| 1 | Revision history for Perl extension AnyEvent. |
1 | Revision history for Perl extension AnyEvent. |
|
|
2 | |
|
|
3 | - downgrade-or-fail in AnyEvent::Handle::push_write, to |
|
|
4 | diagnose encoding failures earlier and more succinctly. |
|
|
5 | (this works around bugs in perl, throwing away encoding info |
|
|
6 | when passing scalar data to extensions). |
|
|
7 | |
|
|
8 | 5.21 Thu Nov 19 02:48:47 CET 2009 |
|
|
9 | - fix a problem where socket constants were called with parameters |
|
|
10 | (spotted by David Friedland). |
|
|
11 | - fork_call never use'd POSIX (reported by Daisuke Maki). |
|
|
12 | - improve perl 5.6 compatibility further (but it still won't work |
|
|
13 | unless you rip out everything but the core). |
|
|
14 | - prefer Net::DNS::Resolver over ipconfig, if installed, on win32. |
|
|
15 | uses 10MB of RAM, but doesn't open a console window. *sigh*. |
|
|
16 | |
|
|
17 | 5.202 Wed Oct 14 22:35:44 CEST 2009 |
|
|
18 | - AnyEvent::DNS would unexpectedly clobber $_ under windows |
|
|
19 | (analysed by Matthias Waldorf). |
|
|
20 | - AnyEvent::Handle::run_cmd can now pass the PID of the |
|
|
21 | newly-created process, which is much less useful than it might |
|
|
22 | sound (based on patch by Yann Kerherve). |
|
|
23 | |
|
|
24 | 5.201 Tue Sep 29 12:09:25 CEST 2009 |
|
|
25 | - AnyEvent:Handle::on_starttls/on_stoptls methods were broken |
|
|
26 | (reported by Torsten Foertsch). |
|
|
27 | - common::sense 2.0 could cause tcp_server to throw an exception |
|
|
28 | (analysed by elmex). |
|
|
29 | |
|
|
30 | 5.2 Mon Sep 14 07:04:49 CEST 2009 |
|
|
31 | - INCOMPATIBLE CHANGE: do no longer support register_read_type |
|
|
32 | and register_write_type in AnyEvent::Handle, instead support |
|
|
33 | package names (the facility was mostly abused). |
|
|
34 | - implement "packagename-as-read/write type" support in |
|
|
35 | AnyEvent::Handle. |
|
|
36 | - AnyEvent::Handle: new options "keepalive" and "oobinline". |
|
|
37 | - oobinline set by default to avoid security issues. |
|
|
38 | - the pure-perl event loop backend wrongly detected times() fallback |
|
|
39 | support (spotted by Pavel Boldin). |
|
|
40 | |
|
|
41 | 5.12 Tue Sep 1 20:26:50 CEST 2009 |
|
|
42 | - be more lenient when parsing resolv.conf files, as some people |
|
|
43 | use hashmarks as comment indicator inside directives (reported |
|
|
44 | by Michael S. Fischer). |
|
|
45 | - use same set of warnings as common::sense 2.0. |
|
|
46 | - fix a potential 32 bit overflow issue due to perl having problems |
|
|
47 | with large hex constants in the Perl backend. |
|
|
48 | - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS |
|
|
49 | by default. |
|
|
50 | - fix some prototypes. |
|
|
51 | - (optionally) require Async::Interrupt 1.02. |
|
|
52 | - replace opendns fallbacks by verizon/level3 public dns servers. |
|
|
53 | |
|
|
54 | 5.112 Fri Aug 21 13:59:04 CEST 2009 |
|
|
55 | - AnyEvent::Handle could cause hard crashes in Net::SSLeay when |
|
|
56 | using starttls with an illegal/uncreatable tls context. |
|
|
57 | - port to broken win32 perls. |
|
|
58 | |
|
|
59 | 5.111 Sun Aug 16 18:44:36 CEST 2009 |
|
|
60 | - perl errornously requires sockaddr_un structures returned by the |
|
|
61 | kernel to have a specific length. work around this bug by padding |
|
|
62 | sockaddr_un structures (found to fail on OS X, as perl only uses |
|
|
63 | the correct code for GNU/Linux, but likely to fail on other |
|
|
64 | OSes as well). |
|
|
65 | |
|
|
66 | 5.11 Wed Aug 12 17:49:37 CEST 2009 |
|
|
67 | - fixed a bug in run_cmd with input from filename. |
|
|
68 | - create a json coder object when none is given in write direction |
|
|
69 | to reduce dependency on JSON. |
|
|
70 | |
|
|
71 | 5.1 Tue Aug 11 03:17:32 CEST 2009 |
|
|
72 | - tcp_connect now reports the correct error in $!, not an unrelated |
|
|
73 | one (was broken since 4.91). |
|
|
74 | - AE::Handle did not properly clear rtimeout/wtimeout watchers |
|
|
75 | when retrying a tcp_connect. |
|
|
76 | - new functions: AnyEvent::Util::run_cmd and close_all_fds_except. |
|
|
77 | - fixed wrong documentation in AE::Socket. |
|
|
78 | |
|
|
79 | 5.01 Mon Aug 10 03:16:32 CEST 2009 |
|
|
80 | - last release broke EV child watchers :/. |
|
|
81 | |
|
|
82 | 5.0 Sun Aug 9 17:32:09 CEST 2009 |
|
|
83 | - officially introduce and document the AE API. |
|
|
84 | - lowering the timeout at runtime did not have immediate effect |
|
|
85 | in AnyEvent::Handle. |
|
|
86 | - AE::Handle now has separate and independent read and write timeout |
|
|
87 | settings. |
|
|
88 | - probe for Guard module when AnyEvent::Util is loaded, not at |
|
|
89 | runtime (Event doesn't like eval "use" from callbacks, and it avoids |
|
|
90 | exporting the probe function). |
|
|
91 | - Impl::IOAsync's io watchers did not use the filehandle emulation |
|
|
92 | code to work around it's watcher limits. |
|
|
93 | - work around IO::Async's problems with overloaded objects. |
|
|
94 | - try to disable wearnings inside IO::Async, it's just too buggy. |
|
|
95 | - do no longer use Event by default, only EV or Perl, Event is too buggy. |
|
|
96 | - hackishly provide fast AE:: interface when EV or Perl is the backend. |
|
|
97 | - inherit VERSION from AnyEvent::VERSION in most modules. |
|
|
98 | |
|
|
99 | 4.91 Thu Aug 6 15:42:45 CEST 2009 |
|
|
100 | - AE::Handle::starttls could get out of sync when the read buffer |
|
|
101 | already contains some TLS handshake. |
|
|
102 | - AE::Handle did not properly free the TLS session, causing bigger |
|
|
103 | memory leaks in Net::SSLeay than expected (reported by toaster). |
|
|
104 | - AE::Socket::tcp_connect will now artificially delay invoking |
|
|
105 | the callback to avoid returning after invoking the callback. |
|
|
106 | - convert many internal watcher uses to AE API. |
|
|
107 | |
|
|
108 | 4.9 Sat Aug 1 11:07:01 CEST 2009 |
|
|
109 | - Glib silently fails when registering a timeout with |
|
|
110 | a negative value, so avoid that. |
|
|
111 | - call condvar callback immediately when it is set |
|
|
112 | after the condition is already signalled. |
|
|
113 | - check rbuf_max condition only after trying to consume |
|
|
114 | data. |
|
|
115 | - normalise signal numbers to names when using the ae's |
|
|
116 | signal handling, but do not document this (yet). |
|
|
117 | - pure perl signal emulation did not properly set nonblocking |
|
|
118 | mode on the signal pipe on win32 (this is unlikely to have caused |
|
|
119 | any issues). |
|
|
120 | - new module: AnyEvent::Debug. |
|
|
121 | - AnyEvent::Strict now checks that a signal was specified |
|
|
122 | by name (not name or number). |
|
|
123 | - reduce memory footprint in typical cases by ~50kb by |
|
|
124 | only compiling signal code when necessary. |
|
|
125 | - add AnyEvent::Handle->rbuf_max. |
|
|
126 | - grab the AE namespace for future new API, implement stubs for |
|
|
127 | said future/alternative API. |
|
|
128 | - new function: AnyEvent::Socket::format_hostport. |
|
|
129 | |
|
|
130 | 4.881 Tue Jul 28 12:51:53 CEST 2009 |
|
|
131 | - work around a bug in local in pre-5.10 perls, causing |
|
|
132 | AnyEvent::Handle to recurse when it should not |
|
|
133 | (analysed by elmex). |
|
|
134 | |
|
|
135 | 4.88 Tue Jul 28 04:04:37 CEST 2009 |
|
|
136 | - re-bless the handle into a dummy package after calling |
|
|
137 | AnyEvent::Handle::destroy, so the user does not need to check |
|
|
138 | for errors after every push_write etc. |
|
|
139 | - do not attempt to run t/02_signals.t on obviously broken |
|
|
140 | platforms. |
|
|
141 | |
|
|
142 | 4.87 Sun Jul 26 02:06:16 CEST 2009 |
|
|
143 | - do not attempt to linger when there is no longer a valid fh. |
|
|
144 | - enforce tls mode to be either accept or connect, do not simply |
|
|
145 | segfault in Net::SSLeay. |
|
|
146 | - AnyEvent::Handle can now call tcp_connect itself (new parameters |
|
|
147 | connect, on_prepare, on_connect and on_connect_error). Updated |
|
|
148 | tutorial accordingly. |
|
|
149 | - add AnyEvent::Impl::Irssi backend. |
|
|
150 | |
|
|
151 | 4.86 Mon Jul 20 23:52:29 CEST 2009 |
|
|
152 | - since the verbose warning is not enough, explicitly document |
|
|
153 | that versions before 1.33 of Net::SSLeay are not secure. |
|
|
154 | - work around signal handling races in Event and (...) Event::Lib. |
|
|
155 | - try to align signal-race timer to full-second boundaries. |
|
|
156 | - work around Tk not liking negative timeouts. |
|
|
157 | - don't complain of different grades of environmental unfriendlyness |
|
|
158 | in IO::Async. |
|
|
159 | |
|
|
160 | 4.85 Sat Jul 18 06:16:14 CEST 2009 |
|
|
161 | - nail the signal race problem in perl once and for all |
|
|
162 | (see $AnyEvent::MAX_SIGNAL_LATENCY). |
|
|
163 | - take advantage of Async::Interrupt if it is available. |
|
|
164 | - load Time::HiRes and Guard modules on demand only. |
|
|
165 | - add optional/recommended modules section to AnyEvent |
|
|
166 | documentation. |
|
|
167 | - reduce memory usage considerably (and reduce startup penalty) |
|
|
168 | by not using "strict", "warnings" and "overload" modules. |
|
|
169 | - work around buggy windows/openbsd perls and provide EBADMSG |
|
|
170 | and EPROTO ourselves when missing. |
|
|
171 | - improve perl 5.6 compatibility of the core event loop. |
|
|
172 | - made Net::SSLeay version 1.33 a soft requirement. |
|
|
173 | |
|
|
174 | 4.83 Fri Jul 17 16:56:26 CEST 2009 |
|
|
175 | - implement AnyEvent::Socket::getprotobyname. |
|
|
176 | - AnyEvent::CondVar's will now detect recursive blocking |
|
|
177 | waits and will croak, as too many people fall into |
|
|
178 | this trap. |
|
|
179 | - AnyEvent::Handle will now call ->destroy on itself after |
|
|
180 | executing the on_error callback, instead of doing some |
|
|
181 | half-baked internal shutdown, for fatal errors. |
|
|
182 | - clarify on_eof behaviour w.r.t. the read queue and |
|
|
183 | on_read callbacks. |
|
|
184 | - ignore some possible spurious wake-ups in tcp_connect. |
|
|
185 | |
|
|
186 | 4.82 Sat Jul 11 00:34:55 CEST 2009 |
|
|
187 | - POE and Event backends didn't accept some callable objects as |
|
|
188 | callbacks. |
|
|
189 | - use Config module instead of POSIX module to detect signal names |
|
|
190 | in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX |
|
|
191 | module doesn't even have all POSIX signals :/. |
|
|
192 | - use more workarounds around the many refcnt/corruption bugs in |
|
|
193 | Event::Lib. |
|
|
194 | - work around a race condition in perl's select, causing t/03_child.t |
|
|
195 | to rarely fail. |
| 2 | |
196 | |
| 3 | 4.81 Thu Jul 9 10:30:30 CEST 2009 |
197 | 4.81 Thu Jul 9 10:30:30 CEST 2009 |
| 4 | - AnyEvent::Handle didn't properly diagnose write errors |
198 | - AnyEvent::Handle didn't properly diagnose write errors |
| 5 | (it expected -1 from syswrite, how lame... :). |
199 | (it expected -1 from syswrite, how lame... :). |
| 6 | - support file descriptors in addition to file handles |
200 | - support file descriptors in addition to file handles |
| … | |
… | |
| 169 | - warn about not loaidng AnyEvent::Impl::POE early enough |
363 | - warn about not loaidng AnyEvent::Impl::POE early enough |
| 170 | (patch by Adam Rosenstein). |
364 | (patch by Adam Rosenstein). |
| 171 | - fixed a great number of bugs and corner cases in AnyEvent::Handle: |
365 | - fixed a great number of bugs and corner cases in AnyEvent::Handle: |
| 172 | - fix a bug in where in SSL connect mode, the |
366 | - fix a bug in where in SSL connect mode, the |
| 173 | client would first wait for some data by the server and |
367 | client would first wait for some data by the server and |
| 174 | otherwise hang. (reported and analyzed in an absolutely |
368 | otherwise hang. (reported and analysed in an absolutely |
| 175 | exemplary manner by Adam Rosenstein). |
369 | exemplary manner by Adam Rosenstein). |
| 176 | - fix a bug in where SSL EOF would not be treated |
370 | - fix a bug in where SSL EOF would not be treated |
| 177 | as stream EOF, putting the connection into a hung state. |
371 | as stream EOF, putting the connection into a hung state. |
| 178 | (reported and analyzed in an absolutely exemplary manner by Adam |
372 | (reported and analysed in an absolutely exemplary manner by Adam |
| 179 | Rosenstein). |
373 | Rosenstein). |
| 180 | - fix a potential segfault when the TLS context |
374 | - fix a potential segfault when the TLS context |
| 181 | would go missing in a read callback (e.g. due to stoptls) - |
375 | would go missing in a read callback (e.g. due to stoptls) - |
| 182 | Net::SSLeay of course makes no type checking whatsoever. |
376 | Net::SSLeay of course makes no type checking whatsoever. |
| 183 | - AnyEvent::Handle will not stop the read watcher in TLS mode. |
377 | - AnyEvent::Handle will not stop the read watcher in TLS mode. |
