SQLite

Timeline
Login

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

100 check-ins occurring around 0288fa5d25886f6f.

2005-09-22
15:45
Optionally call fdatasync() instead of fsync() only if _POSIX_SYNCHRONIZED_IO is positive, which should only be the case on operating systems that actually support fdatasync(). (CVS 2732) (check-in: a9b341dc user: drh tags: trunk)
2005-09-20
18:13
Add VM code comments on the group-by processing. Extra group-by test case. (CVS 2731) (check-in: 655e75ac user: drh tags: trunk)
17:42
Code cleanup and simplification. Three new Mem opcodes added. The sqlite3VdbeJumpHere function added. (CVS 2730) (check-in: 2471957f user: drh tags: trunk)
13:55
Cleanup the implementation and the documentation of the new 0x200 P1 flag on the VDBE comparison operands. (CVS 2729) (check-in: f22d0f64 user: drh tags: trunk)
13:12
NULLs compare equal to each other when computing GROUP BY categories. (CVS 2728) (check-in: d9b0c970 user: drh tags: trunk)
08:47
Update comments in where.c. No code changes. (CVS 2727) (check-in: bb84d27e user: drh tags: trunk)
01:36
Documentation updates. (CVS 2726) (check-in: 6ec8883c user: drh tags: trunk)
2005-09-19
21:05
ON-clause terms in a LEFT JOIN that restrict only the left table, should not really restrict the left table but instead rows that fail to meet the condition to be joined with NULL rows from the right table. (CVS 2725) (check-in: ea10f978 user: drh tags: trunk)
19:05
Set the default maximum page size to 32768, not 8192. (CVS 2724) (check-in: b32e9ec2 user: drh tags: trunk)
17:35
Do not optimize views with an ORDER BY clause if they are used in a UNION ALL that also has an ORDER BY clause. Ticket #1444. (CVS 2723) (check-in: 6cc57fcf user: drh tags: trunk)
15:37
Fix uninitialized value on a UNION ALL select with an ORDER BY clause. (CVS 2722) (check-in: 92126a21 user: drh tags: trunk)
13:15
The sqlite3_query_plan debugging variable now only appears with SQLITE_TEST=1. (CVS 2721) (check-in: 41e226d2 user: drh tags: trunk)
12:53
Work around limitations of MSVC++ 6. Tickets #1429, #1437, and #1440. (CVS 2720) (check-in: b2d1803c user: drh tags: trunk)
12:37
Change some debugging #defines in order to get SSE working again after recent upgrades. (CVS 2719) (check-in: 86eb7d83 user: drh tags: trunk)
2005-09-17
19:28
Download page reports sizes in MiB and KiB instead of just bytes. (CVS 2718) (check-in: e985f02d user: drh tags: trunk)
19:14
Update to the download page on the website. (CVS 2717) (check-in: 91bc0d00 user: drh tags: trunk)
18:49
Version 3.2.6 (CVS 2716) (check-in: 1cdfe667 user: drh tags: trunk)
18:34
The sqlite3ReallocOrFree routine should set its pointer to NULL when it fails. (CVS 2715) (check-in: 0e85af44 user: drh tags: trunk)
18:02
Fix up the busy test so that it can be run multiple times in a row without giving false errors on 2nd and subsequent runs. (CVS 2714) (check-in: 7a7ba73f user: drh tags: trunk)
17:58
Fix a memory leak that occurs after malloc fails. (CVS 2713) (check-in: 85b931bf user: drh tags: trunk)
17:05
Fix the crash.test script to accomodate recent changes. (CVS 2712) (check-in: d594608f user: drh tags: trunk)
16:48
Fix a bug in the lock2 test script introduced by check-in (2709) (CVS 2711) (check-in: c3bf1743 user: drh tags: trunk)
16:36
Fix a problem with the lock_status pragma and the UTF-16 encoding. (CVS 2710) (check-in: 1a737b45 user: drh tags: trunk)
15:20
Changes to make corruption errors easier to track down. (CVS 2709) (check-in: c0733000 user: drh tags: trunk)
15:17
Update the webpage description of 3.2.6 changes to talk about ticket #1432. (CVS 2708) (check-in: 0e23c28b user: drh tags: trunk)
13:29
Bug fix in the ORDER BY optimizer. Ticket #1435. (CVS 2707) (check-in: 553b7ba8 user: drh tags: trunk)
13:07
Make sure dependencies on the right-hand side of IN operators are checked correctly. Ticket #1433. (CVS 2706) (check-in: 21740794 user: drh tags: trunk)
02:34
Updates to the FAQ. (CVS 2705) (check-in: 0eaf430d user: drh tags: trunk)
2005-09-16
17:16
Do not journal the locking page which doing an autovacuum. Similar to the problem of #1432 except that this one occurs on autovacuum instead of vacuum. An assert() has been added to catch any future incidents of this type. (CVS 2704) (check-in: 5b6dc12b user: drh tags: trunk)
11:32
Do not journal the file locking page when doing a database copy as part of a VACUUM. Ticket #1432. (CVS 2703) (check-in: 248f7797 user: drh tags: trunk)
10:18
Prevent the sqlite3_pagecount() function from ever returning the pending-byte page. (CVS 2702) (check-in: 35a88935 user: danielk1977 tags: trunk)
10:13
Move the definition of sqlite3_pending_byte from test2.c to os_common.h. (CVS 2701) (check-in: bedf702f user: danielk1977 tags: trunk)
09:52
Fix an auto-vacuum problem with the PENDING_BYTE page. Also link the Tcl variable sqlite_pending_byte to the internal pending-byte location when in test mode. (CVS 2700) (check-in: 9115e062 user: danielk1977 tags: trunk)
02:55
Documentation changes in preparation for the release of 3.2.6. (CVS 2698) (check-in: 243f455c user: drh tags: trunk)
02:48
Undo check-in (2694). (CVS 2697) (check-in: 6017b196 user: drh tags: trunk)
02:38
Fix a whole host of newly discovered memory leaks the occur after a failure of realloc(). (CVS 2696) (check-in: 4686d649 user: drh tags: trunk)
00:27
Fix a memory leak: a failure to deallocate the P3 parameter on a call to sqlite3VdbeChangeP3. (CVS 2695) (check-in: 714254cb user: drh tags: trunk)
2005-09-15
21:24
The table_info pragma now gives the order of columns in the primary key. (CVS 2694) (check-in: 9b60f48d user: drh tags: trunk)
2005-09-13
16:12
Correct the sense of a test for SQLITE_DEBUG on the resent NDEBUG change. Ticket #1425 (CVS 2693) (check-in: 81fdffdf user: drh tags: trunk)
07:00
Updates to the documentation on the TCL bindings. (CVS 2692) (check-in: 94aa2d32 user: drh tags: trunk)
00:02
Fix a comment typo in the previous check-in. (CVS 2691) (check-in: 49c95280 user: drh tags: trunk)
00:00
Automatically set the NDEBUG preprocessor macro unless SQLITE_DEBUG is set. In this way, assert() statements become opt-in rather than opt-out. (CVS 2690) (check-in: 2ec2bd85 user: drh tags: trunk)
2005-09-12
23:03
Bug fix in the new COUNT(DISTINCT) logic. Ticket #1420. (CVS 2689) (check-in: 37116550 user: drh tags: trunk)
2005-09-11
11:56
COUNT(DISTINCT x) is now fully functional, though it could benefit from additional testing. (CVS 2688) (check-in: 2f397bd8 user: drh tags: trunk)
02:03
The ".dump" command from the shell correctly saves the state of the sqlite_sequence and sqlite_stat1 tables, if they exist. Ticket #1419. (CVS 2687) (check-in: 3f191cf4 user: drh tags: trunk)
2005-09-10
22:40
The shell does not output the sqlite_stat1 table on .dump or .schema. The ANALYZE command now gathers statistics on tables that have only a single index because this sometimes helps when reordering tables in a join. (CVS 2686) (check-in: 26565b89 user: drh tags: trunk)
16:46
Add the experimental EXPLAIN QUERY PLAN diagnostic capability. (CVS 2685) (check-in: 986efb7b user: drh tags: trunk)
15:35
Avoid a memory leak and/or assertion failure when parsing a table declaration that contains a duplicate column name. Ticket #1418. (CVS 2684) (check-in: f4342774 user: drh tags: trunk)
15:28
Use of the CROSS keyword in a join prevents table reordering. Ticket #1414. (CVS 2683) (check-in: 415b8b24 user: drh tags: trunk)
2005-09-09
10:46
Enhanced disk-full tests. (CVS 2682) (check-in: 0288fa5d user: drh tags: trunk)
10:17
Detect errors returned by SetFilePointer on windows. (CVS 2681) (check-in: bc8c33f9 user: drh tags: trunk)
01:33
Infrastructure for the DISTINCT keyword in aggregate functions. But it does not work yet. If you try to use it you get an error message. (CVS 2680) (check-in: 4d62e36f user: drh tags: trunk)
01:32
Defend against disk I/O errors that happen during an sqlite3OsSeek(). (CVS 2679) (check-in: 461e3a0a user: drh tags: trunk)
2005-09-08
20:37
SUM returns NULL when it has no inputs. Ticket #1413. (CVS 2678) (check-in: 62818594 user: drh tags: trunk)
19:45
A SUM() of all NULLs returns NULL. A SUM() of nothing return 0. A SUM() of a mixture of NULLs and numbers returns the sum of the numbers. Ticket #1413. (CVS 2677) (check-in: 2e6230ed user: drh tags: trunk)
19:01
Remove a C++ism that snuck in on one of the reason changes. (CVS 2676) (check-in: 48f6a331 user: drh tags: trunk)
14:17
Remove a few unused variables detected by Borland C. Ticket #1412. (CVS 2675) (check-in: cdfe372a user: drh tags: trunk)
12:57
Aggregates with GROUP BY can sometimes return no result rows. (CVS 2674) (check-in: 0c3e368d user: drh tags: trunk)
12:38
While doing a transaction comment, use fdatasync() instead of fsync() in cases there the file size is unchanged. (CVS 2673) (check-in: 3c555a87 user: drh tags: trunk)
12:37
Fix another test to account for SUM() returning integer rather than float. (CVS 2672) (check-in: bc723235 user: drh tags: trunk)
11:04
Fix typos in documentation. Ticket #1239. (CVS 2671) (check-in: 73fbb8d9 user: drh tags: trunk)
10:58
Fix a comment in sqlite.h. Ticket #1321. (CVS 2670) (check-in: ed4e9e75 user: drh tags: trunk)
10:37
The SUM() aggregate function returns an integer result if all inputs are integers. Any single non-integer input causes the result to be a floating point value. (CVS 2669) (check-in: 21adf4bd user: drh tags: trunk)
02:00
Changes to comments only in sqliteInt.h. No changes to code. (CVS 2668) (check-in: cc2a6165 user: drh tags: trunk)
01:58
Optimizations and refinements. Improvements to test coverage. (CVS 2667) (check-in: 7283f7c2 user: drh tags: trunk)
00:13
All regression tests now pass. But I am sure there must still be problems. New tests need to be added. (CVS 2666) (check-in: bcc7d722 user: drh tags: trunk)
2005-09-07
23:05
More bug fixes. All of the "quick" tests pass. The full test suite still shows problems. (CVS 2665) (check-in: a1b6d910 user: drh tags: trunk)
22:48
More bug fixes. But there are still tests that fail. (CVS 2664) (check-in: 7e85a162 user: drh tags: trunk)
22:09
Bug fixes in aggregate processing. Fewer tests fail. (CVS 2663) (check-in: c3ac5859 user: drh tags: trunk)
21:22
Rewrite the aggregate handling logic so that it runs in O(1) space. This is the first cut at the code. Many regression tests fail. (CVS 2662) (check-in: 17039ec3 user: drh tags: trunk)
2005-09-06
21:40
Changes to reduce the amount of stack space required. (CVS 2661) (check-in: b86bd70f user: drh tags: trunk)
20:36
Cleanup the processing of MEM_Agg elements. (CVS 2660) (check-in: 7ecf3654 user: drh tags: trunk)
10:26
Fix ANALYZE so that it works even if the empty_result_callbacks pragma is enabled. Ticket #1409. (CVS 2659) (check-in: bfdaef12 user: drh tags: trunk)
2005-09-05
20:35
Make due with "sub()" instead of "gsub()" in awk scripts so that builds will work on Solaris which will runs an archiac version of AWK. (CVS 2658) (check-in: 449a7895 user: drh tags: trunk)
20:06
Fix bugs introduced by checkin (2656) and do some modest code enhancements. (CVS 2657) (check-in: 7b56763a user: drh tags: trunk)
19:08
Use the unicode API to win32 where available. Tickets #1407, #1396, #1331, #1243, #1206 (CVS 2656) (check-in: 3ec58c67 user: drh tags: trunk)
2005-09-01
17:47
Fix over-aggressive optimization of ORDER BY as reported on the mailing list. (CVS 2655) (check-in: efbb4bc8 user: drh tags: trunk)
12:16
All regression tests now pass with the new bounded-memory sort code. There is still lots of opportunity for optimization, however. (CVS 2654) (check-in: 81259a01 user: drh tags: trunk)
03:07
Sorting is now done using a sorting index rather than loading the entire result set into memory and doing a merge sort. The old merge sort technique was a carry-over from SQLite version 1. The new method uses a bounded amount of memory and scales to much larger result sets. There are still errors: some 39 regression tests fail. (CVS 2653) (check-in: 09db0a24 user: drh tags: trunk)
2005-08-31
18:20
{quote: KeyInfo} generation moved to a common subroutine. (CVS 2652) (check-in: a25801df user: drh tags: trunk)
13:48
Updates to the query optimizer overview document. (CVS 2651) (check-in: b1dceef0 user: drh tags: trunk)
13:13
Explicit typecasts to silence nuisance compiler warnings. Ticket #1398. (CVS 2650) (check-in: 90712ea7 user: drh tags: trunk)
03:13
Refinements to the optimizer overview and integration into the website. (CVS 2649) (check-in: ccd12e9e user: drh tags: trunk)
02:46
Update the FAQ to include an entry about binary versus decimal numbers. (CVS 2648) (check-in: 0bbe73fc user: drh tags: trunk)
01:49
Adding the first cut of the optimizer overview document. (CVS 2647) (check-in: 4c8d0a4c user: drh tags: trunk)
2005-08-30
22:44
Adjust the makefile for new optimizer documentation. (CVS 2646) (check-in: 14668c85 user: drh tags: trunk)
22:44
Initial work on documentation describing the optimizer. (CVS 2645) (check-in: 5cebd7ba user: drh tags: trunk)
20:12
The CSV output mode does not sign-extend bytes where the high-order bit is set. Ticket #1397. (CVS 2644) (check-in: 528df777 user: drh tags: trunk)
19:30
Fix printf.c so that it handles 10.0 correctly. (CVS 2643) (check-in: 0f7a53f7 user: drh tags: trunk)
00:54
Minor code enhancements. (CVS 2642) (check-in: 4ab994a8 user: drh tags: trunk)
00:10
Update the home page to brag about 150KiB code footprint when optional features are omitted. (CVS 2641) (check-in: 0e1ac128 user: drh tags: trunk)
2005-08-29
23:06
Make the shell aware of SQLITE_OMIT_MEMORYDB. Ticket #1395. (CVS 2640) (check-in: 71a71cef user: drh tags: trunk)
23:00
Increase resolution of time-of-day on unix. Add an experimental sqlite3_profile() API. (CVS 2639) (check-in: ed2ca087 user: drh tags: trunk)
16:40
Initialize a local variable to avoid a nuisance compiler warning. Ticket #1394. (CVS 2638) (check-in: 9b914901 user: drh tags: trunk)
2005-08-28
17:00
The LIKE optimization does the right thing when collating sequences are present. LIKE expressions where the left-hand side has COLLATE NOCASE are optimized in the default case. (CVS 2637) (check-in: ef84ff79 user: drh tags: trunk)
01:38
Improvements to the formatting and layout of the code in the previous checkin. (CVS 2636) (check-in: 73b430de user: drh tags: trunk)
01:34
Add an assert to verify that floating point values and integers use the same byte order. Add a comment to the SELECT code generator. (CVS 2635) (check-in: 07b48921 user: drh tags: trunk)
2005-08-27
17:23
Version 3.2.5 (CVS 2634) (check-in: b2415a74 user: drh tags: trunk)
17:22
Change the wording of the news line for release 3.2.5. (CVS 2633) (check-in: 70db46d3 user: drh tags: trunk)
17:19
Get the date right. Today is the 27th, not the 25th or the 28th. (CVS 2632) (check-in: 905b9d6c user: drh tags: trunk)