| … | |
… | |
| 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? (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 | TODO: fallocate, if it wouldn't be so useless. |
|
|
9 | TODO: IO::AIO::mmap? aio_mincore? |
| 8 | |
10 | |
| 9 | - avoid some degenerate sorting cases on systems without d_ino, |
11 | - (libeio) max_poll_time was not properly converted to ticks. |
| 10 | or when the dentries are mostly sorted by inode already (common |
12 | - clarify cancel_subs description. |
| 11 | with freshly unpacked tarfiles for example). |
13 | - IO::AIO::sendfile did not actually return the return value |
|
|
14 | from the sendfile call. |
|
|
15 | - implement aio_msync, aio_mtouch. |
|
|
16 | - (libeio) tentatively support darwin in sendfile. |
|
|
17 | |
|
|
18 | 3.31 Thu Nov 12 02:14:29 CET 2009 |
|
|
19 | - fix result status documentation of aio_copy, aio_move. |
|
|
20 | - speed up object creation considerably by avoiding |
|
|
21 | hash lookups. |
|
|
22 | |
|
|
23 | 3.3 Wed Aug 5 13:52:58 CEST 2009 |
|
|
24 | - use common::sense. |
|
|
25 | - use common schmorp.h header. |
|
|
26 | - allow integers (file descriptors) in addition to file handles. |
|
|
27 | - take advantage of linux' eventfd (but it seems slower on smp?). |
|
|
28 | - use poll not select on posix platforms. |
|
|
29 | |
|
|
30 | 3.261 Wed Jul 1 10:11:51 CEST 2009 |
|
|
31 | - more 0S X workarounds (patch by Tokuhiro Matsuno). |
|
|
32 | |
|
|
33 | 3.26 Tue Jun 30 09:33:26 CEST 2009 |
|
|
34 | - 0S X of course claims to be posix 2008 but lacks posix_fadvise. |
|
|
35 | try to work around this horribly broken OS in a somewhat hackish |
|
|
36 | way. might help other os's too. |
|
|
37 | |
|
|
38 | 3.25 Sat Jun 27 05:18:26 CEST 2009 |
|
|
39 | - added IO::AIO::fadvise and IO::AIO::sendfile. |
|
|
40 | - (libeio) replaced quicksort+insertion sort by a tuned radix |
|
|
41 | sort + insertion sort, resulting in comparable runtime (usually |
|
|
42 | faster) to the old version, but without any ill side effects on |
|
|
43 | degenerated (for quicksort) data. |
| 12 | - correctly sort dirs by inodes only if we have perfect type |
44 | - (libeio) correctly sort dirs by inodes if we have perfect type |
| 13 | knowledge and DIRS_FIRST is used. |
45 | knowledge and DIRS_FIRST is used. |
| 14 | - speed up directory entry sorting further by some 20%. |
46 | - (libeio) internally the wrong DT_* constants were used, but no |
|
|
47 | known system has differing EIO_DT_*/DT_* values, so not an issue. |
|
|
48 | - removed a (harmless) assert that was left over on the code |
|
|
49 | but should not have been. |
|
|
50 | - use more correct types (change IO length from IV to VAL64, |
|
|
51 | which makes a difference on systems without 64 bit perls). |
| 15 | |
52 | |
| 16 | 3.23 Sat Jun 13 16:57:58 CEST 2009 |
53 | 3.23 Sat Jun 13 16:57:58 CEST 2009 |
| 17 | - fix off-by-one bug in aio_readdir that was introduced in the |
54 | - fix off-by-one bug in aio_readdir that was introduced in the |
| 18 | rushed 3.22 release. |
55 | rushed 3.22 release. |
| 19 | |
56 | |
