| 1 | Revision history for Perl extension AnyEvent. |
1 | Revision history for Perl extension AnyEvent. |
| 2 | |
2 | |
| 3 | TODO: implement condvars... |
3 | 5.25 Sat Mar 13 00:23:14 CET 2010 |
|
|
4 | - fix a race condition in AnyEvent::Handle that would cause |
|
|
5 | a "bio_read: unsupported method" error (in _another_ |
|
|
6 | ssl connection) after user code threw an exception. |
|
|
7 | - speed up AnyEvent::Socket::format_address and ::format_ipv6. |
|
|
8 | - the AnyEvent::Util::fh_nonblocking function worked only |
|
|
9 | by lucky accident on win32. |
|
|
10 | - smaller and faster AnyEvent::Util::fh_nonblocking. |
|
|
11 | - when the (required!) Time::HiRes module is missing, AnyEvent |
|
|
12 | did not fall back to built-in time properly. |
|
|
13 | - do not load Fcntl at runtime, saving memory and loading time. |
|
|
14 | - precompile a number of constants and use them instead of runtime |
|
|
15 | detection and eval. |
|
|
16 | - free detection code after detection and similar memory optimisations. |
|
|
17 | - Perl backend timer interval best effort drift has been improved |
|
|
18 | (ame algorithm as EV). |
|
|
19 | |
|
|
20 | 5.24 Tue Jan 5 11:39:43 CET 2010 |
|
|
21 | - cygwin never reports errors from failed connects. 1.5 just gives |
|
|
22 | you continous readyness and EAGAIN, 1.7 is even more broken and |
|
|
23 | just hangs. work around both issues in a cygwin-specific hack. |
|
|
24 | - improve idle watcher documentation slightly. |
|
|
25 | |
|
|
26 | 5.23 Sun Dec 20 23:48:00 CET 2009 |
|
|
27 | - support IDNs in resolve_sockaddr, and therefore in tcp_connect. |
|
|
28 | - implement punycode_encode/decode, idn_nameprep, |
|
|
29 | idn_to_ascii and idn_to_unicode operations in AnyEvent::Util. |
|
|
30 | - provide $AE::VERSION. |
|
|
31 | - removed traces of "no strict 'refs'". |
|
|
32 | |
|
|
33 | 5.22 Sat Dec 5 03:51:13 CET 2009 |
|
|
34 | - downgrade-or-fail in AnyEvent::Handle::push_write, to |
|
|
35 | diagnose encoding failures earlier and more succinctly. |
|
|
36 | (this works around bugs in perl, throwing away encoding info |
|
|
37 | when passing scalar data to extensions). |
|
|
38 | - add more examples to AnyEvent::Socket manpage. |
|
|
39 | - upgrade internal warning set to the same as common::sense 2.03. |
|
|
40 | - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS |
|
|
41 | (John Beppu). |
|
|
42 | - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented. |
|
|
43 | - add google public dns servers to fallback server set and make sure |
|
|
44 | we load-balance properly between the three sets. also add all |
|
|
45 | fallback dns servers, not just a random one, to each dns config. |
|
|
46 | |
|
|
47 | 5.21 Thu Nov 19 02:48:47 CET 2009 |
|
|
48 | - fix a problem where socket constants were called with parameters |
|
|
49 | (spotted by David Friedland). |
|
|
50 | - fork_call never use'd POSIX (reported by Daisuke Maki). |
|
|
51 | - improve perl 5.6 compatibility further (but it still won't work |
|
|
52 | unless you rip out everything but the core). |
|
|
53 | - prefer Net::DNS::Resolver over ipconfig, if installed, on win32. |
|
|
54 | uses 10MB of RAM, but doesn't open a console window. *sigh*. |
|
|
55 | |
|
|
56 | 5.202 Wed Oct 14 22:35:44 CEST 2009 |
|
|
57 | - AnyEvent::DNS would unexpectedly clobber $_ under windows |
|
|
58 | (analysed by Matthias Waldorf). |
|
|
59 | - AnyEvent::Handle::run_cmd can now pass the PID of the |
|
|
60 | newly-created process, which is much less useful than it might |
|
|
61 | sound (based on patch by Yann Kerherve). |
|
|
62 | |
|
|
63 | 5.201 Tue Sep 29 12:09:25 CEST 2009 |
|
|
64 | - AnyEvent:Handle::on_starttls/on_stoptls methods were broken |
|
|
65 | (reported by Torsten Foertsch). |
|
|
66 | - common::sense 2.0 could cause tcp_server to throw an exception |
|
|
67 | (analysed by elmex). |
|
|
68 | |
|
|
69 | 5.2 Mon Sep 14 07:04:49 CEST 2009 |
|
|
70 | - INCOMPATIBLE CHANGE: do no longer support register_read_type |
|
|
71 | and register_write_type in AnyEvent::Handle, instead support |
|
|
72 | package names (the facility was mostly abused). |
|
|
73 | - implement "packagename-as-read/write type" support in |
|
|
74 | AnyEvent::Handle. |
|
|
75 | - AnyEvent::Handle: new options "keepalive" and "oobinline". |
|
|
76 | - oobinline set by default to avoid security issues. |
|
|
77 | - the pure-perl event loop backend wrongly detected times() fallback |
|
|
78 | support (spotted by Pavel Boldin). |
|
|
79 | |
|
|
80 | 5.12 Tue Sep 1 20:26:50 CEST 2009 |
|
|
81 | - be more lenient when parsing resolv.conf files, as some people |
|
|
82 | use hashmarks as comment indicator inside directives (reported |
|
|
83 | by Michael S. Fischer). |
|
|
84 | - use same set of warnings as common::sense 2.0. |
|
|
85 | - fix a potential 32 bit overflow issue due to perl having problems |
|
|
86 | with large hex constants in the Perl backend. |
|
|
87 | - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS |
|
|
88 | by default. |
|
|
89 | - fix some prototypes. |
|
|
90 | - (optionally) require Async::Interrupt 1.02. |
|
|
91 | - replace opendns fallbacks by verizon/level3 public dns servers. |
|
|
92 | |
|
|
93 | 5.112 Fri Aug 21 13:59:04 CEST 2009 |
|
|
94 | - AnyEvent::Handle could cause hard crashes in Net::SSLeay when |
|
|
95 | using starttls with an illegal/uncreatable tls context. |
|
|
96 | - port to broken win32 perls. |
|
|
97 | |
|
|
98 | 5.111 Sun Aug 16 18:44:36 CEST 2009 |
|
|
99 | - perl errornously requires sockaddr_un structures returned by the |
|
|
100 | kernel to have a specific length. work around this bug by padding |
|
|
101 | sockaddr_un structures (found to fail on OS X, as perl only uses |
|
|
102 | the correct code for GNU/Linux, but likely to fail on other |
|
|
103 | OSes as well). |
|
|
104 | |
|
|
105 | 5.11 Wed Aug 12 17:49:37 CEST 2009 |
|
|
106 | - fixed a bug in run_cmd with input from filename. |
|
|
107 | - create a json coder object when none is given in write direction |
|
|
108 | to reduce dependency on JSON. |
|
|
109 | |
|
|
110 | 5.1 Tue Aug 11 03:17:32 CEST 2009 |
|
|
111 | - tcp_connect now reports the correct error in $!, not an unrelated |
|
|
112 | one (was broken since 4.91). |
|
|
113 | - AE::Handle did not properly clear rtimeout/wtimeout watchers |
|
|
114 | when retrying a tcp_connect. |
|
|
115 | - new functions: AnyEvent::Util::run_cmd and close_all_fds_except. |
|
|
116 | - fixed wrong documentation in AE::Socket. |
|
|
117 | |
|
|
118 | 5.01 Mon Aug 10 03:16:32 CEST 2009 |
|
|
119 | - last release broke EV child watchers :/. |
|
|
120 | |
|
|
121 | 5.0 Sun Aug 9 17:32:09 CEST 2009 |
|
|
122 | - officially introduce and document the AE API. |
|
|
123 | - lowering the timeout at runtime did not have immediate effect |
|
|
124 | in AnyEvent::Handle. |
|
|
125 | - AE::Handle now has separate and independent read and write timeout |
|
|
126 | settings. |
|
|
127 | - probe for Guard module when AnyEvent::Util is loaded, not at |
|
|
128 | runtime (Event doesn't like eval "use" from callbacks, and it avoids |
|
|
129 | exporting the probe function). |
|
|
130 | - Impl::IOAsync's io watchers did not use the filehandle emulation |
|
|
131 | code to work around it's watcher limits. |
|
|
132 | - work around IO::Async's problems with overloaded objects. |
|
|
133 | - try to disable wearnings inside IO::Async, it's just too buggy. |
|
|
134 | - do no longer use Event by default, only EV or Perl, Event is too buggy. |
|
|
135 | - hackishly provide fast AE:: interface when EV or Perl is the backend. |
|
|
136 | - inherit VERSION from AnyEvent::VERSION in most modules. |
|
|
137 | |
|
|
138 | 4.91 Thu Aug 6 15:42:45 CEST 2009 |
|
|
139 | - AE::Handle::starttls could get out of sync when the read buffer |
|
|
140 | already contains some TLS handshake. |
|
|
141 | - AE::Handle did not properly free the TLS session, causing bigger |
|
|
142 | memory leaks in Net::SSLeay than expected (reported by toaster). |
|
|
143 | - AE::Socket::tcp_connect will now artificially delay invoking |
|
|
144 | the callback to avoid returning after invoking the callback. |
|
|
145 | - convert many internal watcher uses to AE API. |
|
|
146 | |
|
|
147 | 4.9 Sat Aug 1 11:07:01 CEST 2009 |
|
|
148 | - Glib silently fails when registering a timeout with |
|
|
149 | a negative value, so avoid that. |
|
|
150 | - call condvar callback immediately when it is set |
|
|
151 | after the condition is already signalled. |
|
|
152 | - check rbuf_max condition only after trying to consume |
|
|
153 | data. |
|
|
154 | - normalise signal numbers to names when using the ae's |
|
|
155 | signal handling, but do not document this (yet). |
|
|
156 | - pure perl signal emulation did not properly set nonblocking |
|
|
157 | mode on the signal pipe on win32 (this is unlikely to have caused |
|
|
158 | any issues). |
|
|
159 | - new module: AnyEvent::Debug. |
|
|
160 | - AnyEvent::Strict now checks that a signal was specified |
|
|
161 | by name (not name or number). |
|
|
162 | - reduce memory footprint in typical cases by ~50kb by |
|
|
163 | only compiling signal code when necessary. |
|
|
164 | - add AnyEvent::Handle->rbuf_max. |
|
|
165 | - grab the AE namespace for future new API, implement stubs for |
|
|
166 | said future/alternative API. |
|
|
167 | - new function: AnyEvent::Socket::format_hostport. |
|
|
168 | |
|
|
169 | 4.881 Tue Jul 28 12:51:53 CEST 2009 |
|
|
170 | - work around a bug in local in pre-5.10 perls, causing |
|
|
171 | AnyEvent::Handle to recurse when it should not |
|
|
172 | (analysed by elmex). |
|
|
173 | |
|
|
174 | 4.88 Tue Jul 28 04:04:37 CEST 2009 |
|
|
175 | - re-bless the handle into a dummy package after calling |
|
|
176 | AnyEvent::Handle::destroy, so the user does not need to check |
|
|
177 | for errors after every push_write etc. |
|
|
178 | - do not attempt to run t/02_signals.t on obviously broken |
|
|
179 | platforms. |
|
|
180 | |
|
|
181 | 4.87 Sun Jul 26 02:06:16 CEST 2009 |
|
|
182 | - do not attempt to linger when there is no longer a valid fh. |
|
|
183 | - enforce tls mode to be either accept or connect, do not simply |
|
|
184 | segfault in Net::SSLeay. |
|
|
185 | - AnyEvent::Handle can now call tcp_connect itself (new parameters |
|
|
186 | connect, on_prepare, on_connect and on_connect_error). Updated |
|
|
187 | tutorial accordingly. |
|
|
188 | - add AnyEvent::Impl::Irssi backend. |
|
|
189 | |
|
|
190 | 4.86 Mon Jul 20 23:52:29 CEST 2009 |
|
|
191 | - since the verbose warning is not enough, explicitly document |
|
|
192 | that versions before 1.33 of Net::SSLeay are not secure. |
|
|
193 | - work around signal handling races in Event and (...) Event::Lib. |
|
|
194 | - try to align signal-race timer to full-second boundaries. |
|
|
195 | - work around Tk not liking negative timeouts. |
|
|
196 | - don't complain of different grades of environmental unfriendlyness |
|
|
197 | in IO::Async. |
|
|
198 | |
|
|
199 | 4.85 Sat Jul 18 06:16:14 CEST 2009 |
|
|
200 | - nail the signal race problem in perl once and for all |
|
|
201 | (see $AnyEvent::MAX_SIGNAL_LATENCY). |
|
|
202 | - take advantage of Async::Interrupt if it is available. |
|
|
203 | - load Time::HiRes and Guard modules on demand only. |
|
|
204 | - add optional/recommended modules section to AnyEvent |
|
|
205 | documentation. |
|
|
206 | - reduce memory usage considerably (and reduce startup penalty) |
|
|
207 | by not using "strict", "warnings" and "overload" modules. |
|
|
208 | - work around buggy windows/openbsd perls and provide EBADMSG |
|
|
209 | and EPROTO ourselves when missing. |
|
|
210 | - improve perl 5.6 compatibility of the core event loop. |
|
|
211 | - made Net::SSLeay version 1.33 a soft requirement. |
|
|
212 | |
|
|
213 | 4.83 Fri Jul 17 16:56:26 CEST 2009 |
|
|
214 | - implement AnyEvent::Socket::getprotobyname. |
|
|
215 | - AnyEvent::CondVar's will now detect recursive blocking |
|
|
216 | waits and will croak, as too many people fall into |
|
|
217 | this trap. |
|
|
218 | - AnyEvent::Handle will now call ->destroy on itself after |
|
|
219 | executing the on_error callback, instead of doing some |
|
|
220 | half-baked internal shutdown, for fatal errors. |
|
|
221 | - clarify on_eof behaviour w.r.t. the read queue and |
|
|
222 | on_read callbacks. |
|
|
223 | - ignore some possible spurious wake-ups in tcp_connect. |
|
|
224 | |
|
|
225 | 4.82 Sat Jul 11 00:34:55 CEST 2009 |
|
|
226 | - POE and Event backends didn't accept some callable objects as |
|
|
227 | callbacks. |
|
|
228 | - use Config module instead of POSIX module to detect signal names |
|
|
229 | in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX |
|
|
230 | module doesn't even have all POSIX signals :/. |
|
|
231 | - use more workarounds around the many refcnt/corruption bugs in |
|
|
232 | Event::Lib. |
|
|
233 | - work around a race condition in perl's select, causing t/03_child.t |
|
|
234 | to rarely fail. |
|
|
235 | |
|
|
236 | 4.81 Thu Jul 9 10:30:30 CEST 2009 |
|
|
237 | - AnyEvent::Handle didn't properly diagnose write errors |
|
|
238 | (it expected -1 from syswrite, how lame... :). |
|
|
239 | - support file descriptors in addition to file handles |
|
|
240 | in AnyEvent->io. |
|
|
241 | - new env variables: PERL_ANYEVENT_RESOLV_CONF, |
|
|
242 | PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE |
|
|
243 | and PERL_ANYEVENT_CA_PATH. |
|
|
244 | - provide a sensible synopsis section for AnyEvent::TLS. |
|
|
245 | - add a "supported backends" section to the manpage. |
|
|
246 | - added simple io watcher test to testsuite, using a |
|
|
247 | portable_socketpair. |
|
|
248 | - tried to improve the stability of the Event::Lib backend, |
|
|
249 | YMMV. |
|
|
250 | |
|
|
251 | 4.8 Mon Jul 6 23:45:16 CEST 2009 |
|
|
252 | - AnyEvent::DNS did not properly follow CNAME records with |
|
|
253 | uppercase targets. |
|
|
254 | - AnyEvent::DNS would errornously return AAAA records |
|
|
255 | with v4 mapped addresses (a faulty record) as ipv4 addresses, |
|
|
256 | causing AnyEvent::Socket to throw an exception. |
|
|
257 | - added new module AnyEvent::TLS for easier SSL/TLS context |
|
|
258 | creation, with many options including hostname verification, |
|
|
259 | secure default configuration, lots of documentation and, |
|
|
260 | predefined diffie-hellman keys for perfect forward security |
|
|
261 | and much more. get it while it's still fresh! |
|
|
262 | - use AnyEvent::TLS in AnyEvent::Handle for context management. |
|
|
263 | - load AnyEvent::Handle only on demand in AnyEvent::DNS, |
|
|
264 | so AnyEvent::Socket users have smaller memory footprint |
|
|
265 | in the common case. |
|
|
266 | - add AnyEvent::Handle->push_shutdown method. |
|
|
267 | - add an additional error message parameter to AnyEvent::Handle's |
|
|
268 | on_error callback (for TLS, $! is still available). |
|
|
269 | - add AnyEvent::Handle on_starttls/on_stoptls callbacks. |
|
|
270 | - make AnyEvent::Handle more robust against early conenction |
|
|
271 | failures (during new), and return C<undef> in that case |
|
|
272 | from the constructor. |
|
|
273 | - AnyEvent::Handle will now try to load only JSON::XS first, |
|
|
274 | then fall back to JSON. |
|
|
275 | - format_ipv4/format_ipv6 are now exported by default, for symmetry, |
|
|
276 | and because it was documented that way. |
|
|
277 | |
|
|
278 | 4.451 Fri Jul 3 00:28:58 CEST 2009 |
|
|
279 | - do not clear rbuf when shutting down an AnyEvent::Handle |
|
|
280 | object - doing so breaks AnyEvent::HTTP. |
|
|
281 | |
|
|
282 | 4.45 Mon Jun 29 22:59:26 CEST 2009 |
|
|
283 | - a write error could cause AnyEvent::Handle to create |
|
|
284 | an I/O watcher with an undefined $fh. |
|
|
285 | - special-case mapped ipv4 addresses in both |
|
|
286 | AnyEvent::Socket::format_address and parse_address, |
|
|
287 | to treat them just like ipv4 addresses. |
|
|
288 | - updated and overhauled the AnyEvent::Intro doc. |
|
|
289 | - implement AnyEvent::Socket::format_ipv4/ipv6. |
|
|
290 | - slightly speed up portable_pipe/socketpair. |
|
|
291 | - expand condvar begin/end documentation. |
|
|
292 | |
|
|
293 | 4.42 Fri Jun 26 08:32:18 CEST 2009 |
|
|
294 | - preliminary and neccesarily incomplete support for IO::Async. |
|
|
295 | - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in |
|
|
296 | case it was set to IGNORE, to ensure we can catch child statuses |
|
|
297 | even when the calling env acts stupidly. |
|
|
298 | - updated benchmarks with IO::Async, which performs very well. |
|
|
299 | |
|
|
300 | 4.412 Wed Jun 24 01:35:57 CEST 2009 |
|
|
301 | - support an "untaint" attribute for AnyEvent::DNS |
|
|
302 | and set it on the default resolver. |
|
|
303 | - implement and document AnyEvent::Impl::Perl::loop. |
|
|
304 | - remove all anyevent-env variables from %ENV |
|
|
305 | when running in tainted mode. |
|
|
306 | - mention and extend the IO::Lambda benchmark. |
|
|
307 | |
|
|
308 | 4.411 Sun Jun 7 18:48:02 CEST 2009 |
|
|
309 | - do not try to use F_SETFD on windows. |
|
|
310 | |
|
|
311 | 4.41 Thu May 14 06:40:11 CEST 2009 |
|
|
312 | - work around issues in older perls (5.8.0?) when |
|
|
313 | a signal handler is deleted from the %SIG hash. |
|
|
314 | - use POSIX::_exit in child test, to avoid running |
|
|
315 | destructors. |
|
|
316 | - speed up CHLD handling by relying on SIGCHLD being |
|
|
317 | synchronously delivered, even when we roll our own |
|
|
318 | implementation. |
|
|
319 | - AnyEvent::DNS: add the "dname" resource record name |
|
|
320 | for cosmetic reasons. |
|
|
321 | |
|
|
322 | 4.4 Sun Apr 26 20:12:33 CEST 2009 |
|
|
323 | - implemented idle watchers, where applicable. |
|
|
324 | - AnyEvent->time died when Event backend was in use. |
|
|
325 | - fix a memleak in the Tk backend. |
|
|
326 | - sped up Tk timer handling. |
|
|
327 | - clip negative "after" values to 0 in AnyEvent::Impl::Event |
|
|
328 | to avoid spamming out warning messages. |
|
|
329 | - fix Qt timers without interval. |
|
|
330 | - avoid Qt zero-timeout special case, in old libqt's. |
|
|
331 | |
|
|
332 | 4.352 Mon Apr 20 16:31:11 CEST 2009 |
|
|
333 | - fix AnyEvent::Strict error messages for child watchers. |
|
|
334 | - fix/update Makefile.PL "recommends" versions. |
|
|
335 | - add AnyEvent->now_update. |
|
|
336 | |
|
|
337 | 4.351 Sat Apr 11 07:56:14 CEST 2009 |
|
|
338 | - actually make the signal pipe work under win32. |
|
|
339 | - localise $! in signal handler to avoid changing $!, although |
|
|
340 | perl itself does not seem to save/restore errno either. |
|
|
341 | - set the cloexec flag on the signal pipe (normally set by perl too). |
|
|
342 | |
|
|
343 | 4.35 Fri Mar 27 11:48:20 CET 2009 |
|
|
344 | - event models relying on AnyEvent's signal watcher emulation did |
|
|
345 | invoke the callback asynchronously, contradicting documentation |
|
|
346 | and causing signals to get lost (this includes AnyEvent's own |
|
|
347 | event loop). AnyEvent now uses the standard pipe trick to make |
|
|
348 | callback execution synchronous to the event loop. |
|
|
349 | - AnyEvent::Handle didn't free TLS context data on DESTROY |
|
|
350 | (patch by Pavel Shaydo). |
|
|
351 | - work around the ever-incompatibly-changing API of MakeMaker. |
|
|
352 | - document that changing global variables without restoring them |
|
|
353 | is a bad idea in Perl (noted by Adam Rosenstein). |
|
|
354 | - AnyEvent::Strict now barfs if ->io is passed a file. |
|
|
355 | |
|
|
356 | 4.34 Thu Feb 12 18:32:45 CET 2009 |
|
|
357 | - separately buffer TLS read data, as otherwise the read queue could |
|
|
358 | deadlock as receiving data is not expected while draining the |
|
|
359 | read queue (which cna only happen with TLS). |
|
|
360 | - raise EBADMSG error on JSON decoding errors. |
|
|
361 | - fix some minor manpage bugs (reported by Maximilian Gaß). |
|
|
362 | - speed up select bitmask parsing quite a bit in the pure perl |
|
|
363 | backend. |
|
|
364 | - use CORE::select instead of just select in AnyEvent::Impl::Perl. |
|
|
365 | |
|
|
366 | 4.331 Tue Jan 6 21:07:25 CET 2009 |
|
|
367 | - socketpair fails on many vista machines because vista has |
|
|
368 | completely broken accept/getpeername and getsockname functions, |
|
|
369 | so we provide our own socketpair emulation that kind of works |
|
|
370 | (AnyEvent::Util::portable_pipe). |
|
|
371 | - new function: AnyEvent::Util::portable_socketpair. |
|
|
372 | - take advantage of the Guard module if it exists. |
|
|
373 | |
|
|
374 | 4.33 Fri Nov 21 02:35:40 CET 2008 |
|
|
375 | - AnyEvent::Strict did errornously flag a fileno of 0 as |
|
|
376 | illegal. |
|
|
377 | - reduce memory usage and slightly speed up the pure perl backend |
|
|
378 | by only storing the file descriptor, not the file handle. |
|
|
379 | - add missing autocork method to AnyEvent::Handle (reported |
|
|
380 | by Adam Rosenstein). |
|
|
381 | - AnyEvent::DNS->resolve errornously documented an $rcode result |
|
|
382 | argument, but there isn't (spotted by Henrik Krohns). |
|
|
383 | - the naked truth about Tk - it's basically unmaintained. |
|
|
384 | |
|
|
385 | 4.32 Mon Nov 3 22:46:32 CET 2008 |
|
|
386 | - fix AnyEvent::Socket::resolve_sockaddr to properly support |
|
|
387 | unix sockets again and choose a proper default for |
|
|
388 | the socket type. |
|
|
389 | - fix call to resolve_sockaddr in tcp_connect. |
|
|
390 | |
|
|
391 | 4.31 Thu Oct 30 04:41:48 CET 2008 |
|
|
392 | - implemented AnyEvent::Handle->destroy method. |
|
|
393 | - hint about unexpected effects in TLS mode. |
|
|
394 | - speed up AnyEvent::Impl::Perl by using more arrays |
|
|
395 | instead of hashes without the slightest loss of readability :) |
|
|
396 | - work around a perl argument refcounting bug. |
|
|
397 | - with some perl patches applied, AnyEvent now seems leak-free. |
|
|
398 | |
|
|
399 | 4.3 Fri Oct 3 09:18:43 CEST 2008 |
|
|
400 | - AnyEvent will now install a no-op signal handler for SIGPIPE, |
|
|
401 | unless one has been installed already. |
|
|
402 | - warn about not loaidng AnyEvent::Impl::POE early enough |
|
|
403 | (patch by Adam Rosenstein). |
|
|
404 | - fixed a great number of bugs and corner cases in AnyEvent::Handle: |
|
|
405 | - fix a bug in where in SSL connect mode, the |
|
|
406 | client would first wait for some data by the server and |
|
|
407 | otherwise hang. (reported and analysed in an absolutely |
|
|
408 | exemplary manner by Adam Rosenstein). |
|
|
409 | - fix a bug in where SSL EOF would not be treated |
|
|
410 | as stream EOF, putting the connection into a hung state. |
|
|
411 | (reported and analysed in an absolutely exemplary manner by Adam |
|
|
412 | Rosenstein). |
|
|
413 | - fix a potential segfault when the TLS context |
|
|
414 | would go missing in a read callback (e.g. due to stoptls) - |
|
|
415 | Net::SSLeay of course makes no type checking whatsoever. |
|
|
416 | - AnyEvent::Handle will not stop the read watcher in TLS mode. |
|
|
417 | - AnyEvent::Handle->stoptls will send a TLS close notify |
|
|
418 | instead of simply destroying the stream now. |
|
|
419 | - fix a bug where large blocks of data written in TLS mode |
|
|
420 | would not be sent unless triggered by receives. |
|
|
421 | - on_drain will now take the tls write buffer into account. |
|
|
422 | - SSL operations have been streamlined a bit and should be faster |
|
|
423 | now, more to come, though. |
|
|
424 | - added a FAQ section, will document nontrivial issues in there. |
|
|
425 | |
|
|
426 | 4.234 Mon Sep 29 04:08:13 CEST 2008 |
|
|
427 | - fix child watcher documentation: only child exits will be reported, |
|
|
428 | no trace events. |
|
|
429 | - mention SIGPIPE in AnyEvent::Handle. |
|
|
430 | - perl backend will now use POSIX::times as monotonic clock |
|
|
431 | when available and the monotonic clock option is not. |
|
|
432 | - run even without Time::HiRes in some cases now, but do not |
|
|
433 | provide sub-second accuracy (all *supported* perl versions have |
|
|
434 | Time::HiRes). |
|
|
435 | |
|
|
436 | 4.233 Fri Aug 22 01:48:05 CEST 2008 |
|
|
437 | - fix a bug in the testsuite, causing a hang. |
|
|
438 | - clarified lots of AnyEvent::Handle constructor arguments. |
|
|
439 | |
|
|
440 | 4.232 Thu Aug 21 20:44:25 CEST 2008 |
|
|
441 | - fix a bug in AnyEvent::Handle that could cause two fatal error |
|
|
442 | callbacks on EOF, which caused AnyEvent::HTTP to signal success twice |
|
|
443 | (testcase by Leon Brocard). |
|
|
444 | - clarified on_eof/on_error documentation a bit. |
|
|
445 | |
|
|
446 | 4.231 Tue Jul 29 13:12:15 CEST 2008 |
|
|
447 | - remove some debugging code left in AnyEvent::Util::fork_call |
|
|
448 | (and no, it's impossible to implement with the broken windows |
|
|
449 | perls without resource leaks or worse). |
|
|
450 | |
|
|
451 | 4.23 Tue Jul 29 12:19:59 CEST 2008 |
|
|
452 | - document the first parameter passed to condvar callbacks to be |
|
|
453 | the callback. |
|
|
454 | - add AnyEvent::Socket::{ntoa,aton} aliases. |
|
|
455 | - optimize the AE::Handle->push_read (line) for the default |
|
|
456 | eol marker. |
|
|
457 | - optimize push_read (packstring|storable) for small packets. |
|
|
458 | - invoke on_error callback when no on_eof callback is set. |
|
|
459 | - fix a bug in push_read (storable) of unknown impact. |
|
|
460 | |
|
|
461 | 4.22 Sun Jul 20 16:34:13 CEST 2008 |
|
|
462 | - new function AnyEvent::Socket::parse_hostport. |
|
|
463 | - as the bulkheads at microsoft can't even get getprotobyname reliably |
|
|
464 | working on their shitty fucking broken os we need to hardcode |
|
|
465 | some common protocol numbers in AnyEvent::Socket. How can |
|
|
466 | people even bother with such a pile of shit as windows. |
|
|
467 | |
|
|
468 | 4.21 Thu Jul 17 14:40:05 CEST 2008 |
|
|
469 | - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype |
|
|
470 | which is incompatible to previous invocation syntax. |
|
|
471 | - work around more windows perl bugs in fork_call. I cannot imagine |
|
|
472 | why anybody would bother implementing fork in such an obviously |
|
|
473 | unusable way. |
|
|
474 | - avoid calling $do_retry if the request has already finished in |
|
|
475 | AnyEvent::DNS, thus avoiding an exception (reported by Anatoly |
|
|
476 | K. Sharifulin). |
|
|
477 | |
|
|
478 | 4.2 Sat Jul 12 22:42:11 CEST 2008 |
|
|
479 | - support an "interval" argument for repeating timers. |
|
|
480 | - fix a bug in the pure perl event loop that caused timers to |
|
|
481 | be delayed under some circumstances. |
|
|
482 | - implement optional argument checking via PERL_ANYEVENT_STRICT. |
|
|
483 | - not importing anything from Event causes it to malfunction |
|
|
484 | and not use Time::HiRes, work around this bug. |
|
|
485 | - more bug workarounds against the endlessly broken Tk module, |
|
|
486 | seems to work now on 32 bit machines, but less so on 64 bit. |
|
|
487 | - minor optimisations applied to most backends. |
|
|
488 | |
|
|
489 | 4.161 Fri Jul 4 14:24:48 CEST 2008 |
|
|
490 | - fixed the t/handle/02_write.t, it read after write, with the |
|
|
491 | assumption that all data could be written before read |
|
|
492 | (reported by Jim Razmus). |
|
|
493 | |
|
|
494 | 4.160 Thu Jul 3 04:02:21 CEST 2008 |
|
|
495 | - re-registering signal watchers after unregistering one did |
|
|
496 | not work if a backend falls back on AnyEvent's default |
|
|
497 | implementation (also affected child watchers). |
|
|
498 | - new AnyEvent::Handle options: autocork and no_delay. |
|
|
499 | |
|
|
500 | 4.152 Sun Jun 22 14:15:44 CEST 2008 |
|
|
501 | - allow for 32-bit perls that implement shifts differently |
|
|
502 | on different architectures in parse_ipv4 (reported and |
|
|
503 | analysed by Keiichi DAIBA). |
|
|
504 | |
|
|
505 | 4.151 Fri Jun 6 17:34:24 CEST 2008 |
|
|
506 | - make sure specifying _only_ on_read and never pushing reads |
|
|
507 | works. |
|
|
508 | |
|
|
509 | 4.15 Fri Jun 6 13:00:46 CEST 2008 |
|
|
510 | - the pure perl backend would keep some watchers alive when more than |
|
|
511 | one watcher was registered for the same fd. |
|
|
512 | - new "packstring" and "storable" read and write types |
|
|
513 | for AnyEvent::Handle. |
|
|
514 | - allow on_eof handler to be called after on_error with EPIPE returns. |
|
|
515 | - do not immediately call on_read callback in handle constructor. |
|
|
516 | |
|
|
517 | 4.14 Thu Jun 5 20:29:31 CEST 2008 |
|
|
518 | - Fixed a bug in DNS SRV priority sorting. |
|
|
519 | - AnyEvent::Util::guard now reports runtime errors while |
|
|
520 | executing the guard block as warnings. |
|
|
521 | - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle. |
|
|
522 | |
|
|
523 | 4.13 Thu Jun 5 00:47:59 CEST 2008 |
|
|
524 | - AnyEvent::DNS only followed cname chains with length 2, |
|
|
525 | contrary to documentation. bumped it up to 10, thanks to |
|
|
526 | microsoft, the current broken-dns-config-king. |
|
|
527 | - AnyEvent::DNS didn't check for socket return status |
|
|
528 | but instead relied on perl not creating filehandles in that |
|
|
529 | case - too bad it gives you a bogus file handle (reported |
|
|
530 | and analysed by Vladimir Timofeev). |
|
|
531 | - fix queue management logic in AnyEvent::Handle: |
|
|
532 | when on_read was registered, the queue was empty |
|
|
533 | and no progress could be made AnyEvent::Handle would |
|
|
534 | enter an endless loop. |
|
|
535 | - correctly start reading again when the handle became |
|
|
536 | busy again after idling. |
|
|
537 | - correctly treat tls shut-downs as EOF condition |
|
|
538 | (for the time being, would be nice to have a callback for that). |
|
|
539 | - correctly call eof callback under all conditions |
|
|
540 | (hopefully). |
|
|
541 | - the timeout callback did not expect that $self can go |
|
|
542 | away any time. |
|
|
543 | |
|
|
544 | 4.12 Tue Jun 3 10:58:04 CEST 2008 |
|
|
545 | - include AnyEvent::Intro, a tutorial for anyevent, |
|
|
546 | anyevent::socket and anyevent::handle. |
|
|
547 | - allow more options in on_error. |
|
|
548 | |
|
|
549 | 4.11 Fri May 30 23:42:25 CEST 2008 |
|
|
550 | - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide |
|
|
551 | reverse_lookup and reverse_verify to replace it, support |
|
|
552 | v4mapped and v4compat addresses. |
|
|
553 | - provide more documentation for the resolver class. |
|
|
554 | - really replace longest run of :0: by :: in format_address, |
|
|
555 | also properly convert :: and ::1 again. |
|
|
556 | - support NAPTR record name and decode it. |
|
|
557 | - implement random weight sampling for SRV records, as per |
|
|
558 | rfc 2782. |
|
|
559 | - correctly abort on srv-record targets of ".". |
|
|
560 | - added AnyEvent::DNS::wait_for_slot. |
|
|
561 | - in the unlikely event of a virtual circuit connection |
|
|
562 | being invalidated by a delayed udp reply, AnyEvent::DNS |
|
|
563 | could die. this has been fixed. |
|
|
564 | - plug a probable memleak in the DNS vc code. |
|
|
565 | - use configured timeout also for VC request phase. |
|
|
566 | - implement timeout and max_outstanding methods |
|
|
567 | for resolver class. |
|
|
568 | - update version numbers in all modules. |
|
|
569 | |
|
|
570 | 4.1 Thu May 29 05:45:40 CEST 2008 |
|
|
571 | - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address |
|
|
572 | which mirrors their purpose better (old names still available |
|
|
573 | for a while). Also moved AnyEvent::DNS::addr to |
|
|
574 | AnyEvent::Socket::resolve_sockaddr. |
|
|
575 | - implement AnyEvent->time and AnyEvent->now. |
|
|
576 | - fix IPv6 support in pack_sockaddr. |
|
|
577 | - officially un-experimentalise all newly introduced code. |
|
|
578 | - support unix domain sockets everywhere by specifying |
|
|
579 | a host of "unix/" and the pathname as service. |
|
|
580 | - implement an activity timeout in AnyEvent::Handle. |
|
|
581 | - added a regex read type to AnyEvent::Handle. |
|
|
582 | - added a json read/write type to AnyEvent::Handle. |
|
|
583 | - always croak in on_error in AnyEvent::Handle. |
|
|
584 | - document how to subclass AnyEvent::Handle. |
|
|
585 | - implement AnyEvent::Util::fork_call. |
|
|
586 | - add support for IPv6 nameservers and nameserver statements. |
|
|
587 | - work around _yet_ _another_ windows perl bug in where empty |
|
|
588 | select masks cause errors in the pure perl backend (this |
|
|
589 | workaround cannot be implemented for the other event loops, |
|
|
590 | but EV also works around this). |
|
|
591 | - supply AnyEvent::Socket::address_type and make good use of it. |
|
|
592 | - clarify and fix the Handle documentation w.r.t $self vs. $handle |
|
|
593 | vs. callback arguments. |
|
|
594 | - add some recommends to the META.yml. |
|
|
595 | |
|
|
596 | 4.05 Mon May 26 19:44:06 CEST 2008 |
|
|
597 | - some platforms ignore the rfc and prepend an extra sa_len member to |
|
|
598 | the sockaddr structure, cater for those. also use sockaddr_family, |
|
|
599 | didn't know it was there, but comes in handy for the workaround. |
|
|
600 | - undo the import WIN32 hack. |
|
|
601 | |
|
|
602 | 4.04 Mon May 26 08:03:31 CEST 2008 |
|
|
603 | - try to work around yet another windows bug: failed connects are reported |
|
|
604 | as if it were out-of-band data. windows users: you suck. Only |
|
|
605 | EV, Glib, Event and the pure perl backend can handle this condition. |
|
|
606 | - optimize write algorithm in AnyEvent::Handle. |
|
|
607 | - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly. |
|
|
608 | - tcp_connect now properly iterates through all targets. |
|
|
609 | - check wether IPv6 sockets can actually be created, otherwise, disable ipv6 |
|
|
610 | support entirely. also hardcode some AF_INET6 constants for the |
|
|
611 | sake of perl 5.8. |
|
|
612 | - vastly improved nameserver/suffix detection algorithm for windows |
|
|
613 | (still a hack...) |
|
|
614 | - try to use a fallback dns server if no dns servers could be found. |
|
|
615 | - splatter around some local $SIG{__DIE__}'s to work around bad |
|
|
616 | die handlers. |
|
|
617 | |
|
|
618 | 4.03 Sun May 25 05:05:57 CEST 2008 |
|
|
619 | - work around perl bugs on windows, where perl returns undocumented |
|
|
620 | error codes for sysread, syswrite etc. |
|
|
621 | - fix AnyEvent::Handle::unshift_read to really unshift and not push. |
|
|
622 | - AnyEvent::Handle could sometimes stumble over 0-byte writes. |
|
|
623 | - fix unpack_sockaddr, now works for AF_INET6 sockaddr's. |
|
|
624 | - updated documentation. |
|
|
625 | - lots of workarounds for perl 5.8 quirks. |
|
|
626 | - implement a more flexible type system for both reads and writes |
|
|
627 | in AnyEvent::Handle. |
|
|
628 | - support IPv6 in tcp_server. |
|
|
629 | - pass local hostname and port to the prepare callback in tcp_server. |
|
|
630 | - make types extendable in AnyEvent::Handle. |
|
|
631 | - croak, not die, in AnyEvent::Handle; |
|
|
632 | |
|
|
633 | 4.0 Sat May 24 19:58:08 CEST 2008 |
|
|
634 | - added (experimental) AnyEvent::DNS module. |
|
|
635 | - added (experimental) AnyEvent::Socket module. |
|
|
636 | - added (experiemntal) support for TLS/SSL in AnyEvent::Handle. |
|
|
637 | - make condvar->end call ->send as documented. |
|
|
638 | - make condvar constructor support the documented cb => argument. |
|
|
639 | - overload condvars so you can use them instead of code references. |
|
|
640 | - support more of the POSIX forms instead of a pure dotted quad for |
|
|
641 | verbatim ip addresses. |
|
|
642 | - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. |
|
|
643 | - try very hard not to freeze on broken windows perls. |
|
|
644 | - make on_eof optional in AnyEvent::Handle. |
|
|
645 | - support service names in tcp_connect and tcp_server. |
|
|
646 | - work around netbsd/openbsd bug causing unix sockets not to |
|
|
647 | become writable until empty. |
|
|
648 | - only make a dummy read for error detection when getpeername |
|
|
649 | returns ENOTCONN. |
|
|
650 | - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal |
|
|
651 | override. |
|
|
652 | - new AnyEvent::Guard::cancel. |
|
|
653 | |
|
|
654 | 3.5 Sat May 17 23:17:14 CEST 2008 |
|
|
655 | - gracefully deal with systems not defining CLOCK_MONOTONIC |
|
|
656 | (as opposed to not supporting the clock itself, which worked in 3.41). |
|
|
657 | - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server |
|
|
658 | helpers, to replace the ill-designed AnyEvent::Socket class. |
|
|
659 | - added some examples (eg/connect, eg/handle). |
|
|
660 | - added AnyEvent::Util::guard. |
|
|
661 | - AnyEvent::Util now exports most of its functions. |
|
|
662 | - fixed a bug in *_read_chunk, where the callback was called with $self as |
|
|
663 | first and second argument. |
|
|
664 | - simplified and fixed regex handling in AnyEvent::Handle::*_read_line. |
|
|
665 | - fix manpage condvars examples to use new syntax. |
|
|
666 | - considerably improve inet_aton with EV::ADNS backend to support |
|
|
667 | cname chains (common dns configuration bug, e.g. with www.google.de) |
|
|
668 | and aaaa records if no a record could be found. |
|
|
669 | |
|
|
670 | 3.41 Sun May 11 19:53:13 CEST 2008 |
|
|
671 | - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. |
|
|
672 | |
|
|
673 | 3.4 Sun May 11 00:29:25 CEST 2008 |
|
|
674 | - complete re-implementation of condvars with a hopefully much more useful |
|
|
675 | API (backward compatible functions exist but are undocumented). |
|
|
676 | - AnyEvent::Handle fully rewritten with a hopefully more powerful API |
|
|
677 | (still subject to change). |
| 4 | - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
678 | - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
|
|
679 | - provide AnyEvent::post_detect and @AnyEvent::post_detect, which |
|
|
680 | allows module authors to avoid forcing event loop detection. |
|
|
681 | (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example). |
|
|
682 | - remove coro backends: Coro now provides generic support for AnyEvent |
|
|
683 | via Coro::AnyEvent. |
| 5 | |
684 | |
| 6 | 3.3 Mon Apr 28 09:51:06 CEST 2008 |
685 | 3.3 Mon Apr 28 09:51:06 CEST 2008 |
| 7 | - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. |
686 | - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. |
| 8 | - fix a bug in the pure perl backend that kept watchers alive |
687 | - fix a bug in the pure perl backend that kept watchers alive |
| 9 | when multiple watchers were registered for the same fh/poll combo. |
688 | when multiple watchers were registered for the same fh/poll combo. |
