| 1 | Revision history for IO::AIO |
1 | Revision history for IO::AIO |
|
|
2 | |
|
|
3 | - (experimental) sendfile support, including sendfile emulation |
|
|
4 | when sendfile isn't available. Linux, FreeBSD and HP-UX might |
|
|
5 | work (sendfilev support is unavailable on my solaris 8 machine...). |
|
|
6 | |
|
|
7 | 1.5 Sat Aug 20 02:32:19 CEST 2005 |
|
|
8 | - create the aio threads on demand only. |
|
|
9 | - somebody nagged about C89 compatibility, so this release should |
|
|
10 | be C89 compatible, but no guarantees for future ones. |
|
|
11 | - abort when the configure script fails. |
|
|
12 | |
|
|
13 | 1.4 Wed Aug 17 08:07:27 CEST 2005 |
|
|
14 | - forget to recreate the pipe after forking, this could cause |
|
|
15 | deadlocks. |
|
|
16 | |
|
|
17 | 1.3 Wed Aug 17 07:25:54 CEST 2005 |
|
|
18 | - properly propagate exceptions from callbacks, instead of silently |
|
|
19 | eating them. |
|
|
20 | - use a different fork algorithm that avoids recursive callback |
|
|
21 | invocations and reduces fork latency in busy processes. |
|
|
22 | - fix a bug where the buffer scalar in aio_read got assigned the wrong |
|
|
23 | length after reading. |
|
|
24 | - fix a bug where the child would process some outstanding requests |
|
|
25 | from the parent. |
|
|
26 | - make a copy of the callback scalar, to avoid some very uncommon |
|
|
27 | (but valid) usages where the callback scalar changes. |
|
|
28 | - the callback was never freed in aio_read/aio_write. |
|
|
29 | - aio_read/aio_write will now set the readonly flag on the data |
|
|
30 | scalar for better error reporting. |
|
|
31 | |
|
|
32 | 1.2 Wed Aug 17 01:22:58 CEST 2005 |
|
|
33 | - verified on cygwin, linux/amd64+ia32+alpha, solaris. |
|
|
34 | - use SvPVbyte for filenames now. You *need* to encode your strings to |
|
|
35 | the external filename encoding first. |
|
|
36 | - fix the bug where filehandles could leak (again). |
|
|
37 | - clarify freebsd situation: make it work and tell me how, or go away |
|
|
38 | with your underdocumented, broken and outdated OS. |
|
|
39 | - fork'ing should now be officially supported (where linux and |
|
|
40 | posix contradict each other, I followed posix). |
|
|
41 | - only call read() once per poll, not twice (speedup). |
|
|
42 | - new functions: aio_rmdir, aio_symlink. |
|
|
43 | - force callback arguments to be undef or a reference, for better |
|
|
44 | error catching. |
|
|
45 | |
|
|
46 | 1.1 Sun Aug 7 05:27:23 CEST 2005 |
|
|
47 | - added configure script to detect availability of some calls. |
|
|
48 | (imperfect, as we cannot detect things like -lrt due to missing |
|
|
49 | feedback yet). |
|
|
50 | - emulate fdatasync via fsync, pread/pwrite via unsafe read/write |
|
|
51 | and readahead via manual reading on platforms lacking those. |
|
|
52 | - ported t/03_errors.t+framework by Brad Fitzpatrick |
|
|
53 | from Linux::AIO. |
|
|
54 | - removed XOPEN_SOURCE because stupid solaris 9 doesn't |
|
|
55 | compile with it enabled. Probably breaks others. Oh well. |
|
|
56 | - try to be more careful with errno on non-threaded perls. |
|
|
57 | - cygwins perl insisted on a matching open mode when morphing |
|
|
58 | the fd to an fh. |
|
|
59 | - convert sv's to "byte" encoding in aio_read/aio_write. Might |
|
|
60 | do that for pathnames, too, in the future, so make sure you |
|
|
61 | properly encode your pathnames. |
|
|
62 | |
|
|
63 | 1.0 Fri Jul 22 10:43:30 CEST 2005 |
|
|
64 | - aio_open did not reliably deliver (open) filehandles. |
|
|
65 | - all aio_*-functions keep a reference to the filehandle so |
|
|
66 | it doesn't go away while the aio requests is working on it. |
|
|
67 | - disable warnings in AIO.pm. |
| 2 | |
68 | |
| 3 | 0.9 Wed Jul 20 23:55:04 CEST 2005 |
69 | 0.9 Wed Jul 20 23:55:04 CEST 2005 |
| 4 | - allow recursive calls to poll_cb, use less locking calls, too. |
70 | - allow recursive calls to poll_cb, use less locking calls, too. |
| 5 | - fix the Gtk2-example in the SYNOPSIS. |
71 | - fix the Gtk2-example in the SYNOPSIS. |
| 6 | - use pread/pwrite not pread64/pwrite64. this was simply a leftover |
72 | - use pread/pwrite not pread64/pwrite64. this was simply a leftover |
