… | |
… | |
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. |
8 | TODO: aio_mincore? |
9 | TODO: IO::AIO::mmap? aio_mincore? |
9 | TODO: getxattr etc.? |
|
|
10 | TODO: F_DUPFD_CLOEXEC |
10 | |
11 | |
|
|
12 | - new request aio_seek. |
|
|
13 | - fix stat structure usage on windows, which caused bogus stat results. |
|
|
14 | - (libeio) make readahead emulation behave more like actual readahead by never failing. |
|
|
15 | |
|
|
16 | 4.12 Fri Dec 30 08:51:25 CET 2011 |
|
|
17 | - realpath would return a random value if the name was too long, |
|
|
18 | instead of -1. |
|
|
19 | - port to c89. |
|
|
20 | |
|
|
21 | 4.11 Mon Oct 10 00:24:11 CEST 2011 |
|
|
22 | - libeio didn't compile on !linux (Paul Howarth). |
|
|
23 | |
|
|
24 | 4.1 Sun Oct 9 10:24:11 CEST 2011 |
|
|
25 | - IO::AIO did access uninitialised memory on unsuccessful stats. |
|
|
26 | - (libeio) added syncfs syscall wrapper. |
|
|
27 | - (libeio) set thread name on linux (ps -L/Hcx, top, gdb). |
|
|
28 | - (libeio) support multiple indepenent working directories. |
|
|
29 | - applied speling corrections by Nicholas Bamber. |
|
|
30 | - tune treescan #threads and #outstanding requests a bit. |
|
|
31 | - reduce the number of file descriptors in use at any one time |
|
|
32 | in treescan, to avoid running out of them (and using too much ram). |
|
|
33 | - take advantage of working directoriy abstractions |
|
|
34 | in aio_scandir and treescan. |
|
|
35 | - reduce compiled size by ~10% by not inlining some key functionality. |
|
|
36 | - added --progress switch to treescan. |
|
|
37 | |
|
|
38 | 4.0 Mon Jul 18 05:01:10 CEST 2011 |
|
|
39 | - INCOMPATIBLE CHANGE: fork is no longer supported (indeed, it never was), |
|
|
40 | see FORK BEHAVIOUR in manpage for details. |
|
|
41 | - passes testsuite on win32 now and may actually work |
|
|
42 | (activestate 5.10.1 / MSVC6, YMMV). |
|
|
43 | - (libeio) fix a deadlock where a wakeup signal could be missed when |
|
|
44 | a timeout occured at the same time. |
|
|
45 | - (libeio) added realpath. |
|
|
46 | - (libeio) added fallocate. |
|
|
47 | - (libeio) disabling sendfile on darwin unfortunately broke the emulation. |
|
|
48 | - (libeio) do not acquire any locks when forking. |
|
|
49 | - (libeio) use fewer time() syscalls when waiting for new requests. |
|
|
50 | |
|
|
51 | 3.93 Wed Jun 29 23:44:18 CEST 2011 |
|
|
52 | - ECB.H WAS MISSING. |
|
|
53 | |
|
|
54 | 3.92 Wed Jun 29 14:45:41 CEST 2011 |
|
|
55 | - ecb.h was missing. |
|
|
56 | |
|
|
57 | 3.91 Wed Jun 29 13:24:42 CEST 2011 |
|
|
58 | - (libeio) work around a Linux (and likely FreeBSD and other |
|
|
59 | kernels) bug where sendfile would not transfer all the requested |
|
|
60 | bytes on large transfers, using a heuristic. |
|
|
61 | - buggy sendfile caused aio_move/copy to sometimes fail for big |
|
|
62 | files (fortunately it checks that the whole file has been |
|
|
63 | transferred...) |
|
|
64 | - use libecb for higher performance and higher portability. |
|
|
65 | - (libeio) disable sendfile on darwin, it's too broken. |
|
|
66 | - disable fork tests on !linux, as only linux supports mixing |
|
|
67 | pthread and fork in perl. |
|
|
68 | - document the fact that fork doesn't work anymore when using |
|
|
69 | this module, on anything !GNU/Linux. |
|
|
70 | - increase timeout in t/04_fork.t, as too many CPAN-tester setups |
|
|
71 | run on a 0.1MHz cpu. Or so. |
|
|
72 | |
|
|
73 | 3.9 Fri May 27 02:43:47 CEST 2011 |
|
|
74 | - (libeio) fix memory corruption in aio_readdirx for the flags |
|
|
75 | combination READDIR_STAT_ORDER | READDIR_DIRS_FIRST. |
|
|
76 | - add lots of newer POSIX and GNU/Linux-specific open |
|
|
77 | flags. |
|
|
78 | |
|
|
79 | 3.8 Sun Mar 27 12:25:33 CEST 2011 |
|
|
80 | - use nonstandard but maybe-working-on-bsd fork technique. |
|
|
81 | - support a max_idle value of 0. |
|
|
82 | - support setting of idle timeout value (IO::AIO::idle_timeout). |
|
|
83 | |
|
|
84 | 3.72 Fri Feb 11 04:25:38 CET 2011 |
|
|
85 | - use _POSIX_MEMLOCK_RANGE to detect mlock/munlock. |
|
|
86 | - aio_mknod always used a dev_t value of 0. |
|
|
87 | - new treescan option: --grep. |
|
|
88 | - add more S_IF macros, and major/minor/makedev "macros". |
|
|
89 | |
|
|
90 | 3.71 Thu Dec 30 08:18:46 CET 2010 |
|
|
91 | - the numerical result value passed to callbacks did not stringify |
|
|
92 | correctly, due to internal reuse and failure to reset the sv flags. |
|
|
93 | - actually test for posix_[mf]advise, as at least uClibc defines |
|
|
94 | _POSIX_ADVISORY_INFO without actually having any of the required |
|
|
95 | funcitonality. ugh. |
|
|
96 | |
|
|
97 | 3.7 Mon Nov 1 23:00:34 CET 2010 |
|
|
98 | - implement/add madvise, mmap, munmap, aio_mlockall, munlockall, |
|
|
99 | aio_mlock, munlock, msync and mtouch to @EXPORT_OK. |
|
|
100 | - document the sad state of affairs w.r.t. pthread on many bsds. |
|
|
101 | - do not enable mmap on systems without _POSIX_MAPPED_FILES |
|
|
102 | (openbsd 4.8). |
|
|
103 | - do not leak memory in IO::AIO::mmap when the scalar already |
|
|
104 | had string-data. |
|
|
105 | - add O_RDWR, O_APPEND and O_EXCL symbols and semi-document |
|
|
106 | them. |
|
|
107 | - cache the result IV, for a minor speedup in the common case. |
|
|
108 | - croak when an mmapped-scalar changes location, to detect |
|
|
109 | user-errors better. |
|
|
110 | - fix aio_readlink prototype. |
|
|
111 | |
|
|
112 | 3.65 Wed Mar 31 02:45:05 CEST 2010 |
|
|
113 | - actually use PATH_MAX instead of NAME_MAX for readlink, as to not |
|
|
114 | cut off long pathnames in aio_readlink (based on patch by |
|
|
115 | Rasmus Andersson). |
|
|
116 | - a double fork partially killed the event pipe (great testcase |
|
|
117 | by dormando). affects IO::AIO, BDB and Async::Interrupt. |
|
|
118 | - suck steve's dick and rename our symbols to avoid colliding |
|
|
119 | with nonstandard identifier spam on the broken os x pseudo-os. |
|
|
120 | affects IO::AIO and BDB. |
|
|
121 | |
|
|
122 | 3.6 Mon Jan 11 00:43:39 CET 2010 |
|
|
123 | - (libeio) more fixes for the freebsd/apple sendfile - broken by |
|
|
124 | manpage standards, but apparently correct accoridng to actual |
|
|
125 | kernel sources. |
|
|
126 | - add IO::AIO::mmap/munmap support. |
|
|
127 | - add IO::AIO::m(un)lockall support. |
|
|
128 | - clean up manpage. |
|
|
129 | |
|
|
130 | 3.5 Thu Jan 7 21:25:04 CET 2010 |
|
|
131 | - (liebio) fix freebsd sendfile (Vanilla Hsu). |
|
|
132 | - (libeio) also fix darwin sendfile that suffered from a similar bug. |
|
|
133 | - add aio_statvfs as interface to statvfs/fstatvfs. |
|
|
134 | - work around buggy symlink() on cygwin in the testsuite. |
|
|
135 | - wtf. freebsd-8.0 actually passes the testsuite! I knew |
|
|
136 | they could implement fork and semaphores one day! cheers! |
|
|
137 | |
|
|
138 | 3.4 Sat Jan 2 15:13:04 CET 2010 |
11 | - (libeio) max_poll_time was not properly converted to ticks. |
139 | - (libeio) max_poll_time was not properly converted to ticks. |
12 | - clarify cancel_subs description. |
140 | - clarify cancel_subs description. |
13 | - IO::AIO::sendfile did not actually return the return value |
141 | - IO::AIO::sendfile did not actually return the return value |
14 | from the sendfile call. |
142 | from the sendfile call. |
15 | - implement aio_msync, aio_mtouch. |
143 | - implement aio_msync, aio_mtouch. |