ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent/Changes
Revision: 1.217
Committed: Tue Feb 10 13:58:49 2009 UTC (15 years, 3 months ago) by root
Branch: MAIN
Changes since 1.216: +1 -0 lines
Log Message:
*** empty log message ***

File Contents

# Content
1 Revision history for Perl extension AnyEvent.
2
3 TODO: ssl-bug (write can caquse read data to become available, in_drain causes skip, no progress detected).
4 - raise EBADMSG error on JSON decoding errors.
5 - fix some minor manpage bugs (reported by Maximilian Gaß).
6 - speed up select bitmask parsing quite a bit in the pure perl
7 backend.
8 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
9
10 4.331 Tue Jan 6 21:07:25 CET 2009
11 - socketpair fails on many vista machines because vista has
12 completely broken accept/getpeername and getsockname functions,
13 so we provide our own socketpair emulation that kind of works
14 (AnyEvent::Util::portable_pipe).
15 - new function: AnyEvent::Util::portable_socketpair.
16 - take advantage of the Guard module if it exists.
17
18 4.33 Fri Nov 21 02:35:40 CET 2008
19 - AnyEvent::Strict did errornously flag a fileno of 0 as
20 illegal.
21 - reduce memory usage and slightly speed up the pure perl backend
22 by only storing the file descriptor, not the file handle.
23 - add missing autocork method to AnyEvent::Handle (reported
24 by Adam Rosenstein).
25 - AnyEvent::DNS->resolve errornously documented an $rcode result
26 argument, but there isn't (spotted by Henrik Krohns).
27 - the naked truth about Tk - it's basically unmaintained.
28
29 4.32 Mon Nov 3 22:46:32 CET 2008
30 - fix AnyEvent::Socket::resolve_sockaddr to properly support
31 unix sockets again and choose a proper default for
32 the socket type.
33 - fix call to resolve_sockaddr in tcp_connect.
34
35 4.31 Thu Oct 30 04:41:48 CET 2008
36 - implemented AnyEvent::Handle->destroy method.
37 - hint about unexpected effects in TLS mode.
38 - speed up AnyEvent::Impl::Perl by using more arrays
39 instead of hashes without the slightest loss of readability :)
40 - work around a perl argument refcounting bug.
41 - with some perl patches applied, AnyEvent now seems leak-free.
42
43 4.3 Fri Oct 3 09:18:43 CEST 2008
44 - AnyEvent will now install a no-op signal handler for SIGPIPE,
45 unless one has been installed already.
46 - warn about not loaidng AnyEvent::Impl::POE early enough
47 (patch by Adam Rosenstein).
48 - fixed a great number of bugs and corner cases in AnyEvent::Handle:
49 - fix a bug in where in SSL connect mode, the
50 client would first wait for some data by the server and
51 otherwise hang. (reported and analyzed in an absolutely
52 exemplary manner by Adam Rosenstein).
53 - fix a bug in where SSL EOF would not be treated
54 as stream EOF, putting the connection into a hung state.
55 (reported and analyzed in an absolutely exemplary manner by Adam
56 Rosenstein).
57 - fix a potential segfault when the TLS context
58 would go missing in a read callback (e.g. due to stoptls) -
59 Net::SSLeay of course makes no type checking whatsoever.
60 - AnyEvent::Handle will not stop the read watcher in TLS mode.
61 - AnyEvent::Handle->stoptls will send a TLS close notify
62 instead of simply destroying the stream now.
63 - fix a bug where large blocks of data written in TLS mode
64 would not be sent unless triggered by receives.
65 - on_drain will now take the tls write buffer into account.
66 - SSL operations have been streamlined a bit and should be faster
67 now, more to come, though.
68 - added a FAQ section, will document nontrivial issues in there.
69
70 4.234 Mon Sep 29 04:08:13 CEST 2008
71 - fix child watcher documentation: only child exits will be reported,
72 no trace events.
73 - mention SIGPIPE in AnyEvent::Handle.
74 - perl backend will now use POSIX::times as monotonic clock
75 when available and the monotonic clock option is not.
76 - run even without Time::HiRes in some cases now, but do not
77 provide sub-second accuracy (all *supported* perl versions have
78 Time::HiRes).
79
80 4.233 Fri Aug 22 01:48:05 CEST 2008
81 - fix a bug in the testsuite, causing a hang.
82 - clarified lots of AnyEvent::Handle constructor arguments.
83
84 4.232 Thu Aug 21 20:44:25 CEST 2008
85 - fix a bug in AnyEvent::Handle that could cause two fatal error
86 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
87 (testcase by Leon Brocard).
88 - clarified on_eof/on_error documentation a bit.
89
90 4.231 Tue Jul 29 13:12:15 CEST 2008
91 - remove some debugging code left in AnyEvent::Util::fork_call
92 (and no, it's impossible to implement with the broken windows
93 perls without resource leaks or worse).
94
95 4.23 Tue Jul 29 12:19:59 CEST 2008
96 - document the first parameter passed to condvar callbacks to be
97 the callback.
98 - add AnyEvent::Socket::{ntoa,aton} aliases.
99 - optimize the AE::Handle->push_read (line) for the default
100 eol marker.
101 - optimize push_read (packstring|storable) for small packets.
102 - invoke on_error callback when no on_eof callback is set.
103 - fix a bug in push_read (storable) of unknown impact.
104
105 4.22 Sun Jul 20 16:34:13 CEST 2008
106 - new function AnyEvent::Socket::parse_hostport.
107 - as the bulkheads at microsoft can't even get getprotobyname reliably
108 working on their shitty fucking broken os we need to hardcode
109 some common protocol numbers in AnyEvent::Socket. How can
110 people even bother with such a pile of shit as windows.
111
112 4.21 Thu Jul 17 14:40:05 CEST 2008
113 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
114 which is incompatible to previous invocation syntax.
115 - work around more windows perl bugs in fork_call. I cannot imagine
116 why anybody would bother implementing fork in such an obviously
117 unusable way.
118 - avoid calling $do_retry if the request has already finished in
119 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
120 K. Sharifulin).
121
122 4.2 Sat Jul 12 22:42:11 CEST 2008
123 - support an "interval" argument for repeating timers.
124 - fix a bug in the pure perl event loop that caused timers to
125 be delayed under some circumstances.
126 - implement optional argument checking via PERL_ANYEVENT_STRICT.
127 - not importing anything from Event causes it to malfunction
128 and not use Time::HiRes, work around this bug.
129 - more bug workarounds against the endlessly broken Tk module,
130 seems to work now on 32 bit machines, but less so on 64 bit.
131 - minor optimisations applied to most backends.
132
133 4.161 Fri Jul 4 14:24:48 CEST 2008
134 - fixed the t/handle/02_write.t, it read after write, with the
135 assumption that all data could be written before read
136 (reported by Jim Razmus).
137
138 4.160 Thu Jul 3 04:02:21 CEST 2008
139 - re-registering signal watchers after unregistering one did
140 not work if a backend falls back on AnyEvent's default
141 implementation (also affected child watchers).
142 - new AnyEvent::Handle options: autocork and no_delay.
143
144 4.152 Sun Jun 22 14:15:44 CEST 2008
145 - allow for 32-bit perls that implement shifts differently
146 on different architectures in parse_ipv4 (reported and
147 analysed by Keiichi DAIBA).
148
149 4.151 Fri Jun 6 17:34:24 CEST 2008
150 - make sure specifying _only_ on_read and never pushing reads
151 works.
152
153 4.15 Fri Jun 6 13:00:46 CEST 2008
154 - the pure perl backend would keep some watchers alive when more than
155 one watcher was registered for the same fd.
156 - new "packstring" and "storable" read and write types
157 for AnyEvent::Handle.
158 - allow on_eof handler to be called after on_error with EPIPE returns.
159 - do not immediately call on_read callback in handle constructor.
160
161 4.14 Thu Jun 5 20:29:31 CEST 2008
162 - Fixed a bug in DNS SRV priority sorting.
163 - AnyEvent::Util::guard now reports runtime errors while
164 executing the guard block as warnings.
165 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
166
167 4.13 Thu Jun 5 00:47:59 CEST 2008
168 - AnyEvent::DNS only followed cname chains with length 2,
169 contrary to documentation. bumped it up to 10, thanks to
170 microsoft, the current broken-dns-config-king.
171 - AnyEvent::DNS didn't check for socket return status
172 but instead relied on perl not creating filehandles in that
173 case - too bad it gives you a bogus file handle (reported
174 and analysed by Vladimir Timofeev).
175 - fix queue management logic in AnyEvent::Handle:
176 when on_read was registered, the queue was empty
177 and no progress could be made AnyEvent::Handle would
178 enter an endless loop.
179 - correctly start reading again when the handle became
180 busy again after idling.
181 - correctly treat tls shut-downs as EOF condition
182 (for the time being, would be nice to have a callback for that).
183 - correctly call eof callback under all conditions
184 (hopefully).
185 - the timeout callback did not expect that $self can go
186 away any time.
187
188 4.12 Tue Jun 3 10:58:04 CEST 2008
189 - include AnyEvent::Intro, a tutorial for anyevent,
190 anyevent::socket and anyevent::handle.
191 - allow more options in on_error.
192
193 4.11 Fri May 30 23:42:25 CEST 2008
194 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
195 reverse_lookup and reverse_verify to replace it, support
196 v4mapped and v4compat addresses.
197 - provide more documentation for the resolver class.
198 - really replace longest run of :0: by :: in format_address,
199 also properly convert :: and ::1 again.
200 - support NAPTR record name and decode it.
201 - implement random weight sampling for SRV records, as per
202 rfc 2782.
203 - correctly abort on srv-record targets of ".".
204 - added AnyEvent::DNS::wait_for_slot.
205 - in the unlikely event of a virtual circuit connection
206 being invalidated by a delayed udp reply, AnyEvent::DNS
207 could die. this has been fixed.
208 - plug a probable memleak in the DNS vc code.
209 - use configured timeout also for VC request phase.
210 - implement timeout and max_outstanding methods
211 for resolver class.
212 - update version numbers in all modules.
213
214 4.1 Thu May 29 05:45:40 CEST 2008
215 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
216 which mirrors their purpose better (old names still available
217 for a while). Also moved AnyEvent::DNS::addr to
218 AnyEvent::Socket::resolve_sockaddr.
219 - implement AnyEvent->time and AnyEvent->now.
220 - fix IPv6 support in pack_sockaddr.
221 - officially un-experimentalise all newly introduced code.
222 - support unix domain sockets everywhere by specifying
223 a host of "unix/" and the pathname as service.
224 - implement an activity timeout in AnyEvent::Handle.
225 - added a regex read type to AnyEvent::Handle.
226 - added a json read/write type to AnyEvent::Handle.
227 - always croak in on_error in AnyEvent::Handle.
228 - document how to subclass AnyEvent::Handle.
229 - implement AnyEvent::Util::fork_call.
230 - add support for IPv6 nameservers and nameserver statements.
231 - work around _yet_ _another_ windows perl bug in where empty
232 select masks cause errors in the pure perl backend (this
233 workaround cannot be implemented for the other event loops,
234 but EV also works around this).
235 - supply AnyEvent::Socket::address_type and make good use of it.
236 - clarify and fix the Handle documentation w.r.t $self vs. $handle
237 vs. callback arguments.
238 - add some recommends to the META.yml.
239
240 4.05 Mon May 26 19:44:06 CEST 2008
241 - some platforms ignore the rfc and prepend an extra sa_len member to
242 the sockaddr structure, cater for those. also use sockaddr_family,
243 didn't know it was there, but comes in handy for the workaround.
244 - undo the import WIN32 hack.
245
246 4.04 Mon May 26 08:03:31 CEST 2008
247 - try to work around yet another windows bug: failed connects are reported
248 as if it were out-of-band data. windows users: you suck. Only
249 EV, Glib, Event and the pure perl backend can handle this condition.
250 - optimize write algorithm in AnyEvent::Handle.
251 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
252 - tcp_connect now properly iterates through all targets.
253 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
254 support entirely. also hardcode some AF_INET6 constants for the
255 sake of perl 5.8.
256 - vastly improved nameserver/suffix detection algorithm for windows
257 (still a hack...)
258 - try to use a fallback dns server if no dns servers could be found.
259 - splatter around some local $SIG{__DIE__}'s to work around bad
260 die handlers.
261
262 4.03 Sun May 25 05:05:57 CEST 2008
263 - work around perl bugs on windows, where perl returns undocumented
264 error codes for sysread, syswrite etc.
265 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
266 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
267 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
268 - updated documentation.
269 - lots of workarounds for perl 5.8 quirks.
270 - implement a more flexible type system for both reads and writes
271 in AnyEvent::Handle.
272 - support IPv6 in tcp_server.
273 - pass local hostname and port to the prepare callback in tcp_server.
274 - make types extendable in AnyEvent::Handle.
275 - croak, not die, in AnyEvent::Handle;
276
277 4.0 Sat May 24 19:58:08 CEST 2008
278 - added (experimental) AnyEvent::DNS module.
279 - added (experimental) AnyEvent::Socket module.
280 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
281 - make condvar->end call ->send as documented.
282 - make condvar constructor support the documented cb => argument.
283 - overload condvars so you can use them instead of code references.
284 - support more of the POSIX forms instead of a pure dotted quad for
285 verbatim ip addresses.
286 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
287 - try very hard not to freeze on broken windows perls.
288 - make on_eof optional in AnyEvent::Handle.
289 - support service names in tcp_connect and tcp_server.
290 - work around netbsd/openbsd bug causing unix sockets not to
291 become writable until empty.
292 - only make a dummy read for error detection when getpeername
293 returns ENOTCONN.
294 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
295 override.
296 - new AnyEvent::Guard::cancel.
297
298 3.5 Sat May 17 23:17:14 CEST 2008
299 - gracefully deal with systems not defining CLOCK_MONOTONIC
300 (as opposed to not supporting the clock itself, which worked in 3.41).
301 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
302 helpers, to replace the ill-designed AnyEvent::Socket class.
303 - added some examples (eg/connect, eg/handle).
304 - added AnyEvent::Util::guard.
305 - AnyEvent::Util now exports most of its functions.
306 - fixed a bug in *_read_chunk, where the callback was called with $self as
307 first and second argument.
308 - simplified and fixed regex handling in AnyEvent::Handle::*_read_line.
309 - fix manpage condvars examples to use new syntax.
310 - considerably improve inet_aton with EV::ADNS backend to support
311 cname chains (common dns configuration bug, e.g. with www.google.de)
312 and aaaa records if no a record could be found.
313
314 3.41 Sun May 11 19:53:13 CEST 2008
315 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
316
317 3.4 Sun May 11 00:29:25 CEST 2008
318 - complete re-implementation of condvars with a hopefully much more useful
319 API (backward compatible functions exist but are undocumented).
320 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
321 (still subject to change).
322 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
323 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which
324 allows module authors to avoid forcing event loop detection.
325 (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example).
326 - remove coro backends: Coro now provides generic support for AnyEvent
327 via Coro::AnyEvent.
328
329 3.3 Mon Apr 28 09:51:06 CEST 2008
330 - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules.
331 - fix a bug in the pure perl backend that kept watchers alive
332 when multiple watchers were registered for the same fh/poll combo.
333 - add a benchmark section showing AnyEvent overhead and comparing
334 the different event loops with each other.
335 - prefer pure perl over tk when autoprobing, as it's about as fast,
336 but doesn't crash with many watchers.
337 - declare Qt support non-experimental.
338 - clarify the confusing section about the file descriptor being
339 kept alive.
340 - document the race between loading of an event module and
341 child processes exiting.
342 - support POE as "backend" (with some caveats, POE is not generic
343 enough, and darn slow).
344 - support Wx and Prima through POE.
345 - optimise perl backend to use 20% less memory and take advantage
346 of typical timeout behaviour. It can now compete with
347 select/poll-based C event loops in most cases (it is usually
348 faster than Event and Glib :).
349 - roughly cut EV memory use in half and increase its speed by 30%,
350 by removing undocumented functionality. Did something similar
351 to other event loops where possible.
352
353 3.2 Thu Apr 24 10:10:40 CEST 2008
354 - do not die when anyevent watchers are destroyed while running callbacks
355 in the pureperl backend (could only happen when two watchers are registered
356 for the same fh/poll combo).
357 - support autoloading for child watchers, was broken in all versions.
358 - implement PERL_ANYEVENT_MODEL env variable.
359 - (experimental) implement interface to the (very crashy/buggy) Event::Lib module.
360 - (experimental) implement interface to the Qt module (cannot be autoprobed).
361 - this release is IO:AnyEvent-proof.
362
363 3.12 Tue Apr 22 07:11:46 CEST 2008
364 - reinstate AUTHOR section that got lost somehow.
365 - do not hang in the testsuite with badly broken perls (activestate,
366 strawberry...), but instead diagnose the problem and continue.
367 - use INT instead of CHLD in an attempt to support broken windows perls
368 better (this decreases the test quality, unfortunately).
369 - do not send the signal to the process group
370 (no problem for CHLD, bad for INT :).
371
372 3.11 Sat Apr 19 06:57:31 CEST 2008
373 - major documentation rework.
374 - document the fact that child watchers only watch for zombies.
375 - fix the child watcher example.
376
377 3.1 Wed Apr 16 17:09:01 CEST 2008
378 - work around recurring bugs in Tk by dup'ing filehandles, the
379 only method with good success chances on Tk (the bugs apparently
380 don't get fixed anytime soon).
381 - lift the restriction of only one watcher per fh direction
382 (as the Tk bug workaround also lifts it and only Tk imposed
383 such strong limits).
384 - changed probe order to prefer coro adaptors.
385 - explain why recursion into the event loop is not supported
386 unless the backend supports it (only Coro::EV does without
387 any restrictions...).
388 - add simple manpages for all backend modules.
389
390 3.0 Mon Apr 7 21:30:23 CEST 2008
391 - Coro::Signal changed semantics, roll our own, also cleaning
392 up the Coro implementation in general.
393 - rename Coro backend to CoroEvent.
394 - add some decision helping paragraph to the manpage that should
395 help people to decide whether AnyEvent is the right thing for them.
396
397 2.9 Mon Jan 28 13:31:54 CET 2008
398 - update for EV 3.0 API changes.
399
400 2.8 Sun Nov 25 15:06:03 CET 2007
401 - waitpid can validly return 0. accept this fact of life
402 instead of reporting it to any watchers.
403
404 2.7 Fri Nov 23 11:41:14 CET 2007
405 - force use of AnyEvent::Impl::Perl in testsuite, there is
406 too much breakage outside AnyEvent.
407 - deliver signals synchronously in AnyEvent::Impl::Perl
408 or any other event loop that relies on AnyEvents child
409 watcher emulation. *Could* help with hanging testsuite
410 (except when Event or EV are installed).
411
412 2.6 Fri Nov 9 20:36:35 CET 2007
413 - fix bug in testsuite.
414 - move EV adaptor modules to AnyEvent.
415 - add Coro+EV adaptor module.
416
417 2.55 Tue Nov 6 17:41:32 CET 2007
418 - add EV to the list of supported event models.
419 - do not auto-reset pid watchers, pass pid and status to them.
420 - allow a pid of zero to watch for all children in child watchers.
421
422 2.54 Wed Jul 18 17:36:23 CEST 2007
423 - work around a perl bug that results in
424 BEGIN not safe after errors--compilation aborted without
425 any discernible reason or error message by once
426 again not use'ing strict.
427
428 2.53 Sun Jul 8 10:51:53 CEST 2007
429 - make Glib interface work again (spotted by elmex).
430
431 2.52 Wed Mar 7 18:36:16 CET 2007
432 - child watchers stopped working permanently when
433 all current child watchers were destroyed.
434
435 2.51 Mon Dec 11 21:33:24 CET 2006
436 - work around bugs in perl where eval "require Module" returns
437 true even if the module couldn't be loaded :/.
438
439 2.5 Mon Dec 11 02:15:28 CET 2006
440 - avoid $AUTOLOAD because many perls corrupt it.
441 - AnyEvent::detect forces autodetection.
442 - implement signal watchers (experimental)
443 (Tk does not support async signals (<= 804.027 at least),
444 so they might get delayed indefinitely on Tk).
445 - implement child watchers (experimental).
446 - moved default condvar implementation into base module, simplifying
447 most implementation modules.
448
449 2.1 Fri Nov 24 15:50:48 CET 2006
450 - better docs.
451 - simple testsuite.
452 - added AnyEvent->one_event method for special purposes.
453
454 2.0 Wed Nov 1 02:21:30 CET 2006
455 - INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both.
456 - INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore.
457 - EXPERIMENTAL:
458 new pure-perl model added, AnyEvent now always finds a suitable
459 event model.
460 - improved documentation.
461 - reduced cpu overhead.
462
463 1.02 Fri Jan 13 14:15:40 CET 2006
464 - add COPYING to clarify license.
465
466 1.01 Sun Jan 8 05:48:09 CET 2006
467 - correctly cancel Event timers, avoid memory leaks.
468
469 1.0 Sun Jan 8 05:40:58 CET 2006
470 - allow third-party interfaces to be added at runtime.
471 - fix version number.
472
473 0.9 Fri Dec 30 02:25:41 CET 2005
474 - fixed errors in the documentation.
475 - going back to using $VERSION as "load check",
476 as suggested by Jörn Reder. Let's see why I stopped using
477 this method.
478 - add 'hup' to the list of events to receive for glib
479 and generate 'rw' events for it, as some versions of
480 glib seem to require this, while this breaks other
481 (older) versions of glib.
482 - implemented PERL_ANYEVENT_VERBOSE env variable.
483
484 0.3 Sun Dec 4 10:43:33 CET 2005
485 - no changes to the code.
486 - fixed documentation.
487 - clarified some corner cases.
488 - vastly improved documentation with a long example.
489
490 0.2 Thu Dec 1 22:18:49 CET 2005
491 - check whether a valid method was called, to avoid endless loops.
492 - move condvars into their own class (cleaner, safer).
493 - don't die when some modules are not available.
494 - Tk is too broken to allow for words. Will work once Tk works.
495
496 0.1 Thu Dec 1 19:44:57 CET 2005
497 - spiced up and cleaned up for release.
498
499 0.01 Tue Sep 21 18:12:09 CEST 2004
500 - original version; copied from Convert::Scalar.
501