/ Timeline
Login

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

100 check-ins occurring around ad84ca24ff760f7a.

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)
17:47
Fix virtual table support for SQLITE_OPEN_REUSABLE_SCHEMA connections. (check-in: 3ca8856a user: dan tags: reuse-schema)
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)
19:30
Add test cases and fix problems on this branch. (check-in: 2b2e9f81 user: dan tags: reuse-schema)
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)
14:59
Merge the fix for ticket [4e8e4857d32d401f], so that this branch now contains release 3.27.1 plus the extra patch to preserve ROWID values on VACUUM. (check-in: 0cdae60e user: drh tags: apple-osx)
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)
13:17
Version 3.27.1 (check-in: 0eca3dd3 user: drh tags: release, branch-3.27, version-3.27.1)
12:46
Cherrypick the fix for ticket [4e8e4857d32d401f] from trunk. (check-in: d5d944d7 user: drh tags: branch-3.27)
12:44
Increase the version number to 3.27.1. (check-in: b94d5428 user: drh tags: branch-3.27)
04:15
Do not do the optimization that attempts to pull expression values from an index on that expression when processing a multi-index OR (see check-in [a47efb7c8520a0111]) because the expression transformations that are applied become invalid when the processing moves off of the current index and on to the next index. Fix for ticket [4e8e4857d32d401f]. (check-in: 440a7cda user: drh tags: trunk)
2019-02-07
19:28
Pull in the change that causes VACUUM to preserve rowid values. (check-in: c9af2f71 user: drh tags: apple-osx)
19:07
Change VACUUM so that it preserves ROWID values. (check-in: 49e5d11d user: drh tags: trunk)
18:49
Merge the offical 3.27.0 release. (check-in: 62829164 user: drh tags: apple-osx)
17:02
Version 3.27.0 (check-in: 97744701 user: drh tags: trunk, release, version-3.27.0)
15:09
Experimental change to VACUUM so that it preserves existing ROWID values. (Closed-Leaf check-in: f264015f user: drh tags: vacuum-preserves-rowid)
00:17
In the fuzzcheck test program, make sure the isspace() and related routines are not given values outside the range of -1 to 255. (check-in: f2983287 user: drh tags: trunk)
2019-02-06
20:49
Ensure all bytes of the PgHdr1 structure are initialized. This avoids a valgrind error when running corruptI.test. (check-in: a505e34d user: dan tags: trunk)
20:12
Update an assert() in test_journal.c to take the new "VACUUM INTO" functionality into account. (check-in: 0e6249cb user: dan tags: trunk)
19:42
Avoid attempting to delete a file while it is still open in walfault2.test. (check-in: 9d54a63d user: dan tags: trunk)
18:08
Update test file dbfuzz001.test so that it works with SQLITE_ENABLE_OVERSIZE_CELL_CHECK builds. (check-in: 4371a0c4 user: dan tags: trunk)
15:23
Add -DSQLITE_ENABLE_DESERIALIZE to the default testing configuration in test/releasetest.tcl. (check-in: 433d6ef6 user: dan tags: trunk)
13:48
Do not attempt to test the shell ".trace" command in SQLITE_OMIT_TRACE builds. (check-in: 3839e909 user: dan tags: trunk)
13:41
Fix a problem preventing testfixture from being built with SQLITE_OMIT_VIRTUALTABLE. (check-in: d2cca36a user: dan tags: trunk)
01:18
When extending the size of SrcList objects, do not attempt to use extra space on the end of the allocation, as that provides scarcely any performance benefit but does greatly complicate cross-platform testing. (check-in: 3087a0c3 user: drh tags: trunk)
00:55
If the query flattener detects an error, cause the SELECT code generator to abort immediately. (check-in: 3d3b142f user: drh tags: trunk)
00:11
Back out check-in [008112bcef561a8], reenabling deprecated PRAGMAs when the SQLITE_OMIT_DEPRECATED compile-time option is used, because way too many TH3 tests depend on those deprecated PRAGMAs. (check-in: 9dbf512d user: drh tags: trunk)
2019-02-05
20:51
Omit the ".progress" command from the shell if it is compiled with the SQLITE_OMIT_PROGRESS_CALLBACK option. (check-in: 4d0a949f user: drh tags: trunk)
19:52
Fix a harmless compiler warning in the memtrace.c extension. (check-in: 2f468da4 user: drh tags: trunk)
19:51
Merge latest trunk into this branch. (check-in: c089cc4f user: dan tags: reuse-schema)
19:48
Fix compiler warnings on Windows. (check-in: 4978ee8b user: drh tags: trunk)
19:15
Fix memory leaks on this branch. (check-in: e9c5e189 user: dan tags: reuse-schema)
16:53
Update the tester.tcl --malloctrace option so that it uses eu-addr2line instead of addr2line. (check-in: 79c07387 user: dan tags: trunk)
14:36
The IS NOT NULL operator does not imply that the operand is never NULL. Fix for ticket [5948e09b8c415bc45da5cf] (check-in: d840e9bb user: drh tags: trunk)
12:16
Keep the temp_store_directory and data_store_directory pragmas even when compiling with SQLITE_OMIT_DEPRECATED. (check-in: 93386a7c user: drh tags: trunk)
08:55
Update test file "resetdb.test" so that it works with the "prepare" permutation. (check-in: 95d33812 user: dan tags: trunk)
2019-02-04
21:10
Early detection of implausibly sized records to avoid unnecessary large memory allocations. (check-in: 2c8769c6 user: drh tags: trunk)
21:02
Fix a problem with reloading the schema on this branch. (check-in: 5dfbef83 user: dan tags: reuse-schema)
19:52
Mention the new -memtrace command-line option in the -help output of the CLI. (check-in: ada91aef user: drh tags: trunk)
19:50
Add the -memtrace option to dbfuzz2. (check-in: 67fecbc7 user: drh tags: trunk)
19:45
Enhance dbfuzz2 so that with the -v option it shows the return code and error message for any failing SQL statements. (check-in: 3a127ef9 user: drh tags: trunk)
19:12
Ensure that the sqlite3_exec() callback gets the correct number of columns for a query, even if the schema changes out from under it. (check-in: a16ffb5a user: drh tags: trunk)
16:42
Do not run shmlock.test as part of the inmemory_journal permutation. (check-in: d3fea69c user: dan tags: trunk)
16:36
Update the test/fuzzdata8.db database with new interesting cases from dbsqlfuzz. (check-in: e467e4c8 user: drh tags: trunk)
16:25
Fix problems with SAVEPOINT and related commands within transactions on a corrupt database that include fts5 or some other virtual tables. (check-in: b0da6e0d user: dan tags: trunk)
14:53
Add the ".progress" command to the CLI. (check-in: ad84ca24 user: drh tags: trunk)
14:00
Fix another potential buffer overread in fts5. (check-in: 14fba4cc user: dan tags: trunk)
11:54
Fix another corruption case in fts5. (check-in: e608085d user: dan tags: trunk)
2019-02-03
07:46
Fix an assertion failure in fts3 triggered by a corrupt database. (check-in: 560ebd42 user: dan tags: trunk)
2019-02-02
21:02
Try new approach ensuring that each Schema object is only used by one connection/database at any one time. (check-in: 9e8e5f52 user: dan tags: reuse-schema)
18:24
New testcase added to test/fuzzdata8.db. (check-in: 2f5c1925 user: drh tags: trunk)
15:59
Honor key query parameters for SEE on the URI filename for ATTACH and VACUUM INTO. (check-in: 2e01096b user: drh tags: trunk)
15:05
Give a link to the URI Filename document in the documentation to the various sqlite3_uri_parameter() interfaces. (check-in: 69b0b2ee user: drh tags: trunk)
13:47
Fix a buffer overread in fts5 debugging scalar function fts5_decode(). (check-in: 54f2399f user: dan tags: trunk)
01:27
Fix harmless compiler warning. (check-in: dddda685 user: mistachkin tags: trunk)
2019-02-01
21:08
Slight adjustment to the printf formatter large memory allocation detector so that it does not overestimate the amount of space needed for oversize %d conversions. (check-in: 1aee70d6 user: drh tags: trunk)
20:29
Prevent the printf formatter from doing large memory allocations - larger than either the size of the static buffer for interfaces like sqlite3_snprintf(), or larger than SQLITE_LIMIT_LENGTH for interfaces that are associated with a database connection. This helps to prevent DOS attacks on products that let hostile sources inject arbitrary SQL. It also helps fuzzers run faster and more effectively. (check-in: 179e5d46 user: drh tags: trunk)
18:46
Performance improvement in the parsing of options to %-formats in the printf implementation. (check-in: 40d8f8ae user: drh tags: trunk)
15:06
Ensure that the Walker.pParse structure is initialized when walking the source tree to gather the WINDOW clauses on a query with window functions. (check-in: 4ca9d5d5 user: drh tags: trunk)
14:54
New test cases added to test/fuzzdata8.db. (check-in: e5924939 user: drh tags: trunk)
14:50
Improve the strict enforcement of cell sizes in balancing from check-in [12713f320b2c1def] so that it also works with table-btrees in addition to index-btrees. (check-in: ef27e7a0 user: drh tags: trunk)
14:40
Fix an assert() in fts5 that could fail if the database is corrupt. (check-in: 55f06aa3 user: dan tags: trunk)
13:34
Fix another corrupt-database-handling problem in fts5. (check-in: 9876ae1b user: dan tags: trunk)
2019-01-31
16:09
Remove an incorrect assert() from the deserialize in-memory database VFS. (check-in: 80151d7e user: drh tags: trunk)
15:38
Add the SQLITE_CONFIG_MEMDB_MAXSIZE configuration option for configuring the default maximum size of an in-memory database created using sqlite3_deserialize(). This is necessary to make the interface reasonably testable. (check-in: cb72ee04 user: drh tags: trunk)
14:37
Fix another buffer overrun that could occur when quering a corrupt database using an fts5vocab table. (check-in: b80119bd user: dan tags: trunk)
02:42
If the amalgamation file "sqlite3.c" is generated using the SQLITE_ENABLE_UPDATE_DELETE_LIMIT option, but then subsequently (mis-)compiled without that option, and then a particular syntax error could cause a memory leak, but for this fix. (check-in: 1bb17c86 user: drh tags: trunk)
01:39
Improvements to "PRAGMA parser_trace=ON": Make it a flag pragma, and then use the flag to show the complete SQL text at the beginning of the parse. (check-in: 507c4353 user: drh tags: trunk)
2019-01-30
19:50
Change the error message slightly for when two or more indexes have the same rootpage. (check-in: 916c52da user: drh tags: trunk)
19:12
When parsing the schema, check for duplicate index root page numbers before adding the index to the schema hash table. (check-in: 6b360bc0 user: dan tags: trunk)
18:47
Add test case for the previous commit. (check-in: 197edb23 user: dan tags: trunk)
18:33
Detect if two indexes of the same table share a common rootpage while parsing the schema, and throw an error immediately. (check-in: f21c6f13 user: drh tags: trunk)
16:58
Fix an off-by-one error when parsing the names of indexes that do not have arguments in the index_usage utility. (check-in: dc794d8f user: drh tags: trunk)
15:47
Add the --progress, --using, and -q options to the index_usage utility program. (check-in: a5e6be7c user: drh tags: trunk)
14:01
Enhancements to the index_usage utility program. (check-in: 19c739b4 user: drh tags: trunk)
12:15
Fix another buffer overread in fts5 that may occur when accessing a corrupt database. (check-in: 760d1437 user: dan tags: trunk)
2019-01-29
19:17
Merge enhancements from trunk. (check-in: 31efcc37 user: drh tags: apple-osx)
16:54
Fix harmless compiler warnings. (check-in: 9a93c68a user: drh tags: trunk)
16:47
Load all the latest dbsqlfuzz finds into test/fuzzdata8.db. (check-in: e744d2dd user: drh tags: trunk)
16:41
More robust handling of corrupt database file in the rebalance operation of the btree logic. (check-in: 97704cb7 user: drh tags: trunk)
16:34
Avoid an integer overflow in the fts5 snippet() function triggered by a corrupt database record. (check-in: 7c862c46 user: dan tags: trunk)
15:30
Avoid a crash when the fts5 snippet() function (or similar) is used with a special query pattern, like '*id' or '*reads'. (check-in: 9d58a157 user: dan tags: trunk)
11:42
Fix a buffer overrun triggered by a merge operation on a corrupt fts5 database. (check-in: 7ee3cd5b user: dan tags: trunk)
02:37
Set a low limit on the length of strings and blobs in the OSSFuzz module, in an effort to avoid timeouts when the fuzzer does things like "randomblob(1e12)". (check-in: 119d1609 user: drh tags: trunk)
2019-01-28
19:06
Fix the query planner so that it is able to use an index on a CAST expression. (check-in: 2c886f3d user: drh tags: trunk)
18:58
Fix a performance regression caused by the previous commit. (Closed-Leaf check-in: c4db0ad1 user: drh tags: index-on-cast)
18:08
Make indexes on CAST(...) expressions work. (check-in: 3ef711d9 user: dan tags: index-on-cast)
16:50
Fix a buffer overread in fts3 that could occur when accessing a corrupt database. (check-in: a9faf903 user: dan tags: trunk)
13:27
Do not assume that text parameters passed to fts4aux queries do not contain embedded nul characters. (check-in: df04859a user: dan tags: trunk)
11:54
Add a new test case to fuzzdata8.db. (check-in: ce8e279a user: drh tags: trunk)
00:42
Change the definition of the BtCursor.skipNext field so that it is undefined for states CURSOR_VALID and CURSOR_INVALID. (check-in: 6089d5db user: drh tags: trunk)
2019-01-27
19:50
Improvements to the ".eqp trace" command in the CLI so that it ensures that the schema has been read prior to activing the tracing feature, to avoid cluttering the output with a trace of the schema parse. (check-in: 42687d45 user: drh tags: trunk)
02:45
New test cases form dbsqlfuzz (check-in: 9cf8ebd1 user: drh tags: trunk)
02:41
Window functions that can abort should indicate this, so that if they are used in DML statement, a statement journal will be used. (check-in: 0ea05a0e user: drh tags: trunk)
01:11
Update fuzzcheck so that with the --load-dbsql options it screens its inputs to ensure that they are valid dbsqlfuzz cases. Add new dbsqlfuzz finds to the test/fuzzdata8.db database. (check-in: 004f7d9b user: drh tags: trunk)
2019-01-26
23:34
Ensure that the btree cursor is correctly set up prior to backing it up as part of a delete operation on a divider cell in an index. (check-in: a3ea1a82 user: drh tags: trunk)