ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent/Changes
(Generate patch)

Comparing AnyEvent/Changes (file contents):
Revision 1.41 by root, Wed Apr 16 14:39:00 2008 UTC vs.
Revision 1.290 by root, Sat Jul 18 05:19:09 2009 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
34.85 Sat Jul 18 06:16:14 CEST 2009
4 - nail the signal race problem in perl once and for all
5 (see $AnyEvent::MAX_SIGNAL_LATENCY).
6 - take advantage of Async::Interrupt if it is available.
7 - load Time::HiRes and Guard modules on demand only.
8 - add optional/recommended modules section to AnyEvent
9 documentation.
10 - reduce memory usage considerably (and reduce startup penalty)
11 by not using "strict", "warnings" and "overload" modules.
12 - work around buggy windows/openbsd perls and provide EBADMSG
13 and EPROTO ourselves when missing.
14 - improve perl 5.6 compatibility of the core event loop.
15 - made Net::SSLeay version 1.33 a soft requirement.
16
174.83 Fri Jul 17 16:56:26 CEST 2009
18 - implement AnyEvent::Socket::getprotobyname.
19 - AnyEvent::CondVar's will now detect recursive blocking
20 waits and will croak, as too many people fall into
21 this trap.
22 - AnyEvent::Handle will now call ->destroy on itself after
23 executing the on_error callback, instead of doing some
24 half-baked internal shutdown, for fatal errors.
25 - clarify on_eof behaviour w.r.t. the read queue and
26 on_read callbacks.
27 - ignore some possible spurious wake-ups in tcp_connect.
28
294.82 Sat Jul 11 00:34:55 CEST 2009
30 - POE and Event backends didn't accept some callable objects as
31 callbacks.
32 - use Config module instead of POSIX module to detect signal names
33 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
34 module doesn't even have all POSIX signals :/.
35 - use more workarounds around the many refcnt/corruption bugs in
36 Event::Lib.
37 - work around a race condition in perl's select, causing t/03_child.t
38 to rarely fail.
39
404.81 Thu Jul 9 10:30:30 CEST 2009
41 - AnyEvent::Handle didn't properly diagnose write errors
42 (it expected -1 from syswrite, how lame... :).
43 - support file descriptors in addition to file handles
44 in AnyEvent->io.
45 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
46 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
47 and PERL_ANYEVENT_CA_PATH.
48 - provide a sensible synopsis section for AnyEvent::TLS.
49 - add a "supported backends" section to the manpage.
50 - added simple io watcher test to testsuite, using a
51 portable_socketpair.
52 - tried to improve the stability of the Event::Lib backend,
53 YMMV.
54
554.8 Mon Jul 6 23:45:16 CEST 2009
56 - AnyEvent::DNS did not properly follow CNAME records with
57 uppercase targets.
58 - AnyEvent::DNS would errornously return AAAA records
59 with v4 mapped addresses (a faulty record) as ipv4 addresses,
60 causing AnyEvent::Socket to throw an exception.
61 - added new module AnyEvent::TLS for easier SSL/TLS context
62 creation, with many options including hostname verification,
63 secure default configuration, lots of documentation and,
64 predefined diffie-hellman keys for perfect forward security
65 and much more. get it while it's still fresh!
66 - use AnyEvent::TLS in AnyEvent::Handle for context management.
67 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
68 so AnyEvent::Socket users have smaller memory footprint
69 in the common case.
70 - add AnyEvent::Handle->push_shutdown method.
71 - add an additional error message parameter to AnyEvent::Handle's
72 on_error callback (for TLS, $! is still available).
73 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
74 - make AnyEvent::Handle more robust against early conenction
75 failures (during new), and return C<undef> in that case
76 from the constructor.
77 - AnyEvent::Handle will now try to load only JSON::XS first,
78 then fall back to JSON.
79 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
80 and because it was documented that way.
81
824.451 Fri Jul 3 00:28:58 CEST 2009
83 - do not clear rbuf when shutting down an AnyEvent::Handle
84 object - doing so breaks AnyEvent::HTTP.
85
864.45 Mon Jun 29 22:59:26 CEST 2009
87 - a write error could cause AnyEvent::Handle to create
88 an I/O watcher with an undefined $fh.
89 - special-case mapped ipv4 addresses in both
90 AnyEvent::Socket::format_address and parse_address,
91 to treat them just like ipv4 addresses.
92 - updated and overhauled the AnyEvent::Intro doc.
93 - implement AnyEvent::Socket::format_ipv4/ipv6.
94 - slightly speed up portable_pipe/socketpair.
95 - expand condvar begin/end documentation.
96
974.42 Fri Jun 26 08:32:18 CEST 2009
98 - preliminary and neccesarily incomplete support for IO::Async.
99 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
100 case it was set to IGNORE, to ensure we can catch child statuses
101 even when the calling env acts stupidly.
102 - updated benchmarks with IO::Async, which performs very well.
103
1044.412 Wed Jun 24 01:35:57 CEST 2009
105 - support an "untaint" attribute for AnyEvent::DNS
106 and set it on the default resolver.
107 - implement and document AnyEvent::Impl::Perl::loop.
108 - remove all anyevent-env variables from %ENV
109 when running in tainted mode.
110 - mention and extend the IO::Lambda benchmark.
111
1124.411 Sun Jun 7 18:48:02 CEST 2009
113 - do not try to use F_SETFD on windows.
114
1154.41 Thu May 14 06:40:11 CEST 2009
116 - work around issues in older perls (5.8.0?) when
117 a signal handler is deleted from the %SIG hash.
118 - use POSIX::_exit in child test, to avoid running
119 destructors.
120 - speed up CHLD handling by relying on SIGCHLD being
121 synchronously delivered, even when we roll our own
122 implementation.
123 - AnyEvent::DNS: add the "dname" resource record name
124 for cosmetic reasons.
125
1264.4 Sun Apr 26 20:12:33 CEST 2009
127 - implemented idle watchers, where applicable.
128 - AnyEvent->time died when Event backend was in use.
129 - fix a memleak in the Tk backend.
130 - sped up Tk timer handling.
131 - clip negative "after" values to 0 in AnyEvent::Impl::Event
132 to avoid spamming out warning messages.
133 - fix Qt timers without interval.
134 - avoid Qt zero-timeout special case, in old libqt's.
135
1364.352 Mon Apr 20 16:31:11 CEST 2009
137 - fix AnyEvent::Strict error messages for child watchers.
138 - fix/update Makefile.PL "recommends" versions.
139 - add AnyEvent->now_update.
140
1414.351 Sat Apr 11 07:56:14 CEST 2009
142 - actually make the signal pipe work under win32.
143 - localise $! in signal handler to avoid changing $!, although
144 perl itself does not seem to save/restore errno either.
145 - set the cloexec flag on the signal pipe (normally set by perl too).
146
1474.35 Fri Mar 27 11:48:20 CET 2009
148 - event models relying on AnyEvent's signal watcher emulation did
149 invoke the callback asynchronously, contradicting documentation
150 and causing signals to get lost (this includes AnyEvent's own
151 event loop). AnyEvent now uses the standard pipe trick to make
152 callback execution synchronous to the event loop.
153 - AnyEvent::Handle didn't free TLS context data on DESTROY
154 (patch by Pavel Shaydo).
155 - work around the ever-incompatibly-changing API of MakeMaker.
156 - document that changing global variables without restoring them
157 is a bad idea in Perl (noted by Adam Rosenstein).
158 - AnyEvent::Strict now barfs if ->io is passed a file.
159
1604.34 Thu Feb 12 18:32:45 CET 2009
161 - separately buffer TLS read data, as otherwise the read queue could
162 deadlock as receiving data is not expected while draining the
163 read queue (which cna only happen with TLS).
164 - raise EBADMSG error on JSON decoding errors.
165 - fix some minor manpage bugs (reported by Maximilian Gaß).
166 - speed up select bitmask parsing quite a bit in the pure perl
167 backend.
168 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
169
1704.331 Tue Jan 6 21:07:25 CET 2009
171 - socketpair fails on many vista machines because vista has
172 completely broken accept/getpeername and getsockname functions,
173 so we provide our own socketpair emulation that kind of works
174 (AnyEvent::Util::portable_pipe).
175 - new function: AnyEvent::Util::portable_socketpair.
176 - take advantage of the Guard module if it exists.
177
1784.33 Fri Nov 21 02:35:40 CET 2008
179 - AnyEvent::Strict did errornously flag a fileno of 0 as
180 illegal.
181 - reduce memory usage and slightly speed up the pure perl backend
182 by only storing the file descriptor, not the file handle.
183 - add missing autocork method to AnyEvent::Handle (reported
184 by Adam Rosenstein).
185 - AnyEvent::DNS->resolve errornously documented an $rcode result
186 argument, but there isn't (spotted by Henrik Krohns).
187 - the naked truth about Tk - it's basically unmaintained.
188
1894.32 Mon Nov 3 22:46:32 CET 2008
190 - fix AnyEvent::Socket::resolve_sockaddr to properly support
191 unix sockets again and choose a proper default for
192 the socket type.
193 - fix call to resolve_sockaddr in tcp_connect.
194
1954.31 Thu Oct 30 04:41:48 CET 2008
196 - implemented AnyEvent::Handle->destroy method.
197 - hint about unexpected effects in TLS mode.
198 - speed up AnyEvent::Impl::Perl by using more arrays
199 instead of hashes without the slightest loss of readability :)
200 - work around a perl argument refcounting bug.
201 - with some perl patches applied, AnyEvent now seems leak-free.
202
2034.3 Fri Oct 3 09:18:43 CEST 2008
204 - AnyEvent will now install a no-op signal handler for SIGPIPE,
205 unless one has been installed already.
206 - warn about not loaidng AnyEvent::Impl::POE early enough
207 (patch by Adam Rosenstein).
208 - fixed a great number of bugs and corner cases in AnyEvent::Handle:
209 - fix a bug in where in SSL connect mode, the
210 client would first wait for some data by the server and
211 otherwise hang. (reported and analyzed in an absolutely
212 exemplary manner by Adam Rosenstein).
213 - fix a bug in where SSL EOF would not be treated
214 as stream EOF, putting the connection into a hung state.
215 (reported and analyzed in an absolutely exemplary manner by Adam
216 Rosenstein).
217 - fix a potential segfault when the TLS context
218 would go missing in a read callback (e.g. due to stoptls) -
219 Net::SSLeay of course makes no type checking whatsoever.
220 - AnyEvent::Handle will not stop the read watcher in TLS mode.
221 - AnyEvent::Handle->stoptls will send a TLS close notify
222 instead of simply destroying the stream now.
223 - fix a bug where large blocks of data written in TLS mode
224 would not be sent unless triggered by receives.
225 - on_drain will now take the tls write buffer into account.
226 - SSL operations have been streamlined a bit and should be faster
227 now, more to come, though.
228 - added a FAQ section, will document nontrivial issues in there.
229
2304.234 Mon Sep 29 04:08:13 CEST 2008
231 - fix child watcher documentation: only child exits will be reported,
232 no trace events.
233 - mention SIGPIPE in AnyEvent::Handle.
234 - perl backend will now use POSIX::times as monotonic clock
235 when available and the monotonic clock option is not.
236 - run even without Time::HiRes in some cases now, but do not
237 provide sub-second accuracy (all *supported* perl versions have
238 Time::HiRes).
239
2404.233 Fri Aug 22 01:48:05 CEST 2008
241 - fix a bug in the testsuite, causing a hang.
242 - clarified lots of AnyEvent::Handle constructor arguments.
243
2444.232 Thu Aug 21 20:44:25 CEST 2008
245 - fix a bug in AnyEvent::Handle that could cause two fatal error
246 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
247 (testcase by Leon Brocard).
248 - clarified on_eof/on_error documentation a bit.
249
2504.231 Tue Jul 29 13:12:15 CEST 2008
251 - remove some debugging code left in AnyEvent::Util::fork_call
252 (and no, it's impossible to implement with the broken windows
253 perls without resource leaks or worse).
254
2554.23 Tue Jul 29 12:19:59 CEST 2008
256 - document the first parameter passed to condvar callbacks to be
257 the callback.
258 - add AnyEvent::Socket::{ntoa,aton} aliases.
259 - optimize the AE::Handle->push_read (line) for the default
260 eol marker.
261 - optimize push_read (packstring|storable) for small packets.
262 - invoke on_error callback when no on_eof callback is set.
263 - fix a bug in push_read (storable) of unknown impact.
264
2654.22 Sun Jul 20 16:34:13 CEST 2008
266 - new function AnyEvent::Socket::parse_hostport.
267 - as the bulkheads at microsoft can't even get getprotobyname reliably
268 working on their shitty fucking broken os we need to hardcode
269 some common protocol numbers in AnyEvent::Socket. How can
270 people even bother with such a pile of shit as windows.
271
2724.21 Thu Jul 17 14:40:05 CEST 2008
273 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
274 which is incompatible to previous invocation syntax.
275 - work around more windows perl bugs in fork_call. I cannot imagine
276 why anybody would bother implementing fork in such an obviously
277 unusable way.
278 - avoid calling $do_retry if the request has already finished in
279 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
280 K. Sharifulin).
281
2824.2 Sat Jul 12 22:42:11 CEST 2008
283 - support an "interval" argument for repeating timers.
284 - fix a bug in the pure perl event loop that caused timers to
285 be delayed under some circumstances.
286 - implement optional argument checking via PERL_ANYEVENT_STRICT.
287 - not importing anything from Event causes it to malfunction
288 and not use Time::HiRes, work around this bug.
289 - more bug workarounds against the endlessly broken Tk module,
290 seems to work now on 32 bit machines, but less so on 64 bit.
291 - minor optimisations applied to most backends.
292
2934.161 Fri Jul 4 14:24:48 CEST 2008
294 - fixed the t/handle/02_write.t, it read after write, with the
295 assumption that all data could be written before read
296 (reported by Jim Razmus).
297
2984.160 Thu Jul 3 04:02:21 CEST 2008
299 - re-registering signal watchers after unregistering one did
300 not work if a backend falls back on AnyEvent's default
301 implementation (also affected child watchers).
302 - new AnyEvent::Handle options: autocork and no_delay.
303
3044.152 Sun Jun 22 14:15:44 CEST 2008
305 - allow for 32-bit perls that implement shifts differently
306 on different architectures in parse_ipv4 (reported and
307 analysed by Keiichi DAIBA).
308
3094.151 Fri Jun 6 17:34:24 CEST 2008
310 - make sure specifying _only_ on_read and never pushing reads
311 works.
312
3134.15 Fri Jun 6 13:00:46 CEST 2008
314 - the pure perl backend would keep some watchers alive when more than
315 one watcher was registered for the same fd.
316 - new "packstring" and "storable" read and write types
317 for AnyEvent::Handle.
318 - allow on_eof handler to be called after on_error with EPIPE returns.
319 - do not immediately call on_read callback in handle constructor.
320
3214.14 Thu Jun 5 20:29:31 CEST 2008
322 - Fixed a bug in DNS SRV priority sorting.
323 - AnyEvent::Util::guard now reports runtime errors while
324 executing the guard block as warnings.
325 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
326
3274.13 Thu Jun 5 00:47:59 CEST 2008
328 - AnyEvent::DNS only followed cname chains with length 2,
329 contrary to documentation. bumped it up to 10, thanks to
330 microsoft, the current broken-dns-config-king.
331 - AnyEvent::DNS didn't check for socket return status
332 but instead relied on perl not creating filehandles in that
333 case - too bad it gives you a bogus file handle (reported
334 and analysed by Vladimir Timofeev).
335 - fix queue management logic in AnyEvent::Handle:
336 when on_read was registered, the queue was empty
337 and no progress could be made AnyEvent::Handle would
338 enter an endless loop.
339 - correctly start reading again when the handle became
340 busy again after idling.
341 - correctly treat tls shut-downs as EOF condition
342 (for the time being, would be nice to have a callback for that).
343 - correctly call eof callback under all conditions
344 (hopefully).
345 - the timeout callback did not expect that $self can go
346 away any time.
347
3484.12 Tue Jun 3 10:58:04 CEST 2008
349 - include AnyEvent::Intro, a tutorial for anyevent,
350 anyevent::socket and anyevent::handle.
351 - allow more options in on_error.
352
3534.11 Fri May 30 23:42:25 CEST 2008
354 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
355 reverse_lookup and reverse_verify to replace it, support
356 v4mapped and v4compat addresses.
357 - provide more documentation for the resolver class.
358 - really replace longest run of :0: by :: in format_address,
359 also properly convert :: and ::1 again.
360 - support NAPTR record name and decode it.
361 - implement random weight sampling for SRV records, as per
362 rfc 2782.
363 - correctly abort on srv-record targets of ".".
364 - added AnyEvent::DNS::wait_for_slot.
365 - in the unlikely event of a virtual circuit connection
366 being invalidated by a delayed udp reply, AnyEvent::DNS
367 could die. this has been fixed.
368 - plug a probable memleak in the DNS vc code.
369 - use configured timeout also for VC request phase.
370 - implement timeout and max_outstanding methods
371 for resolver class.
372 - update version numbers in all modules.
373
3744.1 Thu May 29 05:45:40 CEST 2008
375 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
376 which mirrors their purpose better (old names still available
377 for a while). Also moved AnyEvent::DNS::addr to
378 AnyEvent::Socket::resolve_sockaddr.
379 - implement AnyEvent->time and AnyEvent->now.
380 - fix IPv6 support in pack_sockaddr.
381 - officially un-experimentalise all newly introduced code.
382 - support unix domain sockets everywhere by specifying
383 a host of "unix/" and the pathname as service.
384 - implement an activity timeout in AnyEvent::Handle.
385 - added a regex read type to AnyEvent::Handle.
386 - added a json read/write type to AnyEvent::Handle.
387 - always croak in on_error in AnyEvent::Handle.
388 - document how to subclass AnyEvent::Handle.
389 - implement AnyEvent::Util::fork_call.
390 - add support for IPv6 nameservers and nameserver statements.
391 - work around _yet_ _another_ windows perl bug in where empty
392 select masks cause errors in the pure perl backend (this
393 workaround cannot be implemented for the other event loops,
394 but EV also works around this).
395 - supply AnyEvent::Socket::address_type and make good use of it.
396 - clarify and fix the Handle documentation w.r.t $self vs. $handle
397 vs. callback arguments.
398 - add some recommends to the META.yml.
399
4004.05 Mon May 26 19:44:06 CEST 2008
401 - some platforms ignore the rfc and prepend an extra sa_len member to
402 the sockaddr structure, cater for those. also use sockaddr_family,
403 didn't know it was there, but comes in handy for the workaround.
404 - undo the import WIN32 hack.
405
4064.04 Mon May 26 08:03:31 CEST 2008
407 - try to work around yet another windows bug: failed connects are reported
408 as if it were out-of-band data. windows users: you suck. Only
409 EV, Glib, Event and the pure perl backend can handle this condition.
410 - optimize write algorithm in AnyEvent::Handle.
411 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
412 - tcp_connect now properly iterates through all targets.
413 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
414 support entirely. also hardcode some AF_INET6 constants for the
415 sake of perl 5.8.
416 - vastly improved nameserver/suffix detection algorithm for windows
417 (still a hack...)
418 - try to use a fallback dns server if no dns servers could be found.
419 - splatter around some local $SIG{__DIE__}'s to work around bad
420 die handlers.
421
4224.03 Sun May 25 05:05:57 CEST 2008
423 - work around perl bugs on windows, where perl returns undocumented
424 error codes for sysread, syswrite etc.
425 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
426 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
427 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
428 - updated documentation.
429 - lots of workarounds for perl 5.8 quirks.
430 - implement a more flexible type system for both reads and writes
431 in AnyEvent::Handle.
432 - support IPv6 in tcp_server.
433 - pass local hostname and port to the prepare callback in tcp_server.
434 - make types extendable in AnyEvent::Handle.
435 - croak, not die, in AnyEvent::Handle;
436
4374.0 Sat May 24 19:58:08 CEST 2008
438 - added (experimental) AnyEvent::DNS module.
439 - added (experimental) AnyEvent::Socket module.
440 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
441 - make condvar->end call ->send as documented.
442 - make condvar constructor support the documented cb => argument.
443 - overload condvars so you can use them instead of code references.
444 - support more of the POSIX forms instead of a pure dotted quad for
445 verbatim ip addresses.
446 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
447 - try very hard not to freeze on broken windows perls.
448 - make on_eof optional in AnyEvent::Handle.
449 - support service names in tcp_connect and tcp_server.
450 - work around netbsd/openbsd bug causing unix sockets not to
451 become writable until empty.
452 - only make a dummy read for error detection when getpeername
453 returns ENOTCONN.
454 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
455 override.
456 - new AnyEvent::Guard::cancel.
457
4583.5 Sat May 17 23:17:14 CEST 2008
459 - gracefully deal with systems not defining CLOCK_MONOTONIC
460 (as opposed to not supporting the clock itself, which worked in 3.41).
461 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
462 helpers, to replace the ill-designed AnyEvent::Socket class.
463 - added some examples (eg/connect, eg/handle).
464 - added AnyEvent::Util::guard.
465 - AnyEvent::Util now exports most of its functions.
466 - fixed a bug in *_read_chunk, where the callback was called with $self as
467 first and second argument.
468 - simplified and fixed regex handling in AnyEvent::Handle::*_read_line.
469 - fix manpage condvars examples to use new syntax.
470 - considerably improve inet_aton with EV::ADNS backend to support
471 cname chains (common dns configuration bug, e.g. with www.google.de)
472 and aaaa records if no a record could be found.
473
4743.41 Sun May 11 19:53:13 CEST 2008
475 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
476
4773.4 Sun May 11 00:29:25 CEST 2008
478 - complete re-implementation of condvars with a hopefully much more useful
479 API (backward compatible functions exist but are undocumented).
480 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
481 (still subject to change).
482 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
483 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which
484 allows module authors to avoid forcing event loop detection.
485 (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example).
486 - remove coro backends: Coro now provides generic support for AnyEvent
487 via Coro::AnyEvent.
488
4893.3 Mon Apr 28 09:51:06 CEST 2008
490 - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules.
491 - fix a bug in the pure perl backend that kept watchers alive
492 when multiple watchers were registered for the same fh/poll combo.
493 - add a benchmark section showing AnyEvent overhead and comparing
494 the different event loops with each other.
495 - prefer pure perl over tk when autoprobing, as it's about as fast,
496 but doesn't crash with many watchers.
497 - declare Qt support non-experimental.
498 - clarify the confusing section about the file descriptor being
499 kept alive.
500 - document the race between loading of an event module and
501 child processes exiting.
502 - support POE as "backend" (with some caveats, POE is not generic
503 enough, and darn slow).
504 - support Wx and Prima through POE.
505 - optimise perl backend to use 20% less memory and take advantage
506 of typical timeout behaviour. It can now compete with
507 select/poll-based C event loops in most cases (it is usually
508 faster than Event and Glib :).
509 - roughly cut EV memory use in half and increase its speed by 30%,
510 by removing undocumented functionality. Did something similar
511 to other event loops where possible.
512
5133.2 Thu Apr 24 10:10:40 CEST 2008
514 - do not die when anyevent watchers are destroyed while running callbacks
515 in the pureperl backend (could only happen when two watchers are registered
516 for the same fh/poll combo).
517 - support autoloading for child watchers, was broken in all versions.
518 - implement PERL_ANYEVENT_MODEL env variable.
519 - (experimental) implement interface to the (very crashy/buggy) Event::Lib module.
520 - (experimental) implement interface to the Qt module (cannot be autoprobed).
521 - this release is IO:AnyEvent-proof.
522
5233.12 Tue Apr 22 07:11:46 CEST 2008
524 - reinstate AUTHOR section that got lost somehow.
525 - do not hang in the testsuite with badly broken perls (activestate,
526 strawberry...), but instead diagnose the problem and continue.
527 - use INT instead of CHLD in an attempt to support broken windows perls
528 better (this decreases the test quality, unfortunately).
529 - do not send the signal to the process group
530 (no problem for CHLD, bad for INT :).
531
5323.11 Sat Apr 19 06:57:31 CEST 2008
533 - major documentation rework.
534 - document the fact that child watchers only watch for zombies.
535 - fix the child watcher example.
536
5373.1 Wed Apr 16 17:09:01 CEST 2008
3 - work around recurring bugs in Tk by dup'ing filehandles, the 538 - work around recurring bugs in Tk by dup'ing filehandles, the
4 only method with good success chances to work on Tk. 539 only method with good success chances on Tk (the bugs apparently
540 don't get fixed anytime soon).
5 - lift the restriction of only one watcher per fh direction. 541 - lift the restriction of only one watcher per fh direction
542 (as the Tk bug workaround also lifts it and only Tk imposed
543 such strong limits).
6 - changed probe order to prefer coro adaptors. 544 - changed probe order to prefer coro adaptors.
7 - explain why recursion into the event loop is not supported 545 - explain why recursion into the event loop is not supported
8 unless the backend supports it (only Coro::EV does without 546 unless the backend supports it (only Coro::EV does without
9 any restrictions...). 547 any restrictions...).
10 - add simple manpages for all backend modules. 548 - add simple manpages for all backend modules.
… …
123.0 Mon Apr 7 21:30:23 CEST 2008 5503.0 Mon Apr 7 21:30:23 CEST 2008
13 - Coro::Signal changed semantics, roll our own, also cleaning 551 - Coro::Signal changed semantics, roll our own, also cleaning
14 up the Coro implementation in general. 552 up the Coro implementation in general.
15 - rename Coro backend to CoroEvent. 553 - rename Coro backend to CoroEvent.
16 - add some decision helping paragraph to the manpage that should 554 - add some decision helping paragraph to the manpage that should
17 help people to decide wether AnyEvent is the right thing for them. 555 help people to decide whether AnyEvent is the right thing for them.
18 556
192.9 Mon Jan 28 13:31:54 CET 2008 5572.9 Mon Jan 28 13:31:54 CET 2008
20 - update for EV 3.0 API changes. 558 - update for EV 3.0 API changes.
21 559
222.8 Sun Nov 25 15:06:03 CET 2007 5602.8 Sun Nov 25 15:06:03 CET 2007
… …
108 - fixed documentation. 646 - fixed documentation.
109 - clarified some corner cases. 647 - clarified some corner cases.
110 - vastly improved documentation with a long example. 648 - vastly improved documentation with a long example.
111 649
1120.2 Thu Dec 1 22:18:49 CET 2005 6500.2 Thu Dec 1 22:18:49 CET 2005
113 - check wether a valid method was called, to avoid endless loops. 651 - check whether a valid method was called, to avoid endless loops.
114 - move condvars into their own class (cleaner, safer). 652 - move condvars into their own class (cleaner, safer).
115 - don't die when some modules are not available. 653 - don't die when some modules are not available.
116 - Tk is too broken to allow for words. Will work once Tk works. 654 - Tk is too broken to allow for words. Will work once Tk works.
117 655
1180.1 Thu Dec 1 19:44:57 CET 2005 6560.1 Thu Dec 1 19:44:57 CET 2005

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines