SQLite4
Timeline
Not logged in

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

50 ancestors of [8ede88c1df]

2013-02-14
15:32
Fold in Peter Reid's fixes and enhancements to the sqlite4_num object. check-in: 8ede88c1df user: drh tags: trunk
01:51
Fix sqlite4_num_to_text when formatting an integer with a negative exponent. It was leaving an uninitialized byte in the destination string. Leaf check-in: 3cac6cdb86 user: peterreid tags: num_work
2013-02-13
12:28
Change the API for binding strings and blobs so that the destructor now carries an extra argument which can be used as the sqlite4_env pointer, thus allowing functions like sqlite4_free() to be used as a string or blob destructor. check-in: 56335097b1 user: drh tags: trunk
2013-02-11
18:33
Fix the data decoder so that it correctly handles +Inf and -Inf. check-in: c04b47fd62 user: drh tags: trunk
16:04
Fix a memory leak in DETACH. check-in: 9b03a19a30 user: drh tags: trunk
15:51
Fix a memory leak in the test harness. check-in: 108d62cab9 user: drh tags: trunk
15:17
Remove two MySQL-isms: Strings in double-quotes and identifiers quoted by grave accents. check-in: eec75c074c user: drh tags: trunk
14:03
Add test fixture functions for sqlite4_num_isinf and sqlite4_num_isnan, and some related test cases. check-in: 58a9cfcb39 user: peterreid tags: num_work
04:09
Num comparison says that +0 = -0. Before, 1-1 > (-1)-(-1). check-in: 13d342510f user: peterreid tags: num_work
2013-02-10
22:03
Add sqlite4_num_add/sub/mul/div to testfixture and some tests that use them. check-in: 91bfc9e7d7 user: peterreid tags: num_work
19:21
Fix defect in which sqlite4_num_from_text returns NaN if it uses exactly as many bytes as the caller allows. check-in: 24ee36b76f user: peterreid tags: num_work
04:26
Add a failing test and supporting functions. When the buffer is ended by a passed-in byte count instead of a trailing 0, sqlite4_num_from_text returns NaN check-in: 2cff3b428a user: peterreid tags: num_work
04:06
Add a few sqlite4_num tests. check-in: 6cc07d19e4 user: peterreid tags: num_work
2013-02-09
16:55
Have worker clients and writers that discard an old in-memory tree update a read-lock slot before concluding their work or write transaction. This is required for read-only clients - which cannot set the value of their own read-lock slot. check-in: 798d9e23be user: dan tags: trunk
05:36
Create sqlite4_num_isinf check-in: 555cdfbf52 user: peterreid tags: num_work
05:35
Fix sqlite4_num_from_text in 0.000(many zeros)00123 case. check-in: 97728c04e8 user: peterreid tags: num_work
05:33
Make sqlite4_num's exponent signed. check-in: b320c13e42 user: peterreid tags: num_work
05:32
Create sqlite4_num_from_int64. check-in: 2552bc97ef user: peterreid tags: num_work
05:31
Prevent reading out of bounds in sqlite4_num_from_text. Parsing 'in' (prefix of 'inf') with byte count 2 triggered the bug. check-in: 408475b607 user: peterreid tags: num_work
05:29
Create sqlite4_num_isnan. check-in: cc1dc90b47 user: peterreid tags: num_work
05:27
Make LARGEST_UINT64 unsigned. This was causing adjustExponent to fail when compiled with msvc. check-in: 9ff6991d9f user: peterreid tags: num_work
2013-02-08
15:22
Avoid extending the database file when truncating it to the minimum number of blocks required during system shutdown. check-in: 9afc42d70d user: dan tags: trunk
14:39
Changes to lsm_unix.c to build on android: (a) do not use fdatasync() on android and (b) account for the fact that usleep() returns void on android. check-in: 302b222395 user: dan tags: trunk
11:30
Merge compression-id branch with trunk. check-in: 76297939d3 user: dan tags: trunk
2013-02-07
19:50
Add the INFO_COMPRESSION_ID request. And the factory method for providing compression/encryption functions. Leaf check-in: bb85de9cd3 user: dan tags: compression-id
2013-02-06
19:43
Add API to register a compression-factory method with an lsm handle. check-in: 60908fd4d1 user: dan tags: compression-id
19:03
Add a field to the database header to identify the compression scheme in use. check-in: 3bf1db9709 user: dan tags: compression-id
11:58
Fix bug to do with block redirection. check-in: 7cc153f523 user: dan tags: trunk
2013-02-05
18:32
Fix lsm_checkpoint() and some lsm_info() calls to match documented behaviour. check-in: 89e314d98b user: dan tags: trunk
18:06
Fixes to lsm_work() to bring it into line with documentation: (a) third argument is a number of KB, not a number of db pages, and (b) passing -1 as the third argument means "do as much work as possible". check-in: 2f8966cd9c user: dan tags: trunk
16:03
Change the BLOCK_SIZE, AUTOFLUSH and AUTOCHECKPOINT lsm_config() options to work in KB instead of bytes. This matches the user guide. check-in: 340c9dcd22 user: dan tags: trunk
09:52
Add test file lsm3.test, which should have been added a few days ago. check-in: 5dfd8651df user: dan tags: trunk
2013-02-02
16:45
Fix LSM single-process mode so that it holds an exclusive lock on the database file - preventing connections from within external processes. check-in: d6bd08ca0e user: dan tags: trunk
2013-02-01
19:49
Simplifications and clarifications to lsmusr.wiki. check-in: 33eca2e1f4 user: dan tags: trunk
2013-01-31
05:58
Add the definition of sqlite4_stricmp() to sqlite.h.in. Avoid multiple declarations the u8 and similar typedefs in the amalgmation. check-in: d966049dd6 user: drh tags: trunk
05:36
Change all occurrences of sqlite4StrICmp to sqlite4_stricmp and change sqlite4StrNICmp to sqlite4_strnicmp. check-in: b6d0a62048 user: drh tags: trunk
05:27
Publish routines sqlite4_stricmp() and sqlite4_strnicmp() that do simple unicode case folding. Use these routines for all internal case-insensitive string comparisons. check-in: 19e9278b9c user: drh tags: trunk
2013-01-28
08:56
Merge block-redirects branch with trunk. This changes the lsm file format. check-in: 647229e983 user: dan tags: trunk
2013-01-26
20:18
Handle calls to lsm_work() with the nPage parameter set to not greater than zero. Remove some dead code. Leaf check-in: 520f3729b8 user: dan tags: block-redirects
19:17
Fix problems with redirected blocks in compressed databases. check-in: 930b7e4507 user: dan tags: block-redirects
17:24
Remove an inaccurate assert() from lsm_sorted.c. check-in: 259878e357 user: dan tags: block-redirects
16:44
Fix a problem to do with an lsm_seek(LSM_SEEK_LE) on a level that is currently undergoing a merge. Add more complex assert statements to catch any similar problems. check-in: ca4dc40190 user: dan tags: block-redirects
2013-01-25
21:23
Fix an off-by-one error in the value data decoder logic that prevented DEFAULT VALUE from working correctly in some instances. Also fix some nearby code formatting issues. check-in: 98d35b06cd user: drh tags: trunk
20:46
Fix a problem in the range-delete code. check-in: 96badcb933 user: dan tags: block-redirects
2013-01-24
18:24
Fix a string type compile warning. check-in: 11b2bf1628 user: stephan tags: trunk
16:45
Improve the integrity-check assert that ensures all blocks in the file are accounted for. Roll any in-memory free-list records into the end of the main segment when executing lsm_work(nmerge=1). check-in: e5edba9caa user: dan tags: block-redirects
2013-01-23
18:13
Keep the contents of Segment structures up to date at all times, so that none of the page numbers contained within are subject to redirection. check-in: a89abc2117 user: dan tags: block-redirects
2013-01-22
20:07
Several block-redirect related bugfixes. check-in: a56a334333 user: dan tags: block-redirects
2013-01-21
19:50
Add tests for block-redirects to lsmtest. check-in: eec16b0f2f user: dan tags: block-redirects
16:53
If a free-list-only segment is generated while a merge of the top-level segment is underway, add the new segment to the merge inputs immediately. Also, if auto-checkpoints are enabled, schedule a checkpoint after each block is moved within an lsm_work(nmerge=1) call. check-in: 89b4286682 user: dan tags: block-redirects
14:23
Fix LSM_INFO_PAGE_DUMP so that it works with redirected blocks. check-in: deb0ccacd4 user: dan tags: block-redirects