ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent/Changes
Revision: 1.299
Committed: Fri Jul 24 23:45:54 2009 UTC (14 years, 9 months ago) by root
Branch: MAIN
Changes since 1.298: +2 -1 lines
Log Message:
*** empty log message ***

File Contents

# Content
1 Revision history for Perl extension AnyEvent.
2
3 TODO: maybe support "peerip" verificstion, i.e. getpeername diretcly as
4 additional peername?
5
6 - do not attempt to linger when there is no longer a valid fh.
7 - enforce tls mode to be either accept or connect, do not simply
8 segfault in Net::SSLeay.
9 - AnyEvent::Handle can now call tcp_connect itself (new parameters
10 connect, on_prepare, on_connect and on_connect_error). Updated
11 tutorial accordingly.
12 - add AnyEvent::Impl::Irssi backend.
13
14 4.86 Mon Jul 20 23:52:29 CEST 2009
15 - since the verbose warning is not enough, explicitly document
16 that versions before 1.33 of Net::SSLeay are not secure.
17 - work around signal handling races in Event and (...) Event::Lib.
18 - try to align signal-race timer to full-second boundaries.
19 - work around Tk not liking negative timeouts.
20 - don't complain of different grades of environmental unfriendlyness
21 in IO::ASync.
22
23 4.85 Sat Jul 18 06:16:14 CEST 2009
24 - nail the signal race problem in perl once and for all
25 (see $AnyEvent::MAX_SIGNAL_LATENCY).
26 - take advantage of Async::Interrupt if it is available.
27 - load Time::HiRes and Guard modules on demand only.
28 - add optional/recommended modules section to AnyEvent
29 documentation.
30 - reduce memory usage considerably (and reduce startup penalty)
31 by not using "strict", "warnings" and "overload" modules.
32 - work around buggy windows/openbsd perls and provide EBADMSG
33 and EPROTO ourselves when missing.
34 - improve perl 5.6 compatibility of the core event loop.
35 - made Net::SSLeay version 1.33 a soft requirement.
36
37 4.83 Fri Jul 17 16:56:26 CEST 2009
38 - implement AnyEvent::Socket::getprotobyname.
39 - AnyEvent::CondVar's will now detect recursive blocking
40 waits and will croak, as too many people fall into
41 this trap.
42 - AnyEvent::Handle will now call ->destroy on itself after
43 executing the on_error callback, instead of doing some
44 half-baked internal shutdown, for fatal errors.
45 - clarify on_eof behaviour w.r.t. the read queue and
46 on_read callbacks.
47 - ignore some possible spurious wake-ups in tcp_connect.
48
49 4.82 Sat Jul 11 00:34:55 CEST 2009
50 - POE and Event backends didn't accept some callable objects as
51 callbacks.
52 - use Config module instead of POSIX module to detect signal names
53 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
54 module doesn't even have all POSIX signals :/.
55 - use more workarounds around the many refcnt/corruption bugs in
56 Event::Lib.
57 - work around a race condition in perl's select, causing t/03_child.t
58 to rarely fail.
59
60 4.81 Thu Jul 9 10:30:30 CEST 2009
61 - AnyEvent::Handle didn't properly diagnose write errors
62 (it expected -1 from syswrite, how lame... :).
63 - support file descriptors in addition to file handles
64 in AnyEvent->io.
65 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
66 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
67 and PERL_ANYEVENT_CA_PATH.
68 - provide a sensible synopsis section for AnyEvent::TLS.
69 - add a "supported backends" section to the manpage.
70 - added simple io watcher test to testsuite, using a
71 portable_socketpair.
72 - tried to improve the stability of the Event::Lib backend,
73 YMMV.
74
75 4.8 Mon Jul 6 23:45:16 CEST 2009
76 - AnyEvent::DNS did not properly follow CNAME records with
77 uppercase targets.
78 - AnyEvent::DNS would errornously return AAAA records
79 with v4 mapped addresses (a faulty record) as ipv4 addresses,
80 causing AnyEvent::Socket to throw an exception.
81 - added new module AnyEvent::TLS for easier SSL/TLS context
82 creation, with many options including hostname verification,
83 secure default configuration, lots of documentation and,
84 predefined diffie-hellman keys for perfect forward security
85 and much more. get it while it's still fresh!
86 - use AnyEvent::TLS in AnyEvent::Handle for context management.
87 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
88 so AnyEvent::Socket users have smaller memory footprint
89 in the common case.
90 - add AnyEvent::Handle->push_shutdown method.
91 - add an additional error message parameter to AnyEvent::Handle's
92 on_error callback (for TLS, $! is still available).
93 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
94 - make AnyEvent::Handle more robust against early conenction
95 failures (during new), and return C<undef> in that case
96 from the constructor.
97 - AnyEvent::Handle will now try to load only JSON::XS first,
98 then fall back to JSON.
99 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
100 and because it was documented that way.
101
102 4.451 Fri Jul 3 00:28:58 CEST 2009
103 - do not clear rbuf when shutting down an AnyEvent::Handle
104 object - doing so breaks AnyEvent::HTTP.
105
106 4.45 Mon Jun 29 22:59:26 CEST 2009
107 - a write error could cause AnyEvent::Handle to create
108 an I/O watcher with an undefined $fh.
109 - special-case mapped ipv4 addresses in both
110 AnyEvent::Socket::format_address and parse_address,
111 to treat them just like ipv4 addresses.
112 - updated and overhauled the AnyEvent::Intro doc.
113 - implement AnyEvent::Socket::format_ipv4/ipv6.
114 - slightly speed up portable_pipe/socketpair.
115 - expand condvar begin/end documentation.
116
117 4.42 Fri Jun 26 08:32:18 CEST 2009
118 - preliminary and neccesarily incomplete support for IO::Async.
119 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
120 case it was set to IGNORE, to ensure we can catch child statuses
121 even when the calling env acts stupidly.
122 - updated benchmarks with IO::Async, which performs very well.
123
124 4.412 Wed Jun 24 01:35:57 CEST 2009
125 - support an "untaint" attribute for AnyEvent::DNS
126 and set it on the default resolver.
127 - implement and document AnyEvent::Impl::Perl::loop.
128 - remove all anyevent-env variables from %ENV
129 when running in tainted mode.
130 - mention and extend the IO::Lambda benchmark.
131
132 4.411 Sun Jun 7 18:48:02 CEST 2009
133 - do not try to use F_SETFD on windows.
134
135 4.41 Thu May 14 06:40:11 CEST 2009
136 - work around issues in older perls (5.8.0?) when
137 a signal handler is deleted from the %SIG hash.
138 - use POSIX::_exit in child test, to avoid running
139 destructors.
140 - speed up CHLD handling by relying on SIGCHLD being
141 synchronously delivered, even when we roll our own
142 implementation.
143 - AnyEvent::DNS: add the "dname" resource record name
144 for cosmetic reasons.
145
146 4.4 Sun Apr 26 20:12:33 CEST 2009
147 - implemented idle watchers, where applicable.
148 - AnyEvent->time died when Event backend was in use.
149 - fix a memleak in the Tk backend.
150 - sped up Tk timer handling.
151 - clip negative "after" values to 0 in AnyEvent::Impl::Event
152 to avoid spamming out warning messages.
153 - fix Qt timers without interval.
154 - avoid Qt zero-timeout special case, in old libqt's.
155
156 4.352 Mon Apr 20 16:31:11 CEST 2009
157 - fix AnyEvent::Strict error messages for child watchers.
158 - fix/update Makefile.PL "recommends" versions.
159 - add AnyEvent->now_update.
160
161 4.351 Sat Apr 11 07:56:14 CEST 2009
162 - actually make the signal pipe work under win32.
163 - localise $! in signal handler to avoid changing $!, although
164 perl itself does not seem to save/restore errno either.
165 - set the cloexec flag on the signal pipe (normally set by perl too).
166
167 4.35 Fri Mar 27 11:48:20 CET 2009
168 - event models relying on AnyEvent's signal watcher emulation did
169 invoke the callback asynchronously, contradicting documentation
170 and causing signals to get lost (this includes AnyEvent's own
171 event loop). AnyEvent now uses the standard pipe trick to make
172 callback execution synchronous to the event loop.
173 - AnyEvent::Handle didn't free TLS context data on DESTROY
174 (patch by Pavel Shaydo).
175 - work around the ever-incompatibly-changing API of MakeMaker.
176 - document that changing global variables without restoring them
177 is a bad idea in Perl (noted by Adam Rosenstein).
178 - AnyEvent::Strict now barfs if ->io is passed a file.
179
180 4.34 Thu Feb 12 18:32:45 CET 2009
181 - separately buffer TLS read data, as otherwise the read queue could
182 deadlock as receiving data is not expected while draining the
183 read queue (which cna only happen with TLS).
184 - raise EBADMSG error on JSON decoding errors.
185 - fix some minor manpage bugs (reported by Maximilian Gaß).
186 - speed up select bitmask parsing quite a bit in the pure perl
187 backend.
188 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
189
190 4.331 Tue Jan 6 21:07:25 CET 2009
191 - socketpair fails on many vista machines because vista has
192 completely broken accept/getpeername and getsockname functions,
193 so we provide our own socketpair emulation that kind of works
194 (AnyEvent::Util::portable_pipe).
195 - new function: AnyEvent::Util::portable_socketpair.
196 - take advantage of the Guard module if it exists.
197
198 4.33 Fri Nov 21 02:35:40 CET 2008
199 - AnyEvent::Strict did errornously flag a fileno of 0 as
200 illegal.
201 - reduce memory usage and slightly speed up the pure perl backend
202 by only storing the file descriptor, not the file handle.
203 - add missing autocork method to AnyEvent::Handle (reported
204 by Adam Rosenstein).
205 - AnyEvent::DNS->resolve errornously documented an $rcode result
206 argument, but there isn't (spotted by Henrik Krohns).
207 - the naked truth about Tk - it's basically unmaintained.
208
209 4.32 Mon Nov 3 22:46:32 CET 2008
210 - fix AnyEvent::Socket::resolve_sockaddr to properly support
211 unix sockets again and choose a proper default for
212 the socket type.
213 - fix call to resolve_sockaddr in tcp_connect.
214
215 4.31 Thu Oct 30 04:41:48 CET 2008
216 - implemented AnyEvent::Handle->destroy method.
217 - hint about unexpected effects in TLS mode.
218 - speed up AnyEvent::Impl::Perl by using more arrays
219 instead of hashes without the slightest loss of readability :)
220 - work around a perl argument refcounting bug.
221 - with some perl patches applied, AnyEvent now seems leak-free.
222
223 4.3 Fri Oct 3 09:18:43 CEST 2008
224 - AnyEvent will now install a no-op signal handler for SIGPIPE,
225 unless one has been installed already.
226 - warn about not loaidng AnyEvent::Impl::POE early enough
227 (patch by Adam Rosenstein).
228 - fixed a great number of bugs and corner cases in AnyEvent::Handle:
229 - fix a bug in where in SSL connect mode, the
230 client would first wait for some data by the server and
231 otherwise hang. (reported and analyzed in an absolutely
232 exemplary manner by Adam Rosenstein).
233 - fix a bug in where SSL EOF would not be treated
234 as stream EOF, putting the connection into a hung state.
235 (reported and analyzed in an absolutely exemplary manner by Adam
236 Rosenstein).
237 - fix a potential segfault when the TLS context
238 would go missing in a read callback (e.g. due to stoptls) -
239 Net::SSLeay of course makes no type checking whatsoever.
240 - AnyEvent::Handle will not stop the read watcher in TLS mode.
241 - AnyEvent::Handle->stoptls will send a TLS close notify
242 instead of simply destroying the stream now.
243 - fix a bug where large blocks of data written in TLS mode
244 would not be sent unless triggered by receives.
245 - on_drain will now take the tls write buffer into account.
246 - SSL operations have been streamlined a bit and should be faster
247 now, more to come, though.
248 - added a FAQ section, will document nontrivial issues in there.
249
250 4.234 Mon Sep 29 04:08:13 CEST 2008
251 - fix child watcher documentation: only child exits will be reported,
252 no trace events.
253 - mention SIGPIPE in AnyEvent::Handle.
254 - perl backend will now use POSIX::times as monotonic clock
255 when available and the monotonic clock option is not.
256 - run even without Time::HiRes in some cases now, but do not
257 provide sub-second accuracy (all *supported* perl versions have
258 Time::HiRes).
259
260 4.233 Fri Aug 22 01:48:05 CEST 2008
261 - fix a bug in the testsuite, causing a hang.
262 - clarified lots of AnyEvent::Handle constructor arguments.
263
264 4.232 Thu Aug 21 20:44:25 CEST 2008
265 - fix a bug in AnyEvent::Handle that could cause two fatal error
266 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
267 (testcase by Leon Brocard).
268 - clarified on_eof/on_error documentation a bit.
269
270 4.231 Tue Jul 29 13:12:15 CEST 2008
271 - remove some debugging code left in AnyEvent::Util::fork_call
272 (and no, it's impossible to implement with the broken windows
273 perls without resource leaks or worse).
274
275 4.23 Tue Jul 29 12:19:59 CEST 2008
276 - document the first parameter passed to condvar callbacks to be
277 the callback.
278 - add AnyEvent::Socket::{ntoa,aton} aliases.
279 - optimize the AE::Handle->push_read (line) for the default
280 eol marker.
281 - optimize push_read (packstring|storable) for small packets.
282 - invoke on_error callback when no on_eof callback is set.
283 - fix a bug in push_read (storable) of unknown impact.
284
285 4.22 Sun Jul 20 16:34:13 CEST 2008
286 - new function AnyEvent::Socket::parse_hostport.
287 - as the bulkheads at microsoft can't even get getprotobyname reliably
288 working on their shitty fucking broken os we need to hardcode
289 some common protocol numbers in AnyEvent::Socket. How can
290 people even bother with such a pile of shit as windows.
291
292 4.21 Thu Jul 17 14:40:05 CEST 2008
293 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
294 which is incompatible to previous invocation syntax.
295 - work around more windows perl bugs in fork_call. I cannot imagine
296 why anybody would bother implementing fork in such an obviously
297 unusable way.
298 - avoid calling $do_retry if the request has already finished in
299 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
300 K. Sharifulin).
301
302 4.2 Sat Jul 12 22:42:11 CEST 2008
303 - support an "interval" argument for repeating timers.
304 - fix a bug in the pure perl event loop that caused timers to
305 be delayed under some circumstances.
306 - implement optional argument checking via PERL_ANYEVENT_STRICT.
307 - not importing anything from Event causes it to malfunction
308 and not use Time::HiRes, work around this bug.
309 - more bug workarounds against the endlessly broken Tk module,
310 seems to work now on 32 bit machines, but less so on 64 bit.
311 - minor optimisations applied to most backends.
312
313 4.161 Fri Jul 4 14:24:48 CEST 2008
314 - fixed the t/handle/02_write.t, it read after write, with the
315 assumption that all data could be written before read
316 (reported by Jim Razmus).
317
318 4.160 Thu Jul 3 04:02:21 CEST 2008
319 - re-registering signal watchers after unregistering one did
320 not work if a backend falls back on AnyEvent's default
321 implementation (also affected child watchers).
322 - new AnyEvent::Handle options: autocork and no_delay.
323
324 4.152 Sun Jun 22 14:15:44 CEST 2008
325 - allow for 32-bit perls that implement shifts differently
326 on different architectures in parse_ipv4 (reported and
327 analysed by Keiichi DAIBA).
328
329 4.151 Fri Jun 6 17:34:24 CEST 2008
330 - make sure specifying _only_ on_read and never pushing reads
331 works.
332
333 4.15 Fri Jun 6 13:00:46 CEST 2008
334 - the pure perl backend would keep some watchers alive when more than
335 one watcher was registered for the same fd.
336 - new "packstring" and "storable" read and write types
337 for AnyEvent::Handle.
338 - allow on_eof handler to be called after on_error with EPIPE returns.
339 - do not immediately call on_read callback in handle constructor.
340
341 4.14 Thu Jun 5 20:29:31 CEST 2008
342 - Fixed a bug in DNS SRV priority sorting.
343 - AnyEvent::Util::guard now reports runtime errors while
344 executing the guard block as warnings.
345 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
346
347 4.13 Thu Jun 5 00:47:59 CEST 2008
348 - AnyEvent::DNS only followed cname chains with length 2,
349 contrary to documentation. bumped it up to 10, thanks to
350 microsoft, the current broken-dns-config-king.
351 - AnyEvent::DNS didn't check for socket return status
352 but instead relied on perl not creating filehandles in that
353 case - too bad it gives you a bogus file handle (reported
354 and analysed by Vladimir Timofeev).
355 - fix queue management logic in AnyEvent::Handle:
356 when on_read was registered, the queue was empty
357 and no progress could be made AnyEvent::Handle would
358 enter an endless loop.
359 - correctly start reading again when the handle became
360 busy again after idling.
361 - correctly treat tls shut-downs as EOF condition
362 (for the time being, would be nice to have a callback for that).
363 - correctly call eof callback under all conditions
364 (hopefully).
365 - the timeout callback did not expect that $self can go
366 away any time.
367
368 4.12 Tue Jun 3 10:58:04 CEST 2008
369 - include AnyEvent::Intro, a tutorial for anyevent,
370 anyevent::socket and anyevent::handle.
371 - allow more options in on_error.
372
373 4.11 Fri May 30 23:42:25 CEST 2008
374 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
375 reverse_lookup and reverse_verify to replace it, support
376 v4mapped and v4compat addresses.
377 - provide more documentation for the resolver class.
378 - really replace longest run of :0: by :: in format_address,
379 also properly convert :: and ::1 again.
380 - support NAPTR record name and decode it.
381 - implement random weight sampling for SRV records, as per
382 rfc 2782.
383 - correctly abort on srv-record targets of ".".
384 - added AnyEvent::DNS::wait_for_slot.
385 - in the unlikely event of a virtual circuit connection
386 being invalidated by a delayed udp reply, AnyEvent::DNS
387 could die. this has been fixed.
388 - plug a probable memleak in the DNS vc code.
389 - use configured timeout also for VC request phase.
390 - implement timeout and max_outstanding methods
391 for resolver class.
392 - update version numbers in all modules.
393
394 4.1 Thu May 29 05:45:40 CEST 2008
395 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
396 which mirrors their purpose better (old names still available
397 for a while). Also moved AnyEvent::DNS::addr to
398 AnyEvent::Socket::resolve_sockaddr.
399 - implement AnyEvent->time and AnyEvent->now.
400 - fix IPv6 support in pack_sockaddr.
401 - officially un-experimentalise all newly introduced code.
402 - support unix domain sockets everywhere by specifying
403 a host of "unix/" and the pathname as service.
404 - implement an activity timeout in AnyEvent::Handle.
405 - added a regex read type to AnyEvent::Handle.
406 - added a json read/write type to AnyEvent::Handle.
407 - always croak in on_error in AnyEvent::Handle.
408 - document how to subclass AnyEvent::Handle.
409 - implement AnyEvent::Util::fork_call.
410 - add support for IPv6 nameservers and nameserver statements.
411 - work around _yet_ _another_ windows perl bug in where empty
412 select masks cause errors in the pure perl backend (this
413 workaround cannot be implemented for the other event loops,
414 but EV also works around this).
415 - supply AnyEvent::Socket::address_type and make good use of it.
416 - clarify and fix the Handle documentation w.r.t $self vs. $handle
417 vs. callback arguments.
418 - add some recommends to the META.yml.
419
420 4.05 Mon May 26 19:44:06 CEST 2008
421 - some platforms ignore the rfc and prepend an extra sa_len member to
422 the sockaddr structure, cater for those. also use sockaddr_family,
423 didn't know it was there, but comes in handy for the workaround.
424 - undo the import WIN32 hack.
425
426 4.04 Mon May 26 08:03:31 CEST 2008
427 - try to work around yet another windows bug: failed connects are reported
428 as if it were out-of-band data. windows users: you suck. Only
429 EV, Glib, Event and the pure perl backend can handle this condition.
430 - optimize write algorithm in AnyEvent::Handle.
431 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
432 - tcp_connect now properly iterates through all targets.
433 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
434 support entirely. also hardcode some AF_INET6 constants for the
435 sake of perl 5.8.
436 - vastly improved nameserver/suffix detection algorithm for windows
437 (still a hack...)
438 - try to use a fallback dns server if no dns servers could be found.
439 - splatter around some local $SIG{__DIE__}'s to work around bad
440 die handlers.
441
442 4.03 Sun May 25 05:05:57 CEST 2008
443 - work around perl bugs on windows, where perl returns undocumented
444 error codes for sysread, syswrite etc.
445 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
446 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
447 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
448 - updated documentation.
449 - lots of workarounds for perl 5.8 quirks.
450 - implement a more flexible type system for both reads and writes
451 in AnyEvent::Handle.
452 - support IPv6 in tcp_server.
453 - pass local hostname and port to the prepare callback in tcp_server.
454 - make types extendable in AnyEvent::Handle.
455 - croak, not die, in AnyEvent::Handle;
456
457 4.0 Sat May 24 19:58:08 CEST 2008
458 - added (experimental) AnyEvent::DNS module.
459 - added (experimental) AnyEvent::Socket module.
460 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
461 - make condvar->end call ->send as documented.
462 - make condvar constructor support the documented cb => argument.
463 - overload condvars so you can use them instead of code references.
464 - support more of the POSIX forms instead of a pure dotted quad for
465 verbatim ip addresses.
466 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
467 - try very hard not to freeze on broken windows perls.
468 - make on_eof optional in AnyEvent::Handle.
469 - support service names in tcp_connect and tcp_server.
470 - work around netbsd/openbsd bug causing unix sockets not to
471 become writable until empty.
472 - only make a dummy read for error detection when getpeername
473 returns ENOTCONN.
474 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
475 override.
476 - new AnyEvent::Guard::cancel.
477
478 3.5 Sat May 17 23:17:14 CEST 2008
479 - gracefully deal with systems not defining CLOCK_MONOTONIC
480 (as opposed to not supporting the clock itself, which worked in 3.41).
481 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
482 helpers, to replace the ill-designed AnyEvent::Socket class.
483 - added some examples (eg/connect, eg/handle).
484 - added AnyEvent::Util::guard.
485 - AnyEvent::Util now exports most of its functions.
486 - fixed a bug in *_read_chunk, where the callback was called with $self as
487 first and second argument.
488 - simplified and fixed regex handling in AnyEvent::Handle::*_read_line.
489 - fix manpage condvars examples to use new syntax.
490 - considerably improve inet_aton with EV::ADNS backend to support
491 cname chains (common dns configuration bug, e.g. with www.google.de)
492 and aaaa records if no a record could be found.
493
494 3.41 Sun May 11 19:53:13 CEST 2008
495 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
496
497 3.4 Sun May 11 00:29:25 CEST 2008
498 - complete re-implementation of condvars with a hopefully much more useful
499 API (backward compatible functions exist but are undocumented).
500 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
501 (still subject to change).
502 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
503 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which
504 allows module authors to avoid forcing event loop detection.
505 (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example).
506 - remove coro backends: Coro now provides generic support for AnyEvent
507 via Coro::AnyEvent.
508
509 3.3 Mon Apr 28 09:51:06 CEST 2008
510 - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules.
511 - fix a bug in the pure perl backend that kept watchers alive
512 when multiple watchers were registered for the same fh/poll combo.
513 - add a benchmark section showing AnyEvent overhead and comparing
514 the different event loops with each other.
515 - prefer pure perl over tk when autoprobing, as it's about as fast,
516 but doesn't crash with many watchers.
517 - declare Qt support non-experimental.
518 - clarify the confusing section about the file descriptor being
519 kept alive.
520 - document the race between loading of an event module and
521 child processes exiting.
522 - support POE as "backend" (with some caveats, POE is not generic
523 enough, and darn slow).
524 - support Wx and Prima through POE.
525 - optimise perl backend to use 20% less memory and take advantage
526 of typical timeout behaviour. It can now compete with
527 select/poll-based C event loops in most cases (it is usually
528 faster than Event and Glib :).
529 - roughly cut EV memory use in half and increase its speed by 30%,
530 by removing undocumented functionality. Did something similar
531 to other event loops where possible.
532
533 3.2 Thu Apr 24 10:10:40 CEST 2008
534 - do not die when anyevent watchers are destroyed while running callbacks
535 in the pureperl backend (could only happen when two watchers are registered
536 for the same fh/poll combo).
537 - support autoloading for child watchers, was broken in all versions.
538 - implement PERL_ANYEVENT_MODEL env variable.
539 - (experimental) implement interface to the (very crashy/buggy) Event::Lib module.
540 - (experimental) implement interface to the Qt module (cannot be autoprobed).
541 - this release is IO:AnyEvent-proof.
542
543 3.12 Tue Apr 22 07:11:46 CEST 2008
544 - reinstate AUTHOR section that got lost somehow.
545 - do not hang in the testsuite with badly broken perls (activestate,
546 strawberry...), but instead diagnose the problem and continue.
547 - use INT instead of CHLD in an attempt to support broken windows perls
548 better (this decreases the test quality, unfortunately).
549 - do not send the signal to the process group
550 (no problem for CHLD, bad for INT :).
551
552 3.11 Sat Apr 19 06:57:31 CEST 2008
553 - major documentation rework.
554 - document the fact that child watchers only watch for zombies.
555 - fix the child watcher example.
556
557 3.1 Wed Apr 16 17:09:01 CEST 2008
558 - work around recurring bugs in Tk by dup'ing filehandles, the
559 only method with good success chances on Tk (the bugs apparently
560 don't get fixed anytime soon).
561 - lift the restriction of only one watcher per fh direction
562 (as the Tk bug workaround also lifts it and only Tk imposed
563 such strong limits).
564 - changed probe order to prefer coro adaptors.
565 - explain why recursion into the event loop is not supported
566 unless the backend supports it (only Coro::EV does without
567 any restrictions...).
568 - add simple manpages for all backend modules.
569
570 3.0 Mon Apr 7 21:30:23 CEST 2008
571 - Coro::Signal changed semantics, roll our own, also cleaning
572 up the Coro implementation in general.
573 - rename Coro backend to CoroEvent.
574 - add some decision helping paragraph to the manpage that should
575 help people to decide whether AnyEvent is the right thing for them.
576
577 2.9 Mon Jan 28 13:31:54 CET 2008
578 - update for EV 3.0 API changes.
579
580 2.8 Sun Nov 25 15:06:03 CET 2007
581 - waitpid can validly return 0. accept this fact of life
582 instead of reporting it to any watchers.
583
584 2.7 Fri Nov 23 11:41:14 CET 2007
585 - force use of AnyEvent::Impl::Perl in testsuite, there is
586 too much breakage outside AnyEvent.
587 - deliver signals synchronously in AnyEvent::Impl::Perl
588 or any other event loop that relies on AnyEvents child
589 watcher emulation. *Could* help with hanging testsuite
590 (except when Event or EV are installed).
591
592 2.6 Fri Nov 9 20:36:35 CET 2007
593 - fix bug in testsuite.
594 - move EV adaptor modules to AnyEvent.
595 - add Coro+EV adaptor module.
596
597 2.55 Tue Nov 6 17:41:32 CET 2007
598 - add EV to the list of supported event models.
599 - do not auto-reset pid watchers, pass pid and status to them.
600 - allow a pid of zero to watch for all children in child watchers.
601
602 2.54 Wed Jul 18 17:36:23 CEST 2007
603 - work around a perl bug that results in
604 BEGIN not safe after errors--compilation aborted without
605 any discernible reason or error message by once
606 again not use'ing strict.
607
608 2.53 Sun Jul 8 10:51:53 CEST 2007
609 - make Glib interface work again (spotted by elmex).
610
611 2.52 Wed Mar 7 18:36:16 CET 2007
612 - child watchers stopped working permanently when
613 all current child watchers were destroyed.
614
615 2.51 Mon Dec 11 21:33:24 CET 2006
616 - work around bugs in perl where eval "require Module" returns
617 true even if the module couldn't be loaded :/.
618
619 2.5 Mon Dec 11 02:15:28 CET 2006
620 - avoid $AUTOLOAD because many perls corrupt it.
621 - AnyEvent::detect forces autodetection.
622 - implement signal watchers (experimental)
623 (Tk does not support async signals (<= 804.027 at least),
624 so they might get delayed indefinitely on Tk).
625 - implement child watchers (experimental).
626 - moved default condvar implementation into base module, simplifying
627 most implementation modules.
628
629 2.1 Fri Nov 24 15:50:48 CET 2006
630 - better docs.
631 - simple testsuite.
632 - added AnyEvent->one_event method for special purposes.
633
634 2.0 Wed Nov 1 02:21:30 CET 2006
635 - INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both.
636 - INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore.
637 - EXPERIMENTAL:
638 new pure-perl model added, AnyEvent now always finds a suitable
639 event model.
640 - improved documentation.
641 - reduced cpu overhead.
642
643 1.02 Fri Jan 13 14:15:40 CET 2006
644 - add COPYING to clarify license.
645
646 1.01 Sun Jan 8 05:48:09 CET 2006
647 - correctly cancel Event timers, avoid memory leaks.
648
649 1.0 Sun Jan 8 05:40:58 CET 2006
650 - allow third-party interfaces to be added at runtime.
651 - fix version number.
652
653 0.9 Fri Dec 30 02:25:41 CET 2005
654 - fixed errors in the documentation.
655 - going back to using $VERSION as "load check",
656 as suggested by Jörn Reder. Let's see why I stopped using
657 this method.
658 - add 'hup' to the list of events to receive for glib
659 and generate 'rw' events for it, as some versions of
660 glib seem to require this, while this breaks other
661 (older) versions of glib.
662 - implemented PERL_ANYEVENT_VERBOSE env variable.
663
664 0.3 Sun Dec 4 10:43:33 CET 2005
665 - no changes to the code.
666 - fixed documentation.
667 - clarified some corner cases.
668 - vastly improved documentation with a long example.
669
670 0.2 Thu Dec 1 22:18:49 CET 2005
671 - check whether a valid method was called, to avoid endless loops.
672 - move condvars into their own class (cleaner, safer).
673 - don't die when some modules are not available.
674 - Tk is too broken to allow for words. Will work once Tk works.
675
676 0.1 Thu Dec 1 19:44:57 CET 2005
677 - spiced up and cleaned up for release.
678
679 0.01 Tue Sep 21 18:12:09 CEST 2004
680 - original version; copied from Convert::Scalar.
681