| 1 | Revision history for Perl extension AnyEvent. |
1 | Revision history for Perl extension AnyEvent. |
| 2 | |
2 | |
| 3 | TODO: release a:i 1.0 |
3 | - fixed a bug in run_cmd with input from filename. |
|
|
4 | |
|
|
5 | 5.1 Tue Aug 11 03:17:32 CEST 2009 |
|
|
6 | - tcp_connect now reports the correct error in $!, not an unrelated |
|
|
7 | one (was broken since 4.91). |
|
|
8 | - AE::Handle did not properly clear rtimeout/wtimeout watchers |
|
|
9 | when retrying a tcp_connect. |
|
|
10 | - new functions: AnyEvent::Util::run_cmd and close_all_fds_except. |
|
|
11 | - fixed wrong documentation in AE::Socket. |
|
|
12 | |
|
|
13 | 5.01 Mon Aug 10 03:16:32 CEST 2009 |
|
|
14 | - last release broke EV child watchers :/. |
|
|
15 | |
|
|
16 | 5.0 Sun Aug 9 17:32:09 CEST 2009 |
|
|
17 | - officially introduce and document the AE API. |
|
|
18 | - lowering the timeout at runtime did not have immediate effect |
|
|
19 | in AnyEvent::Handle. |
|
|
20 | - AE::Handle now has separate and independent read and write timeout |
|
|
21 | settings. |
|
|
22 | - probe for Guard module when AnyEvent::Util is loaded, not at |
|
|
23 | runtime (Event doesn't like eval "use" from callbacks, and it avoids |
|
|
24 | exporting the probe function). |
|
|
25 | - Impl::IOAsync's io watchers did not use the filehandle emulation |
|
|
26 | code to work around it's watcher limits. |
|
|
27 | - work around IO::Async's problems with overloaded objects. |
|
|
28 | - try to disable wearnings inside IO::Async, it's just too buggy. |
|
|
29 | - do no longer use Event by default, only EV or Perl, Event is too buggy. |
|
|
30 | - hackishly provide fast AE:: interface when EV or Perl is the backend. |
|
|
31 | - inherit VERSION from AnyEvent::VERSION in most modules. |
|
|
32 | |
|
|
33 | 4.91 Thu Aug 6 15:42:45 CEST 2009 |
|
|
34 | - AE::Handle::starttls could get out of sync when the read buffer |
|
|
35 | already contains some TLS handshake. |
|
|
36 | - AE::Handle did not properly free the TLS session, causing bigger |
|
|
37 | memory leaks in Net::SSLeay than expected (reported by toaster). |
|
|
38 | - AE::Socket::tcp_connect will now artificially delay invoking |
|
|
39 | the callback to avoid returning after invoking the callback. |
|
|
40 | - convert many internal watcher uses to AE API. |
|
|
41 | |
|
|
42 | 4.9 Sat Aug 1 11:07:01 CEST 2009 |
|
|
43 | - Glib silently fails when registering a timeout with |
|
|
44 | a negative value, so avoid that. |
|
|
45 | - call condvar callback immediately when it is set |
|
|
46 | after the condition is already signalled. |
|
|
47 | - check rbuf_max condition only after trying to consume |
|
|
48 | data. |
| 4 | - normalise signal numbers to names when using the ae's |
49 | - normalise signal numbers to names when using the ae's |
| 5 | signal handling, but do not document this (yet). |
50 | signal handling, but do not document this (yet). |
|
|
51 | - pure perl signal emulation did not properly set nonblocking |
|
|
52 | mode on the signal pipe on win32 (this is unlikely to have caused |
|
|
53 | any issues). |
|
|
54 | - new module: AnyEvent::Debug. |
|
|
55 | - AnyEvent::Strict now checks that a signal was specified |
|
|
56 | by name (not name or number). |
|
|
57 | - reduce memory footprint in typical cases by ~50kb by |
|
|
58 | only compiling signal code when necessary. |
|
|
59 | - add AnyEvent::Handle->rbuf_max. |
|
|
60 | - grab the AE namespace for future new API, implement stubs for |
|
|
61 | said future/alternative API. |
|
|
62 | - new function: AnyEvent::Socket::format_hostport. |
| 6 | |
63 | |
| 7 | 4.881 Tue Jul 28 12:51:53 CEST 2009 |
64 | 4.881 Tue Jul 28 12:51:53 CEST 2009 |
| 8 | - work around a bug in local in pre-5.10 perls, causing |
65 | - work around a bug in local in pre-5.10 perls, causing |
| 9 | AnyEvent::Handle to recurse when it should not |
66 | AnyEvent::Handle to recurse when it should not |
| 10 | (analyzed by elmex). |
67 | (analyzed by elmex). |
