Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
History for test/sysfault.test
|
2015-12-03
| ||
| 01:48 | [c9f2b0d8] part of check-in [4f7f3550] A unix VFS change replaces fsync() with fstat() when using SQLITE_NO_SYNC, so set PRAGMA synchronous=OFF in the sysfault-3 test to avoid erroneously causing errors in xSync. (check-in: [4f7f3550] user: drh branch: trunk, size: 7810) | |
|
2013-04-15
| ||
| 17:03 | [fa776e60] part of check-in [ea1404a1] Refactoring the mmap interface. The controlling pragma is now "mmap_size" instead of "mmap_limit". Also change SQLITE_CONFIG_MMAP_LIMIT and SQLITE_FCNTL_MMAP_LIMIT to SQLITE_CONFIG_MMAP_SIZE and SQLITE_FCNTL_MMAP_SIZE, respecctively. The default mmap_size is now always 0, meaning that memory mapped I/O is off by default. There is a new compile-time option SQLITE_MAX_MMAP_SIZE that determines a hard upper bound on the mmap_size. Setting SQLITE_MAX_MMAP_SIZE to zero disables the memory-mapped I/O logic and causes it to be omitted from the build. An extra argument is added to SQLITE_CONFIG_MMAP_SIZE that can optionally lower the SQLITE_MAX_MMAP_SIZE at start-time. The SQLITE_MAX_MMAP_SIZE is zero for platforms where we know that it does not work, meaning that it cannot be turned on by mistake on those platforms. (check-in: [ea1404a1] user: drh branch: trunk, size: 7782) | |
|
2013-03-25
| ||
| 23:09 | [503f7271] part of check-in [1b37c4ef] Memory-mapped I/O is now on by default. The "PRAGMA mmap_limit(N)" can be used to issue a hint to the VFS to limit mmap space to N bytes. The VFS is free to ignore that hint if desired. However, if "PRAGMA mmap_limit(0)" is used, xFetch is never called. (check-in: [1b37c4ef] user: drh branch: experimental-mmap, size: 7783) | |
| 19:05 | [3258d6c8] part of check-in [6ec7367d] Add a test that simulates an error in mmap(). (check-in: [6ec7367d] user: dan branch: experimental-mmap, size: 7783) | |
|
2011-04-01
| ||
| 11:56 | [c79441d8] part of check-in [ff6dfe6e] In os_unix.c, do not return SQLITE_BUSY to SQLite following an error in fcntl(F_UNLCK), regardless of the value of errno. (check-in: [ff6dfe6e] user: dan branch: trunk, size: 7158) | |
|
2011-03-30
| ||
| 19:08 | [a2c3ca66] part of check-in [a84f7711] Further tests for os_unix.c. (check-in: [a84f7711] user: dan branch: trunk, size: 7039) | |
|
2011-03-29
| ||
| 18:28 | [be42aa42] part of check-in [3d2de011] Add tests to syscall.test and sysfault.test. (check-in: [3d2de011] user: dan branch: trunk, size: 4626) | |
| 15:40 | [359ea90a] part of check-in [dbe569a0] Fix a problem whereby following an IO error in CommitPhaseTwo() of a multi-file transaction the b-tree layer could be left in TRANS_WRITE state, causing problems later on. (check-in: [dbe569a0] user: dan branch: trunk, size: 3955) | |
|
2011-03-28
| ||
| 19:10 | [63144f00] part of check-in [0e1d20df] Add test infrastructure (and some tests) to inject faults into os_unix.c using the new xSetSystemCall interface. (check-in: [0e1d20df] user: dan branch: trunk, size: 2010) Added | |