| 1 | Revision history for Perl extension EV |
1 | Revision history for Perl extension EV |
| 2 | |
2 | |
| 3 | Changes marked with (libev) are changes in libev, and have more |
3 | Changes marked with (libev) are changes in libev, and have more |
| 4 | documentation in the libev Changes file. |
4 | documentation in the libev Changes file. |
| 5 | |
5 | |
|
|
6 | 4.22 Sun Dec 20 02:34:39 CET 2015 |
|
|
7 | - (libev) when epoll detects unremovable fds in the fd set, rebuild |
|
|
8 | only the epoll descriptor, not the signal pipe, to avoid |
|
|
9 | SIGPIPE in ev_async_send. This doesn't solve it on fork, |
|
|
10 | so document what needs to be done in ev_loop_fork |
|
|
11 | (analyzed by Benjamin Mahler). |
|
|
12 | - (libev) remove superfluous sys/timeb.h include on win32 |
|
|
13 | (analyzed by Jason Madden). |
|
|
14 | - updated libecb. |
|
|
15 | |
|
|
16 | 4.21 Mon Jul 13 21:47:33 CEST 2015 |
|
|
17 | - allow argument in timer->again. |
|
|
18 | - document timer->remaining. |
|
|
19 | - document default repeat value for timer->set. |
|
|
20 | |
|
|
21 | 4.20 Sat Jun 20 13:07:34 CEST 2015 |
|
|
22 | - added stability canary support. |
|
|
23 | - truly rely on automatic configuration for clock_gettime and others. |
|
|
24 | unfortunately, this doesn't help with activeperl and similar perls, |
|
|
25 | which define _POSIX_TIMERS without actually implementing any of it. |
|
|
26 | - (libev, ecb) make it compile as C++ again. |
|
|
27 | - (libev) fix a potential aliasing issue when accessing watcher cbs. |
|
|
28 | |
|
|
29 | 4.18 Sat Sep 6 20:37:23 CEST 2014 |
|
|
30 | - use slightly better weay to find includes, to support multiarch |
|
|
31 | on newer perls. |
|
|
32 | |
|
|
33 | 4.17 Fri Apr 11 06:22:38 CEST 2014 |
|
|
34 | - perl5porters broke Async::Interrupt, BDB, EV, IO::AIO, OpenCL |
|
|
35 | without warning by switching the meaning of USE_SOCKETS_AS_HANDLES |
|
|
36 | in 5.18. What's so attractive about giving a shit about backwards |
|
|
37 | compatibility - I will never understand. |
|
|
38 | |
|
|
39 | 4.16 Sat Mar 8 16:49:05 CET 2014 |
|
|
40 | - make sure EV::CHECK is available at all times. |
|
|
41 | - avoid a spurious warning when perl -w overrides "no warnings". |
|
|
42 | - use perl's signal implementation on WIN32. |
|
|
43 | - ensure extern "C" function pointers are used for externally-visible |
|
|
44 | pointers. |
|
|
45 | - (libev) mark event pipe fd as cloexec after a fork (analyzed by Sami Farin). |
|
|
46 | - (ecb) support m68k, m88k and sh (patch by Miod Vallat). |
|
|
47 | - (libev) in the absence of autoconf, do not use the clock syscall |
|
|
48 | on glibc >= 2.17 (avoids the syscall AND -lrt on systems |
|
|
49 | doing clock_gettime in userspace). |
|
|
50 | - (ecb) work around memory barriers and volatile apparently both being |
|
|
51 | broken in visual studio 2008 and later (analysed and patch by Nicolas Noble). |
|
|
52 | |
|
|
53 | 4.15 Fri Mar 1 12:15:53 CET 2013 |
|
|
54 | - (libev) upgrade to 4.15 - too many changes to list. |
|
|
55 | - EV::run now returns a boolean. |
|
|
56 | - API version 5:1. |
|
|
57 | - document that cleanup watchers are not available via perl. |
|
|
58 | - cast I32 to int in error message printf. |
|
|
59 | - remove dependencies on librt and libpthreads on GNU/Linux. |
|
|
60 | |
|
|
61 | 4.11 Sat Feb 4 19:56:26 CET 2012 |
|
|
62 | - (libev) implement memory fences for (obsolete) llvm-gcc. |
|
|
63 | |
|
|
64 | 4.10 Thu Jan 19 18:54:23 CET 2012 |
|
|
65 | - (libev) fix a race where the workaround against the epoll fork |
|
|
66 | bugs caused signals to not be handled anymore. |
|
|
67 | - (libev) correct backend_fudge for most backends, and implement |
|
|
68 | a windows specific workaround to avoid looping because we call |
|
|
69 | both select and Sleep, both with different time resolutions. |
| 6 | - e_new wasn'tr declared static (causing very minor .so bloat). |
70 | - e_new wasn't declared static (causing very minor .so bloat). |
|
|
71 | - replace more old api names by new ones. |
| 7 | |
72 | |
| 8 | 4.03 Tue Jan 11 14:51:05 CET 2011 |
73 | 4.03 Tue Jan 11 14:51:05 CET 2011 |
| 9 | - do not avoid the clock_gettime call on GNU/Linux anymore, as EV |
74 | - do not avoid the clock_gettime call on GNU/Linux anymore, as EV |
| 10 | links against -lpthread anyways - as a result, EV might now take |
75 | links against -lpthread anyways - as a result, EV might now take |
| 11 | advantage of fast userspace clock_gettime implementations, |
76 | advantage of fast userspace clock_gettime implementations, |
