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

Comparing AnyEvent/Changes (file contents):
Revision 1.215 by root, Sat Feb 7 22:18:02 2009 UTC vs.
Revision 1.529 by root, Thu Mar 22 18:00:35 2012 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: document TCP_*** constants
4TODO: add debug/trace logging to some modules
5TODO: AnyEvent::IO
6
7 - documented Rocco fabricating statements about AnyEvent and me.
8 - the default logging level was not properly documented in a variety of
9 places, this has hopefully been rectified.
10 - updated uts46data.pl for unicode 6.1.0.
11 - made log messages genearted by AnyEvent submodules not
12 include the package name anymore, as it will be logged
13 by default already.
14
156.14 Tue Jan 31 20:00:24 CET 2012
16 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
17 probably an editing mistake (reported by Darin McBride).
18
196.13 Thu Jan 12 07:27:01 CET 2012
20 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
21 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
22 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
23 regex-matching inside the callback would change the parameter, despite
24 $1 being dynamically scoped per-block. this perl bug is now being
25 worked around (testcase by Cindy Wang).
26
276.12 Mon Dec 12 13:21:10 CET 2011
28 - $! was clobbered by subsequent calls in tcp_connect, due to
29 the postpone not saving/restoring it (analyzed by Richard Garnish).
30
316.11 Tue Nov 22 10:36:05 CET 2011
32 - Tk cannot create windows when tainted unless you set a title, so
33 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
34 - escape any nonprintable/nonascii characters when stringifying
35 backtraces.
36 - log the reason loading an interface module fails to load at level debug.
37
386.1 Tue Oct 4 19:44:30 CEST 2011
39 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
40 and above), and some messages inside AnyEvent have been elevated
41 to higher log levels to print by default.
42 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
43 - work around yet another signal race bug, newly introduced by POE. My
44 how that module sucks.
45 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
46 support ipv6-only operation even with defaults (also use ::1 instead
47 of 127.0.0.1).
48 - give better guidelines on how to chose a logging level.
49 - AnyEvent::Log can now cap (limit) log message priority by context.
50 - AE::log will now save and restore $! and $@, so callers don't have to.
51 - AE::Strict now checks that registered fds are still valid at regular
52 intervals.
53 - skip frequently called 'require' calls if the module appears loaded.
54 - increase signal latency test timer considerably for poor little osx,
55 which can't be bothered with basic posix functionality such as signal
56 delivery.
57 - mention some debugging aids in the tutorial.
58 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
59 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
60 of the context as title.
61 - AnyEvent::Strict would not restore read-write access to $_ when a callback
62 died - this is a user bug, but since AE::Strict exists to help dbeugging...
63 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
64 to work with it - the author is well aware of the issues but refuses
65 to fix them. This extreme step has been taken because the module actively
66 attacks the goals of AnyEvent and makes it harder for module authors to use
67 AnyEvent.
68
696.02 Fri Aug 26 20:08:31 CEST 2011
70 - log_to_syslog now accepts facility strings.
71 - log_to_syslog would not allow facilities - "$facility|$level"
72 *literally* means that string. sheesh.
73 - the CPAN parser got confused and thought we provide Sys::Syslog, try
74 to work around.
75 - renamed AnyEvent::Impl::FLTK2 to FLTK.
76
776.01 Fri Aug 26 07:04:11 CEST 2011
78 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
79 the TTL as fourth element - this affects ->resolve, ->request
80 and ->dns_unpack indirectly. Convenience functions such as
81 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
82 AnyEvent::Socket.
83 - INCOMPATIBLE CHANGE: backend authors now should not implement
84 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
85 methods.
86 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
87 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
88 - DEPRECIATION: please change
89 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
90 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
91 in your programs, for future compatibility.
92 - *blush* idle watcher emulation was borked.
93 - AnyEvent::Strict now checks the AE::xxx functions as well.
94 - querying too long domain names would barf deep inside AnyEvent::DNS - since
95 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
96 for too long domain names explicitly and passes an undef to the callback
97 instead (reported by James Bromberger).
98 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
99 - new functions: AnyEvent::postpone, AnyEvent::log.
100 - new module: AnyEvent::Log, for simple logging needs.
101 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
102 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
103 a debug shell.
104 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
105 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
106 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
107 /etc/hosts can be found.
108 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
109 configuration.
110 - env variables can now be specified using AE_xyz instead of
111 PERL_ANYEVENT_xyz, subject to some rules.
112 - lots of new AnyEvent::Debug functionality.
113 - AnyEvent::Strict now checks for modifications $_ by the callback
114 (by setting it to readonly while executing the callback, if possible).
115 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
116 Invest extra overhead to work around this peculiar behaviour.
117 - split the pure perl backend into a loop module and an impl
118 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
119 - some read types (regex, netstring, json, storable) could
120 errornously keep an AE::Handle object alive on their own
121 (reported by Mohammad Toossi).
122 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
123 supports /etc/hosts overrides.
124 - add more workarounds around child watcher bugs in IO::Async
125 and (newly introduced) in POE.
126 - use glib child watchers and try to work around its limitations,
127 if possible (based on a patch by Kevin Ryde).
128 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
129 - many read types in AnyEvent::Handle would malfunction during
130 transitions to TLS because of a "delete $self->{rbuf}".
131 (reported by Gerald Galster).
132 - improve condvar blocking wait performance (probably only noticable
133 with faster backends), also make it easier for event loops that
134 do not support blocking waits.
135 - PERL_ANYEVENT_MODEL now also allows full module names, although
136 the usefulness of this feature is not yet clear.
137 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
138 - add t/80_ssltext.t.
139 - no longer use AUTOLOAD for the initial loading (for no very good reason).
140 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
141 - do not run event loop detection multiple times when programs call their cached
142 copies of AnyEvent::detect.
143 - parse_hostport now accepts absolute paths als unix domain sockets.
144 - suppress the idiotic warning inside POE by patching the POE::Kernel
145 run flags instead of asking the user to jump through contortions.
146 - add Task::Weaken as dependency (as recommends only at this time).
147 - replace PF_UNSPEC by 0, for good old minix.
148
1496.0 (not officially released, but was on CPAN for about 15 minutes :/)
150
1515.34 Thu May 12 09:59:41 CEST 2011
152 - AE::TLS didn't support both cert_file and key at the same time,
153 due to some copy&paste bug (reported and analyzed by vti).
154 - improve AE::Handle timeout documentation slightly.
155
1565.33 Mon Feb 7 21:51:41 CET 2011
157 - never appeared on CPAN, for unknown reasons.
158 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
159 object when it's already gone (analyzed by Ryan Bullock).
160 - update idna algorithm(s) to UTS#46 version 6.0.0.
161 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
162 uts46 says it's an error, and testsuite makes more sense.
163 - remove leading dots in idn_to_ascii - testsuite says to remove,
164 uts46 says it's an error, and testsuite makes more sense.
165 - fixed a bug in idn ignored character processing.
166 - passes the uts46 6.0.0 testsuite except for three cases which
167 apparently are bugs in the testsuite itself.
168 - passes the uts46 6.0.1 testsuite except for eleven cases which
169 apparently are bugs in the testsuite itself.
170
1715.31 Sun Jan 23 11:49:19 CET 2011
172 - work around a POE bug where POE would first poll for new events and
173 invoke handlers BEFORE handling the message queue.
174 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
175 - stop_read now does just that in TLS mode. see its
176 documentation for caveats.
177 - update to common::sense 3.4 warning set (some warnings in perl impose
178 an arbitrary 32kb limit on regex matches, which affects
179 AnyEvent::Handle).
180 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
181 - new experimental AE::Handle::resettls method.
182
1835.3 Fri Dec 31 05:48:13 CET 2010
184 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
185 issues with IO::Async have been solved in that version, including
186 the availabilty of a default loop!
187 - implement new wbuf_max parameter for AnyEvent::Handle.
188 - added Cocoa::EventLoop backend by Daisuke Murase.
189 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
190 - clarify handle on_prepare documentation.
191 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
192 to avoid leaking them to any child processes.
193 - try to force files into the arch-dependend lib directory, for the
194 half-"turly obsessive compulsive person"(s) that depend on it.
195
1965.29 Sun Dec 5 10:49:21 CET 2010
197 - convert EV backend to EV 4.00 API (so better upgrade EV too).
198 - AE::Handle: implement dynamic read_size adjustment, add
199 max_read_size parameter, reduce default read_size to 2048.
200 - add some FAQ entries for servers and/or on_eof vs. on_error.
201 - work around OS bugs (cygwin again) when lingering in AE::Handle.
202 - reduce memory usage slightly when connect was used in AE::Handle.
203 - wrok around more segfaults in Tk.
204 - document tls_ctx => undef as valid.
205 - detect WNOHANG value at compiletime.
206 - include some probably linux-only support for building AnyEvent
207 as part of the perl core.
208 - improve compatibility of t/08_idna.t to perl 5.8.
209 - make AnyEvent::Strict truly optional (in case somebody builds a
210 minimal perl...).
211
2125.28 Wed Oct 13 04:14:23 CEST 2010
213 - due to a glitch, AnyEvent's internal getprotobyname wasn't
214 actually used in tcp_server.
215 - implement AnyEvent::CondVar->new method, which was documented but
216 missing.
217 - support autoloading of read types in unshift_read.
218 - AnyEvent::Strict: do the test for invalid fh arguments differently,
219 resulting in more informative error messages.
220 - new AnyEvent::FAQ manpage.
221 - clarify that you can actually call push_read etc. in on_read.
222 - put keys and certs in /tmp, not into the current directory,
223 and document this in the SECURITY CONSIDERATIONS sections
224 of AnyEvent::TLS.
225 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
226 requested by Ron Isaacson.
227 - document $AnyEvent::DNS::RESOLVER variable.
228 - incorporated major typo patches by Abhijit Menon-Sen.
229 - AnyEvent::Handle now croaks when negative timeouts are passed.
230 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
231 undocumented, but who knows when they might come in handy.
232
2335.271 Tue Jun 8 12:05:46 CEST 2010
234 - backport to perl 5.8.x.
235
2365.27 Sun Jun 6 12:12:05 CEST 2010
237 - postpone differently in AnyEvent::Socket now, as
238 when not, canceling the connection attempt might fail
239 (found by Felix Antonius Wilhelm Ostmann).
240 - explicitly check for non-stream sockets in AE::Handle, too many
241 clueless people fell into the trap of this somehow working.
242 - simplified and reworked the "OTHER MODULES" section.
243 - better/more condvar examples.
244
2455.261 Wed Apr 28 16:13:36 CEST 2010
246 - AF_INET6 was not properly used from Socket6 during configuration
247 time (found by Andy Grundman).
248
2495.26 Mon Apr 12 04:49:35 CEST 2010
250 - don't generate spurious readyness notifications when select
251 returns EINTR in the pure perl backend (analysed and testcase
252 by Bas Denissen).
253 - give same examples in SYNOPSIS sections of AnyEvent and AE.
254 - provider faster implementations for AE::time/now/now_update for
255 EV and Perl backends.
256 - provide faster AE::cv and AE::time implementations for all
257 backends.
258 - no longer support lower/mixed-case signal names (this was
259 never documented nor universally supported).
260 - some more memory µ-optimisations, and somewhat less messy
261 code to cope with both AnyEvent and AE APIs.
262
2635.251 Sat Mar 13 00:58:21 CET 2010
264 - make test failed if it was run before make install on perl <=
265 5.8.8, and older versions of perl make our live very difficult,
266 so write constants.pl during Makefile.PL time.
267
2685.25 Sat Mar 13 00:23:14 CET 2010
269 - fix a race condition in AnyEvent::Handle that would cause
270 a "bio_read: unsupported method" error (in _another_
271 ssl connection) after user code threw an exception.
272 - added AnyEvent::Handle->destroyed method.
273 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
274 - the AnyEvent::Util::fh_nonblocking function worked only
275 by lucky accident on win32.
276 - smaller and faster AnyEvent::Util::fh_nonblocking.
277 - when the (required!) Time::HiRes module is missing, AnyEvent
278 did not fall back to built-in time properly.
279 - do not load Fcntl at runtime, saving memory and loading time.
280 - precompile a number of constants and use them instead of runtime
281 detection and eval.
282 - free detection code after detection and similar memory optimisations.
283 - Perl backend timer interval best effort drift has been improved
284 (same algorithm as EV).
285 - update unicode idna mapping table.
286
2875.24 Tue Jan 5 11:39:43 CET 2010
288 - cygwin never reports errors from failed connects. 1.5 just gives
289 you continous readyness and EAGAIN, 1.7 is even more broken and
290 just hangs. work around both issues in a cygwin-specific hack.
291 - improve idle watcher documentation slightly.
292
2935.23 Sun Dec 20 23:48:00 CET 2009
294 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
295 - implement punycode_encode/decode, idn_nameprep,
296 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
297 - provide $AE::VERSION.
298 - removed traces of "no strict 'refs'".
299
3005.22 Sat Dec 5 03:51:13 CET 2009
301 - downgrade-or-fail in AnyEvent::Handle::push_write, to
302 diagnose encoding failures earlier and more succinctly.
303 (this works around bugs in perl, throwing away encoding info
304 when passing scalar data to extensions).
305 - add more examples to AnyEvent::Socket manpage.
306 - upgrade internal warning set to the same as common::sense 2.03.
307 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
308 (John Beppu).
309 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
310 - add google public dns servers to fallback server set and make sure
311 we load-balance properly between the three sets. also add all
312 fallback dns servers, not just a random one, to each dns config.
313
3145.21 Thu Nov 19 02:48:47 CET 2009
315 - fix a problem where socket constants were called with parameters
316 (spotted by David Friedland).
317 - fork_call never use'd POSIX (reported by Daisuke Maki).
318 - improve perl 5.6 compatibility further (but it still won't work
319 unless you rip out everything but the core).
320 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
321 uses 10MB of RAM, but doesn't open a console window. *sigh*.
322
3235.202 Wed Oct 14 22:35:44 CEST 2009
324 - AnyEvent::DNS would unexpectedly clobber $_ under windows
325 (analysed by Matthias Waldorf).
326 - AnyEvent::Handle::run_cmd can now pass the PID of the
327 newly-created process, which is much less useful than it might
328 sound (based on patch by Yann Kerherve).
329
3305.201 Tue Sep 29 12:09:25 CEST 2009
331 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
332 (reported by Torsten Foertsch).
333 - common::sense 2.0 could cause tcp_server to throw an exception
334 (analysed by elmex).
335
3365.2 Mon Sep 14 07:04:49 CEST 2009
337 - INCOMPATIBLE CHANGE: do no longer support register_read_type
338 and register_write_type in AnyEvent::Handle, instead support
339 package names (the facility was mostly abused).
340 - implement "packagename-as-read/write type" support in
341 AnyEvent::Handle.
342 - AnyEvent::Handle: new options "keepalive" and "oobinline".
343 - oobinline set by default to avoid security issues.
344 - the pure-perl event loop backend wrongly detected times() fallback
345 support (spotted by Pavel Boldin).
346
3475.12 Tue Sep 1 20:26:50 CEST 2009
348 - be more lenient when parsing resolv.conf files, as some people
349 use hashmarks as comment indicator inside directives (reported
350 by Michael S. Fischer).
351 - use same set of warnings as common::sense 2.0.
352 - fix a potential 32 bit overflow issue due to perl having problems
353 with large hex constants in the Perl backend.
354 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
355 by default.
356 - fix some prototypes.
357 - (optionally) require Async::Interrupt 1.02.
358 - replace opendns fallbacks by verizon/level3 public dns servers.
359
3605.112 Fri Aug 21 13:59:04 CEST 2009
361 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
362 using starttls with an illegal/uncreatable tls context.
363 - port to broken win32 perls.
364
3655.111 Sun Aug 16 18:44:36 CEST 2009
366 - perl errornously requires sockaddr_un structures returned by the
367 kernel to have a specific length. work around this bug by padding
368 sockaddr_un structures (found to fail on OS X, as perl only uses
369 the correct code for GNU/Linux, but likely to fail on other
370 OSes as well).
371
3725.11 Wed Aug 12 17:49:37 CEST 2009
373 - fixed a bug in run_cmd with input from filename.
374 - create a json coder object when none is given in write direction
375 to reduce dependency on JSON.
376
3775.1 Tue Aug 11 03:17:32 CEST 2009
378 - tcp_connect now reports the correct error in $!, not an unrelated
379 one (was broken since 4.91).
380 - AE::Handle did not properly clear rtimeout/wtimeout watchers
381 when retrying a tcp_connect.
382 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
383 - fixed wrong documentation in AE::Socket.
384
3855.01 Mon Aug 10 03:16:32 CEST 2009
386 - last release broke EV child watchers :/.
387
3885.0 Sun Aug 9 17:32:09 CEST 2009
389 - officially introduce and document the AE API.
390 - lowering the timeout at runtime did not have immediate effect
391 in AnyEvent::Handle.
392 - AE::Handle now has separate and independent read and write timeout
393 settings.
394 - probe for Guard module when AnyEvent::Util is loaded, not at
395 runtime (Event doesn't like eval "use" from callbacks, and it avoids
396 exporting the probe function).
397 - Impl::IOAsync's io watchers did not use the filehandle emulation
398 code to work around it's watcher limits.
399 - work around IO::Async's problems with overloaded objects.
400 - try to disable wearnings inside IO::Async, it's just too buggy.
401 - do no longer use Event by default, only EV or Perl, Event is too buggy.
402 - hackishly provide fast AE:: interface when EV or Perl is the backend.
403 - inherit VERSION from AnyEvent::VERSION in most modules.
404
4054.91 Thu Aug 6 15:42:45 CEST 2009
406 - AE::Handle::starttls could get out of sync when the read buffer
407 already contains some TLS handshake.
408 - AE::Handle did not properly free the TLS session, causing bigger
409 memory leaks in Net::SSLeay than expected (reported by toaster).
410 - AE::Socket::tcp_connect will now artificially delay invoking
411 the callback to avoid returning after invoking the callback.
412 - convert many internal watcher uses to AE API.
413
4144.9 Sat Aug 1 11:07:01 CEST 2009
415 - Glib silently fails when registering a timeout with
416 a negative value, so avoid that.
417 - call condvar callback immediately when it is set
418 after the condition is already signalled.
419 - check rbuf_max condition only after trying to consume
420 data.
421 - normalise signal numbers to names when using the ae's
422 signal handling, but do not document this (yet).
423 - pure perl signal emulation did not properly set nonblocking
424 mode on the signal pipe on win32 (this is unlikely to have caused
425 any issues).
426 - new module: AnyEvent::Debug.
427 - AnyEvent::Strict now checks that a signal was specified
428 by name (not name or number).
429 - reduce memory footprint in typical cases by ~50kb by
430 only compiling signal code when necessary.
431 - add AnyEvent::Handle->rbuf_max.
432 - grab the AE namespace for future new API, implement stubs for
433 said future/alternative API.
434 - new function: AnyEvent::Socket::format_hostport.
435
4364.881 Tue Jul 28 12:51:53 CEST 2009
437 - work around a bug in local in pre-5.10 perls, causing
438 AnyEvent::Handle to recurse when it should not
439 (analysed by elmex).
440
4414.88 Tue Jul 28 04:04:37 CEST 2009
442 - re-bless the handle into a dummy package after calling
443 AnyEvent::Handle::destroy, so the user does not need to check
444 for errors after every push_write etc.
445 - do not attempt to run t/02_signals.t on obviously broken
446 platforms.
447
4484.87 Sun Jul 26 02:06:16 CEST 2009
449 - do not attempt to linger when there is no longer a valid fh.
450 - enforce tls mode to be either accept or connect, do not simply
451 segfault in Net::SSLeay.
452 - AnyEvent::Handle can now call tcp_connect itself (new parameters
453 connect, on_prepare, on_connect and on_connect_error). Updated
454 tutorial accordingly.
455 - add AnyEvent::Impl::Irssi backend.
456
4574.86 Mon Jul 20 23:52:29 CEST 2009
458 - since the verbose warning is not enough, explicitly document
459 that versions before 1.33 of Net::SSLeay are not secure.
460 - work around signal handling races in Event and (...) Event::Lib.
461 - try to align signal-race timer to full-second boundaries.
462 - work around Tk not liking negative timeouts.
463 - don't complain of different grades of environmental unfriendlyness
464 in IO::Async.
465
4664.85 Sat Jul 18 06:16:14 CEST 2009
467 - nail the signal race problem in perl once and for all
468 (see $AnyEvent::MAX_SIGNAL_LATENCY).
469 - take advantage of Async::Interrupt if it is available.
470 - load Time::HiRes and Guard modules on demand only.
471 - add optional/recommended modules section to AnyEvent
472 documentation.
473 - reduce memory usage considerably (and reduce startup penalty)
474 by not using "strict", "warnings" and "overload" modules.
475 - work around buggy windows/openbsd perls and provide EBADMSG
476 and EPROTO ourselves when missing.
477 - improve perl 5.6 compatibility of the core event loop.
478 - made Net::SSLeay version 1.33 a soft requirement.
479
4804.83 Fri Jul 17 16:56:26 CEST 2009
481 - implement AnyEvent::Socket::getprotobyname.
482 - AnyEvent::CondVar's will now detect recursive blocking
483 waits and will croak, as too many people fall into
484 this trap.
485 - AnyEvent::Handle will now call ->destroy on itself after
486 executing the on_error callback, instead of doing some
487 half-baked internal shutdown, for fatal errors.
488 - clarify on_eof behaviour w.r.t. the read queue and
489 on_read callbacks.
490 - ignore some possible spurious wake-ups in tcp_connect.
491
4924.82 Sat Jul 11 00:34:55 CEST 2009
493 - POE and Event backends didn't accept some callable objects as
494 callbacks.
495 - use Config module instead of POSIX module to detect signal names
496 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
497 module doesn't even have all POSIX signals :/.
498 - use more workarounds around the many refcnt/corruption bugs in
499 Event::Lib.
500 - work around a race condition in perl's select, causing t/03_child.t
501 to rarely fail.
502
5034.81 Thu Jul 9 10:30:30 CEST 2009
504 - AnyEvent::Handle didn't properly diagnose write errors
505 (it expected -1 from syswrite, how lame... :).
506 - support file descriptors in addition to file handles
507 in AnyEvent->io.
508 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
509 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
510 and PERL_ANYEVENT_CA_PATH.
511 - provide a sensible synopsis section for AnyEvent::TLS.
512 - add a "supported backends" section to the manpage.
513 - added simple io watcher test to testsuite, using a
514 portable_socketpair.
515 - tried to improve the stability of the Event::Lib backend,
516 YMMV.
517
5184.8 Mon Jul 6 23:45:16 CEST 2009
519 - AnyEvent::DNS did not properly follow CNAME records with
520 uppercase targets.
521 - AnyEvent::DNS would errornously return AAAA records
522 with v4 mapped addresses (a faulty record) as ipv4 addresses,
523 causing AnyEvent::Socket to throw an exception.
524 - added new module AnyEvent::TLS for easier SSL/TLS context
525 creation, with many options including hostname verification,
526 secure default configuration, lots of documentation and,
527 predefined diffie-hellman keys for perfect forward security
528 and much more. get it while it's still fresh!
529 - use AnyEvent::TLS in AnyEvent::Handle for context management.
530 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
531 so AnyEvent::Socket users have smaller memory footprint
532 in the common case.
533 - add AnyEvent::Handle->push_shutdown method.
534 - add an additional error message parameter to AnyEvent::Handle's
535 on_error callback (for TLS, $! is still available).
536 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
537 - make AnyEvent::Handle more robust against early conenction
538 failures (during new), and return C<undef> in that case
539 from the constructor.
540 - AnyEvent::Handle will now try to load only JSON::XS first,
541 then fall back to JSON.
542 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
543 and because it was documented that way.
544
5454.451 Fri Jul 3 00:28:58 CEST 2009
546 - do not clear rbuf when shutting down an AnyEvent::Handle
547 object - doing so breaks AnyEvent::HTTP.
548
5494.45 Mon Jun 29 22:59:26 CEST 2009
550 - a write error could cause AnyEvent::Handle to create
551 an I/O watcher with an undefined $fh.
552 - special-case mapped ipv4 addresses in both
553 AnyEvent::Socket::format_address and parse_address,
554 to treat them just like ipv4 addresses.
555 - updated and overhauled the AnyEvent::Intro doc.
556 - implement AnyEvent::Socket::format_ipv4/ipv6.
557 - slightly speed up portable_pipe/socketpair.
558 - expand condvar begin/end documentation.
559
5604.42 Fri Jun 26 08:32:18 CEST 2009
561 - preliminary and neccesarily incomplete support for IO::Async.
562 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
563 case it was set to IGNORE, to ensure we can catch child statuses
564 even when the calling env acts stupidly.
565 - updated benchmarks with IO::Async, which performs very well.
566
5674.412 Wed Jun 24 01:35:57 CEST 2009
568 - support an "untaint" attribute for AnyEvent::DNS
569 and set it on the default resolver.
570 - implement and document AnyEvent::Impl::Perl::loop.
571 - remove all anyevent-env variables from %ENV
572 when running in tainted mode.
573 - mention and extend the IO::Lambda benchmark.
574
5754.411 Sun Jun 7 18:48:02 CEST 2009
576 - do not try to use F_SETFD on windows.
577
5784.41 Thu May 14 06:40:11 CEST 2009
579 - work around issues in older perls (5.8.0?) when
580 a signal handler is deleted from the %SIG hash.
581 - use POSIX::_exit in child test, to avoid running
582 destructors.
583 - speed up CHLD handling by relying on SIGCHLD being
584 synchronously delivered, even when we roll our own
585 implementation.
586 - AnyEvent::DNS: add the "dname" resource record name
587 for cosmetic reasons.
588
5894.4 Sun Apr 26 20:12:33 CEST 2009
590 - implemented idle watchers, where applicable.
591 - AnyEvent->time died when Event backend was in use.
592 - fix a memleak in the Tk backend.
593 - sped up Tk timer handling.
594 - clip negative "after" values to 0 in AnyEvent::Impl::Event
595 to avoid spamming out warning messages.
596 - fix Qt timers without interval.
597 - avoid Qt zero-timeout special case, in old libqt's.
598
5994.352 Mon Apr 20 16:31:11 CEST 2009
600 - fix AnyEvent::Strict error messages for child watchers.
601 - fix/update Makefile.PL "recommends" versions.
602 - add AnyEvent->now_update.
603
6044.351 Sat Apr 11 07:56:14 CEST 2009
605 - actually make the signal pipe work under win32.
606 - localise $! in signal handler to avoid changing $!, although
607 perl itself does not seem to save/restore errno either.
608 - set the cloexec flag on the signal pipe (normally set by perl too).
609
6104.35 Fri Mar 27 11:48:20 CET 2009
611 - event models relying on AnyEvent's signal watcher emulation did
612 invoke the callback asynchronously, contradicting documentation
613 and causing signals to get lost (this includes AnyEvent's own
614 event loop). AnyEvent now uses the standard pipe trick to make
615 callback execution synchronous to the event loop.
616 - AnyEvent::Handle didn't free TLS context data on DESTROY
617 (patch by Pavel Shaydo).
618 - work around the ever-incompatibly-changing API of MakeMaker.
619 - document that changing global variables without restoring them
620 is a bad idea in Perl (noted by Adam Rosenstein).
621 - AnyEvent::Strict now barfs if ->io is passed a file.
622
6234.34 Thu Feb 12 18:32:45 CET 2009
624 - separately buffer TLS read data, as otherwise the read queue could
625 deadlock as receiving data is not expected while draining the
626 read queue (which cna only happen with TLS).
3 - raise EBADMSG error on JSON decoding errors. 627 - raise EBADMSG error on JSON decoding errors.
4 - fix some minor manpage bugs (reported by Maximilian Gaß). 628 - fix some minor manpage bugs (reported by Maximilian Gaß).
5 - speed up select bitmask parsing quite a bit in the pure perl 629 - speed up select bitmask parsing quite a bit in the pure perl
6 backend. 630 backend.
631 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
7 632
84.331 Tue Jan 6 21:07:25 CET 2009 6334.331 Tue Jan 6 21:07:25 CET 2009
9 - socketpair fails on many vista machines because vista has 634 - socketpair fails on many vista machines because vista has
10 completely broken accept/getpeername and getsockname functions, 635 completely broken accept/getpeername and getsockname functions,
11 so we provide our own socketpair emulation that kind of works 636 so we provide our own socketpair emulation that kind of works
… …
414.3 Fri Oct 3 09:18:43 CEST 2008 6664.3 Fri Oct 3 09:18:43 CEST 2008
42 - AnyEvent will now install a no-op signal handler for SIGPIPE, 667 - AnyEvent will now install a no-op signal handler for SIGPIPE,
43 unless one has been installed already. 668 unless one has been installed already.
44 - warn about not loaidng AnyEvent::Impl::POE early enough 669 - warn about not loaidng AnyEvent::Impl::POE early enough
45 (patch by Adam Rosenstein). 670 (patch by Adam Rosenstein).
46 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 671 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
47 - fix a bug in where in SSL connect mode, the 672 - fix a bug in where in SSL connect mode, the
48 client would first wait for some data by the server and 673 client would first wait for some data by the server and
49 otherwise hang. (reported and analyzed in an absolutely 674 otherwise hang. (reported and analysed in an absolutely
50 exemplary manner by Adam Rosenstein). 675 exemplary manner by Adam Rosenstein).
51 - fix a bug in where SSL EOF would not be treated 676 - fix a bug in where SSL EOF would not be treated
52 as stream EOF, putting the connection into a hung state. 677 as stream EOF, putting the connection into a hung state.
53 (reported and analyzed in an absolutely exemplary manner by Adam 678 (reported and analysed in an absolutely exemplary manner by Adam
54 Rosenstein). 679 Rosenstein).
55 - fix a potential segfault when the TLS context 680 - fix a potential segfault when the TLS context
56 would go missing in a read callback (e.g. due to stoptls) - 681 would go missing in a read callback (e.g. due to stoptls) -
57 Net::SSLeay of course makes no type checking whatsoever. 682 Net::SSLeay of course makes no type checking whatsoever.
58 - AnyEvent::Handle will not stop the read watcher in TLS mode. 683 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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