| 1 | Revision history for IO::AIO |
1 | Revision history for IO::AIO |
| 2 | |
2 | |
| 3 | TODO: better autoconf.pm that can return LIBS etc. |
3 | TODO: better autoconf.pm that can return LIBS etc. |
| 4 | TODO: aio_cptree/mvtree |
4 | TODO: aio_cptree/mvtree |
| 5 | TODO: reduce condvar fairness: schedule hot-cache-threads first? |
5 | TODO: reduce condvar fairness: schedule hot-cache-threads first? |
| 6 | TODO: splice/tee/vmsplice/sync_file_range? (http://kerneltrap.org/node/6505 http://lwn.net/Articles/178199/) |
6 | TODO: splice/tee/vmsplice? (http://kerneltrap.org/node/6505 http://lwn.net/Articles/178199/) |
| 7 | TODO: aio_fcntl, at least for file-locking |
7 | TODO: aio_fcntl, at least for file-locking |
|
|
8 | |
|
|
9 | - speed up readdirx sort algorithm slightly. |
|
|
10 | |
|
|
11 | 3.21 Fri Jun 12 18:45:53 CEST 2009 |
|
|
12 | - new options --dirs and --files for treescan. |
|
|
13 | - install bin/treescan by default. |
|
|
14 | - (libeio) aio_readdir can now be cancelled while executing. |
|
|
15 | - fix a printf format string for 64 bit systems (could lead |
|
|
16 | to problems on big endian 64 bit systems). |
|
|
17 | - do not use qsort() but our own algorithm: glibc initialises |
|
|
18 | the whole locale and I/O subsystem inside qsort, causing |
|
|
19 | a stack overflow on 32 bit machines. The new sort uses much less |
|
|
20 | stack and is more than twice as fast in typical situations. |
|
|
21 | |
|
|
22 | 3.2 Sun Jun 7 20:30:05 CEST 2009 |
|
|
23 | - (libeio) pwrite emulation was even more flawed than intended and did |
|
|
24 | not restore the file offset. |
|
|
25 | - add aio_readdirx, which can return inode and filetype and sort |
|
|
26 | the names in various ways. |
|
|
27 | - unfortunately, utime, chmod, chown on an open file that has just |
|
|
28 | been written can easily block, which caused aio_copy to block the |
|
|
29 | process. no more! |
|
|
30 | - no longer rely on dst path in aio_copy when futime is available. |
|
|
31 | |
|
|
32 | 3.19 Tue Apr 21 22:05:21 CEST 2009 |
|
|
33 | - more perl 5.10 workarounds for aio_read and write. |
|
|
34 | - aio_write no longer modifies the sv (if possible). |
|
|
35 | - aio_read now works correctly with magic values. |
|
|
36 | |
|
|
37 | 3.18 Sun Apr 19 21:17:32 CEST 2009 |
|
|
38 | - better diagnostics when some aio_* functions get passed an illegal fh. |
|
|
39 | - try to avoid crashes due to incompatible 5.10 API changes. grmbl. |
|
|
40 | |
|
|
41 | 3.17 Thu Nov 20 08:45:36 CET 2008 |
|
|
42 | - (libeio) added aio_sync_file_range (untested). |
|
|
43 | - add aio_busy to @IO::AIO::AIO_REQ. |
|
|
44 | |
|
|
45 | 3.16 Wed Oct 22 18:28:01 CEST 2008 |
|
|
46 | - use SvREFCNT_inc instead of SvREFCNT_inc_NN in a non-speed critical |
|
|
47 | part to improve portability to perl 5.8 (reported by szymon). |
|
|
48 | |
|
|
49 | 3.15 Mon Oct 13 00:39:55 CEST 2008 |
|
|
50 | - automatic removal of feeders was broken. |
|
|
51 | - (libeio) use a more robust method to detect whether a feeder |
|
|
52 | has added something to the group or not. |
|
|
53 | |
|
|
54 | 3.1 Thu Oct 2 13:34:40 CEST 2008 |
|
|
55 | - pre-deref the passed callback object, for increased speed |
|
|
56 | and decreased memory usage. |
|
|
57 | - call on_next_submit callback even for aio_group - impact unknown, |
|
|
58 | but seems more correct. |
|
|
59 | - $req->cb now returns the original callback. |
|
|
60 | - (libeio) pass EIO_GROUP/aio_group requests through the queue without |
|
|
61 | requiring a thread switch. |
|
|
62 | - (libeio) status code of aio_futime and aio_utime was always passed 0. |
|
|
63 | - do some other ยต-optimisations. |
|
|
64 | |
|
|
65 | 3.07 Sat Aug 2 16:06:13 CEST 2008 |
|
|
66 | - do not include LIBS in autoconf tests. |
| 8 | |
67 | |
| 9 | 3.06 Tue Jul 15 12:41:32 CEST 2008 |
68 | 3.06 Tue Jul 15 12:41:32 CEST 2008 |
| 10 | - move preadwritelock definition before it's first reference, |
69 | - move preadwritelock definition before it's first reference, |
| 11 | to make it compile again on broken platforms (they still exist) |
70 | to make it compile again on broken platforms (they still exist) |
| 12 | or badly cofngiured perls (redhat...), reported by Rob Bloodgood. |
71 | or badly cofngiured perls (redhat...), reported by Rob Bloodgood. |
