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

Comparing AnyEvent/Changes (file contents):
Revision 1.30 by root, Sat Nov 3 09:29:50 2007 UTC vs.
Revision 1.406 by root, Tue Aug 31 22:56:20 2010 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3MAYBE: http://svn.xantus.org/shortbus/trunk/cometd-perl/lab/aio.pl 3TODO: document $AnyEvent::DNS::RESOLVER
4 (poe makes their own "anyevent", unfortunately). 4TODO: maybe remove orderign description and make this a private feature?
5TODO: anyevenmt_read_type for unshift_read
6TODO: condvar->new method is documented and should be there.
7TODO: http://github.com/sanko/net-bittorrent/commit/e1d06941d7217011f5f06ce7152012d5e1b44b9a
8TODO: check for -w? somehow?
9 - clarify that you can actually call push_read etc. in on_read.
10 - incorporated major typo patches by Abhijit Menon-Sen.
5 11
62.55 125.271 Tue Jun 8 12:05:46 CEST 2010
13 - backport to perl 5.8.x.
14
155.27 Sun Jun 6 12:12:05 CEST 2010
16 - postpone differently in AnyEvent::Socket now, as
17 when not, canceling the connection attempt might fail
18 (found by Felix Antonius Wilhelm Ostmann).
19 - explicitly check for non-stream sockets in AE::Handle, too many
20 clueless people fell into the trap of this somehow working.
21 - simplified and reworked the "OTHER MODULES" section.
22 - better/more condvar examples.
23
245.261 Wed Apr 28 16:13:36 CEST 2010
25 - AF_INET6 was not properly used from Socket6 during configuration
26 time (found by Andy Grundman).
27
285.26 Mon Apr 12 04:49:35 CEST 2010
29 - don't generate spurious readyness notifications when select
30 returns EINTR in the pure perl backend (analysed and testcase
31 by Bas Denissen).
32 - give same examples in SYNOPSIS sections of AnyEvent and AE.
33 - provider faster implementations for AE::time/now/now_update for
34 EV and Perl backends.
35 - provide faster AE::cv and AE::time implementations for all
36 backends.
37 - no longer support lower/mixed-case signal names (this was
38 never documented nor universally supported).
39 - some more memory ยต-optimisations, and somewhat less messy
40 code to cope with both AnyEvent and AE APIs.
41
425.251 Sat Mar 13 00:58:21 CET 2010
43 - make test failed if it was run before make install on perl <=
44 5.8.8, and older versions of perl make our live very difficult,
45 so write constants.pl during Makefile.PL time.
46
475.25 Sat Mar 13 00:23:14 CET 2010
48 - fix a race condition in AnyEvent::Handle that would cause
49 a "bio_read: unsupported method" error (in _another_
50 ssl connection) after user code threw an exception.
51 - added AnyEvent::Handle->destroyed method.
52 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
53 - the AnyEvent::Util::fh_nonblocking function worked only
54 by lucky accident on win32.
55 - smaller and faster AnyEvent::Util::fh_nonblocking.
56 - when the (required!) Time::HiRes module is missing, AnyEvent
57 did not fall back to built-in time properly.
58 - do not load Fcntl at runtime, saving memory and loading time.
59 - precompile a number of constants and use them instead of runtime
60 detection and eval.
61 - free detection code after detection and similar memory optimisations.
62 - Perl backend timer interval best effort drift has been improved
63 (same algorithm as EV).
64 - update unicode idna mapping table.
65
665.24 Tue Jan 5 11:39:43 CET 2010
67 - cygwin never reports errors from failed connects. 1.5 just gives
68 you continous readyness and EAGAIN, 1.7 is even more broken and
69 just hangs. work around both issues in a cygwin-specific hack.
70 - improve idle watcher documentation slightly.
71
725.23 Sun Dec 20 23:48:00 CET 2009
73 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
74 - implement punycode_encode/decode, idn_nameprep,
75 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
76 - provide $AE::VERSION.
77 - removed traces of "no strict 'refs'".
78
795.22 Sat Dec 5 03:51:13 CET 2009
80 - downgrade-or-fail in AnyEvent::Handle::push_write, to
81 diagnose encoding failures earlier and more succinctly.
82 (this works around bugs in perl, throwing away encoding info
83 when passing scalar data to extensions).
84 - add more examples to AnyEvent::Socket manpage.
85 - upgrade internal warning set to the same as common::sense 2.03.
86 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
87 (John Beppu).
88 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
89 - add google public dns servers to fallback server set and make sure
90 we load-balance properly between the three sets. also add all
91 fallback dns servers, not just a random one, to each dns config.
92
935.21 Thu Nov 19 02:48:47 CET 2009
94 - fix a problem where socket constants were called with parameters
95 (spotted by David Friedland).
96 - fork_call never use'd POSIX (reported by Daisuke Maki).
97 - improve perl 5.6 compatibility further (but it still won't work
98 unless you rip out everything but the core).
99 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
100 uses 10MB of RAM, but doesn't open a console window. *sigh*.
101
1025.202 Wed Oct 14 22:35:44 CEST 2009
103 - AnyEvent::DNS would unexpectedly clobber $_ under windows
104 (analysed by Matthias Waldorf).
105 - AnyEvent::Handle::run_cmd can now pass the PID of the
106 newly-created process, which is much less useful than it might
107 sound (based on patch by Yann Kerherve).
108
1095.201 Tue Sep 29 12:09:25 CEST 2009
110 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
111 (reported by Torsten Foertsch).
112 - common::sense 2.0 could cause tcp_server to throw an exception
113 (analysed by elmex).
114
1155.2 Mon Sep 14 07:04:49 CEST 2009
116 - INCOMPATIBLE CHANGE: do no longer support register_read_type
117 and register_write_type in AnyEvent::Handle, instead support
118 package names (the facility was mostly abused).
119 - implement "packagename-as-read/write type" support in
120 AnyEvent::Handle.
121 - AnyEvent::Handle: new options "keepalive" and "oobinline".
122 - oobinline set by default to avoid security issues.
123 - the pure-perl event loop backend wrongly detected times() fallback
124 support (spotted by Pavel Boldin).
125
1265.12 Tue Sep 1 20:26:50 CEST 2009
127 - be more lenient when parsing resolv.conf files, as some people
128 use hashmarks as comment indicator inside directives (reported
129 by Michael S. Fischer).
130 - use same set of warnings as common::sense 2.0.
131 - fix a potential 32 bit overflow issue due to perl having problems
132 with large hex constants in the Perl backend.
133 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
134 by default.
135 - fix some prototypes.
136 - (optionally) require Async::Interrupt 1.02.
137 - replace opendns fallbacks by verizon/level3 public dns servers.
138
1395.112 Fri Aug 21 13:59:04 CEST 2009
140 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
141 using starttls with an illegal/uncreatable tls context.
142 - port to broken win32 perls.
143
1445.111 Sun Aug 16 18:44:36 CEST 2009
145 - perl errornously requires sockaddr_un structures returned by the
146 kernel to have a specific length. work around this bug by padding
147 sockaddr_un structures (found to fail on OS X, as perl only uses
148 the correct code for GNU/Linux, but likely to fail on other
149 OSes as well).
150
1515.11 Wed Aug 12 17:49:37 CEST 2009
152 - fixed a bug in run_cmd with input from filename.
153 - create a json coder object when none is given in write direction
154 to reduce dependency on JSON.
155
1565.1 Tue Aug 11 03:17:32 CEST 2009
157 - tcp_connect now reports the correct error in $!, not an unrelated
158 one (was broken since 4.91).
159 - AE::Handle did not properly clear rtimeout/wtimeout watchers
160 when retrying a tcp_connect.
161 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
162 - fixed wrong documentation in AE::Socket.
163
1645.01 Mon Aug 10 03:16:32 CEST 2009
165 - last release broke EV child watchers :/.
166
1675.0 Sun Aug 9 17:32:09 CEST 2009
168 - officially introduce and document the AE API.
169 - lowering the timeout at runtime did not have immediate effect
170 in AnyEvent::Handle.
171 - AE::Handle now has separate and independent read and write timeout
172 settings.
173 - probe for Guard module when AnyEvent::Util is loaded, not at
174 runtime (Event doesn't like eval "use" from callbacks, and it avoids
175 exporting the probe function).
176 - Impl::IOAsync's io watchers did not use the filehandle emulation
177 code to work around it's watcher limits.
178 - work around IO::Async's problems with overloaded objects.
179 - try to disable wearnings inside IO::Async, it's just too buggy.
180 - do no longer use Event by default, only EV or Perl, Event is too buggy.
181 - hackishly provide fast AE:: interface when EV or Perl is the backend.
182 - inherit VERSION from AnyEvent::VERSION in most modules.
183
1844.91 Thu Aug 6 15:42:45 CEST 2009
185 - AE::Handle::starttls could get out of sync when the read buffer
186 already contains some TLS handshake.
187 - AE::Handle did not properly free the TLS session, causing bigger
188 memory leaks in Net::SSLeay than expected (reported by toaster).
189 - AE::Socket::tcp_connect will now artificially delay invoking
190 the callback to avoid returning after invoking the callback.
191 - convert many internal watcher uses to AE API.
192
1934.9 Sat Aug 1 11:07:01 CEST 2009
194 - Glib silently fails when registering a timeout with
195 a negative value, so avoid that.
196 - call condvar callback immediately when it is set
197 after the condition is already signalled.
198 - check rbuf_max condition only after trying to consume
199 data.
200 - normalise signal numbers to names when using the ae's
201 signal handling, but do not document this (yet).
202 - pure perl signal emulation did not properly set nonblocking
203 mode on the signal pipe on win32 (this is unlikely to have caused
204 any issues).
205 - new module: AnyEvent::Debug.
206 - AnyEvent::Strict now checks that a signal was specified
207 by name (not name or number).
208 - reduce memory footprint in typical cases by ~50kb by
209 only compiling signal code when necessary.
210 - add AnyEvent::Handle->rbuf_max.
211 - grab the AE namespace for future new API, implement stubs for
212 said future/alternative API.
213 - new function: AnyEvent::Socket::format_hostport.
214
2154.881 Tue Jul 28 12:51:53 CEST 2009
216 - work around a bug in local in pre-5.10 perls, causing
217 AnyEvent::Handle to recurse when it should not
218 (analysed by elmex).
219
2204.88 Tue Jul 28 04:04:37 CEST 2009
221 - re-bless the handle into a dummy package after calling
222 AnyEvent::Handle::destroy, so the user does not need to check
223 for errors after every push_write etc.
224 - do not attempt to run t/02_signals.t on obviously broken
225 platforms.
226
2274.87 Sun Jul 26 02:06:16 CEST 2009
228 - do not attempt to linger when there is no longer a valid fh.
229 - enforce tls mode to be either accept or connect, do not simply
230 segfault in Net::SSLeay.
231 - AnyEvent::Handle can now call tcp_connect itself (new parameters
232 connect, on_prepare, on_connect and on_connect_error). Updated
233 tutorial accordingly.
234 - add AnyEvent::Impl::Irssi backend.
235
2364.86 Mon Jul 20 23:52:29 CEST 2009
237 - since the verbose warning is not enough, explicitly document
238 that versions before 1.33 of Net::SSLeay are not secure.
239 - work around signal handling races in Event and (...) Event::Lib.
240 - try to align signal-race timer to full-second boundaries.
241 - work around Tk not liking negative timeouts.
242 - don't complain of different grades of environmental unfriendlyness
243 in IO::Async.
244
2454.85 Sat Jul 18 06:16:14 CEST 2009
246 - nail the signal race problem in perl once and for all
247 (see $AnyEvent::MAX_SIGNAL_LATENCY).
248 - take advantage of Async::Interrupt if it is available.
249 - load Time::HiRes and Guard modules on demand only.
250 - add optional/recommended modules section to AnyEvent
251 documentation.
252 - reduce memory usage considerably (and reduce startup penalty)
253 by not using "strict", "warnings" and "overload" modules.
254 - work around buggy windows/openbsd perls and provide EBADMSG
255 and EPROTO ourselves when missing.
256 - improve perl 5.6 compatibility of the core event loop.
257 - made Net::SSLeay version 1.33 a soft requirement.
258
2594.83 Fri Jul 17 16:56:26 CEST 2009
260 - implement AnyEvent::Socket::getprotobyname.
261 - AnyEvent::CondVar's will now detect recursive blocking
262 waits and will croak, as too many people fall into
263 this trap.
264 - AnyEvent::Handle will now call ->destroy on itself after
265 executing the on_error callback, instead of doing some
266 half-baked internal shutdown, for fatal errors.
267 - clarify on_eof behaviour w.r.t. the read queue and
268 on_read callbacks.
269 - ignore some possible spurious wake-ups in tcp_connect.
270
2714.82 Sat Jul 11 00:34:55 CEST 2009
272 - POE and Event backends didn't accept some callable objects as
273 callbacks.
274 - use Config module instead of POSIX module to detect signal names
275 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
276 module doesn't even have all POSIX signals :/.
277 - use more workarounds around the many refcnt/corruption bugs in
278 Event::Lib.
279 - work around a race condition in perl's select, causing t/03_child.t
280 to rarely fail.
281
2824.81 Thu Jul 9 10:30:30 CEST 2009
283 - AnyEvent::Handle didn't properly diagnose write errors
284 (it expected -1 from syswrite, how lame... :).
285 - support file descriptors in addition to file handles
286 in AnyEvent->io.
287 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
288 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
289 and PERL_ANYEVENT_CA_PATH.
290 - provide a sensible synopsis section for AnyEvent::TLS.
291 - add a "supported backends" section to the manpage.
292 - added simple io watcher test to testsuite, using a
293 portable_socketpair.
294 - tried to improve the stability of the Event::Lib backend,
295 YMMV.
296
2974.8 Mon Jul 6 23:45:16 CEST 2009
298 - AnyEvent::DNS did not properly follow CNAME records with
299 uppercase targets.
300 - AnyEvent::DNS would errornously return AAAA records
301 with v4 mapped addresses (a faulty record) as ipv4 addresses,
302 causing AnyEvent::Socket to throw an exception.
303 - added new module AnyEvent::TLS for easier SSL/TLS context
304 creation, with many options including hostname verification,
305 secure default configuration, lots of documentation and,
306 predefined diffie-hellman keys for perfect forward security
307 and much more. get it while it's still fresh!
308 - use AnyEvent::TLS in AnyEvent::Handle for context management.
309 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
310 so AnyEvent::Socket users have smaller memory footprint
311 in the common case.
312 - add AnyEvent::Handle->push_shutdown method.
313 - add an additional error message parameter to AnyEvent::Handle's
314 on_error callback (for TLS, $! is still available).
315 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
316 - make AnyEvent::Handle more robust against early conenction
317 failures (during new), and return C<undef> in that case
318 from the constructor.
319 - AnyEvent::Handle will now try to load only JSON::XS first,
320 then fall back to JSON.
321 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
322 and because it was documented that way.
323
3244.451 Fri Jul 3 00:28:58 CEST 2009
325 - do not clear rbuf when shutting down an AnyEvent::Handle
326 object - doing so breaks AnyEvent::HTTP.
327
3284.45 Mon Jun 29 22:59:26 CEST 2009
329 - a write error could cause AnyEvent::Handle to create
330 an I/O watcher with an undefined $fh.
331 - special-case mapped ipv4 addresses in both
332 AnyEvent::Socket::format_address and parse_address,
333 to treat them just like ipv4 addresses.
334 - updated and overhauled the AnyEvent::Intro doc.
335 - implement AnyEvent::Socket::format_ipv4/ipv6.
336 - slightly speed up portable_pipe/socketpair.
337 - expand condvar begin/end documentation.
338
3394.42 Fri Jun 26 08:32:18 CEST 2009
340 - preliminary and neccesarily incomplete support for IO::Async.
341 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
342 case it was set to IGNORE, to ensure we can catch child statuses
343 even when the calling env acts stupidly.
344 - updated benchmarks with IO::Async, which performs very well.
345
3464.412 Wed Jun 24 01:35:57 CEST 2009
347 - support an "untaint" attribute for AnyEvent::DNS
348 and set it on the default resolver.
349 - implement and document AnyEvent::Impl::Perl::loop.
350 - remove all anyevent-env variables from %ENV
351 when running in tainted mode.
352 - mention and extend the IO::Lambda benchmark.
353
3544.411 Sun Jun 7 18:48:02 CEST 2009
355 - do not try to use F_SETFD on windows.
356
3574.41 Thu May 14 06:40:11 CEST 2009
358 - work around issues in older perls (5.8.0?) when
359 a signal handler is deleted from the %SIG hash.
360 - use POSIX::_exit in child test, to avoid running
361 destructors.
362 - speed up CHLD handling by relying on SIGCHLD being
363 synchronously delivered, even when we roll our own
364 implementation.
365 - AnyEvent::DNS: add the "dname" resource record name
366 for cosmetic reasons.
367
3684.4 Sun Apr 26 20:12:33 CEST 2009
369 - implemented idle watchers, where applicable.
370 - AnyEvent->time died when Event backend was in use.
371 - fix a memleak in the Tk backend.
372 - sped up Tk timer handling.
373 - clip negative "after" values to 0 in AnyEvent::Impl::Event
374 to avoid spamming out warning messages.
375 - fix Qt timers without interval.
376 - avoid Qt zero-timeout special case, in old libqt's.
377
3784.352 Mon Apr 20 16:31:11 CEST 2009
379 - fix AnyEvent::Strict error messages for child watchers.
380 - fix/update Makefile.PL "recommends" versions.
381 - add AnyEvent->now_update.
382
3834.351 Sat Apr 11 07:56:14 CEST 2009
384 - actually make the signal pipe work under win32.
385 - localise $! in signal handler to avoid changing $!, although
386 perl itself does not seem to save/restore errno either.
387 - set the cloexec flag on the signal pipe (normally set by perl too).
388
3894.35 Fri Mar 27 11:48:20 CET 2009
390 - event models relying on AnyEvent's signal watcher emulation did
391 invoke the callback asynchronously, contradicting documentation
392 and causing signals to get lost (this includes AnyEvent's own
393 event loop). AnyEvent now uses the standard pipe trick to make
394 callback execution synchronous to the event loop.
395 - AnyEvent::Handle didn't free TLS context data on DESTROY
396 (patch by Pavel Shaydo).
397 - work around the ever-incompatibly-changing API of MakeMaker.
398 - document that changing global variables without restoring them
399 is a bad idea in Perl (noted by Adam Rosenstein).
400 - AnyEvent::Strict now barfs if ->io is passed a file.
401
4024.34 Thu Feb 12 18:32:45 CET 2009
403 - separately buffer TLS read data, as otherwise the read queue could
404 deadlock as receiving data is not expected while draining the
405 read queue (which cna only happen with TLS).
406 - raise EBADMSG error on JSON decoding errors.
407 - fix some minor manpage bugs (reported by Maximilian GaรŸ).
408 - speed up select bitmask parsing quite a bit in the pure perl
409 backend.
410 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
411
4124.331 Tue Jan 6 21:07:25 CET 2009
413 - socketpair fails on many vista machines because vista has
414 completely broken accept/getpeername and getsockname functions,
415 so we provide our own socketpair emulation that kind of works
416 (AnyEvent::Util::portable_pipe).
417 - new function: AnyEvent::Util::portable_socketpair.
418 - take advantage of the Guard module if it exists.
419
4204.33 Fri Nov 21 02:35:40 CET 2008
421 - AnyEvent::Strict did errornously flag a fileno of 0 as
422 illegal.
423 - reduce memory usage and slightly speed up the pure perl backend
424 by only storing the file descriptor, not the file handle.
425 - add missing autocork method to AnyEvent::Handle (reported
426 by Adam Rosenstein).
427 - AnyEvent::DNS->resolve errornously documented an $rcode result
428 argument, but there isn't (spotted by Henrik Krohns).
429 - the naked truth about Tk - it's basically unmaintained.
430
4314.32 Mon Nov 3 22:46:32 CET 2008
432 - fix AnyEvent::Socket::resolve_sockaddr to properly support
433 unix sockets again and choose a proper default for
434 the socket type.
435 - fix call to resolve_sockaddr in tcp_connect.
436
4374.31 Thu Oct 30 04:41:48 CET 2008
438 - implemented AnyEvent::Handle->destroy method.
439 - hint about unexpected effects in TLS mode.
440 - speed up AnyEvent::Impl::Perl by using more arrays
441 instead of hashes without the slightest loss of readability :)
442 - work around a perl argument refcounting bug.
443 - with some perl patches applied, AnyEvent now seems leak-free.
444
4454.3 Fri Oct 3 09:18:43 CEST 2008
446 - AnyEvent will now install a no-op signal handler for SIGPIPE,
447 unless one has been installed already.
448 - warn about not loaidng AnyEvent::Impl::POE early enough
449 (patch by Adam Rosenstein).
450 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
451 - fix a bug in where in SSL connect mode, the
452 client would first wait for some data by the server and
453 otherwise hang. (reported and analysed in an absolutely
454 exemplary manner by Adam Rosenstein).
455 - fix a bug in where SSL EOF would not be treated
456 as stream EOF, putting the connection into a hung state.
457 (reported and analysed in an absolutely exemplary manner by Adam
458 Rosenstein).
459 - fix a potential segfault when the TLS context
460 would go missing in a read callback (e.g. due to stoptls) -
461 Net::SSLeay of course makes no type checking whatsoever.
462 - AnyEvent::Handle will not stop the read watcher in TLS mode.
463 - AnyEvent::Handle->stoptls will send a TLS close notify
464 instead of simply destroying the stream now.
465 - fix a bug where large blocks of data written in TLS mode
466 would not be sent unless triggered by receives.
467 - on_drain will now take the tls write buffer into account.
468 - SSL operations have been streamlined a bit and should be faster
469 now, more to come, though.
470 - added a FAQ section, will document nontrivial issues in there.
471
4724.234 Mon Sep 29 04:08:13 CEST 2008
473 - fix child watcher documentation: only child exits will be reported,
474 no trace events.
475 - mention SIGPIPE in AnyEvent::Handle.
476 - perl backend will now use POSIX::times as monotonic clock
477 when available and the monotonic clock option is not.
478 - run even without Time::HiRes in some cases now, but do not
479 provide sub-second accuracy (all *supported* perl versions have
480 Time::HiRes).
481
4824.233 Fri Aug 22 01:48:05 CEST 2008
483 - fix a bug in the testsuite, causing a hang.
484 - clarified lots of AnyEvent::Handle constructor arguments.
485
4864.232 Thu Aug 21 20:44:25 CEST 2008
487 - fix a bug in AnyEvent::Handle that could cause two fatal error
488 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
489 (testcase by Leon Brocard).
490 - clarified on_eof/on_error documentation a bit.
491
4924.231 Tue Jul 29 13:12:15 CEST 2008
493 - remove some debugging code left in AnyEvent::Util::fork_call
494 (and no, it's impossible to implement with the broken windows
495 perls without resource leaks or worse).
496
4974.23 Tue Jul 29 12:19:59 CEST 2008
498 - document the first parameter passed to condvar callbacks to be
499 the callback.
500 - add AnyEvent::Socket::{ntoa,aton} aliases.
501 - optimize the AE::Handle->push_read (line) for the default
502 eol marker.
503 - optimize push_read (packstring|storable) for small packets.
504 - invoke on_error callback when no on_eof callback is set.
505 - fix a bug in push_read (storable) of unknown impact.
506
5074.22 Sun Jul 20 16:34:13 CEST 2008
508 - new function AnyEvent::Socket::parse_hostport.
509 - as the bulkheads at microsoft can't even get getprotobyname reliably
510 working on their shitty fucking broken os we need to hardcode
511 some common protocol numbers in AnyEvent::Socket. How can
512 people even bother with such a pile of shit as windows.
513
5144.21 Thu Jul 17 14:40:05 CEST 2008
515 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
516 which is incompatible to previous invocation syntax.
517 - work around more windows perl bugs in fork_call. I cannot imagine
518 why anybody would bother implementing fork in such an obviously
519 unusable way.
520 - avoid calling $do_retry if the request has already finished in
521 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
522 K. Sharifulin).
523
5244.2 Sat Jul 12 22:42:11 CEST 2008
525 - support an "interval" argument for repeating timers.
526 - fix a bug in the pure perl event loop that caused timers to
527 be delayed under some circumstances.
528 - implement optional argument checking via PERL_ANYEVENT_STRICT.
529 - not importing anything from Event causes it to malfunction
530 and not use Time::HiRes, work around this bug.
531 - more bug workarounds against the endlessly broken Tk module,
532 seems to work now on 32 bit machines, but less so on 64 bit.
533 - minor optimisations applied to most backends.
534
5354.161 Fri Jul 4 14:24:48 CEST 2008
536 - fixed the t/handle/02_write.t, it read after write, with the
537 assumption that all data could be written before read
538 (reported by Jim Razmus).
539
5404.160 Thu Jul 3 04:02:21 CEST 2008
541 - re-registering signal watchers after unregistering one did
542 not work if a backend falls back on AnyEvent's default
543 implementation (also affected child watchers).
544 - new AnyEvent::Handle options: autocork and no_delay.
545
5464.152 Sun Jun 22 14:15:44 CEST 2008
547 - allow for 32-bit perls that implement shifts differently
548 on different architectures in parse_ipv4 (reported and
549 analysed by Keiichi DAIBA).
550
5514.151 Fri Jun 6 17:34:24 CEST 2008
552 - make sure specifying _only_ on_read and never pushing reads
553 works.
554
5554.15 Fri Jun 6 13:00:46 CEST 2008
556 - the pure perl backend would keep some watchers alive when more than
557 one watcher was registered for the same fd.
558 - new "packstring" and "storable" read and write types
559 for AnyEvent::Handle.
560 - allow on_eof handler to be called after on_error with EPIPE returns.
561 - do not immediately call on_read callback in handle constructor.
562
5634.14 Thu Jun 5 20:29:31 CEST 2008
564 - Fixed a bug in DNS SRV priority sorting.
565 - AnyEvent::Util::guard now reports runtime errors while
566 executing the guard block as warnings.
567 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
568
5694.13 Thu Jun 5 00:47:59 CEST 2008
570 - AnyEvent::DNS only followed cname chains with length 2,
571 contrary to documentation. bumped it up to 10, thanks to
572 microsoft, the current broken-dns-config-king.
573 - AnyEvent::DNS didn't check for socket return status
574 but instead relied on perl not creating filehandles in that
575 case - too bad it gives you a bogus file handle (reported
576 and analysed by Vladimir Timofeev).
577 - fix queue management logic in AnyEvent::Handle:
578 when on_read was registered, the queue was empty
579 and no progress could be made AnyEvent::Handle would
580 enter an endless loop.
581 - correctly start reading again when the handle became
582 busy again after idling.
583 - correctly treat tls shut-downs as EOF condition
584 (for the time being, would be nice to have a callback for that).
585 - correctly call eof callback under all conditions
586 (hopefully).
587 - the timeout callback did not expect that $self can go
588 away any time.
589
5904.12 Tue Jun 3 10:58:04 CEST 2008
591 - include AnyEvent::Intro, a tutorial for anyevent,
592 anyevent::socket and anyevent::handle.
593 - allow more options in on_error.
594
5954.11 Fri May 30 23:42:25 CEST 2008
596 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
597 reverse_lookup and reverse_verify to replace it, support
598 v4mapped and v4compat addresses.
599 - provide more documentation for the resolver class.
600 - really replace longest run of :0: by :: in format_address,
601 also properly convert :: and ::1 again.
602 - support NAPTR record name and decode it.
603 - implement random weight sampling for SRV records, as per
604 rfc 2782.
605 - correctly abort on srv-record targets of ".".
606 - added AnyEvent::DNS::wait_for_slot.
607 - in the unlikely event of a virtual circuit connection
608 being invalidated by a delayed udp reply, AnyEvent::DNS
609 could die. this has been fixed.
610 - plug a probable memleak in the DNS vc code.
611 - use configured timeout also for VC request phase.
612 - implement timeout and max_outstanding methods
613 for resolver class.
614 - update version numbers in all modules.
615
6164.1 Thu May 29 05:45:40 CEST 2008
617 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
618 which mirrors their purpose better (old names still available
619 for a while). Also moved AnyEvent::DNS::addr to
620 AnyEvent::Socket::resolve_sockaddr.
621 - implement AnyEvent->time and AnyEvent->now.
622 - fix IPv6 support in pack_sockaddr.
623 - officially un-experimentalise all newly introduced code.
624 - support unix domain sockets everywhere by specifying
625 a host of "unix/" and the pathname as service.
626 - implement an activity timeout in AnyEvent::Handle.
627 - added a regex read type to AnyEvent::Handle.
628 - added a json read/write type to AnyEvent::Handle.
629 - always croak in on_error in AnyEvent::Handle.
630 - document how to subclass AnyEvent::Handle.
631 - implement AnyEvent::Util::fork_call.
632 - add support for IPv6 nameservers and nameserver statements.
633 - work around _yet_ _another_ windows perl bug in where empty
634 select masks cause errors in the pure perl backend (this
635 workaround cannot be implemented for the other event loops,
636 but EV also works around this).
637 - supply AnyEvent::Socket::address_type and make good use of it.
638 - clarify and fix the Handle documentation w.r.t $self vs. $handle
639 vs. callback arguments.
640 - add some recommends to the META.yml.
641
6424.05 Mon May 26 19:44:06 CEST 2008
643 - some platforms ignore the rfc and prepend an extra sa_len member to
644 the sockaddr structure, cater for those. also use sockaddr_family,
645 didn't know it was there, but comes in handy for the workaround.
646 - undo the import WIN32 hack.
647
6484.04 Mon May 26 08:03:31 CEST 2008
649 - try to work around yet another windows bug: failed connects are reported
650 as if it were out-of-band data. windows users: you suck. Only
651 EV, Glib, Event and the pure perl backend can handle this condition.
652 - optimize write algorithm in AnyEvent::Handle.
653 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
654 - tcp_connect now properly iterates through all targets.
655 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
656 support entirely. also hardcode some AF_INET6 constants for the
657 sake of perl 5.8.
658 - vastly improved nameserver/suffix detection algorithm for windows
659 (still a hack...)
660 - try to use a fallback dns server if no dns servers could be found.
661 - splatter around some local $SIG{__DIE__}'s to work around bad
662 die handlers.
663
6644.03 Sun May 25 05:05:57 CEST 2008
665 - work around perl bugs on windows, where perl returns undocumented
666 error codes for sysread, syswrite etc.
667 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
668 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
669 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
670 - updated documentation.
671 - lots of workarounds for perl 5.8 quirks.
672 - implement a more flexible type system for both reads and writes
673 in AnyEvent::Handle.
674 - support IPv6 in tcp_server.
675 - pass local hostname and port to the prepare callback in tcp_server.
676 - make types extendable in AnyEvent::Handle.
677 - croak, not die, in AnyEvent::Handle;
678
6794.0 Sat May 24 19:58:08 CEST 2008
680 - added (experimental) AnyEvent::DNS module.
681 - added (experimental) AnyEvent::Socket module.
682 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
683 - make condvar->end call ->send as documented.
684 - make condvar constructor support the documented cb => argument.
685 - overload condvars so you can use them instead of code references.
686 - support more of the POSIX forms instead of a pure dotted quad for
687 verbatim ip addresses.
688 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
689 - try very hard not to freeze on broken windows perls.
690 - make on_eof optional in AnyEvent::Handle.
691 - support service names in tcp_connect and tcp_server.
692 - work around netbsd/openbsd bug causing unix sockets not to
693 become writable until empty.
694 - only make a dummy read for error detection when getpeername
695 returns ENOTCONN.
696 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
697 override.
698 - new AnyEvent::Guard::cancel.
699
7003.5 Sat May 17 23:17:14 CEST 2008
701 - gracefully deal with systems not defining CLOCK_MONOTONIC
702 (as opposed to not supporting the clock itself, which worked in 3.41).
703 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
704 helpers, to replace the ill-designed AnyEvent::Socket class.
705 - added some examples (eg/connect, eg/handle).
706 - added AnyEvent::Util::guard.
707 - AnyEvent::Util now exports most of its functions.
708 - fixed a bug in *_read_chunk, where the callback was called with $self as
709 first and second argument.
710 - simplified and fixed regex handling in AnyEvent::Handle::*_read_line.
711 - fix manpage condvars examples to use new syntax.
712 - considerably improve inet_aton with EV::ADNS backend to support
713 cname chains (common dns configuration bug, e.g. with www.google.de)
714 and aaaa records if no a record could be found.
715
7163.41 Sun May 11 19:53:13 CEST 2008
717 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
718
7193.4 Sun May 11 00:29:25 CEST 2008
720 - complete re-implementation of condvars with a hopefully much more useful
721 API (backward compatible functions exist but are undocumented).
722 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
723 (still subject to change).
724 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
725 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which
726 allows module authors to avoid forcing event loop detection.
727 (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example).
728 - remove coro backends: Coro now provides generic support for AnyEvent
729 via Coro::AnyEvent.
730
7313.3 Mon Apr 28 09:51:06 CEST 2008
732 - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules.
733 - fix a bug in the pure perl backend that kept watchers alive
734 when multiple watchers were registered for the same fh/poll combo.
735 - add a benchmark section showing AnyEvent overhead and comparing
736 the different event loops with each other.
737 - prefer pure perl over tk when autoprobing, as it's about as fast,
738 but doesn't crash with many watchers.
739 - declare Qt support non-experimental.
740 - clarify the confusing section about the file descriptor being
741 kept alive.
742 - document the race between loading of an event module and
743 child processes exiting.
744 - support POE as "backend" (with some caveats, POE is not generic
745 enough, and darn slow).
746 - support Wx and Prima through POE.
747 - optimise perl backend to use 20% less memory and take advantage
748 of typical timeout behaviour. It can now compete with
749 select/poll-based C event loops in most cases (it is usually
750 faster than Event and Glib :).
751 - roughly cut EV memory use in half and increase its speed by 30%,
752 by removing undocumented functionality. Did something similar
753 to other event loops where possible.
754
7553.2 Thu Apr 24 10:10:40 CEST 2008
756 - do not die when anyevent watchers are destroyed while running callbacks
757 in the pureperl backend (could only happen when two watchers are registered
758 for the same fh/poll combo).
759 - support autoloading for child watchers, was broken in all versions.
760 - implement PERL_ANYEVENT_MODEL env variable.
761 - (experimental) implement interface to the (very crashy/buggy) Event::Lib module.
762 - (experimental) implement interface to the Qt module (cannot be autoprobed).
763 - this release is IO:AnyEvent-proof.
764
7653.12 Tue Apr 22 07:11:46 CEST 2008
766 - reinstate AUTHOR section that got lost somehow.
767 - do not hang in the testsuite with badly broken perls (activestate,
768 strawberry...), but instead diagnose the problem and continue.
769 - use INT instead of CHLD in an attempt to support broken windows perls
770 better (this decreases the test quality, unfortunately).
771 - do not send the signal to the process group
772 (no problem for CHLD, bad for INT :).
773
7743.11 Sat Apr 19 06:57:31 CEST 2008
775 - major documentation rework.
776 - document the fact that child watchers only watch for zombies.
777 - fix the child watcher example.
778
7793.1 Wed Apr 16 17:09:01 CEST 2008
780 - work around recurring bugs in Tk by dup'ing filehandles, the
781 only method with good success chances on Tk (the bugs apparently
782 don't get fixed anytime soon).
783 - lift the restriction of only one watcher per fh direction
784 (as the Tk bug workaround also lifts it and only Tk imposed
785 such strong limits).
786 - changed probe order to prefer coro adaptors.
787 - explain why recursion into the event loop is not supported
788 unless the backend supports it (only Coro::EV does without
789 any restrictions...).
790 - add simple manpages for all backend modules.
791
7923.0 Mon Apr 7 21:30:23 CEST 2008
793 - Coro::Signal changed semantics, roll our own, also cleaning
794 up the Coro implementation in general.
795 - rename Coro backend to CoroEvent.
796 - add some decision helping paragraph to the manpage that should
797 help people to decide whether AnyEvent is the right thing for them.
798
7992.9 Mon Jan 28 13:31:54 CET 2008
800 - update for EV 3.0 API changes.
801
8022.8 Sun Nov 25 15:06:03 CET 2007
803 - waitpid can validly return 0. accept this fact of life
804 instead of reporting it to any watchers.
805
8062.7 Fri Nov 23 11:41:14 CET 2007
807 - force use of AnyEvent::Impl::Perl in testsuite, there is
808 too much breakage outside AnyEvent.
809 - deliver signals synchronously in AnyEvent::Impl::Perl
810 or any other event loop that relies on AnyEvents child
811 watcher emulation. *Could* help with hanging testsuite
812 (except when Event or EV are installed).
813
8142.6 Fri Nov 9 20:36:35 CET 2007
815 - fix bug in testsuite.
816 - move EV adaptor modules to AnyEvent.
817 - add Coro+EV adaptor module.
818
8192.55 Tue Nov 6 17:41:32 CET 2007
7 - add EV to the list of supported event models. 820 - add EV to the list of supported event models.
8 - do not auto-reset pid watchers, pass pid and status to them. 821 - do not auto-reset pid watchers, pass pid and status to them.
9 - allow a pid of zero to watch for all children in child watchers. 822 - allow a pid of zero to watch for all children in child watchers.
10 823
112.54 Wed Jul 18 17:36:23 CEST 2007 8242.54 Wed Jul 18 17:36:23 CEST 2007
… …
75 - fixed documentation. 888 - fixed documentation.
76 - clarified some corner cases. 889 - clarified some corner cases.
77 - vastly improved documentation with a long example. 890 - vastly improved documentation with a long example.
78 891
790.2 Thu Dec 1 22:18:49 CET 2005 8920.2 Thu Dec 1 22:18:49 CET 2005
80 - check wether a valid method was called, to avoid endless loops. 893 - check whether a valid method was called, to avoid endless loops.
81 - move condvars into their own class (cleaner, safer). 894 - move condvars into their own class (cleaner, safer).
82 - don't die when some modules are not available. 895 - don't die when some modules are not available.
83 - Tk is too broken to allow for words. Will work once Tk works. 896 - Tk is too broken to allow for words. Will work once Tk works.
84 897
850.1 Thu Dec 1 19:44:57 CET 2005 8980.1 Thu Dec 1 19:44:57 CET 2005

Diff Legend

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