| 1 | Revision history for Perl extension AnyEvent. |
1 | Revision history for Perl extension AnyEvent. |
| 2 | |
2 | |
|
|
3 | TODO: inet_aton does not actually deliver ipv6 addresses |
|
|
4 | - downgrade-or-fail in AnyEvent::Handle::push_write, to |
|
|
5 | diagnose encoding failures earlier and more succinctly. |
|
|
6 | (this works around bugs in perl, throwing away encoding info |
|
|
7 | when passing scalar data to extensions). |
|
|
8 | - add more examples to AnyEvent::Socket manpage. |
|
|
9 | - upgrade internal warning set to the same as common::sense 2.03. |
|
|
10 | - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS |
|
|
11 | (John Beppu). |
|
|
12 | - add google public dns servers to fallback server set and make sure |
|
|
13 | we load-balance properly between the three sets. also add all |
|
|
14 | fallback dns servers, not just a random one, to each dns config. |
|
|
15 | |
|
|
16 | 5.21 Thu Nov 19 02:48:47 CET 2009 |
|
|
17 | - fix a problem where socket constants were called with parameters |
|
|
18 | (spotted by David Friedland). |
|
|
19 | - fork_call never use'd POSIX (reported by Daisuke Maki). |
|
|
20 | - improve perl 5.6 compatibility further (but it still won't work |
|
|
21 | unless you rip out everything but the core). |
|
|
22 | - prefer Net::DNS::Resolver over ipconfig, if installed, on win32. |
|
|
23 | uses 10MB of RAM, but doesn't open a console window. *sigh*. |
|
|
24 | |
|
|
25 | 5.202 Wed Oct 14 22:35:44 CEST 2009 |
|
|
26 | - AnyEvent::DNS would unexpectedly clobber $_ under windows |
|
|
27 | (analysed by Matthias Waldorf). |
|
|
28 | - AnyEvent::Handle::run_cmd can now pass the PID of the |
|
|
29 | newly-created process, which is much less useful than it might |
|
|
30 | sound (based on patch by Yann Kerherve). |
|
|
31 | |
|
|
32 | 5.201 Tue Sep 29 12:09:25 CEST 2009 |
|
|
33 | - AnyEvent:Handle::on_starttls/on_stoptls methods were broken |
|
|
34 | (reported by Torsten Foertsch). |
|
|
35 | - common::sense 2.0 could cause tcp_server to throw an exception |
|
|
36 | (analysed by elmex). |
|
|
37 | |
|
|
38 | 5.2 Mon Sep 14 07:04:49 CEST 2009 |
|
|
39 | - INCOMPATIBLE CHANGE: do no longer support register_read_type |
|
|
40 | and register_write_type in AnyEvent::Handle, instead support |
|
|
41 | package names (the facility was mostly abused). |
|
|
42 | - implement "packagename-as-read/write type" support in |
|
|
43 | AnyEvent::Handle. |
|
|
44 | - AnyEvent::Handle: new options "keepalive" and "oobinline". |
|
|
45 | - oobinline set by default to avoid security issues. |
|
|
46 | - the pure-perl event loop backend wrongly detected times() fallback |
|
|
47 | support (spotted by Pavel Boldin). |
|
|
48 | |
|
|
49 | 5.12 Tue Sep 1 20:26:50 CEST 2009 |
|
|
50 | - be more lenient when parsing resolv.conf files, as some people |
|
|
51 | use hashmarks as comment indicator inside directives (reported |
|
|
52 | by Michael S. Fischer). |
|
|
53 | - use same set of warnings as common::sense 2.0. |
|
|
54 | - fix a potential 32 bit overflow issue due to perl having problems |
|
|
55 | with large hex constants in the Perl backend. |
|
|
56 | - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS |
|
|
57 | by default. |
|
|
58 | - fix some prototypes. |
|
|
59 | - (optionally) require Async::Interrupt 1.02. |
|
|
60 | - replace opendns fallbacks by verizon/level3 public dns servers. |
|
|
61 | |
|
|
62 | 5.112 Fri Aug 21 13:59:04 CEST 2009 |
|
|
63 | - AnyEvent::Handle could cause hard crashes in Net::SSLeay when |
|
|
64 | using starttls with an illegal/uncreatable tls context. |
|
|
65 | - port to broken win32 perls. |
|
|
66 | |
|
|
67 | 5.111 Sun Aug 16 18:44:36 CEST 2009 |
|
|
68 | - perl errornously requires sockaddr_un structures returned by the |
|
|
69 | kernel to have a specific length. work around this bug by padding |
|
|
70 | sockaddr_un structures (found to fail on OS X, as perl only uses |
|
|
71 | the correct code for GNU/Linux, but likely to fail on other |
|
|
72 | OSes as well). |
|
|
73 | |
|
|
74 | 5.11 Wed Aug 12 17:49:37 CEST 2009 |
|
|
75 | - fixed a bug in run_cmd with input from filename. |
|
|
76 | - create a json coder object when none is given in write direction |
|
|
77 | to reduce dependency on JSON. |
|
|
78 | |
|
|
79 | 5.1 Tue Aug 11 03:17:32 CEST 2009 |
|
|
80 | - tcp_connect now reports the correct error in $!, not an unrelated |
|
|
81 | one (was broken since 4.91). |
|
|
82 | - AE::Handle did not properly clear rtimeout/wtimeout watchers |
|
|
83 | when retrying a tcp_connect. |
|
|
84 | - new functions: AnyEvent::Util::run_cmd and close_all_fds_except. |
|
|
85 | - fixed wrong documentation in AE::Socket. |
|
|
86 | |
|
|
87 | 5.01 Mon Aug 10 03:16:32 CEST 2009 |
|
|
88 | - last release broke EV child watchers :/. |
|
|
89 | |
|
|
90 | 5.0 Sun Aug 9 17:32:09 CEST 2009 |
|
|
91 | - officially introduce and document the AE API. |
|
|
92 | - lowering the timeout at runtime did not have immediate effect |
|
|
93 | in AnyEvent::Handle. |
|
|
94 | - AE::Handle now has separate and independent read and write timeout |
|
|
95 | settings. |
|
|
96 | - probe for Guard module when AnyEvent::Util is loaded, not at |
|
|
97 | runtime (Event doesn't like eval "use" from callbacks, and it avoids |
|
|
98 | exporting the probe function). |
|
|
99 | - Impl::IOAsync's io watchers did not use the filehandle emulation |
|
|
100 | code to work around it's watcher limits. |
|
|
101 | - work around IO::Async's problems with overloaded objects. |
|
|
102 | - try to disable wearnings inside IO::Async, it's just too buggy. |
|
|
103 | - do no longer use Event by default, only EV or Perl, Event is too buggy. |
|
|
104 | - hackishly provide fast AE:: interface when EV or Perl is the backend. |
|
|
105 | - inherit VERSION from AnyEvent::VERSION in most modules. |
|
|
106 | |
|
|
107 | 4.91 Thu Aug 6 15:42:45 CEST 2009 |
|
|
108 | - AE::Handle::starttls could get out of sync when the read buffer |
|
|
109 | already contains some TLS handshake. |
|
|
110 | - AE::Handle did not properly free the TLS session, causing bigger |
|
|
111 | memory leaks in Net::SSLeay than expected (reported by toaster). |
|
|
112 | - AE::Socket::tcp_connect will now artificially delay invoking |
|
|
113 | the callback to avoid returning after invoking the callback. |
|
|
114 | - convert many internal watcher uses to AE API. |
|
|
115 | |
|
|
116 | 4.9 Sat Aug 1 11:07:01 CEST 2009 |
|
|
117 | - Glib silently fails when registering a timeout with |
|
|
118 | a negative value, so avoid that. |
|
|
119 | - call condvar callback immediately when it is set |
|
|
120 | after the condition is already signalled. |
|
|
121 | - check rbuf_max condition only after trying to consume |
|
|
122 | data. |
|
|
123 | - normalise signal numbers to names when using the ae's |
|
|
124 | signal handling, but do not document this (yet). |
|
|
125 | - pure perl signal emulation did not properly set nonblocking |
|
|
126 | mode on the signal pipe on win32 (this is unlikely to have caused |
|
|
127 | any issues). |
|
|
128 | - new module: AnyEvent::Debug. |
|
|
129 | - AnyEvent::Strict now checks that a signal was specified |
|
|
130 | by name (not name or number). |
|
|
131 | - reduce memory footprint in typical cases by ~50kb by |
|
|
132 | only compiling signal code when necessary. |
|
|
133 | - add AnyEvent::Handle->rbuf_max. |
|
|
134 | - grab the AE namespace for future new API, implement stubs for |
|
|
135 | said future/alternative API. |
|
|
136 | - new function: AnyEvent::Socket::format_hostport. |
|
|
137 | |
|
|
138 | 4.881 Tue Jul 28 12:51:53 CEST 2009 |
|
|
139 | - work around a bug in local in pre-5.10 perls, causing |
|
|
140 | AnyEvent::Handle to recurse when it should not |
|
|
141 | (analysed by elmex). |
|
|
142 | |
|
|
143 | 4.88 Tue Jul 28 04:04:37 CEST 2009 |
|
|
144 | - re-bless the handle into a dummy package after calling |
|
|
145 | AnyEvent::Handle::destroy, so the user does not need to check |
|
|
146 | for errors after every push_write etc. |
|
|
147 | - do not attempt to run t/02_signals.t on obviously broken |
|
|
148 | platforms. |
|
|
149 | |
|
|
150 | 4.87 Sun Jul 26 02:06:16 CEST 2009 |
|
|
151 | - do not attempt to linger when there is no longer a valid fh. |
|
|
152 | - enforce tls mode to be either accept or connect, do not simply |
|
|
153 | segfault in Net::SSLeay. |
|
|
154 | - AnyEvent::Handle can now call tcp_connect itself (new parameters |
|
|
155 | connect, on_prepare, on_connect and on_connect_error). Updated |
|
|
156 | tutorial accordingly. |
|
|
157 | - add AnyEvent::Impl::Irssi backend. |
|
|
158 | |
|
|
159 | 4.86 Mon Jul 20 23:52:29 CEST 2009 |
|
|
160 | - since the verbose warning is not enough, explicitly document |
|
|
161 | that versions before 1.33 of Net::SSLeay are not secure. |
|
|
162 | - work around signal handling races in Event and (...) Event::Lib. |
|
|
163 | - try to align signal-race timer to full-second boundaries. |
|
|
164 | - work around Tk not liking negative timeouts. |
|
|
165 | - don't complain of different grades of environmental unfriendlyness |
|
|
166 | in IO::Async. |
|
|
167 | |
|
|
168 | 4.85 Sat Jul 18 06:16:14 CEST 2009 |
|
|
169 | - nail the signal race problem in perl once and for all |
|
|
170 | (see $AnyEvent::MAX_SIGNAL_LATENCY). |
|
|
171 | - take advantage of Async::Interrupt if it is available. |
|
|
172 | - load Time::HiRes and Guard modules on demand only. |
|
|
173 | - add optional/recommended modules section to AnyEvent |
|
|
174 | documentation. |
|
|
175 | - reduce memory usage considerably (and reduce startup penalty) |
|
|
176 | by not using "strict", "warnings" and "overload" modules. |
|
|
177 | - work around buggy windows/openbsd perls and provide EBADMSG |
|
|
178 | and EPROTO ourselves when missing. |
|
|
179 | - improve perl 5.6 compatibility of the core event loop. |
|
|
180 | - made Net::SSLeay version 1.33 a soft requirement. |
|
|
181 | |
|
|
182 | 4.83 Fri Jul 17 16:56:26 CEST 2009 |
|
|
183 | - implement AnyEvent::Socket::getprotobyname. |
|
|
184 | - AnyEvent::CondVar's will now detect recursive blocking |
|
|
185 | waits and will croak, as too many people fall into |
|
|
186 | this trap. |
|
|
187 | - AnyEvent::Handle will now call ->destroy on itself after |
|
|
188 | executing the on_error callback, instead of doing some |
|
|
189 | half-baked internal shutdown, for fatal errors. |
|
|
190 | - clarify on_eof behaviour w.r.t. the read queue and |
|
|
191 | on_read callbacks. |
|
|
192 | - ignore some possible spurious wake-ups in tcp_connect. |
|
|
193 | |
|
|
194 | 4.82 Sat Jul 11 00:34:55 CEST 2009 |
|
|
195 | - POE and Event backends didn't accept some callable objects as |
|
|
196 | callbacks. |
|
|
197 | - use Config module instead of POSIX module to detect signal names |
|
|
198 | in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX |
|
|
199 | module doesn't even have all POSIX signals :/. |
|
|
200 | - use more workarounds around the many refcnt/corruption bugs in |
|
|
201 | Event::Lib. |
|
|
202 | - work around a race condition in perl's select, causing t/03_child.t |
|
|
203 | to rarely fail. |
|
|
204 | |
|
|
205 | 4.81 Thu Jul 9 10:30:30 CEST 2009 |
|
|
206 | - AnyEvent::Handle didn't properly diagnose write errors |
|
|
207 | (it expected -1 from syswrite, how lame... :). |
|
|
208 | - support file descriptors in addition to file handles |
|
|
209 | in AnyEvent->io. |
|
|
210 | - new env variables: PERL_ANYEVENT_RESOLV_CONF, |
|
|
211 | PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE |
|
|
212 | and PERL_ANYEVENT_CA_PATH. |
|
|
213 | - provide a sensible synopsis section for AnyEvent::TLS. |
|
|
214 | - add a "supported backends" section to the manpage. |
|
|
215 | - added simple io watcher test to testsuite, using a |
|
|
216 | portable_socketpair. |
|
|
217 | - tried to improve the stability of the Event::Lib backend, |
|
|
218 | YMMV. |
|
|
219 | |
|
|
220 | 4.8 Mon Jul 6 23:45:16 CEST 2009 |
|
|
221 | - AnyEvent::DNS did not properly follow CNAME records with |
|
|
222 | uppercase targets. |
|
|
223 | - AnyEvent::DNS would errornously return AAAA records |
|
|
224 | with v4 mapped addresses (a faulty record) as ipv4 addresses, |
|
|
225 | causing AnyEvent::Socket to throw an exception. |
|
|
226 | - added new module AnyEvent::TLS for easier SSL/TLS context |
|
|
227 | creation, with many options including hostname verification, |
|
|
228 | secure default configuration, lots of documentation and, |
|
|
229 | predefined diffie-hellman keys for perfect forward security |
|
|
230 | and much more. get it while it's still fresh! |
|
|
231 | - use AnyEvent::TLS in AnyEvent::Handle for context management. |
|
|
232 | - load AnyEvent::Handle only on demand in AnyEvent::DNS, |
|
|
233 | so AnyEvent::Socket users have smaller memory footprint |
|
|
234 | in the common case. |
|
|
235 | - add AnyEvent::Handle->push_shutdown method. |
|
|
236 | - add an additional error message parameter to AnyEvent::Handle's |
|
|
237 | on_error callback (for TLS, $! is still available). |
|
|
238 | - add AnyEvent::Handle on_starttls/on_stoptls callbacks. |
|
|
239 | - make AnyEvent::Handle more robust against early conenction |
|
|
240 | failures (during new), and return C<undef> in that case |
|
|
241 | from the constructor. |
|
|
242 | - AnyEvent::Handle will now try to load only JSON::XS first, |
|
|
243 | then fall back to JSON. |
|
|
244 | - format_ipv4/format_ipv6 are now exported by default, for symmetry, |
|
|
245 | and because it was documented that way. |
|
|
246 | |
|
|
247 | 4.451 Fri Jul 3 00:28:58 CEST 2009 |
|
|
248 | - do not clear rbuf when shutting down an AnyEvent::Handle |
|
|
249 | object - doing so breaks AnyEvent::HTTP. |
|
|
250 | |
|
|
251 | 4.45 Mon Jun 29 22:59:26 CEST 2009 |
|
|
252 | - a write error could cause AnyEvent::Handle to create |
|
|
253 | an I/O watcher with an undefined $fh. |
|
|
254 | - special-case mapped ipv4 addresses in both |
|
|
255 | AnyEvent::Socket::format_address and parse_address, |
|
|
256 | to treat them just like ipv4 addresses. |
|
|
257 | - updated and overhauled the AnyEvent::Intro doc. |
|
|
258 | - implement AnyEvent::Socket::format_ipv4/ipv6. |
|
|
259 | - slightly speed up portable_pipe/socketpair. |
|
|
260 | - expand condvar begin/end documentation. |
|
|
261 | |
|
|
262 | 4.42 Fri Jun 26 08:32:18 CEST 2009 |
|
|
263 | - preliminary and neccesarily incomplete support for IO::Async. |
|
|
264 | - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in |
|
|
265 | case it was set to IGNORE, to ensure we can catch child statuses |
|
|
266 | even when the calling env acts stupidly. |
|
|
267 | - updated benchmarks with IO::Async, which performs very well. |
|
|
268 | |
|
|
269 | 4.412 Wed Jun 24 01:35:57 CEST 2009 |
|
|
270 | - support an "untaint" attribute for AnyEvent::DNS |
|
|
271 | and set it on the default resolver. |
|
|
272 | - implement and document AnyEvent::Impl::Perl::loop. |
|
|
273 | - remove all anyevent-env variables from %ENV |
|
|
274 | when running in tainted mode. |
|
|
275 | - mention and extend the IO::Lambda benchmark. |
|
|
276 | |
|
|
277 | 4.411 Sun Jun 7 18:48:02 CEST 2009 |
|
|
278 | - do not try to use F_SETFD on windows. |
|
|
279 | |
|
|
280 | 4.41 Thu May 14 06:40:11 CEST 2009 |
|
|
281 | - work around issues in older perls (5.8.0?) when |
|
|
282 | a signal handler is deleted from the %SIG hash. |
|
|
283 | - use POSIX::_exit in child test, to avoid running |
|
|
284 | destructors. |
|
|
285 | - speed up CHLD handling by relying on SIGCHLD being |
|
|
286 | synchronously delivered, even when we roll our own |
|
|
287 | implementation. |
|
|
288 | - AnyEvent::DNS: add the "dname" resource record name |
|
|
289 | for cosmetic reasons. |
|
|
290 | |
|
|
291 | 4.4 Sun Apr 26 20:12:33 CEST 2009 |
|
|
292 | - implemented idle watchers, where applicable. |
|
|
293 | - AnyEvent->time died when Event backend was in use. |
|
|
294 | - fix a memleak in the Tk backend. |
|
|
295 | - sped up Tk timer handling. |
|
|
296 | - clip negative "after" values to 0 in AnyEvent::Impl::Event |
|
|
297 | to avoid spamming out warning messages. |
|
|
298 | - fix Qt timers without interval. |
|
|
299 | - avoid Qt zero-timeout special case, in old libqt's. |
|
|
300 | |
|
|
301 | 4.352 Mon Apr 20 16:31:11 CEST 2009 |
|
|
302 | - fix AnyEvent::Strict error messages for child watchers. |
|
|
303 | - fix/update Makefile.PL "recommends" versions. |
|
|
304 | - add AnyEvent->now_update. |
|
|
305 | |
|
|
306 | 4.351 Sat Apr 11 07:56:14 CEST 2009 |
|
|
307 | - actually make the signal pipe work under win32. |
|
|
308 | - localise $! in signal handler to avoid changing $!, although |
|
|
309 | perl itself does not seem to save/restore errno either. |
|
|
310 | - set the cloexec flag on the signal pipe (normally set by perl too). |
|
|
311 | |
|
|
312 | 4.35 Fri Mar 27 11:48:20 CET 2009 |
|
|
313 | - event models relying on AnyEvent's signal watcher emulation did |
|
|
314 | invoke the callback asynchronously, contradicting documentation |
|
|
315 | and causing signals to get lost (this includes AnyEvent's own |
|
|
316 | event loop). AnyEvent now uses the standard pipe trick to make |
|
|
317 | callback execution synchronous to the event loop. |
|
|
318 | - AnyEvent::Handle didn't free TLS context data on DESTROY |
|
|
319 | (patch by Pavel Shaydo). |
|
|
320 | - work around the ever-incompatibly-changing API of MakeMaker. |
|
|
321 | - document that changing global variables without restoring them |
|
|
322 | is a bad idea in Perl (noted by Adam Rosenstein). |
|
|
323 | - AnyEvent::Strict now barfs if ->io is passed a file. |
|
|
324 | |
|
|
325 | 4.34 Thu Feb 12 18:32:45 CET 2009 |
|
|
326 | - separately buffer TLS read data, as otherwise the read queue could |
|
|
327 | deadlock as receiving data is not expected while draining the |
|
|
328 | read queue (which cna only happen with TLS). |
| 3 | - raise EBADMSG error on JSON decoding errors. |
329 | - raise EBADMSG error on JSON decoding errors. |
|
|
330 | - fix some minor manpage bugs (reported by Maximilian Gaß). |
|
|
331 | - speed up select bitmask parsing quite a bit in the pure perl |
|
|
332 | backend. |
|
|
333 | - use CORE::select instead of just select in AnyEvent::Impl::Perl. |
| 4 | |
334 | |
| 5 | 4.331 Tue Jan 6 21:07:25 CET 2009 |
335 | 4.331 Tue Jan 6 21:07:25 CET 2009 |
| 6 | - socketpair fails on many vista machines because vista has |
336 | - socketpair fails on many vista machines because vista has |
| 7 | completely broken accept/getpeername and getsockname functions, |
337 | completely broken accept/getpeername and getsockname functions, |
| 8 | so we provide our own socketpair emulation that kind of works |
338 | so we provide our own socketpair emulation that kind of works |
| … | |
… | |
| 41 | - warn about not loaidng AnyEvent::Impl::POE early enough |
371 | - warn about not loaidng AnyEvent::Impl::POE early enough |
| 42 | (patch by Adam Rosenstein). |
372 | (patch by Adam Rosenstein). |
| 43 | - fixed a great number of bugs and corner cases in AnyEvent::Handle: |
373 | - fixed a great number of bugs and corner cases in AnyEvent::Handle: |
| 44 | - fix a bug in where in SSL connect mode, the |
374 | - fix a bug in where in SSL connect mode, the |
| 45 | client would first wait for some data by the server and |
375 | client would first wait for some data by the server and |
| 46 | otherwise hang. (reported and analyzed in an absolutely |
376 | otherwise hang. (reported and analysed in an absolutely |
| 47 | exemplary manner by Adam Rosenstein). |
377 | exemplary manner by Adam Rosenstein). |
| 48 | - fix a bug in where SSL EOF would not be treated |
378 | - fix a bug in where SSL EOF would not be treated |
| 49 | as stream EOF, putting the connection into a hung state. |
379 | as stream EOF, putting the connection into a hung state. |
| 50 | (reported and analyzed in an absolutely exemplary manner by Adam |
380 | (reported and analysed in an absolutely exemplary manner by Adam |
| 51 | Rosenstein). |
381 | Rosenstein). |
| 52 | - fix a potential segfault when the TLS context |
382 | - fix a potential segfault when the TLS context |
| 53 | would go missing in a read callback (e.g. due to stoptls) - |
383 | would go missing in a read callback (e.g. due to stoptls) - |
| 54 | Net::SSLeay of course makes no type checking whatsoever. |
384 | Net::SSLeay of course makes no type checking whatsoever. |
| 55 | - AnyEvent::Handle will not stop the read watcher in TLS mode. |
385 | - AnyEvent::Handle will not stop the read watcher in TLS mode. |
