Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
46 check-ins using file src/sqliteInt.h version 8628c582
|
2019-02-22
| ||
| 21:33 | Check-in [fa792714ae62fa980] is not a valid fix for ticket [df46dfb631f75694], as the new test case in this check-in demonstrates. The fix here causes test cases for the [df46dfb631f75694] bug to fail again, so this check-in is on a branch. A new fix is needed for [df46dfb631f75694]. (check-in: 0d456456 user: drh tags: tkt-df46dfb631) | |
| 19:24 | Internally, remove all references to a Window object that belongs to an expression in an ORDER BY clause if that expression is converted to an alias of a result-set expression. Fix for [4feb3159c6]. (check-in: 579b66ea user: dan tags: trunk) | |
| 16:18 | In sqlite3NestedParse() be sure to detect all SQLITE_NOMEM and SQLITE_TOOBIG errors and to distinguish between them. (check-in: 73056b31 user: drh tags: trunk) | |
| 15:42 | Modify sqlite3_str_finish() and sqlite3VMPrintf() so that they always return NULL on any OOM or SQLITE_LIMIT_LENGTH error. (check-in: e7144ffd user: drh tags: trunk) | |
|
2019-02-21
| ||
| 18:11 | Fix OOM handling in sqlite3NestedParse(). (Leaf check-in: e6e9dd5c user: dan tags: nested-parse-oom) | |
| 16:41 | Detect oversized strings in the OP_String opcode even if the P4 argument is originally UTF8 and has to be converted to UTF16 to match the database file and that conversion causes the string to become shorter and cross below SQLITE_LIMIT_LENGTH threshold. This might fix an OSSFuzz problem that we have been so far unable to reproduce. (check-in: c13d5639 user: drh tags: trunk) | |
|
2019-02-20
| ||
| 19:20 | Changes to the unix VFS that allegedly enable it to build of Fuchsia. We have no way of testing this. (check-in: be21a641 user: drh tags: trunk) | |
| 19:06 | Disable unix-specific features of dbfuzz2 when compiling on windows. (check-in: af84774d user: drh tags: trunk) | |
| 18:13 | Avoid the use of function pointers in columnName(), as function pointers appear to be a source of consternation to LLVM. (check-in: c48f6f39 user: drh tags: trunk) | |
| 13:55 | Progress handler improvements: (1) Invoke the callback after OP_Program opcodes (2) Invoke the callback multiple times in a row to catch up after a long run of no progress checks. (check-in: 0c5db18d user: drh tags: trunk) | |
| 12:52 | When an IN operator drives a query loop, mark it as "CODED" so that it will not be used afterwards for a (pointless) membership test. This is a better fix for ticket [df46dfb631f75694] than the previous fix that is now on a branch as it preserves the full optimization of check-in [e130319317e76119]. (check-in: fa792714 user: drh tags: trunk) | |
|
2019-02-19
| ||
| 20:29 | Enhance the xBestIndex method on delta_parse() to return SQLITE_CONSTRAINT if no delta argument is supplied. (check-in: f16d127c user: drh tags: trunk) | |
| 20:19 | Add the delta_parse(DELTA) table-valued function to the fossildelta extension. (check-in: d91fcc26 user: drh tags: trunk) | |
| 18:39 | Add the fossildelta.c extension in ext/misc with implementations of the Fossil delta functions. (check-in: b80cafa6 user: drh tags: trunk) | |
| 17:45 | Fix a potential memory leak in RBU if the rbu_fossil_delta() SQL function is misused. Misuse never happens in a working RBU system, so this is not a particularly important fix. (check-in: 12517d1b user: drh tags: trunk) | |
| 16:42 | Fix an assert() that might not be true if the database file is corrupt. (check-in: f2d400db user: drh tags: trunk) | |
| 13:51 | New assert() to verify the TF_HasprimaryKey flag. (check-in: 0abace8a user: drh tags: trunk) | |
| 13:08 | Always use the symbolic name SQLITE_IDXTYPE_PRIMARYKEY rather than a hard-coded magic number 2. (check-in: 55d43adc user: drh tags: trunk) | |
|
2019-02-18
| ||
| 12:16 | Fix an assert() that might not be true if the %_segdir table of FTS3 contains corrupt entries. (check-in: a6509ca3 user: drh tags: trunk) | |
|
2019-02-16
| ||
| 22:45 | Fix an assert() that might be false for a corrupt database. (check-in: 0cfb8887 user: drh tags: trunk) | |
|
2019-02-14
| ||
| 15:27 | Improved oversized cell detection when updating ptrmap pages in balance_nonroot(). (check-in: aa61435a user: drh tags: trunk) | |
|
2019-02-12
| ||
| 22:58 | Ensure that the nProgressLimit variable is always initialized in sqlite3VdbeExec(), even if the routine jumps to its exit processing early. (check-in: 167b91df user: drh tags: trunk) | |
| 21:04 | Enhancement the progress callback mechanism so that the progress callback is always invoked at least once at the end of a prepared statement if the opcode count has been exceeded. This makes the progress callback more effective at limiting run times. This check-in also includes and unrelated performance enhancement to OP_Column. (check-in: 68cce272 user: drh tags: trunk) | |
| 16:58 | Further performance improvements to btreeInitPage(). (check-in: 93ae382e user: drh tags: trunk) | |
| 15:51 | Increase the version number to 3.28.0 for the next release cycle. (check-in: 6eb38c59 user: drh tags: trunk) | |
| 12:25 | Register the "vfsstat" virtual table immediately when loading the vfsstat.c extension. (check-in: 9bd92afd user: drh tags: trunk) | |
| 01:28 | New test cases in test/fuzzdata8.db. (check-in: ab2356f5 user: drh tags: trunk) | |
| 01:04 | Defer computing the number of bytes of free space on a btree page until that value is actually needed. (check-in: 177f5f40 user: drh tags: trunk) | |
| 00:58 | Change an assert() into a NEVER(), since the condition is difficult to prove with certainty. Improved comment on the MemPage.nFree field. (Closed-Leaf check-in: fec071b8 user: drh tags: deferred-free-space) | |
|
2019-02-11
| ||
| 22:50 | Do not invoke btreeComputeFreeSpace() when not necessary. (check-in: f11b0ed4 user: drh tags: deferred-free-space) | |
| 19:27 | Improvements to comments. No changes to code. (check-in: 16eb0fc8 user: drh tags: deferred-free-space) | |
| 17:22 | Fix test cases. Avoid unnecessary tests for MemPage.nFree less than zero. (check-in: 480a5090 user: drh tags: deferred-free-space) | |
| 16:12 | Fix another segfault that could occur in fts5 with a corrupted database. (check-in: 09e33738 user: dan tags: trunk) | |
| 13:25 | Fix typos in comments in fts5.h. No changes to code. (check-in: c19e97bc user: dan tags: trunk) | |
| 12:51 | Updates to a comment to keep it in alignment to the changes on this branch. (check-in: 7499f743 user: drh tags: deferred-free-space) | |
| 11:34 | Build and test fixes for SQLITE_OMIT_WINDOWFUNC builds. (check-in: 8824115a user: dan tags: trunk) | |
| 11:04 | Avoid using a "return void_function();" statement in void function sqlite3WalSnapshotUnlock(). It provokes a warning from some compilers. (check-in: 4a4952d3 user: dan tags: trunk) | |
| 01:58 | Add an assert() in an attempt to repro an ASAN warning from OSSFuzz. (check-in: 7b412224 user: drh tags: trunk) | |
|
2019-02-09
| ||
| 22:33 | Fix a page-cache reference leak in the btree balancer when there is a corrupt database. (check-in: 92858991 user: drh tags: deferred-free-space) | |
| 21:06 | Defer computing the MemPage.nFree value of an in-memory btree page until it is actually needed, since for many pages it is never needed. This checkin works sufficiently to prove the concept, but still has issues with exception handling. (check-in: 1d43ee40 user: drh tags: deferred-free-space) | |
| 19:23 | Change a few assert() statements in fts3 that might fail if the database is corrupt. (check-in: db74a56a user: dan tags: trunk) | |
|
2019-02-08
| ||
| 22:34 | Small performance improvement and size reduction for pageFindSlot() - the routine in btree.c that locates a free slot for a cell on a btree page. (check-in: 1969372a user: drh tags: trunk) | |
| 20:55 | Use a fast compiler-provided memory barrier exclusively, if such a memory barrier is available. (Leaf check-in: df66fec9 user: drh tags: faster-memory-barrier) | |
| 17:28 | Further simplifications to sqlite3VdbeMemSetStr(). (check-in: 1d212957 user: drh tags: trunk) | |
| 15:59 | Change the sqlite3VdbeMemGrow() routine so that it no longer guarantees a minimum size of 32 bytes. That minimum is no longer required, and without the extra check for the minimum size, the routine runs faster. (check-in: 5c499da8 user: drh tags: trunk) | |
| 14:55 | Give the sqlite3 object a pointer to the current Parse so that if an OOM occurs, it can automatically set the Parse.rc value to SQLITE_NOMEM. This avoids a frequent extra test of db.mallocFailed in the innermost parser loop. (check-in: 5c663804 user: drh tags: trunk) | |