| 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/rmtree |
|
|
5 | TODO: aio_utime, aio_chmod, aio_chown? should simply stat/open first? |
| 4 | |
6 | |
| 5 | 1.73 |
7 | 2.2 Tue Oct 31 00:31:00 CET 2006 |
| 6 | - AIX isn't being helpful again, hardcode a value for NAME_MAX |
8 | - minor doc updates. |
|
|
9 | - added aio_readlink. |
|
|
10 | - properly zero-terminate aio_read buffer. |
|
|
11 | - fix aio_read/write with nonzero dataoffset. |
|
|
12 | - reduced size of aio request structure. |
|
|
13 | |
|
|
14 | 2.1 Sun Oct 29 02:01:57 CET 2006 |
|
|
15 | - INCOMPATIBLE CHANGE: replace IO::AIO::poll_some by |
|
|
16 | IO::AIO::max_poll_reqs and IO::AIO::max_poll_time. |
|
|
17 | - fix a bug in aio_move where it would fail everytime |
|
|
18 | it would have to copy files manually. |
|
|
19 | - doc updates. |
|
|
20 | - start threads only on (very low) demand. |
|
|
21 | - end superfluous threads automatically after 10s idling. |
|
|
22 | - add IO::AIO::max_idle to control this. |
|
|
23 | - some tuning applied to random places. |
|
|
24 | - add aio_mknod. |
|
|
25 | |
|
|
26 | 2.0 Thu Oct 26 18:27:58 CEST 2006 |
|
|
27 | - minor incompatibility: max_outstanding semantics have changed. |
|
|
28 | - major rewrite, consider all functionality EXPERIMENTAL. |
|
|
29 | - aio requests can be cancelled. |
|
|
30 | - callbacks can be (re-)set later. |
|
|
31 | - aio requests can choose between 9 priorities. |
|
|
32 | - add aio_group: aio requests can be grouped |
|
|
33 | into composite requests. |
|
|
34 | - generator interface for groups. |
|
|
35 | - add IO::AIO::poll_some. |
|
|
36 | - major documentation improvements. |
|
|
37 | - very minor bugfixes. |
|
|
38 | - add aio_nop for dummy requests. |
|
|
39 | - add aio_busy mainly for benchmarking and debugging. |
|
|
40 | - use "fast" mutexes on linux. |
|
|
41 | - use dynamic (and larger) buffers instead of |
|
|
42 | putting them onto the stack. |
|
|
43 | - optimise x86 and amd64 a bit. |
|
|
44 | - better error checking when ENOMEM. |
|
|
45 | - hopefully fix mem and dirp leaks on fork. |
|
|
46 | - warn about broken perl malloc. |
|
|
47 | - compiles on cygwin (but pread is broken on cygwin |
|
|
48 | which is not detected by autoconf, so you have to |
|
|
49 | edit autoconf/config.h yourself to enable emulation). |
|
|
50 | |
|
|
51 | 1.8 Sat Jun 24 22:20:44 CEST 2006 |
|
|
52 | - add and/or document aio_rename, aio_link, aio_symlink and aio_move. |
|
|
53 | |
|
|
54 | 1.73 Wed Mar 1 22:49:32 CET 2006 |
|
|
55 | - codename "jost" |
|
|
56 | - AIX/Solaris(?) aren't being helpful again, hardcode a value |
| 7 | and drive with your eyes closed (reported by wcooley@nakedape.cc). |
57 | for NAME_MAX and drive with your eyes closed (reported by |
|
|
58 | wcooley@nakedape.cc). |
| 8 | |
59 | |
| 9 | 1.72 Thu Feb 2 00:40:44 CET 2006 |
60 | 1.72 Thu Feb 2 00:40:44 CET 2006 |
| 10 | - another minor bug in aio_scandir, again, the callback was called |
61 | - another minor bug in aio_scandir, again, the callback was called |
| 11 | twice when the directory can't be stat'ed. |
62 | twice when the directory can't be stat'ed. |
| 12 | |
63 | |
| … | |
… | |
| 37 | - somebody nagged about C89 compatibility, so this release should |
88 | - somebody nagged about C89 compatibility, so this release should |
| 38 | be C89 compatible, but no guarantees for future ones. |
89 | be C89 compatible, but no guarantees for future ones. |
| 39 | - abort when the configure script fails. |
90 | - abort when the configure script fails. |
| 40 | |
91 | |
| 41 | 1.4 Wed Aug 17 08:07:27 CEST 2005 |
92 | 1.4 Wed Aug 17 08:07:27 CEST 2005 |
| 42 | - forget to recreate the pipe after forking, this could cause |
93 | - forgot to recreate the pipe after forking, this could cause |
| 43 | deadlocks. |
94 | deadlocks. |
| 44 | |
95 | |
| 45 | 1.3 Wed Aug 17 07:25:54 CEST 2005 |
96 | 1.3 Wed Aug 17 07:25:54 CEST 2005 |
| 46 | - properly propagate exceptions from callbacks, instead of silently |
97 | - properly propagate exceptions from callbacks, instead of silently |
| 47 | eating them. |
98 | eating them. |
