| 1 | Revision history for Perl extension AnyEvent. |
1 | Revision history for Perl extension AnyEvent. |
|
|
2 | |
|
|
3 | TODO: _AVOID_, _VERBOSE_ |
|
|
4 | - improve perl 5.6 compatibility of the core event loop. |
|
|
5 | - nail the signal race problem in perl once and for all |
|
|
6 | (see $AnyEvent::MAX_SIGNAL_LATENCY). |
|
|
7 | - take advantage of Async::Interrupt if it is available. |
|
|
8 | - load Time::HiRes and Guard modules on demand only. |
|
|
9 | - add optional/recommended modules section to AnyEvent |
|
|
10 | documentation. |
|
|
11 | - reduce memory usage considerably (and reduce startup penalty) |
|
|
12 | by not using "strict", "warnings" and "overload" modules. |
|
|
13 | - work around buggy windows perls and provide EBADMSG and EPROTO |
|
|
14 | ourselves. |
|
|
15 | |
|
|
16 | 4.83 Fri Jul 17 16:56:26 CEST 2009 |
|
|
17 | - implement AnyEvent::Socket::getprotobyname. |
|
|
18 | - AnyEvent::CondVar's will now detect recursive blocking |
|
|
19 | waits and will croak, as too many people fall into |
|
|
20 | this trap. |
|
|
21 | - AnyEvent::Handle will now call ->destroy on itself after |
|
|
22 | executing the on_error callback, instead of doing some |
|
|
23 | half-baked internal shutdown, for fatal errors. |
|
|
24 | - clarify on_eof behaviour w.r.t. the read queue and |
|
|
25 | on_read callbacks. |
|
|
26 | - ignore some possible spurious wake-ups in tcp_connect. |
|
|
27 | |
|
|
28 | 4.82 Sat Jul 11 00:34:55 CEST 2009 |
|
|
29 | - POE and Event backends didn't accept some callable objects as |
|
|
30 | callbacks. |
|
|
31 | - use Config module instead of POSIX module to detect signal names |
|
|
32 | in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX |
|
|
33 | module doesn't even have all POSIX signals :/. |
|
|
34 | - use more workarounds around the many refcnt/corruption bugs in |
|
|
35 | Event::Lib. |
|
|
36 | - work around a race condition in perl's select, causing t/03_child.t |
|
|
37 | to rarely fail. |
|
|
38 | |
|
|
39 | 4.81 Thu Jul 9 10:30:30 CEST 2009 |
|
|
40 | - AnyEvent::Handle didn't properly diagnose write errors |
|
|
41 | (it expected -1 from syswrite, how lame... :). |
|
|
42 | - support file descriptors in addition to file handles |
|
|
43 | in AnyEvent->io. |
|
|
44 | - new env variables: PERL_ANYEVENT_RESOLV_CONF, |
|
|
45 | PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE |
|
|
46 | and PERL_ANYEVENT_CA_PATH. |
|
|
47 | - provide a sensible synopsis section for AnyEvent::TLS. |
|
|
48 | - add a "supported backends" section to the manpage. |
|
|
49 | - added simple io watcher test to testsuite, using a |
|
|
50 | portable_socketpair. |
|
|
51 | - tried to improve the stability of the Event::Lib backend, |
|
|
52 | YMMV. |
|
|
53 | |
|
|
54 | 4.8 Mon Jul 6 23:45:16 CEST 2009 |
|
|
55 | - AnyEvent::DNS did not properly follow CNAME records with |
|
|
56 | uppercase targets. |
|
|
57 | - AnyEvent::DNS would errornously return AAAA records |
|
|
58 | with v4 mapped addresses (a faulty record) as ipv4 addresses, |
|
|
59 | causing AnyEvent::Socket to throw an exception. |
|
|
60 | - added new module AnyEvent::TLS for easier SSL/TLS context |
|
|
61 | creation, with many options including hostname verification, |
|
|
62 | secure default configuration, lots of documentation and, |
|
|
63 | predefined diffie-hellman keys for perfect forward security |
|
|
64 | and much more. get it while it's still fresh! |
|
|
65 | - use AnyEvent::TLS in AnyEvent::Handle for context management. |
|
|
66 | - load AnyEvent::Handle only on demand in AnyEvent::DNS, |
|
|
67 | so AnyEvent::Socket users have smaller memory footprint |
|
|
68 | in the common case. |
|
|
69 | - add AnyEvent::Handle->push_shutdown method. |
|
|
70 | - add an additional error message parameter to AnyEvent::Handle's |
|
|
71 | on_error callback (for TLS, $! is still available). |
|
|
72 | - add AnyEvent::Handle on_starttls/on_stoptls callbacks. |
|
|
73 | - make AnyEvent::Handle more robust against early conenction |
|
|
74 | failures (during new), and return C<undef> in that case |
|
|
75 | from the constructor. |
|
|
76 | - AnyEvent::Handle will now try to load only JSON::XS first, |
|
|
77 | then fall back to JSON. |
|
|
78 | - format_ipv4/format_ipv6 are now exported by default, for symmetry, |
|
|
79 | and because it was documented that way. |
|
|
80 | |
|
|
81 | 4.451 Fri Jul 3 00:28:58 CEST 2009 |
|
|
82 | - do not clear rbuf when shutting down an AnyEvent::Handle |
|
|
83 | object - doing so breaks AnyEvent::HTTP. |
|
|
84 | |
|
|
85 | 4.45 Mon Jun 29 22:59:26 CEST 2009 |
|
|
86 | - a write error could cause AnyEvent::Handle to create |
|
|
87 | an I/O watcher with an undefined $fh. |
|
|
88 | - special-case mapped ipv4 addresses in both |
|
|
89 | AnyEvent::Socket::format_address and parse_address, |
|
|
90 | to treat them just like ipv4 addresses. |
|
|
91 | - updated and overhauled the AnyEvent::Intro doc. |
|
|
92 | - implement AnyEvent::Socket::format_ipv4/ipv6. |
|
|
93 | - slightly speed up portable_pipe/socketpair. |
|
|
94 | - expand condvar begin/end documentation. |
|
|
95 | |
|
|
96 | 4.42 Fri Jun 26 08:32:18 CEST 2009 |
|
|
97 | - preliminary and neccesarily incomplete support for IO::Async. |
|
|
98 | - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in |
|
|
99 | case it was set to IGNORE, to ensure we can catch child statuses |
|
|
100 | even when the calling env acts stupidly. |
|
|
101 | - updated benchmarks with IO::Async, which performs very well. |
|
|
102 | |
|
|
103 | 4.412 Wed Jun 24 01:35:57 CEST 2009 |
|
|
104 | - support an "untaint" attribute for AnyEvent::DNS |
|
|
105 | and set it on the default resolver. |
|
|
106 | - implement and document AnyEvent::Impl::Perl::loop. |
|
|
107 | - remove all anyevent-env variables from %ENV |
|
|
108 | when running in tainted mode. |
|
|
109 | - mention and extend the IO::Lambda benchmark. |
|
|
110 | |
|
|
111 | 4.411 Sun Jun 7 18:48:02 CEST 2009 |
|
|
112 | - do not try to use F_SETFD on windows. |
|
|
113 | |
|
|
114 | 4.41 Thu May 14 06:40:11 CEST 2009 |
|
|
115 | - work around issues in older perls (5.8.0?) when |
|
|
116 | a signal handler is deleted from the %SIG hash. |
|
|
117 | - use POSIX::_exit in child test, to avoid running |
|
|
118 | destructors. |
|
|
119 | - speed up CHLD handling by relying on SIGCHLD being |
|
|
120 | synchronously delivered, even when we roll our own |
|
|
121 | implementation. |
|
|
122 | - AnyEvent::DNS: add the "dname" resource record name |
|
|
123 | for cosmetic reasons. |
|
|
124 | |
|
|
125 | 4.4 Sun Apr 26 20:12:33 CEST 2009 |
|
|
126 | - implemented idle watchers, where applicable. |
|
|
127 | - AnyEvent->time died when Event backend was in use. |
|
|
128 | - fix a memleak in the Tk backend. |
|
|
129 | - sped up Tk timer handling. |
|
|
130 | - clip negative "after" values to 0 in AnyEvent::Impl::Event |
|
|
131 | to avoid spamming out warning messages. |
|
|
132 | - fix Qt timers without interval. |
|
|
133 | - avoid Qt zero-timeout special case, in old libqt's. |
|
|
134 | |
|
|
135 | 4.352 Mon Apr 20 16:31:11 CEST 2009 |
|
|
136 | - fix AnyEvent::Strict error messages for child watchers. |
|
|
137 | - fix/update Makefile.PL "recommends" versions. |
|
|
138 | - add AnyEvent->now_update. |
|
|
139 | |
|
|
140 | 4.351 Sat Apr 11 07:56:14 CEST 2009 |
|
|
141 | - actually make the signal pipe work under win32. |
|
|
142 | - localise $! in signal handler to avoid changing $!, although |
|
|
143 | perl itself does not seem to save/restore errno either. |
|
|
144 | - set the cloexec flag on the signal pipe (normally set by perl too). |
|
|
145 | |
|
|
146 | 4.35 Fri Mar 27 11:48:20 CET 2009 |
|
|
147 | - event models relying on AnyEvent's signal watcher emulation did |
|
|
148 | invoke the callback asynchronously, contradicting documentation |
|
|
149 | and causing signals to get lost (this includes AnyEvent's own |
|
|
150 | event loop). AnyEvent now uses the standard pipe trick to make |
|
|
151 | callback execution synchronous to the event loop. |
|
|
152 | - AnyEvent::Handle didn't free TLS context data on DESTROY |
|
|
153 | (patch by Pavel Shaydo). |
|
|
154 | - work around the ever-incompatibly-changing API of MakeMaker. |
|
|
155 | - document that changing global variables without restoring them |
|
|
156 | is a bad idea in Perl (noted by Adam Rosenstein). |
|
|
157 | - AnyEvent::Strict now barfs if ->io is passed a file. |
|
|
158 | |
|
|
159 | 4.34 Thu Feb 12 18:32:45 CET 2009 |
|
|
160 | - separately buffer TLS read data, as otherwise the read queue could |
|
|
161 | deadlock as receiving data is not expected while draining the |
|
|
162 | read queue (which cna only happen with TLS). |
|
|
163 | - raise EBADMSG error on JSON decoding errors. |
|
|
164 | - fix some minor manpage bugs (reported by Maximilian Gaß). |
|
|
165 | - speed up select bitmask parsing quite a bit in the pure perl |
|
|
166 | backend. |
|
|
167 | - use CORE::select instead of just select in AnyEvent::Impl::Perl. |
|
|
168 | |
|
|
169 | 4.331 Tue Jan 6 21:07:25 CET 2009 |
|
|
170 | - socketpair fails on many vista machines because vista has |
|
|
171 | completely broken accept/getpeername and getsockname functions, |
|
|
172 | so we provide our own socketpair emulation that kind of works |
|
|
173 | (AnyEvent::Util::portable_pipe). |
|
|
174 | - new function: AnyEvent::Util::portable_socketpair. |
|
|
175 | - take advantage of the Guard module if it exists. |
|
|
176 | |
|
|
177 | 4.33 Fri Nov 21 02:35:40 CET 2008 |
|
|
178 | - AnyEvent::Strict did errornously flag a fileno of 0 as |
|
|
179 | illegal. |
|
|
180 | - reduce memory usage and slightly speed up the pure perl backend |
|
|
181 | by only storing the file descriptor, not the file handle. |
|
|
182 | - add missing autocork method to AnyEvent::Handle (reported |
|
|
183 | by Adam Rosenstein). |
|
|
184 | - AnyEvent::DNS->resolve errornously documented an $rcode result |
|
|
185 | argument, but there isn't (spotted by Henrik Krohns). |
|
|
186 | - the naked truth about Tk - it's basically unmaintained. |
|
|
187 | |
|
|
188 | 4.32 Mon Nov 3 22:46:32 CET 2008 |
|
|
189 | - fix AnyEvent::Socket::resolve_sockaddr to properly support |
|
|
190 | unix sockets again and choose a proper default for |
|
|
191 | the socket type. |
|
|
192 | - fix call to resolve_sockaddr in tcp_connect. |
|
|
193 | |
|
|
194 | 4.31 Thu Oct 30 04:41:48 CET 2008 |
|
|
195 | - implemented AnyEvent::Handle->destroy method. |
|
|
196 | - hint about unexpected effects in TLS mode. |
|
|
197 | - speed up AnyEvent::Impl::Perl by using more arrays |
|
|
198 | instead of hashes without the slightest loss of readability :) |
|
|
199 | - work around a perl argument refcounting bug. |
|
|
200 | - with some perl patches applied, AnyEvent now seems leak-free. |
|
|
201 | |
|
|
202 | 4.3 Fri Oct 3 09:18:43 CEST 2008 |
|
|
203 | - AnyEvent will now install a no-op signal handler for SIGPIPE, |
|
|
204 | unless one has been installed already. |
|
|
205 | - warn about not loaidng AnyEvent::Impl::POE early enough |
|
|
206 | (patch by Adam Rosenstein). |
|
|
207 | - fixed a great number of bugs and corner cases in AnyEvent::Handle: |
|
|
208 | - fix a bug in where in SSL connect mode, the |
|
|
209 | client would first wait for some data by the server and |
|
|
210 | otherwise hang. (reported and analyzed in an absolutely |
|
|
211 | exemplary manner by Adam Rosenstein). |
|
|
212 | - fix a bug in where SSL EOF would not be treated |
|
|
213 | as stream EOF, putting the connection into a hung state. |
|
|
214 | (reported and analyzed in an absolutely exemplary manner by Adam |
|
|
215 | Rosenstein). |
|
|
216 | - fix a potential segfault when the TLS context |
|
|
217 | would go missing in a read callback (e.g. due to stoptls) - |
|
|
218 | Net::SSLeay of course makes no type checking whatsoever. |
|
|
219 | - AnyEvent::Handle will not stop the read watcher in TLS mode. |
|
|
220 | - AnyEvent::Handle->stoptls will send a TLS close notify |
|
|
221 | instead of simply destroying the stream now. |
|
|
222 | - fix a bug where large blocks of data written in TLS mode |
|
|
223 | would not be sent unless triggered by receives. |
|
|
224 | - on_drain will now take the tls write buffer into account. |
|
|
225 | - SSL operations have been streamlined a bit and should be faster |
|
|
226 | now, more to come, though. |
|
|
227 | - added a FAQ section, will document nontrivial issues in there. |
|
|
228 | |
|
|
229 | 4.234 Mon Sep 29 04:08:13 CEST 2008 |
|
|
230 | - fix child watcher documentation: only child exits will be reported, |
|
|
231 | no trace events. |
|
|
232 | - mention SIGPIPE in AnyEvent::Handle. |
|
|
233 | - perl backend will now use POSIX::times as monotonic clock |
|
|
234 | when available and the monotonic clock option is not. |
|
|
235 | - run even without Time::HiRes in some cases now, but do not |
|
|
236 | provide sub-second accuracy (all *supported* perl versions have |
|
|
237 | Time::HiRes). |
|
|
238 | |
|
|
239 | 4.233 Fri Aug 22 01:48:05 CEST 2008 |
|
|
240 | - fix a bug in the testsuite, causing a hang. |
|
|
241 | - clarified lots of AnyEvent::Handle constructor arguments. |
|
|
242 | |
|
|
243 | 4.232 Thu Aug 21 20:44:25 CEST 2008 |
|
|
244 | - fix a bug in AnyEvent::Handle that could cause two fatal error |
|
|
245 | callbacks on EOF, which caused AnyEvent::HTTP to signal success twice |
|
|
246 | (testcase by Leon Brocard). |
|
|
247 | - clarified on_eof/on_error documentation a bit. |
|
|
248 | |
|
|
249 | 4.231 Tue Jul 29 13:12:15 CEST 2008 |
|
|
250 | - remove some debugging code left in AnyEvent::Util::fork_call |
|
|
251 | (and no, it's impossible to implement with the broken windows |
|
|
252 | perls without resource leaks or worse). |
|
|
253 | |
|
|
254 | 4.23 Tue Jul 29 12:19:59 CEST 2008 |
|
|
255 | - document the first parameter passed to condvar callbacks to be |
|
|
256 | the callback. |
|
|
257 | - add AnyEvent::Socket::{ntoa,aton} aliases. |
|
|
258 | - optimize the AE::Handle->push_read (line) for the default |
|
|
259 | eol marker. |
|
|
260 | - optimize push_read (packstring|storable) for small packets. |
|
|
261 | - invoke on_error callback when no on_eof callback is set. |
|
|
262 | - fix a bug in push_read (storable) of unknown impact. |
|
|
263 | |
|
|
264 | 4.22 Sun Jul 20 16:34:13 CEST 2008 |
|
|
265 | - new function AnyEvent::Socket::parse_hostport. |
|
|
266 | - as the bulkheads at microsoft can't even get getprotobyname reliably |
|
|
267 | working on their shitty fucking broken os we need to hardcode |
|
|
268 | some common protocol numbers in AnyEvent::Socket. How can |
|
|
269 | people even bother with such a pile of shit as windows. |
|
|
270 | |
|
|
271 | 4.21 Thu Jul 17 14:40:05 CEST 2008 |
|
|
272 | - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype |
|
|
273 | which is incompatible to previous invocation syntax. |
|
|
274 | - work around more windows perl bugs in fork_call. I cannot imagine |
|
|
275 | why anybody would bother implementing fork in such an obviously |
|
|
276 | unusable way. |
|
|
277 | - avoid calling $do_retry if the request has already finished in |
|
|
278 | AnyEvent::DNS, thus avoiding an exception (reported by Anatoly |
|
|
279 | K. Sharifulin). |
|
|
280 | |
|
|
281 | 4.2 Sat Jul 12 22:42:11 CEST 2008 |
|
|
282 | - support an "interval" argument for repeating timers. |
|
|
283 | - fix a bug in the pure perl event loop that caused timers to |
|
|
284 | be delayed under some circumstances. |
|
|
285 | - implement optional argument checking via PERL_ANYEVENT_STRICT. |
|
|
286 | - not importing anything from Event causes it to malfunction |
|
|
287 | and not use Time::HiRes, work around this bug. |
|
|
288 | - more bug workarounds against the endlessly broken Tk module, |
|
|
289 | seems to work now on 32 bit machines, but less so on 64 bit. |
|
|
290 | - minor optimisations applied to most backends. |
|
|
291 | |
|
|
292 | 4.161 Fri Jul 4 14:24:48 CEST 2008 |
|
|
293 | - fixed the t/handle/02_write.t, it read after write, with the |
|
|
294 | assumption that all data could be written before read |
|
|
295 | (reported by Jim Razmus). |
|
|
296 | |
|
|
297 | 4.160 Thu Jul 3 04:02:21 CEST 2008 |
|
|
298 | - re-registering signal watchers after unregistering one did |
|
|
299 | not work if a backend falls back on AnyEvent's default |
|
|
300 | implementation (also affected child watchers). |
|
|
301 | - new AnyEvent::Handle options: autocork and no_delay. |
|
|
302 | |
|
|
303 | 4.152 Sun Jun 22 14:15:44 CEST 2008 |
|
|
304 | - allow for 32-bit perls that implement shifts differently |
|
|
305 | on different architectures in parse_ipv4 (reported and |
|
|
306 | analysed by Keiichi DAIBA). |
|
|
307 | |
|
|
308 | 4.151 Fri Jun 6 17:34:24 CEST 2008 |
|
|
309 | - make sure specifying _only_ on_read and never pushing reads |
|
|
310 | works. |
|
|
311 | |
|
|
312 | 4.15 Fri Jun 6 13:00:46 CEST 2008 |
|
|
313 | - the pure perl backend would keep some watchers alive when more than |
|
|
314 | one watcher was registered for the same fd. |
|
|
315 | - new "packstring" and "storable" read and write types |
|
|
316 | for AnyEvent::Handle. |
|
|
317 | - allow on_eof handler to be called after on_error with EPIPE returns. |
|
|
318 | - do not immediately call on_read callback in handle constructor. |
|
|
319 | |
|
|
320 | 4.14 Thu Jun 5 20:29:31 CEST 2008 |
|
|
321 | - Fixed a bug in DNS SRV priority sorting. |
|
|
322 | - AnyEvent::Util::guard now reports runtime errors while |
|
|
323 | executing the guard block as warnings. |
|
|
324 | - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle. |
| 2 | |
325 | |
| 3 | 4.13 Thu Jun 5 00:47:59 CEST 2008 |
326 | 4.13 Thu Jun 5 00:47:59 CEST 2008 |
| 4 | - AnyEvent::DNS only followed cname chains with length 2, |
327 | - AnyEvent::DNS only followed cname chains with length 2, |
| 5 | contrary to documentation. bumped it up to 10, thanks to |
328 | contrary to documentation. bumped it up to 10, thanks to |
| 6 | microsoft, the current broken-dns-config-king. |
329 | microsoft, the current broken-dns-config-king. |
| … | |
… | |
| 125 | - support service names in tcp_connect and tcp_server. |
448 | - support service names in tcp_connect and tcp_server. |
| 126 | - work around netbsd/openbsd bug causing unix sockets not to |
449 | - work around netbsd/openbsd bug causing unix sockets not to |
| 127 | become writable until empty. |
450 | become writable until empty. |
| 128 | - only make a dummy read for error detection when getpeername |
451 | - only make a dummy read for error detection when getpeername |
| 129 | returns ENOTCONN. |
452 | returns ENOTCONN. |
| 130 | - defend AnyEvent::Util::inet_aton against Coro::LWP brutal |
453 | - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal |
| 131 | override. |
454 | override. |
| 132 | - new AnyEvent::Guard::cancel. |
455 | - new AnyEvent::Guard::cancel. |
| 133 | |
456 | |
| 134 | 3.5 Sat May 17 23:17:14 CEST 2008 |
457 | 3.5 Sat May 17 23:17:14 CEST 2008 |
| 135 | - gracefully deal with systems not defining CLOCK_MONOTONIC |
458 | - gracefully deal with systems not defining CLOCK_MONOTONIC |
| … | |
… | |
| 149 | |
472 | |
| 150 | 3.41 Sun May 11 19:53:13 CEST 2008 |
473 | 3.41 Sun May 11 19:53:13 CEST 2008 |
| 151 | - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. |
474 | - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. |
| 152 | |
475 | |
| 153 | 3.4 Sun May 11 00:29:25 CEST 2008 |
476 | 3.4 Sun May 11 00:29:25 CEST 2008 |
| 154 | - complete re-implementation of condvars with a hopefully much mroe useful |
477 | - complete re-implementation of condvars with a hopefully much more useful |
| 155 | API (backward compatible functions exist but are undocumented). |
478 | API (backward compatible functions exist but are undocumented). |
| 156 | - AnyEvent::Handle fully rewritten with a hopefully more powerful API |
479 | - AnyEvent::Handle fully rewritten with a hopefully more powerful API |
| 157 | (still subject to change). |
480 | (still subject to change). |
| 158 | - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
481 | - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
| 159 | - provide AnyEvent::post_detect and @AnyEvent::post_detect, which |
482 | - provide AnyEvent::post_detect and @AnyEvent::post_detect, which |
