| 1 | Revision history for IO::AIO |
1 | Revision history for IO::AIO |
| 2 | |
2 | |
| 3 | TODO: better autoconf.pm that can return LIBS etc. |
|
|
| 4 | TODO: aio_cptree/mvtree |
3 | TODO: aio_cptree/mvtree |
| 5 | TODO: reduce condvar fairness: schedule hot-cache-threads first? |
4 | TODO: reduce condvar fairness: schedule hot-cache-threads first? |
| 6 | TODO: splice/tee/vmsplice? (http://kerneltrap.org/node/6505 http://lwn.net/Articles/178199/) |
5 | TODO: vmsplice? (http://kerneltrap.org/node/6505 http://lwn.net/Articles/178199/) |
| 7 | TODO: aio_fcntl, at least for file-locking |
6 | TODO: aio_fcntl, at least for file-locking |
| 8 | TODO: fallocate, if it wouldn't be so useless. |
|
|
| 9 | TODO: aio_mincore? |
7 | TODO: aio_mincore? |
|
|
8 | TODO: getxattr etc.? |
|
|
9 | TODO: F_DUPFD_CLOEXEC |
|
|
10 | TODO: emulation for splice? |
| 10 | |
11 | |
| 11 | TODO: MS_xxx, posix_fadvise |
12 | - make fh argument of IO::AIO::mmap optional, also add mmap/munmap |
|
|
13 | to quick overview. |
|
|
14 | - splice can corrupt data. |
|
|
15 | - (libeio) remove pread/pwrite emulation. |
|
|
16 | |
|
|
17 | 4.15 Tue Apr 10 06:59:00 CEST 2012 |
|
|
18 | - always include linux/types.h for fiemap, for compatibility to |
|
|
19 | ancient systems (Paul Howarth). |
|
|
20 | - experimental support for IO::AIO::splice and ::tee (no aio_...). |
|
|
21 | - provide SEEK_HOLE and SEEK_DATA, if available. |
|
|
22 | - work around (again!) an immensely stupid bug in RHEL, defining |
|
|
23 | autoconf macros in linux system headers (analysed by Paul Howarth). |
|
|
24 | |
|
|
25 | 4.14 Sat Apr 7 02:45:18 CEST 2012 |
|
|
26 | - fix stat structure usage on windows, which caused bogus stat results. |
|
|
27 | - (libeio) make readahead emulation behave more like actual readahead by never failing. |
|
|
28 | - new request aio_seek. |
|
|
29 | - new request aio_fiemap. |
|
|
30 | - autogenerate the #ifdef/#define 0 blocks for symbols we export. |
|
|
31 | |
|
|
32 | 4.12 Fri Dec 30 08:51:25 CET 2011 |
|
|
33 | - realpath would return a random value if the name was too long, |
|
|
34 | instead of -1. |
|
|
35 | - port to c89. |
|
|
36 | |
|
|
37 | 4.11 Mon Oct 10 00:24:11 CEST 2011 |
|
|
38 | - libeio didn't compile on !linux (Paul Howarth). |
|
|
39 | |
|
|
40 | 4.1 Sun Oct 9 10:24:11 CEST 2011 |
|
|
41 | - IO::AIO did access uninitialised memory on unsuccessful stats. |
|
|
42 | - (libeio) added syncfs syscall wrapper. |
|
|
43 | - (libeio) set thread name on linux (ps -L/Hcx, top, gdb). |
|
|
44 | - (libeio) support multiple indepenent working directories. |
|
|
45 | - applied speling corrections by Nicholas Bamber. |
|
|
46 | - tune treescan #threads and #outstanding requests a bit. |
|
|
47 | - reduce the number of file descriptors in use at any one time |
|
|
48 | in treescan, to avoid running out of them (and using too much ram). |
|
|
49 | - take advantage of working directoriy abstractions |
|
|
50 | in aio_scandir and treescan. |
|
|
51 | - reduce compiled size by ~10% by not inlining some key functionality. |
|
|
52 | - added --progress switch to treescan. |
|
|
53 | |
|
|
54 | 4.0 Mon Jul 18 05:01:10 CEST 2011 |
|
|
55 | - INCOMPATIBLE CHANGE: fork is no longer supported (indeed, it never was), |
|
|
56 | see FORK BEHAVIOUR in manpage for details. |
|
|
57 | - passes testsuite on win32 now and may actually work |
|
|
58 | (activestate 5.10.1 / MSVC6, YMMV). |
|
|
59 | - (libeio) fix a deadlock where a wakeup signal could be missed when |
|
|
60 | a timeout occured at the same time. |
|
|
61 | - (libeio) added realpath. |
|
|
62 | - (libeio) added fallocate. |
|
|
63 | - (libeio) disabling sendfile on darwin unfortunately broke the emulation. |
|
|
64 | - (libeio) do not acquire any locks when forking. |
|
|
65 | - (libeio) use fewer time() syscalls when waiting for new requests. |
|
|
66 | |
|
|
67 | 3.93 Wed Jun 29 23:44:18 CEST 2011 |
|
|
68 | - ECB.H WAS MISSING. |
|
|
69 | |
|
|
70 | 3.92 Wed Jun 29 14:45:41 CEST 2011 |
|
|
71 | - ecb.h was missing. |
|
|
72 | |
|
|
73 | 3.91 Wed Jun 29 13:24:42 CEST 2011 |
| 12 | - (libeio) work around a Linux (and likely FreeBSD and other |
74 | - (libeio) work around a Linux (and likely FreeBSD and other |
| 13 | kernels) bug where sendfile would not transfer all the requested |
75 | kernels) bug where sendfile would not transfer all the requested |
| 14 | bytes on large transfers, using a heuristic. |
76 | bytes on large transfers, using a heuristic. |
|
|
77 | - buggy sendfile caused aio_move/copy to sometimes fail for big |
|
|
78 | files (fortunately it checks that the whole file has been |
|
|
79 | transferred...) |
|
|
80 | - use libecb for higher performance and higher portability. |
|
|
81 | - (libeio) disable sendfile on darwin, it's too broken. |
|
|
82 | - disable fork tests on !linux, as only linux supports mixing |
|
|
83 | pthread and fork in perl. |
|
|
84 | - document the fact that fork doesn't work anymore when using |
|
|
85 | this module, on anything !GNU/Linux. |
|
|
86 | - increase timeout in t/04_fork.t, as too many CPAN-tester setups |
|
|
87 | run on a 0.1MHz cpu. Or so. |
| 15 | |
88 | |
| 16 | 3.9 Fri May 27 02:43:47 CEST 2011 |
89 | 3.9 Fri May 27 02:43:47 CEST 2011 |
| 17 | - (libeio) fix memory corruption in aio_readdirx for the flags |
90 | - (libeio) fix memory corruption in aio_readdirx for the flags |
| 18 | combination READDIR_STAT_ORDER | READDIR_DIRS_FIRST. |
91 | combination READDIR_STAT_ORDER | READDIR_DIRS_FIRST. |
| 19 | - add lots of newre POSIX and GNU/Linux-specific open |
92 | - add lots of newer POSIX and GNU/Linux-specific open |
| 20 | flags. |
93 | flags. |
| 21 | |
94 | |
| 22 | 3.8 Sun Mar 27 12:25:33 CEST 2011 |
95 | 3.8 Sun Mar 27 12:25:33 CEST 2011 |
| 23 | - use nonstandard but maybe-working-on-bsd fork technique. |
96 | - use nonstandard but maybe-working-on-bsd fork technique. |
| 24 | - support a max_idle value of 0. |
97 | - support a max_idle value of 0. |
