1 | Revision history for Perl extension Async::Interrupt. |
1 | Revision history for Perl extension Async::Interrupt. |
2 | |
2 | |
|
|
3 | 1.22 Thu Jan 18 17:43:14 CET 2018 |
|
|
4 | - further USE_SOCKETS_AS_HANDLES fixes (Z59). |
|
|
5 | |
|
|
6 | 1.21 Wed Feb 11 20:31:01 CET 2015 |
|
|
7 | - upgrade ecb.h for C11 compatibility. |
|
|
8 | |
|
|
9 | 1.2 Fri Apr 11 06:22:38 CEST 2014 |
|
|
10 | - perl5porters broke Async::Interrupt, BDB, EV, IO::AIO, OpenCL |
|
|
11 | without warning by switching the meaning of USE_SOCKETS_AS_HANDLES |
|
|
12 | in 5.18. What's so attractive about giving a shit about backwards |
|
|
13 | compatibility - I will never understand. |
|
|
14 | |
|
|
15 | 1.1 Wed Apr 25 00:46:08 CEST 2012 |
|
|
16 | - new $async->handle method. |
|
|
17 | - new $async->pipe_drain method. |
|
|
18 | - use memory fences for !x86, and x86 future |
|
|
19 | proofing. use libecb for implementation. |
|
|
20 | |
|
|
21 | 1.05 Sat May 15 02:06:33 CEST 2010 |
|
|
22 | - implement $epipe->signal_func method. |
|
|
23 | |
|
|
24 | 1.04 Wed Mar 31 02:46:49 CEST 2010 |
|
|
25 | - a double fork partially killed the event pipe (great testcase |
|
|
26 | by dormando). affects IO::AIO, BDB and Async::Interrupt. |
|
|
27 | |
|
|
28 | 1.03 Tue Nov 24 14:31:10 CET 2009 |
|
|
29 | - port to loser platform. |
|
|
30 | |
|
|
31 | 1.02 Tue Sep 1 18:41:09 CEST 2009 |
|
|
32 | - prototypes for sig2name/sig2num were missing. |
|
|
33 | |
|
|
34 | 1.01 Wed Aug 5 13:50:59 CEST 2009 |
|
|
35 | - evpipe->wait did immediately return on !windows. |
|
|
36 | |
|
|
37 | 1.0 Thu Jul 30 05:58:55 CEST 2009 |
|
|
38 | - implement signal_hysteresis. |
|
|
39 | - implement scope_block_func. |
|
|
40 | - implement sig2name/sig2num convenience functions. |
|
|
41 | |
|
|
42 | 0.6 Sat Jul 18 07:09:33 CEST 2009 |
|
|
43 | - add autodrain setting. |
|
|
44 | - added Async::Interrupt::EventPipe class. |
|
|
45 | - fixed a potential race issue by removing the internal |
|
|
46 | epipe state. |
|
|
47 | - destroying asyncs inside their handler callback could lead |
|
|
48 | to segfaults. |
|
|
49 | - destroying an interrupt object did not properly remove it |
|
|
50 | from the asyncs list (classical case of side-effect-inside-assert). |
|
|
51 | |
|
|
52 | 0.501 Fri Jul 17 16:58:51 CEST 2009 |
3 | - do no longer errornously do operations on fd 0 if no pipe |
53 | - do no longer errornously do operations on fd 0 if no pipe |
4 | is associated with the interrupt object. |
54 | is associated with the interrupt object. |
5 | |
55 | |
6 | 0.5 Fri Jul 17 03:53:21 CEST 2009 |
56 | 0.5 Fri Jul 17 03:53:21 CEST 2009 |
7 | - INCOMPATIBLE CHANGE: signalling the value 0 is no longer allowed. |
57 | - INCOMPATIBLE CHANGE: signalling the value 0 is no longer allowed. |