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

Comparing AnyEvent/Changes (file contents):
Revision 1.84 by root, Wed May 21 14:34:42 2008 UTC vs.
Revision 1.419 by root, Mon Nov 15 22:29:36 2010 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
33.51 3TODO: document TCP_*** constants
4TODO: catch AE::xxx for anyevent struct and anyevent debug
5TODO: anyevent::trace? or debug tracing all watchers?
6TODO: anyevent strict etc. for AE::
7TODO: anyevent strict checking for $_ changes?
8
9 - convert EV backend to EV 4.00 API.
10 - AE::Handle: implement dynamic read_size adjustment, add
11 max_read_size parameter, reduce default read_size to 2048.
12 - work around OS bugs (cygwin again) when lingering in AE::Handle.
13 - reduce memory usage slightly when connect was used in AE::Handle.
14 - document tls_ctx => undef as valid.
15
165.28 Wed Oct 13 04:14:23 CEST 2010
17 - due to a glitch, AnyEvent's internal getprotobyname wasn't
18 actually used in tcp_server.
19 - implement AnyEvent::CondVar->new method, which was documented but
20 missing.
21 - support autoloading of read types in unshift_read.
22 - AnyEvent::Strict: do the test for invalid fh arguments differently,
23 resulting in more informative error messages.
24 - new AnyEvent::FAQ manpage.
25 - clarify that you can actually call push_read etc. in on_read.
26 - put keys and certs in /tmp, not into the current directory,
27 and document this in the SECURITY CONSIDERATIONS sections
28 of AnyEvent::TLS.
29 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
30 requested by Ron Isaacson.
31 - document $AnyEvent::DNS::RESOLVER variable.
32 - incorporated major typo patches by Abhijit Menon-Sen.
33 - AnyEvent::Handle now croaks when negative timeouts are passed.
34 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
35 undocumented, but who knows when they might come in handy.
36
375.271 Tue Jun 8 12:05:46 CEST 2010
38 - backport to perl 5.8.x.
39
405.27 Sun Jun 6 12:12:05 CEST 2010
41 - postpone differently in AnyEvent::Socket now, as
42 when not, canceling the connection attempt might fail
43 (found by Felix Antonius Wilhelm Ostmann).
44 - explicitly check for non-stream sockets in AE::Handle, too many
45 clueless people fell into the trap of this somehow working.
46 - simplified and reworked the "OTHER MODULES" section.
47 - better/more condvar examples.
48
495.261 Wed Apr 28 16:13:36 CEST 2010
50 - AF_INET6 was not properly used from Socket6 during configuration
51 time (found by Andy Grundman).
52
535.26 Mon Apr 12 04:49:35 CEST 2010
54 - don't generate spurious readyness notifications when select
55 returns EINTR in the pure perl backend (analysed and testcase
56 by Bas Denissen).
57 - give same examples in SYNOPSIS sections of AnyEvent and AE.
58 - provider faster implementations for AE::time/now/now_update for
59 EV and Perl backends.
60 - provide faster AE::cv and AE::time implementations for all
61 backends.
62 - no longer support lower/mixed-case signal names (this was
63 never documented nor universally supported).
64 - some more memory µ-optimisations, and somewhat less messy
65 code to cope with both AnyEvent and AE APIs.
66
675.251 Sat Mar 13 00:58:21 CET 2010
68 - make test failed if it was run before make install on perl <=
69 5.8.8, and older versions of perl make our live very difficult,
70 so write constants.pl during Makefile.PL time.
71
725.25 Sat Mar 13 00:23:14 CET 2010
73 - fix a race condition in AnyEvent::Handle that would cause
74 a "bio_read: unsupported method" error (in _another_
75 ssl connection) after user code threw an exception.
76 - added AnyEvent::Handle->destroyed method.
77 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
78 - the AnyEvent::Util::fh_nonblocking function worked only
79 by lucky accident on win32.
80 - smaller and faster AnyEvent::Util::fh_nonblocking.
81 - when the (required!) Time::HiRes module is missing, AnyEvent
82 did not fall back to built-in time properly.
83 - do not load Fcntl at runtime, saving memory and loading time.
84 - precompile a number of constants and use them instead of runtime
85 detection and eval.
86 - free detection code after detection and similar memory optimisations.
87 - Perl backend timer interval best effort drift has been improved
88 (same algorithm as EV).
89 - update unicode idna mapping table.
90
915.24 Tue Jan 5 11:39:43 CET 2010
92 - cygwin never reports errors from failed connects. 1.5 just gives
93 you continous readyness and EAGAIN, 1.7 is even more broken and
94 just hangs. work around both issues in a cygwin-specific hack.
95 - improve idle watcher documentation slightly.
96
975.23 Sun Dec 20 23:48:00 CET 2009
98 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
99 - implement punycode_encode/decode, idn_nameprep,
100 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
101 - provide $AE::VERSION.
102 - removed traces of "no strict 'refs'".
103
1045.22 Sat Dec 5 03:51:13 CET 2009
105 - downgrade-or-fail in AnyEvent::Handle::push_write, to
106 diagnose encoding failures earlier and more succinctly.
107 (this works around bugs in perl, throwing away encoding info
108 when passing scalar data to extensions).
109 - add more examples to AnyEvent::Socket manpage.
110 - upgrade internal warning set to the same as common::sense 2.03.
111 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
112 (John Beppu).
113 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
114 - add google public dns servers to fallback server set and make sure
115 we load-balance properly between the three sets. also add all
116 fallback dns servers, not just a random one, to each dns config.
117
1185.21 Thu Nov 19 02:48:47 CET 2009
119 - fix a problem where socket constants were called with parameters
120 (spotted by David Friedland).
121 - fork_call never use'd POSIX (reported by Daisuke Maki).
122 - improve perl 5.6 compatibility further (but it still won't work
123 unless you rip out everything but the core).
124 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
125 uses 10MB of RAM, but doesn't open a console window. *sigh*.
126
1275.202 Wed Oct 14 22:35:44 CEST 2009
128 - AnyEvent::DNS would unexpectedly clobber $_ under windows
129 (analysed by Matthias Waldorf).
130 - AnyEvent::Handle::run_cmd can now pass the PID of the
131 newly-created process, which is much less useful than it might
132 sound (based on patch by Yann Kerherve).
133
1345.201 Tue Sep 29 12:09:25 CEST 2009
135 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
136 (reported by Torsten Foertsch).
137 - common::sense 2.0 could cause tcp_server to throw an exception
138 (analysed by elmex).
139
1405.2 Mon Sep 14 07:04:49 CEST 2009
141 - INCOMPATIBLE CHANGE: do no longer support register_read_type
142 and register_write_type in AnyEvent::Handle, instead support
143 package names (the facility was mostly abused).
144 - implement "packagename-as-read/write type" support in
145 AnyEvent::Handle.
146 - AnyEvent::Handle: new options "keepalive" and "oobinline".
147 - oobinline set by default to avoid security issues.
148 - the pure-perl event loop backend wrongly detected times() fallback
149 support (spotted by Pavel Boldin).
150
1515.12 Tue Sep 1 20:26:50 CEST 2009
152 - be more lenient when parsing resolv.conf files, as some people
153 use hashmarks as comment indicator inside directives (reported
154 by Michael S. Fischer).
155 - use same set of warnings as common::sense 2.0.
156 - fix a potential 32 bit overflow issue due to perl having problems
157 with large hex constants in the Perl backend.
158 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
159 by default.
160 - fix some prototypes.
161 - (optionally) require Async::Interrupt 1.02.
162 - replace opendns fallbacks by verizon/level3 public dns servers.
163
1645.112 Fri Aug 21 13:59:04 CEST 2009
165 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
166 using starttls with an illegal/uncreatable tls context.
167 - port to broken win32 perls.
168
1695.111 Sun Aug 16 18:44:36 CEST 2009
170 - perl errornously requires sockaddr_un structures returned by the
171 kernel to have a specific length. work around this bug by padding
172 sockaddr_un structures (found to fail on OS X, as perl only uses
173 the correct code for GNU/Linux, but likely to fail on other
174 OSes as well).
175
1765.11 Wed Aug 12 17:49:37 CEST 2009
177 - fixed a bug in run_cmd with input from filename.
178 - create a json coder object when none is given in write direction
179 to reduce dependency on JSON.
180
1815.1 Tue Aug 11 03:17:32 CEST 2009
182 - tcp_connect now reports the correct error in $!, not an unrelated
183 one (was broken since 4.91).
184 - AE::Handle did not properly clear rtimeout/wtimeout watchers
185 when retrying a tcp_connect.
186 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
187 - fixed wrong documentation in AE::Socket.
188
1895.01 Mon Aug 10 03:16:32 CEST 2009
190 - last release broke EV child watchers :/.
191
1925.0 Sun Aug 9 17:32:09 CEST 2009
193 - officially introduce and document the AE API.
194 - lowering the timeout at runtime did not have immediate effect
195 in AnyEvent::Handle.
196 - AE::Handle now has separate and independent read and write timeout
197 settings.
198 - probe for Guard module when AnyEvent::Util is loaded, not at
199 runtime (Event doesn't like eval "use" from callbacks, and it avoids
200 exporting the probe function).
201 - Impl::IOAsync's io watchers did not use the filehandle emulation
202 code to work around it's watcher limits.
203 - work around IO::Async's problems with overloaded objects.
204 - try to disable wearnings inside IO::Async, it's just too buggy.
205 - do no longer use Event by default, only EV or Perl, Event is too buggy.
206 - hackishly provide fast AE:: interface when EV or Perl is the backend.
207 - inherit VERSION from AnyEvent::VERSION in most modules.
208
2094.91 Thu Aug 6 15:42:45 CEST 2009
210 - AE::Handle::starttls could get out of sync when the read buffer
211 already contains some TLS handshake.
212 - AE::Handle did not properly free the TLS session, causing bigger
213 memory leaks in Net::SSLeay than expected (reported by toaster).
214 - AE::Socket::tcp_connect will now artificially delay invoking
215 the callback to avoid returning after invoking the callback.
216 - convert many internal watcher uses to AE API.
217
2184.9 Sat Aug 1 11:07:01 CEST 2009
219 - Glib silently fails when registering a timeout with
220 a negative value, so avoid that.
221 - call condvar callback immediately when it is set
222 after the condition is already signalled.
223 - check rbuf_max condition only after trying to consume
224 data.
225 - normalise signal numbers to names when using the ae's
226 signal handling, but do not document this (yet).
227 - pure perl signal emulation did not properly set nonblocking
228 mode on the signal pipe on win32 (this is unlikely to have caused
229 any issues).
230 - new module: AnyEvent::Debug.
231 - AnyEvent::Strict now checks that a signal was specified
232 by name (not name or number).
233 - reduce memory footprint in typical cases by ~50kb by
234 only compiling signal code when necessary.
235 - add AnyEvent::Handle->rbuf_max.
236 - grab the AE namespace for future new API, implement stubs for
237 said future/alternative API.
238 - new function: AnyEvent::Socket::format_hostport.
239
2404.881 Tue Jul 28 12:51:53 CEST 2009
241 - work around a bug in local in pre-5.10 perls, causing
242 AnyEvent::Handle to recurse when it should not
243 (analysed by elmex).
244
2454.88 Tue Jul 28 04:04:37 CEST 2009
246 - re-bless the handle into a dummy package after calling
247 AnyEvent::Handle::destroy, so the user does not need to check
248 for errors after every push_write etc.
249 - do not attempt to run t/02_signals.t on obviously broken
250 platforms.
251
2524.87 Sun Jul 26 02:06:16 CEST 2009
253 - do not attempt to linger when there is no longer a valid fh.
254 - enforce tls mode to be either accept or connect, do not simply
255 segfault in Net::SSLeay.
256 - AnyEvent::Handle can now call tcp_connect itself (new parameters
257 connect, on_prepare, on_connect and on_connect_error). Updated
258 tutorial accordingly.
259 - add AnyEvent::Impl::Irssi backend.
260
2614.86 Mon Jul 20 23:52:29 CEST 2009
262 - since the verbose warning is not enough, explicitly document
263 that versions before 1.33 of Net::SSLeay are not secure.
264 - work around signal handling races in Event and (...) Event::Lib.
265 - try to align signal-race timer to full-second boundaries.
266 - work around Tk not liking negative timeouts.
267 - don't complain of different grades of environmental unfriendlyness
268 in IO::Async.
269
2704.85 Sat Jul 18 06:16:14 CEST 2009
271 - nail the signal race problem in perl once and for all
272 (see $AnyEvent::MAX_SIGNAL_LATENCY).
273 - take advantage of Async::Interrupt if it is available.
274 - load Time::HiRes and Guard modules on demand only.
275 - add optional/recommended modules section to AnyEvent
276 documentation.
277 - reduce memory usage considerably (and reduce startup penalty)
278 by not using "strict", "warnings" and "overload" modules.
279 - work around buggy windows/openbsd perls and provide EBADMSG
280 and EPROTO ourselves when missing.
281 - improve perl 5.6 compatibility of the core event loop.
282 - made Net::SSLeay version 1.33 a soft requirement.
283
2844.83 Fri Jul 17 16:56:26 CEST 2009
285 - implement AnyEvent::Socket::getprotobyname.
286 - AnyEvent::CondVar's will now detect recursive blocking
287 waits and will croak, as too many people fall into
288 this trap.
289 - AnyEvent::Handle will now call ->destroy on itself after
290 executing the on_error callback, instead of doing some
291 half-baked internal shutdown, for fatal errors.
292 - clarify on_eof behaviour w.r.t. the read queue and
293 on_read callbacks.
294 - ignore some possible spurious wake-ups in tcp_connect.
295
2964.82 Sat Jul 11 00:34:55 CEST 2009
297 - POE and Event backends didn't accept some callable objects as
298 callbacks.
299 - use Config module instead of POSIX module to detect signal names
300 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
301 module doesn't even have all POSIX signals :/.
302 - use more workarounds around the many refcnt/corruption bugs in
303 Event::Lib.
304 - work around a race condition in perl's select, causing t/03_child.t
305 to rarely fail.
306
3074.81 Thu Jul 9 10:30:30 CEST 2009
308 - AnyEvent::Handle didn't properly diagnose write errors
309 (it expected -1 from syswrite, how lame... :).
310 - support file descriptors in addition to file handles
311 in AnyEvent->io.
312 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
313 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
314 and PERL_ANYEVENT_CA_PATH.
315 - provide a sensible synopsis section for AnyEvent::TLS.
316 - add a "supported backends" section to the manpage.
317 - added simple io watcher test to testsuite, using a
318 portable_socketpair.
319 - tried to improve the stability of the Event::Lib backend,
320 YMMV.
321
3224.8 Mon Jul 6 23:45:16 CEST 2009
323 - AnyEvent::DNS did not properly follow CNAME records with
324 uppercase targets.
325 - AnyEvent::DNS would errornously return AAAA records
326 with v4 mapped addresses (a faulty record) as ipv4 addresses,
327 causing AnyEvent::Socket to throw an exception.
328 - added new module AnyEvent::TLS for easier SSL/TLS context
329 creation, with many options including hostname verification,
330 secure default configuration, lots of documentation and,
331 predefined diffie-hellman keys for perfect forward security
332 and much more. get it while it's still fresh!
333 - use AnyEvent::TLS in AnyEvent::Handle for context management.
334 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
335 so AnyEvent::Socket users have smaller memory footprint
336 in the common case.
337 - add AnyEvent::Handle->push_shutdown method.
338 - add an additional error message parameter to AnyEvent::Handle's
339 on_error callback (for TLS, $! is still available).
340 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
341 - make AnyEvent::Handle more robust against early conenction
342 failures (during new), and return C<undef> in that case
343 from the constructor.
344 - AnyEvent::Handle will now try to load only JSON::XS first,
345 then fall back to JSON.
346 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
347 and because it was documented that way.
348
3494.451 Fri Jul 3 00:28:58 CEST 2009
350 - do not clear rbuf when shutting down an AnyEvent::Handle
351 object - doing so breaks AnyEvent::HTTP.
352
3534.45 Mon Jun 29 22:59:26 CEST 2009
354 - a write error could cause AnyEvent::Handle to create
355 an I/O watcher with an undefined $fh.
356 - special-case mapped ipv4 addresses in both
357 AnyEvent::Socket::format_address and parse_address,
358 to treat them just like ipv4 addresses.
359 - updated and overhauled the AnyEvent::Intro doc.
360 - implement AnyEvent::Socket::format_ipv4/ipv6.
361 - slightly speed up portable_pipe/socketpair.
362 - expand condvar begin/end documentation.
363
3644.42 Fri Jun 26 08:32:18 CEST 2009
365 - preliminary and neccesarily incomplete support for IO::Async.
366 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
367 case it was set to IGNORE, to ensure we can catch child statuses
368 even when the calling env acts stupidly.
369 - updated benchmarks with IO::Async, which performs very well.
370
3714.412 Wed Jun 24 01:35:57 CEST 2009
372 - support an "untaint" attribute for AnyEvent::DNS
373 and set it on the default resolver.
374 - implement and document AnyEvent::Impl::Perl::loop.
375 - remove all anyevent-env variables from %ENV
376 when running in tainted mode.
377 - mention and extend the IO::Lambda benchmark.
378
3794.411 Sun Jun 7 18:48:02 CEST 2009
380 - do not try to use F_SETFD on windows.
381
3824.41 Thu May 14 06:40:11 CEST 2009
383 - work around issues in older perls (5.8.0?) when
384 a signal handler is deleted from the %SIG hash.
385 - use POSIX::_exit in child test, to avoid running
386 destructors.
387 - speed up CHLD handling by relying on SIGCHLD being
388 synchronously delivered, even when we roll our own
389 implementation.
390 - AnyEvent::DNS: add the "dname" resource record name
391 for cosmetic reasons.
392
3934.4 Sun Apr 26 20:12:33 CEST 2009
394 - implemented idle watchers, where applicable.
395 - AnyEvent->time died when Event backend was in use.
396 - fix a memleak in the Tk backend.
397 - sped up Tk timer handling.
398 - clip negative "after" values to 0 in AnyEvent::Impl::Event
399 to avoid spamming out warning messages.
400 - fix Qt timers without interval.
401 - avoid Qt zero-timeout special case, in old libqt's.
402
4034.352 Mon Apr 20 16:31:11 CEST 2009
404 - fix AnyEvent::Strict error messages for child watchers.
405 - fix/update Makefile.PL "recommends" versions.
406 - add AnyEvent->now_update.
407
4084.351 Sat Apr 11 07:56:14 CEST 2009
409 - actually make the signal pipe work under win32.
410 - localise $! in signal handler to avoid changing $!, although
411 perl itself does not seem to save/restore errno either.
412 - set the cloexec flag on the signal pipe (normally set by perl too).
413
4144.35 Fri Mar 27 11:48:20 CET 2009
415 - event models relying on AnyEvent's signal watcher emulation did
416 invoke the callback asynchronously, contradicting documentation
417 and causing signals to get lost (this includes AnyEvent's own
418 event loop). AnyEvent now uses the standard pipe trick to make
419 callback execution synchronous to the event loop.
420 - AnyEvent::Handle didn't free TLS context data on DESTROY
421 (patch by Pavel Shaydo).
422 - work around the ever-incompatibly-changing API of MakeMaker.
423 - document that changing global variables without restoring them
424 is a bad idea in Perl (noted by Adam Rosenstein).
425 - AnyEvent::Strict now barfs if ->io is passed a file.
426
4274.34 Thu Feb 12 18:32:45 CET 2009
428 - separately buffer TLS read data, as otherwise the read queue could
429 deadlock as receiving data is not expected while draining the
430 read queue (which cna only happen with TLS).
431 - raise EBADMSG error on JSON decoding errors.
432 - fix some minor manpage bugs (reported by Maximilian Gaß).
433 - speed up select bitmask parsing quite a bit in the pure perl
434 backend.
435 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
436
4374.331 Tue Jan 6 21:07:25 CET 2009
438 - socketpair fails on many vista machines because vista has
439 completely broken accept/getpeername and getsockname functions,
440 so we provide our own socketpair emulation that kind of works
441 (AnyEvent::Util::portable_pipe).
442 - new function: AnyEvent::Util::portable_socketpair.
443 - take advantage of the Guard module if it exists.
444
4454.33 Fri Nov 21 02:35:40 CET 2008
446 - AnyEvent::Strict did errornously flag a fileno of 0 as
447 illegal.
448 - reduce memory usage and slightly speed up the pure perl backend
449 by only storing the file descriptor, not the file handle.
450 - add missing autocork method to AnyEvent::Handle (reported
451 by Adam Rosenstein).
452 - AnyEvent::DNS->resolve errornously documented an $rcode result
453 argument, but there isn't (spotted by Henrik Krohns).
454 - the naked truth about Tk - it's basically unmaintained.
455
4564.32 Mon Nov 3 22:46:32 CET 2008
457 - fix AnyEvent::Socket::resolve_sockaddr to properly support
458 unix sockets again and choose a proper default for
459 the socket type.
460 - fix call to resolve_sockaddr in tcp_connect.
461
4624.31 Thu Oct 30 04:41:48 CET 2008
463 - implemented AnyEvent::Handle->destroy method.
464 - hint about unexpected effects in TLS mode.
465 - speed up AnyEvent::Impl::Perl by using more arrays
466 instead of hashes without the slightest loss of readability :)
467 - work around a perl argument refcounting bug.
468 - with some perl patches applied, AnyEvent now seems leak-free.
469
4704.3 Fri Oct 3 09:18:43 CEST 2008
471 - AnyEvent will now install a no-op signal handler for SIGPIPE,
472 unless one has been installed already.
473 - warn about not loaidng AnyEvent::Impl::POE early enough
474 (patch by Adam Rosenstein).
475 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
476 - fix a bug in where in SSL connect mode, the
477 client would first wait for some data by the server and
478 otherwise hang. (reported and analysed in an absolutely
479 exemplary manner by Adam Rosenstein).
480 - fix a bug in where SSL EOF would not be treated
481 as stream EOF, putting the connection into a hung state.
482 (reported and analysed in an absolutely exemplary manner by Adam
483 Rosenstein).
484 - fix a potential segfault when the TLS context
485 would go missing in a read callback (e.g. due to stoptls) -
486 Net::SSLeay of course makes no type checking whatsoever.
487 - AnyEvent::Handle will not stop the read watcher in TLS mode.
488 - AnyEvent::Handle->stoptls will send a TLS close notify
489 instead of simply destroying the stream now.
490 - fix a bug where large blocks of data written in TLS mode
491 would not be sent unless triggered by receives.
492 - on_drain will now take the tls write buffer into account.
493 - SSL operations have been streamlined a bit and should be faster
494 now, more to come, though.
495 - added a FAQ section, will document nontrivial issues in there.
496
4974.234 Mon Sep 29 04:08:13 CEST 2008
498 - fix child watcher documentation: only child exits will be reported,
499 no trace events.
500 - mention SIGPIPE in AnyEvent::Handle.
501 - perl backend will now use POSIX::times as monotonic clock
502 when available and the monotonic clock option is not.
503 - run even without Time::HiRes in some cases now, but do not
504 provide sub-second accuracy (all *supported* perl versions have
505 Time::HiRes).
506
5074.233 Fri Aug 22 01:48:05 CEST 2008
508 - fix a bug in the testsuite, causing a hang.
509 - clarified lots of AnyEvent::Handle constructor arguments.
510
5114.232 Thu Aug 21 20:44:25 CEST 2008
512 - fix a bug in AnyEvent::Handle that could cause two fatal error
513 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
514 (testcase by Leon Brocard).
515 - clarified on_eof/on_error documentation a bit.
516
5174.231 Tue Jul 29 13:12:15 CEST 2008
518 - remove some debugging code left in AnyEvent::Util::fork_call
519 (and no, it's impossible to implement with the broken windows
520 perls without resource leaks or worse).
521
5224.23 Tue Jul 29 12:19:59 CEST 2008
523 - document the first parameter passed to condvar callbacks to be
524 the callback.
525 - add AnyEvent::Socket::{ntoa,aton} aliases.
526 - optimize the AE::Handle->push_read (line) for the default
527 eol marker.
528 - optimize push_read (packstring|storable) for small packets.
529 - invoke on_error callback when no on_eof callback is set.
530 - fix a bug in push_read (storable) of unknown impact.
531
5324.22 Sun Jul 20 16:34:13 CEST 2008
533 - new function AnyEvent::Socket::parse_hostport.
534 - as the bulkheads at microsoft can't even get getprotobyname reliably
535 working on their shitty fucking broken os we need to hardcode
536 some common protocol numbers in AnyEvent::Socket. How can
537 people even bother with such a pile of shit as windows.
538
5394.21 Thu Jul 17 14:40:05 CEST 2008
540 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
541 which is incompatible to previous invocation syntax.
542 - work around more windows perl bugs in fork_call. I cannot imagine
543 why anybody would bother implementing fork in such an obviously
544 unusable way.
545 - avoid calling $do_retry if the request has already finished in
546 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
547 K. Sharifulin).
548
5494.2 Sat Jul 12 22:42:11 CEST 2008
550 - support an "interval" argument for repeating timers.
551 - fix a bug in the pure perl event loop that caused timers to
552 be delayed under some circumstances.
553 - implement optional argument checking via PERL_ANYEVENT_STRICT.
554 - not importing anything from Event causes it to malfunction
555 and not use Time::HiRes, work around this bug.
556 - more bug workarounds against the endlessly broken Tk module,
557 seems to work now on 32 bit machines, but less so on 64 bit.
558 - minor optimisations applied to most backends.
559
5604.161 Fri Jul 4 14:24:48 CEST 2008
561 - fixed the t/handle/02_write.t, it read after write, with the
562 assumption that all data could be written before read
563 (reported by Jim Razmus).
564
5654.160 Thu Jul 3 04:02:21 CEST 2008
566 - re-registering signal watchers after unregistering one did
567 not work if a backend falls back on AnyEvent's default
568 implementation (also affected child watchers).
569 - new AnyEvent::Handle options: autocork and no_delay.
570
5714.152 Sun Jun 22 14:15:44 CEST 2008
572 - allow for 32-bit perls that implement shifts differently
573 on different architectures in parse_ipv4 (reported and
574 analysed by Keiichi DAIBA).
575
5764.151 Fri Jun 6 17:34:24 CEST 2008
577 - make sure specifying _only_ on_read and never pushing reads
578 works.
579
5804.15 Fri Jun 6 13:00:46 CEST 2008
581 - the pure perl backend would keep some watchers alive when more than
582 one watcher was registered for the same fd.
583 - new "packstring" and "storable" read and write types
584 for AnyEvent::Handle.
585 - allow on_eof handler to be called after on_error with EPIPE returns.
586 - do not immediately call on_read callback in handle constructor.
587
5884.14 Thu Jun 5 20:29:31 CEST 2008
589 - Fixed a bug in DNS SRV priority sorting.
590 - AnyEvent::Util::guard now reports runtime errors while
591 executing the guard block as warnings.
592 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
593
5944.13 Thu Jun 5 00:47:59 CEST 2008
595 - AnyEvent::DNS only followed cname chains with length 2,
596 contrary to documentation. bumped it up to 10, thanks to
597 microsoft, the current broken-dns-config-king.
598 - AnyEvent::DNS didn't check for socket return status
599 but instead relied on perl not creating filehandles in that
600 case - too bad it gives you a bogus file handle (reported
601 and analysed by Vladimir Timofeev).
602 - fix queue management logic in AnyEvent::Handle:
603 when on_read was registered, the queue was empty
604 and no progress could be made AnyEvent::Handle would
605 enter an endless loop.
606 - correctly start reading again when the handle became
607 busy again after idling.
608 - correctly treat tls shut-downs as EOF condition
609 (for the time being, would be nice to have a callback for that).
610 - correctly call eof callback under all conditions
611 (hopefully).
612 - the timeout callback did not expect that $self can go
613 away any time.
614
6154.12 Tue Jun 3 10:58:04 CEST 2008
616 - include AnyEvent::Intro, a tutorial for anyevent,
617 anyevent::socket and anyevent::handle.
618 - allow more options in on_error.
619
6204.11 Fri May 30 23:42:25 CEST 2008
621 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
622 reverse_lookup and reverse_verify to replace it, support
623 v4mapped and v4compat addresses.
624 - provide more documentation for the resolver class.
625 - really replace longest run of :0: by :: in format_address,
626 also properly convert :: and ::1 again.
627 - support NAPTR record name and decode it.
628 - implement random weight sampling for SRV records, as per
629 rfc 2782.
630 - correctly abort on srv-record targets of ".".
631 - added AnyEvent::DNS::wait_for_slot.
632 - in the unlikely event of a virtual circuit connection
633 being invalidated by a delayed udp reply, AnyEvent::DNS
634 could die. this has been fixed.
635 - plug a probable memleak in the DNS vc code.
636 - use configured timeout also for VC request phase.
637 - implement timeout and max_outstanding methods
638 for resolver class.
639 - update version numbers in all modules.
640
6414.1 Thu May 29 05:45:40 CEST 2008
642 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
643 which mirrors their purpose better (old names still available
644 for a while). Also moved AnyEvent::DNS::addr to
645 AnyEvent::Socket::resolve_sockaddr.
646 - implement AnyEvent->time and AnyEvent->now.
647 - fix IPv6 support in pack_sockaddr.
648 - officially un-experimentalise all newly introduced code.
649 - support unix domain sockets everywhere by specifying
650 a host of "unix/" and the pathname as service.
651 - implement an activity timeout in AnyEvent::Handle.
652 - added a regex read type to AnyEvent::Handle.
653 - added a json read/write type to AnyEvent::Handle.
654 - always croak in on_error in AnyEvent::Handle.
655 - document how to subclass AnyEvent::Handle.
656 - implement AnyEvent::Util::fork_call.
657 - add support for IPv6 nameservers and nameserver statements.
658 - work around _yet_ _another_ windows perl bug in where empty
659 select masks cause errors in the pure perl backend (this
660 workaround cannot be implemented for the other event loops,
661 but EV also works around this).
662 - supply AnyEvent::Socket::address_type and make good use of it.
663 - clarify and fix the Handle documentation w.r.t $self vs. $handle
664 vs. callback arguments.
665 - add some recommends to the META.yml.
666
6674.05 Mon May 26 19:44:06 CEST 2008
668 - some platforms ignore the rfc and prepend an extra sa_len member to
669 the sockaddr structure, cater for those. also use sockaddr_family,
670 didn't know it was there, but comes in handy for the workaround.
671 - undo the import WIN32 hack.
672
6734.04 Mon May 26 08:03:31 CEST 2008
674 - try to work around yet another windows bug: failed connects are reported
675 as if it were out-of-band data. windows users: you suck. Only
676 EV, Glib, Event and the pure perl backend can handle this condition.
677 - optimize write algorithm in AnyEvent::Handle.
678 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
679 - tcp_connect now properly iterates through all targets.
680 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
681 support entirely. also hardcode some AF_INET6 constants for the
682 sake of perl 5.8.
683 - vastly improved nameserver/suffix detection algorithm for windows
684 (still a hack...)
685 - try to use a fallback dns server if no dns servers could be found.
686 - splatter around some local $SIG{__DIE__}'s to work around bad
687 die handlers.
688
6894.03 Sun May 25 05:05:57 CEST 2008
690 - work around perl bugs on windows, where perl returns undocumented
691 error codes for sysread, syswrite etc.
692 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
693 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
694 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
695 - updated documentation.
696 - lots of workarounds for perl 5.8 quirks.
697 - implement a more flexible type system for both reads and writes
698 in AnyEvent::Handle.
699 - support IPv6 in tcp_server.
700 - pass local hostname and port to the prepare callback in tcp_server.
701 - make types extendable in AnyEvent::Handle.
702 - croak, not die, in AnyEvent::Handle;
703
7044.0 Sat May 24 19:58:08 CEST 2008
705 - added (experimental) AnyEvent::DNS module.
706 - added (experimental) AnyEvent::Socket module.
707 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
708 - make condvar->end call ->send as documented.
709 - make condvar constructor support the documented cb => argument.
710 - overload condvars so you can use them instead of code references.
711 - support more of the POSIX forms instead of a pure dotted quad for
712 verbatim ip addresses.
713 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
714 - try very hard not to freeze on broken windows perls.
715 - make on_eof optional in AnyEvent::Handle.
4 - support service names in tcp_connect and tcp_server. 716 - support service names in tcp_connect and tcp_server.
5 - work around netbsd/openbsd bug causing unix sockets not to 717 - work around netbsd/openbsd bug causing unix sockets not to
6 become writable untill empty. 718 become writable until empty.
7 - only make a dummy read for error detection when getpeername 719 - only make a dummy read for error detection when getpeername
8 returns ENOTCONN. 720 returns ENOTCONN.
9 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 721 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
10 override. 722 override.
723 - new AnyEvent::Guard::cancel.
11 724
123.5 Sat May 17 23:17:14 CEST 2008 7253.5 Sat May 17 23:17:14 CEST 2008
13 - gracefully deal with systems not defining CLOCK_MONOTONIC 726 - gracefully deal with systems not defining CLOCK_MONOTONIC
14 (as opposed to not supporting the clock itself, which worked in 3.41). 727 (as opposed to not supporting the clock itself, which worked in 3.41).
15 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server 728 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
… …
27 740
283.41 Sun May 11 19:53:13 CEST 2008 7413.41 Sun May 11 19:53:13 CEST 2008
29 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 742 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
30 743
313.4 Sun May 11 00:29:25 CEST 2008 7443.4 Sun May 11 00:29:25 CEST 2008
32 - complete re-implementation of condvars with a hopefully much mroe useful 745 - complete re-implementation of condvars with a hopefully much more useful
33 API (backward compatible functions exist but are undocumented). 746 API (backward compatible functions exist but are undocumented).
34 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 747 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
35 (still subject to change). 748 (still subject to change).
36 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 749 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
37 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 750 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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