| 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: aio_utime, aio_chmod, aio_chown? should simply stat/open first? |
|
|
| 6 | TODO: reduce condvar fairness: schedule hot-cache-threads first? |
5 | TODO: reduce condvar fairness: schedule hot-cache-threads first? |
| 7 | TODO: fix setsig |
6 | TODO: splice/tee/vmsplice/sync_file_range? (http://kerneltrap.org/node/6505 http://lwn.net/Articles/178199/) |
|
|
7 | TODO: aio_fcntl, at least for file-locking |
|
|
8 | |
|
|
9 | - port to solaris perls configured for old posix |
|
|
10 | (analysed by Jost Krieger). |
|
|
11 | |
|
|
12 | 2.61 Wed Apr 16 18:45:02 CEST 2008 |
|
|
13 | - fix treescan output duplication and improve output ordering. |
|
|
14 | also display files given on the commandline. |
|
|
15 | - use a different algorithm for aio_close that is faster |
|
|
16 | and probably has even lower chances of blocking. |
|
|
17 | - do our own stack memory management for threads - linux |
|
|
18 | allocates outrageous amounts of VM (not so bad unless you |
|
|
19 | use mlockall...), which severely limits the number of threads |
|
|
20 | on 32-bit arches: stack size is the larger of PTHREAD_STACK_MIN |
|
|
21 | and 4096 * sizeof (long) (usually this is 16..64k). |
|
|
22 | |
|
|
23 | 2.6 Sun Mar 30 08:28:11 CEST 2008 |
|
|
24 | - added aio_sync. |
|
|
25 | - added aio_pathsync. |
|
|
26 | - fix prototypes of (void) functions. |
|
|
27 | |
|
|
28 | 2.51 Sat Oct 6 16:04:54 CEST 2007 |
|
|
29 | - perlio isn't generally threadsafe, so aio_close could not work reliably. |
|
|
30 | aio_close now tries a safe hack that might not be that asynchronous |
|
|
31 | (see the manpage for details). |
|
|
32 | - discard callback return values as to not grow the stack endlessly |
|
|
33 | when poll_cb handles many requests without returning. |
|
|
34 | - minor code cleanups. |
|
|
35 | |
|
|
36 | 2.5 Thu Oct 4 14:49:08 CEST 2007 |
|
|
37 | - replaced _fd2fh with faster xs code. |
|
|
38 | - aio_close will now try to do "the right thing" and thus might |
|
|
39 | work sensibly for the very first time. |
|
|
40 | |
|
|
41 | 2.41 Mon Sep 24 21:28:21 CEST 2007 |
|
|
42 | - after fork, invest some work to make sure that the poll_fileno |
|
|
43 | stays stable (by dup2'ing the new fd over the old one), to make |
|
|
44 | it easier for programs/libs that don't control their forking |
|
|
45 | behaviour and cannot recreate their watchers. |
|
|
46 | |
|
|
47 | 2.4 Sun Aug 5 18:44:22 CEST 2007 |
|
|
48 | - add aio_truncate, aio_chmod, aio_chown, aio_utime (linux |
|
|
49 | successfully demonstrated that you can block on futimes...). |
|
|
50 | - allow undef as fileoffset for aio_read/write and use read/write(2) |
|
|
51 | internally (useful for sockets or O_APPEND handles). |
|
|
52 | - allow undef for length in aio_write. |
|
|
53 | - negative dataoffsets work as expected now in aio_read/aio_write. |
|
|
54 | - use NV instead of UV for 32 bit perls and file offsets, as NVs |
|
|
55 | have a larger range then. |
|
|
56 | - shared code between BDB and IO::AIO. |
|
|
57 | - aio_busy was completely broken. now fixed. |
|
|
58 | - readahead emulation now returns something. |
|
|
59 | - correctly set errno to ENOSYS on unimplemented functions |
|
|
60 | (should never happen, though). |
|
|
61 | - large changes to make it partially compile and run on win32, |
|
|
62 | but win32 lacks too much functionality, and perl overrides way |
|
|
63 | too many functions with crashing versions. |
| 8 | |
64 | |
| 9 | 2.33 Tue Jan 23 23:55:41 CET 2007 |
65 | 2.33 Tue Jan 23 23:55:41 CET 2007 |
| 10 | - fix install path (Andreas J. Koenig). |
66 | - fix install path (Andreas J. Koenig). |
| 11 | |
67 | |
| 12 | 2.32 Mon Jan 22 16:56:23 CET 2007 |
68 | 2.32 Mon Jan 22 16:56:23 CET 2007 |
