17:29 | Remove the (undocumented) query-planner control that prevents a "SELECT ALL" subquery in FROM clause from being implemented as a co-routine. This control was added by [a29e117d7ec], where it was called a "stop-gap". (check-in: ff2f5a31a2 user: drh tags: trunk) | |
| 16:56 | Fix over-length source code lines in select.c. No logic changes. (check-in: fd3267ef92 user: drh tags: trunk) | |
| 13:47 | Fix an issue introduced by check-in [4cd2a9672c59] (2017-03-03) that could allow a negative value in the 3rd parameter to memmove() when defragmentPage() is called on a btree page with a corrupted freeblock list. The corruption is now detected early and results in an SQLITE_CORRUPT return before the memmove() is reached. (check-in: 5b9ae69312 user: drh tags: trunk) | |
| 01:58 | Add new routines to simplify dealing with collating sequences in expressions: sqlite3ExprNNCollSeq() and sqlite3ExprCollSeqMatch(). (check-in: 490e488ea9 user: drh tags: trunk) | |
| 01:09 | Indexes on expressions with a COLLATE clause are able to satisfy an ORDER BY with the same COLLATE clause. (check-in: 0413001843 user: drh tags: trunk) | |
| 00:01 | In two places, change the magic number -1 to its proper symbol XN_ROWID. (check-in: 80277d2fc9 user: drh tags: trunk) | |
|
2017-09-27
| ||
| 16:51 | Disable tests for the LIKE optimization when ICU is enabled, since the LIKE optimization only works for the built-in LIKE operator and ICU overrides the built-in. (check-in: f5ef2e1bcd user: drh tags: trunk) | |
|
2017-09-25
| ||
| 09:37 | Add an extra snapshot related test case. (check-in: 24a95e1437 user: dan tags: trunk) | |
|
2017-09-22
| ||
| 16:23 | Use the updated Win32 VFS semantics for winOpen from check-in [5d03c738e9] for WinRT, et al, as well. (check-in: 2c03d8b8f0 user: mistachkin tags: trunk) | |
| 12:52 | Partial backout of check-in [e0af9a904076]. It turns out we do need some extra space at the end of the record blob as an overrun area to use when decoding a maliciously malformed record. (check-in: 403b88a894 user: drh tags: trunk) | |
| 00:24 | Update the configure script so that it looks for tclsh8.7 ahead of tclsh8.6. (check-in: 0a12915b37 user: drh tags: trunk) | |
|
2017-09-21
| ||
| 20:43 | In the Windows VFS, when trying to open a database file read/write, if it fails check to see if the file exists and is read-only and immediately fall back to a read-only open attempt, rather than running the AV retry loop. (check-in: 5d03c738e9 user: drh tags: trunk) | |
| 20:03 | In the Windows VFS, do not emit an SQLITE_CANTOPEN error log message when falling back from SQLITE_OPEN_READWRITE to SQLITE_OPEN_READONLY. Wait until the open fails completely. (check-in: fa3f5bcc23 user: drh tags: trunk) | |
| 19:08 | Fix harmless UB in the ICU extension. (check-in: 68e9a840d9 user: drh tags: trunk) | |
| 13:11 | Fix a potential use of an uninitialized pointer in RTree following an OOM error. (check-in: fd4ec0cdbd user: drh tags: trunk) | |
| 10:24 | Fix a bug in tool/mksourceid.c: The mksourceid.c program was incorrectly including the "# Remove this line" line of the Fossil-generated "manifest" file in the SHA3 hash. That means that all SQLITE_SOURCE_IDs for trunk versions going back to check-in [30966d56] (2017-08-22) are incorrect. (check-in: 65765222ef user: drh tags: trunk) | |
| 01:04 | Revert one performance improvement changes from check-in [3b3e32d4cd07] as it was causing a reference to an uninitialized value. (check-in: f8b1c64d3e user: drh tags: trunk) | |
| 00:49 | Fix the rendering of the P4_INTARRAY argument to the OP_IntegrityCk opcode in the output of EXPLAIN. (check-in: adc12c83dd user: drh tags: trunk) | |
|
2017-09-20
| ||
| 18:47 | Size and performance optimization on sqlite3VdbeMemGrow(). (check-in: 4b3f7eacb8 user: drh tags: trunk) | |
| 18:07 | The BLOB returned by sqlite3VdbeMemFromBtree() does not need to be zero-terminated. (check-in: e0af9a9040 user: drh tags: trunk) | |
| 17:49 | Small size and performance optimization in the bytecode engine. (check-in: 3b3e32d4cd user: drh tags: trunk) | |
| 10:47 | Improved resolution of large integer values in "CAST(x AS NUMERIC)". (check-in: 7f2bd4ff45 user: drh tags: trunk) | |