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

Comparing AnyEvent/Changes (file contents):
Revision 1.180 by root, Sun Jul 27 07:25:39 2008 UTC vs.
Revision 1.533 by root, Wed Mar 28 23:11:16 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
6TODO: PERL_ANYEVENT_IO_MODEL docs
7tODO: maxim souffle handle storable
8TODO: document import tags aeio, also ae_ naming
9TODO: t/io
10lib/AnyEvent/DNS.pm: open my $fh, "<", $resolv_conf
11lib/AnyEvent/DNS.pm: if (open my $fh, "ipconfig /all |") {
12lib/AnyEvent/Log.pm: open my $fh, ">>", $path
13lib/AnyEvent/Log.pm: open my $fh, ">>", $path
14lib/AnyEvent/Socket.pm: open my $fh, "<",
15lib/AnyEvent/TLS.pm: or croak "$dh_file: failed to open DH parameter file: $!";
16
17 - documented Rocco fabricating statements about AnyEvent and me.
18 - new module AnyEvent::IO, that is a frontend to either a pure-perl
19 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
20 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
21 - the default logging level was not properly documented in a variety of
22 places, this has hopefully been rectified.
23 - updated uts46data.pl for unicode 6.1.0.
24 - made log messages genearted by AnyEvent submodules not
25 include the package name anymore, as it will be logged
26 by default already.
27 - upgrade to the trick used by common-sense 3.5 to work
28 around extra warning torture/breakage under perl 5.15.x.
29 - log messages by anyevent are now ucfirst, are usually full
30 sentences and do no longer include the package name.
31
326.14 Tue Jan 31 20:00:24 CET 2012
33 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
34 probably an editing mistake (reported by Darin McBride).
35
366.13 Thu Jan 12 07:27:01 CET 2012
37 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
38 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
39 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
40 regex-matching inside the callback would change the parameter, despite
41 $1 being dynamically scoped per-block. this perl bug is now being
42 worked around (testcase by Cindy Wang).
43
446.12 Mon Dec 12 13:21:10 CET 2011
45 - $! was clobbered by subsequent calls in tcp_connect, due to
46 the postpone not saving/restoring it (analyzed by Richard Garnish).
47
486.11 Tue Nov 22 10:36:05 CET 2011
49 - Tk cannot create windows when tainted unless you set a title, so
50 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
51 - escape any nonprintable/nonascii characters when stringifying
52 backtraces.
53 - log the reason loading an interface module fails to load at level debug.
54
556.1 Tue Oct 4 19:44:30 CEST 2011
56 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
57 and above), and some messages inside AnyEvent have been elevated
58 to higher log levels to print by default.
59 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
60 - work around yet another signal race bug, newly introduced by POE. My
61 how that module sucks.
62 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
63 support ipv6-only operation even with defaults (also use ::1 instead
64 of 127.0.0.1).
65 - give better guidelines on how to chose a logging level.
66 - AnyEvent::Log can now cap (limit) log message priority by context.
67 - AE::log will now save and restore $! and $@, so callers don't have to.
68 - AE::Strict now checks that registered fds are still valid at regular
69 intervals.
70 - skip frequently called 'require' calls if the module appears loaded.
71 - increase signal latency test timer considerably for poor little osx,
72 which can't be bothered with basic posix functionality such as signal
73 delivery.
74 - mention some debugging aids in the tutorial.
75 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
76 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
77 of the context as title.
78 - AnyEvent::Strict would not restore read-write access to $_ when a callback
79 died - this is a user bug, but since AE::Strict exists to help dbeugging...
80 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
81 to work with it - the author is well aware of the issues but refuses
82 to fix them. This extreme step has been taken because the module actively
83 attacks the goals of AnyEvent and makes it harder for module authors to use
84 AnyEvent.
85
866.02 Fri Aug 26 20:08:31 CEST 2011
87 - log_to_syslog now accepts facility strings.
88 - log_to_syslog would not allow facilities - "$facility|$level"
89 *literally* means that string. sheesh.
90 - the CPAN parser got confused and thought we provide Sys::Syslog, try
91 to work around.
92 - renamed AnyEvent::Impl::FLTK2 to FLTK.
93
946.01 Fri Aug 26 07:04:11 CEST 2011
95 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
96 the TTL as fourth element - this affects ->resolve, ->request
97 and ->dns_unpack indirectly. Convenience functions such as
98 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
99 AnyEvent::Socket.
100 - INCOMPATIBLE CHANGE: backend authors now should not implement
101 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
102 methods.
103 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
104 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
105 - DEPRECIATION: please change
106 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
107 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
108 in your programs, for future compatibility.
109 - *blush* idle watcher emulation was borked.
110 - AnyEvent::Strict now checks the AE::xxx functions as well.
111 - querying too long domain names would barf deep inside AnyEvent::DNS - since
112 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
113 for too long domain names explicitly and passes an undef to the callback
114 instead (reported by James Bromberger).
115 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
116 - new functions: AnyEvent::postpone, AnyEvent::log.
117 - new module: AnyEvent::Log, for simple logging needs.
118 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
119 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
120 a debug shell.
121 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
122 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
123 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
124 /etc/hosts can be found.
125 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
126 configuration.
127 - env variables can now be specified using AE_xyz instead of
128 PERL_ANYEVENT_xyz, subject to some rules.
129 - lots of new AnyEvent::Debug functionality.
130 - AnyEvent::Strict now checks for modifications $_ by the callback
131 (by setting it to readonly while executing the callback, if possible).
132 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
133 Invest extra overhead to work around this peculiar behaviour.
134 - split the pure perl backend into a loop module and an impl
135 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
136 - some read types (regex, netstring, json, storable) could
137 errornously keep an AE::Handle object alive on their own
138 (reported by Mohammad Toossi).
139 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
140 supports /etc/hosts overrides.
141 - add more workarounds around child watcher bugs in IO::Async
142 and (newly introduced) in POE.
143 - use glib child watchers and try to work around its limitations,
144 if possible (based on a patch by Kevin Ryde).
145 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
146 - many read types in AnyEvent::Handle would malfunction during
147 transitions to TLS because of a "delete $self->{rbuf}".
148 (reported by Gerald Galster).
149 - improve condvar blocking wait performance (probably only noticable
150 with faster backends), also make it easier for event loops that
151 do not support blocking waits.
152 - PERL_ANYEVENT_MODEL now also allows full module names, although
153 the usefulness of this feature is not yet clear.
154 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
155 - add t/80_ssltext.t.
156 - no longer use AUTOLOAD for the initial loading (for no very good reason).
157 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
158 - do not run event loop detection multiple times when programs call their cached
159 copies of AnyEvent::detect.
160 - parse_hostport now accepts absolute paths als unix domain sockets.
161 - suppress the idiotic warning inside POE by patching the POE::Kernel
162 run flags instead of asking the user to jump through contortions.
163 - add Task::Weaken as dependency (as recommends only at this time).
164 - replace PF_UNSPEC by 0, for good old minix.
165
1666.0 (not officially released, but was on CPAN for about 15 minutes :/)
167
1685.34 Thu May 12 09:59:41 CEST 2011
169 - AE::TLS didn't support both cert_file and key at the same time,
170 due to some copy&paste bug (reported and analyzed by vti).
171 - improve AE::Handle timeout documentation slightly.
172
1735.33 Mon Feb 7 21:51:41 CET 2011
174 - never appeared on CPAN, for unknown reasons.
175 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
176 object when it's already gone (analyzed by Ryan Bullock).
177 - update idna algorithm(s) to UTS#46 version 6.0.0.
178 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
179 uts46 says it's an error, and testsuite makes more sense.
180 - remove leading dots in idn_to_ascii - testsuite says to remove,
181 uts46 says it's an error, and testsuite makes more sense.
182 - fixed a bug in idn ignored character processing.
183 - passes the uts46 6.0.0 testsuite except for three cases which
184 apparently are bugs in the testsuite itself.
185 - passes the uts46 6.0.1 testsuite except for eleven cases which
186 apparently are bugs in the testsuite itself.
187
1885.31 Sun Jan 23 11:49:19 CET 2011
189 - work around a POE bug where POE would first poll for new events and
190 invoke handlers BEFORE handling the message queue.
191 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
192 - stop_read now does just that in TLS mode. see its
193 documentation for caveats.
194 - update to common::sense 3.4 warning set (some warnings in perl impose
195 an arbitrary 32kb limit on regex matches, which affects
196 AnyEvent::Handle).
197 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
198 - new experimental AE::Handle::resettls method.
199
2005.3 Fri Dec 31 05:48:13 CET 2010
201 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
202 issues with IO::Async have been solved in that version, including
203 the availabilty of a default loop!
204 - implement new wbuf_max parameter for AnyEvent::Handle.
205 - added Cocoa::EventLoop backend by Daisuke Murase.
206 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
207 - clarify handle on_prepare documentation.
208 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
209 to avoid leaking them to any child processes.
210 - try to force files into the arch-dependend lib directory, for the
211 half-"turly obsessive compulsive person"(s) that depend on it.
212
2135.29 Sun Dec 5 10:49:21 CET 2010
214 - convert EV backend to EV 4.00 API (so better upgrade EV too).
215 - AE::Handle: implement dynamic read_size adjustment, add
216 max_read_size parameter, reduce default read_size to 2048.
217 - add some FAQ entries for servers and/or on_eof vs. on_error.
218 - work around OS bugs (cygwin again) when lingering in AE::Handle.
219 - reduce memory usage slightly when connect was used in AE::Handle.
220 - wrok around more segfaults in Tk.
221 - document tls_ctx => undef as valid.
222 - detect WNOHANG value at compiletime.
223 - include some probably linux-only support for building AnyEvent
224 as part of the perl core.
225 - improve compatibility of t/08_idna.t to perl 5.8.
226 - make AnyEvent::Strict truly optional (in case somebody builds a
227 minimal perl...).
228
2295.28 Wed Oct 13 04:14:23 CEST 2010
230 - due to a glitch, AnyEvent's internal getprotobyname wasn't
231 actually used in tcp_server.
232 - implement AnyEvent::CondVar->new method, which was documented but
233 missing.
234 - support autoloading of read types in unshift_read.
235 - AnyEvent::Strict: do the test for invalid fh arguments differently,
236 resulting in more informative error messages.
237 - new AnyEvent::FAQ manpage.
238 - clarify that you can actually call push_read etc. in on_read.
239 - put keys and certs in /tmp, not into the current directory,
240 and document this in the SECURITY CONSIDERATIONS sections
241 of AnyEvent::TLS.
242 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
243 requested by Ron Isaacson.
244 - document $AnyEvent::DNS::RESOLVER variable.
245 - incorporated major typo patches by Abhijit Menon-Sen.
246 - AnyEvent::Handle now croaks when negative timeouts are passed.
247 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
248 undocumented, but who knows when they might come in handy.
249
2505.271 Tue Jun 8 12:05:46 CEST 2010
251 - backport to perl 5.8.x.
252
2535.27 Sun Jun 6 12:12:05 CEST 2010
254 - postpone differently in AnyEvent::Socket now, as
255 when not, canceling the connection attempt might fail
256 (found by Felix Antonius Wilhelm Ostmann).
257 - explicitly check for non-stream sockets in AE::Handle, too many
258 clueless people fell into the trap of this somehow working.
259 - simplified and reworked the "OTHER MODULES" section.
260 - better/more condvar examples.
261
2625.261 Wed Apr 28 16:13:36 CEST 2010
263 - AF_INET6 was not properly used from Socket6 during configuration
264 time (found by Andy Grundman).
265
2665.26 Mon Apr 12 04:49:35 CEST 2010
267 - don't generate spurious readyness notifications when select
268 returns EINTR in the pure perl backend (analysed and testcase
269 by Bas Denissen).
270 - give same examples in SYNOPSIS sections of AnyEvent and AE.
271 - provider faster implementations for AE::time/now/now_update for
272 EV and Perl backends.
273 - provide faster AE::cv and AE::time implementations for all
274 backends.
275 - no longer support lower/mixed-case signal names (this was
276 never documented nor universally supported).
277 - some more memory µ-optimisations, and somewhat less messy
278 code to cope with both AnyEvent and AE APIs.
279
2805.251 Sat Mar 13 00:58:21 CET 2010
281 - make test failed if it was run before make install on perl <=
282 5.8.8, and older versions of perl make our live very difficult,
283 so write constants.pl during Makefile.PL time.
284
2855.25 Sat Mar 13 00:23:14 CET 2010
286 - fix a race condition in AnyEvent::Handle that would cause
287 a "bio_read: unsupported method" error (in _another_
288 ssl connection) after user code threw an exception.
289 - added AnyEvent::Handle->destroyed method.
290 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
291 - the AnyEvent::Util::fh_nonblocking function worked only
292 by lucky accident on win32.
293 - smaller and faster AnyEvent::Util::fh_nonblocking.
294 - when the (required!) Time::HiRes module is missing, AnyEvent
295 did not fall back to built-in time properly.
296 - do not load Fcntl at runtime, saving memory and loading time.
297 - precompile a number of constants and use them instead of runtime
298 detection and eval.
299 - free detection code after detection and similar memory optimisations.
300 - Perl backend timer interval best effort drift has been improved
301 (same algorithm as EV).
302 - update unicode idna mapping table.
303
3045.24 Tue Jan 5 11:39:43 CET 2010
305 - cygwin never reports errors from failed connects. 1.5 just gives
306 you continous readyness and EAGAIN, 1.7 is even more broken and
307 just hangs. work around both issues in a cygwin-specific hack.
308 - improve idle watcher documentation slightly.
309
3105.23 Sun Dec 20 23:48:00 CET 2009
311 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
312 - implement punycode_encode/decode, idn_nameprep,
313 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
314 - provide $AE::VERSION.
315 - removed traces of "no strict 'refs'".
316
3175.22 Sat Dec 5 03:51:13 CET 2009
318 - downgrade-or-fail in AnyEvent::Handle::push_write, to
319 diagnose encoding failures earlier and more succinctly.
320 (this works around bugs in perl, throwing away encoding info
321 when passing scalar data to extensions).
322 - add more examples to AnyEvent::Socket manpage.
323 - upgrade internal warning set to the same as common::sense 2.03.
324 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
325 (John Beppu).
326 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
327 - add google public dns servers to fallback server set and make sure
328 we load-balance properly between the three sets. also add all
329 fallback dns servers, not just a random one, to each dns config.
330
3315.21 Thu Nov 19 02:48:47 CET 2009
332 - fix a problem where socket constants were called with parameters
333 (spotted by David Friedland).
334 - fork_call never use'd POSIX (reported by Daisuke Maki).
335 - improve perl 5.6 compatibility further (but it still won't work
336 unless you rip out everything but the core).
337 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
338 uses 10MB of RAM, but doesn't open a console window. *sigh*.
339
3405.202 Wed Oct 14 22:35:44 CEST 2009
341 - AnyEvent::DNS would unexpectedly clobber $_ under windows
342 (analysed by Matthias Waldorf).
343 - AnyEvent::Handle::run_cmd can now pass the PID of the
344 newly-created process, which is much less useful than it might
345 sound (based on patch by Yann Kerherve).
346
3475.201 Tue Sep 29 12:09:25 CEST 2009
348 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
349 (reported by Torsten Foertsch).
350 - common::sense 2.0 could cause tcp_server to throw an exception
351 (analysed by elmex).
352
3535.2 Mon Sep 14 07:04:49 CEST 2009
354 - INCOMPATIBLE CHANGE: do no longer support register_read_type
355 and register_write_type in AnyEvent::Handle, instead support
356 package names (the facility was mostly abused).
357 - implement "packagename-as-read/write type" support in
358 AnyEvent::Handle.
359 - AnyEvent::Handle: new options "keepalive" and "oobinline".
360 - oobinline set by default to avoid security issues.
361 - the pure-perl event loop backend wrongly detected times() fallback
362 support (spotted by Pavel Boldin).
363
3645.12 Tue Sep 1 20:26:50 CEST 2009
365 - be more lenient when parsing resolv.conf files, as some people
366 use hashmarks as comment indicator inside directives (reported
367 by Michael S. Fischer).
368 - use same set of warnings as common::sense 2.0.
369 - fix a potential 32 bit overflow issue due to perl having problems
370 with large hex constants in the Perl backend.
371 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
372 by default.
373 - fix some prototypes.
374 - (optionally) require Async::Interrupt 1.02.
375 - replace opendns fallbacks by verizon/level3 public dns servers.
376
3775.112 Fri Aug 21 13:59:04 CEST 2009
378 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
379 using starttls with an illegal/uncreatable tls context.
380 - port to broken win32 perls.
381
3825.111 Sun Aug 16 18:44:36 CEST 2009
383 - perl errornously requires sockaddr_un structures returned by the
384 kernel to have a specific length. work around this bug by padding
385 sockaddr_un structures (found to fail on OS X, as perl only uses
386 the correct code for GNU/Linux, but likely to fail on other
387 OSes as well).
388
3895.11 Wed Aug 12 17:49:37 CEST 2009
390 - fixed a bug in run_cmd with input from filename.
391 - create a json coder object when none is given in write direction
392 to reduce dependency on JSON.
393
3945.1 Tue Aug 11 03:17:32 CEST 2009
395 - tcp_connect now reports the correct error in $!, not an unrelated
396 one (was broken since 4.91).
397 - AE::Handle did not properly clear rtimeout/wtimeout watchers
398 when retrying a tcp_connect.
399 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
400 - fixed wrong documentation in AE::Socket.
401
4025.01 Mon Aug 10 03:16:32 CEST 2009
403 - last release broke EV child watchers :/.
404
4055.0 Sun Aug 9 17:32:09 CEST 2009
406 - officially introduce and document the AE API.
407 - lowering the timeout at runtime did not have immediate effect
408 in AnyEvent::Handle.
409 - AE::Handle now has separate and independent read and write timeout
410 settings.
411 - probe for Guard module when AnyEvent::Util is loaded, not at
412 runtime (Event doesn't like eval "use" from callbacks, and it avoids
413 exporting the probe function).
414 - Impl::IOAsync's io watchers did not use the filehandle emulation
415 code to work around it's watcher limits.
416 - work around IO::Async's problems with overloaded objects.
417 - try to disable wearnings inside IO::Async, it's just too buggy.
418 - do no longer use Event by default, only EV or Perl, Event is too buggy.
419 - hackishly provide fast AE:: interface when EV or Perl is the backend.
420 - inherit VERSION from AnyEvent::VERSION in most modules.
421
4224.91 Thu Aug 6 15:42:45 CEST 2009
423 - AE::Handle::starttls could get out of sync when the read buffer
424 already contains some TLS handshake.
425 - AE::Handle did not properly free the TLS session, causing bigger
426 memory leaks in Net::SSLeay than expected (reported by toaster).
427 - AE::Socket::tcp_connect will now artificially delay invoking
428 the callback to avoid returning after invoking the callback.
429 - convert many internal watcher uses to AE API.
430
4314.9 Sat Aug 1 11:07:01 CEST 2009
432 - Glib silently fails when registering a timeout with
433 a negative value, so avoid that.
434 - call condvar callback immediately when it is set
435 after the condition is already signalled.
436 - check rbuf_max condition only after trying to consume
437 data.
438 - normalise signal numbers to names when using the ae's
439 signal handling, but do not document this (yet).
440 - pure perl signal emulation did not properly set nonblocking
441 mode on the signal pipe on win32 (this is unlikely to have caused
442 any issues).
443 - new module: AnyEvent::Debug.
444 - AnyEvent::Strict now checks that a signal was specified
445 by name (not name or number).
446 - reduce memory footprint in typical cases by ~50kb by
447 only compiling signal code when necessary.
448 - add AnyEvent::Handle->rbuf_max.
449 - grab the AE namespace for future new API, implement stubs for
450 said future/alternative API.
451 - new function: AnyEvent::Socket::format_hostport.
452
4534.881 Tue Jul 28 12:51:53 CEST 2009
454 - work around a bug in local in pre-5.10 perls, causing
455 AnyEvent::Handle to recurse when it should not
456 (analysed by elmex).
457
4584.88 Tue Jul 28 04:04:37 CEST 2009
459 - re-bless the handle into a dummy package after calling
460 AnyEvent::Handle::destroy, so the user does not need to check
461 for errors after every push_write etc.
462 - do not attempt to run t/02_signals.t on obviously broken
463 platforms.
464
4654.87 Sun Jul 26 02:06:16 CEST 2009
466 - do not attempt to linger when there is no longer a valid fh.
467 - enforce tls mode to be either accept or connect, do not simply
468 segfault in Net::SSLeay.
469 - AnyEvent::Handle can now call tcp_connect itself (new parameters
470 connect, on_prepare, on_connect and on_connect_error). Updated
471 tutorial accordingly.
472 - add AnyEvent::Impl::Irssi backend.
473
4744.86 Mon Jul 20 23:52:29 CEST 2009
475 - since the verbose warning is not enough, explicitly document
476 that versions before 1.33 of Net::SSLeay are not secure.
477 - work around signal handling races in Event and (...) Event::Lib.
478 - try to align signal-race timer to full-second boundaries.
479 - work around Tk not liking negative timeouts.
480 - don't complain of different grades of environmental unfriendlyness
481 in IO::Async.
482
4834.85 Sat Jul 18 06:16:14 CEST 2009
484 - nail the signal race problem in perl once and for all
485 (see $AnyEvent::MAX_SIGNAL_LATENCY).
486 - take advantage of Async::Interrupt if it is available.
487 - load Time::HiRes and Guard modules on demand only.
488 - add optional/recommended modules section to AnyEvent
489 documentation.
490 - reduce memory usage considerably (and reduce startup penalty)
491 by not using "strict", "warnings" and "overload" modules.
492 - work around buggy windows/openbsd perls and provide EBADMSG
493 and EPROTO ourselves when missing.
494 - improve perl 5.6 compatibility of the core event loop.
495 - made Net::SSLeay version 1.33 a soft requirement.
496
4974.83 Fri Jul 17 16:56:26 CEST 2009
498 - implement AnyEvent::Socket::getprotobyname.
499 - AnyEvent::CondVar's will now detect recursive blocking
500 waits and will croak, as too many people fall into
501 this trap.
502 - AnyEvent::Handle will now call ->destroy on itself after
503 executing the on_error callback, instead of doing some
504 half-baked internal shutdown, for fatal errors.
505 - clarify on_eof behaviour w.r.t. the read queue and
506 on_read callbacks.
507 - ignore some possible spurious wake-ups in tcp_connect.
508
5094.82 Sat Jul 11 00:34:55 CEST 2009
510 - POE and Event backends didn't accept some callable objects as
511 callbacks.
512 - use Config module instead of POSIX module to detect signal names
513 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
514 module doesn't even have all POSIX signals :/.
515 - use more workarounds around the many refcnt/corruption bugs in
516 Event::Lib.
517 - work around a race condition in perl's select, causing t/03_child.t
518 to rarely fail.
519
5204.81 Thu Jul 9 10:30:30 CEST 2009
521 - AnyEvent::Handle didn't properly diagnose write errors
522 (it expected -1 from syswrite, how lame... :).
523 - support file descriptors in addition to file handles
524 in AnyEvent->io.
525 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
526 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
527 and PERL_ANYEVENT_CA_PATH.
528 - provide a sensible synopsis section for AnyEvent::TLS.
529 - add a "supported backends" section to the manpage.
530 - added simple io watcher test to testsuite, using a
531 portable_socketpair.
532 - tried to improve the stability of the Event::Lib backend,
533 YMMV.
534
5354.8 Mon Jul 6 23:45:16 CEST 2009
536 - AnyEvent::DNS did not properly follow CNAME records with
537 uppercase targets.
538 - AnyEvent::DNS would errornously return AAAA records
539 with v4 mapped addresses (a faulty record) as ipv4 addresses,
540 causing AnyEvent::Socket to throw an exception.
541 - added new module AnyEvent::TLS for easier SSL/TLS context
542 creation, with many options including hostname verification,
543 secure default configuration, lots of documentation and,
544 predefined diffie-hellman keys for perfect forward security
545 and much more. get it while it's still fresh!
546 - use AnyEvent::TLS in AnyEvent::Handle for context management.
547 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
548 so AnyEvent::Socket users have smaller memory footprint
549 in the common case.
550 - add AnyEvent::Handle->push_shutdown method.
551 - add an additional error message parameter to AnyEvent::Handle's
552 on_error callback (for TLS, $! is still available).
553 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
554 - make AnyEvent::Handle more robust against early conenction
555 failures (during new), and return C<undef> in that case
556 from the constructor.
557 - AnyEvent::Handle will now try to load only JSON::XS first,
558 then fall back to JSON.
559 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
560 and because it was documented that way.
561
5624.451 Fri Jul 3 00:28:58 CEST 2009
563 - do not clear rbuf when shutting down an AnyEvent::Handle
564 object - doing so breaks AnyEvent::HTTP.
565
5664.45 Mon Jun 29 22:59:26 CEST 2009
567 - a write error could cause AnyEvent::Handle to create
568 an I/O watcher with an undefined $fh.
569 - special-case mapped ipv4 addresses in both
570 AnyEvent::Socket::format_address and parse_address,
571 to treat them just like ipv4 addresses.
572 - updated and overhauled the AnyEvent::Intro doc.
573 - implement AnyEvent::Socket::format_ipv4/ipv6.
574 - slightly speed up portable_pipe/socketpair.
575 - expand condvar begin/end documentation.
576
5774.42 Fri Jun 26 08:32:18 CEST 2009
578 - preliminary and neccesarily incomplete support for IO::Async.
579 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
580 case it was set to IGNORE, to ensure we can catch child statuses
581 even when the calling env acts stupidly.
582 - updated benchmarks with IO::Async, which performs very well.
583
5844.412 Wed Jun 24 01:35:57 CEST 2009
585 - support an "untaint" attribute for AnyEvent::DNS
586 and set it on the default resolver.
587 - implement and document AnyEvent::Impl::Perl::loop.
588 - remove all anyevent-env variables from %ENV
589 when running in tainted mode.
590 - mention and extend the IO::Lambda benchmark.
591
5924.411 Sun Jun 7 18:48:02 CEST 2009
593 - do not try to use F_SETFD on windows.
594
5954.41 Thu May 14 06:40:11 CEST 2009
596 - work around issues in older perls (5.8.0?) when
597 a signal handler is deleted from the %SIG hash.
598 - use POSIX::_exit in child test, to avoid running
599 destructors.
600 - speed up CHLD handling by relying on SIGCHLD being
601 synchronously delivered, even when we roll our own
602 implementation.
603 - AnyEvent::DNS: add the "dname" resource record name
604 for cosmetic reasons.
605
6064.4 Sun Apr 26 20:12:33 CEST 2009
607 - implemented idle watchers, where applicable.
608 - AnyEvent->time died when Event backend was in use.
609 - fix a memleak in the Tk backend.
610 - sped up Tk timer handling.
611 - clip negative "after" values to 0 in AnyEvent::Impl::Event
612 to avoid spamming out warning messages.
613 - fix Qt timers without interval.
614 - avoid Qt zero-timeout special case, in old libqt's.
615
6164.352 Mon Apr 20 16:31:11 CEST 2009
617 - fix AnyEvent::Strict error messages for child watchers.
618 - fix/update Makefile.PL "recommends" versions.
619 - add AnyEvent->now_update.
620
6214.351 Sat Apr 11 07:56:14 CEST 2009
622 - actually make the signal pipe work under win32.
623 - localise $! in signal handler to avoid changing $!, although
624 perl itself does not seem to save/restore errno either.
625 - set the cloexec flag on the signal pipe (normally set by perl too).
626
6274.35 Fri Mar 27 11:48:20 CET 2009
628 - event models relying on AnyEvent's signal watcher emulation did
629 invoke the callback asynchronously, contradicting documentation
630 and causing signals to get lost (this includes AnyEvent's own
631 event loop). AnyEvent now uses the standard pipe trick to make
632 callback execution synchronous to the event loop.
633 - AnyEvent::Handle didn't free TLS context data on DESTROY
634 (patch by Pavel Shaydo).
635 - work around the ever-incompatibly-changing API of MakeMaker.
636 - document that changing global variables without restoring them
637 is a bad idea in Perl (noted by Adam Rosenstein).
638 - AnyEvent::Strict now barfs if ->io is passed a file.
639
6404.34 Thu Feb 12 18:32:45 CET 2009
641 - separately buffer TLS read data, as otherwise the read queue could
642 deadlock as receiving data is not expected while draining the
643 read queue (which cna only happen with TLS).
644 - raise EBADMSG error on JSON decoding errors.
645 - fix some minor manpage bugs (reported by Maximilian Gaß).
646 - speed up select bitmask parsing quite a bit in the pure perl
647 backend.
648 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
649
6504.331 Tue Jan 6 21:07:25 CET 2009
651 - socketpair fails on many vista machines because vista has
652 completely broken accept/getpeername and getsockname functions,
653 so we provide our own socketpair emulation that kind of works
654 (AnyEvent::Util::portable_pipe).
655 - new function: AnyEvent::Util::portable_socketpair.
656 - take advantage of the Guard module if it exists.
657
6584.33 Fri Nov 21 02:35:40 CET 2008
659 - AnyEvent::Strict did errornously flag a fileno of 0 as
660 illegal.
661 - reduce memory usage and slightly speed up the pure perl backend
662 by only storing the file descriptor, not the file handle.
663 - add missing autocork method to AnyEvent::Handle (reported
664 by Adam Rosenstein).
665 - AnyEvent::DNS->resolve errornously documented an $rcode result
666 argument, but there isn't (spotted by Henrik Krohns).
667 - the naked truth about Tk - it's basically unmaintained.
668
6694.32 Mon Nov 3 22:46:32 CET 2008
670 - fix AnyEvent::Socket::resolve_sockaddr to properly support
671 unix sockets again and choose a proper default for
672 the socket type.
673 - fix call to resolve_sockaddr in tcp_connect.
674
6754.31 Thu Oct 30 04:41:48 CET 2008
676 - implemented AnyEvent::Handle->destroy method.
677 - hint about unexpected effects in TLS mode.
678 - speed up AnyEvent::Impl::Perl by using more arrays
679 instead of hashes without the slightest loss of readability :)
680 - work around a perl argument refcounting bug.
681 - with some perl patches applied, AnyEvent now seems leak-free.
682
6834.3 Fri Oct 3 09:18:43 CEST 2008
684 - AnyEvent will now install a no-op signal handler for SIGPIPE,
685 unless one has been installed already.
686 - warn about not loaidng AnyEvent::Impl::POE early enough
687 (patch by Adam Rosenstein).
688 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
689 - fix a bug in where in SSL connect mode, the
690 client would first wait for some data by the server and
691 otherwise hang. (reported and analysed in an absolutely
692 exemplary manner by Adam Rosenstein).
693 - fix a bug in where SSL EOF would not be treated
694 as stream EOF, putting the connection into a hung state.
695 (reported and analysed in an absolutely exemplary manner by Adam
696 Rosenstein).
697 - fix a potential segfault when the TLS context
698 would go missing in a read callback (e.g. due to stoptls) -
699 Net::SSLeay of course makes no type checking whatsoever.
700 - AnyEvent::Handle will not stop the read watcher in TLS mode.
701 - AnyEvent::Handle->stoptls will send a TLS close notify
702 instead of simply destroying the stream now.
703 - fix a bug where large blocks of data written in TLS mode
704 would not be sent unless triggered by receives.
705 - on_drain will now take the tls write buffer into account.
706 - SSL operations have been streamlined a bit and should be faster
707 now, more to come, though.
708 - added a FAQ section, will document nontrivial issues in there.
709
7104.234 Mon Sep 29 04:08:13 CEST 2008
711 - fix child watcher documentation: only child exits will be reported,
712 no trace events.
713 - mention SIGPIPE in AnyEvent::Handle.
714 - perl backend will now use POSIX::times as monotonic clock
715 when available and the monotonic clock option is not.
716 - run even without Time::HiRes in some cases now, but do not
717 provide sub-second accuracy (all *supported* perl versions have
718 Time::HiRes).
719
7204.233 Fri Aug 22 01:48:05 CEST 2008
721 - fix a bug in the testsuite, causing a hang.
722 - clarified lots of AnyEvent::Handle constructor arguments.
723
7244.232 Thu Aug 21 20:44:25 CEST 2008
725 - fix a bug in AnyEvent::Handle that could cause two fatal error
726 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
727 (testcase by Leon Brocard).
728 - clarified on_eof/on_error documentation a bit.
729
7304.231 Tue Jul 29 13:12:15 CEST 2008
731 - remove some debugging code left in AnyEvent::Util::fork_call
732 (and no, it's impossible to implement with the broken windows
733 perls without resource leaks or worse).
734
7354.23 Tue Jul 29 12:19:59 CEST 2008
3 - document the first parameter passed to condvar callbacks to be 736 - document the first parameter passed to condvar callbacks to be
4 the callback. 737 the callback.
5 - add AnyEvent::Socket::{ntoa,aton} aliases. 738 - add AnyEvent::Socket::{ntoa,aton} aliases.
6 - optimize the AE::Handle->push_read (line) for the default 739 - optimize the AE::Handle->push_read (line) for the default
7 eol marker. 740 eol marker.
8 - optimize push_read (packstring|storable) for small packets. 741 - optimize push_read (packstring|storable) for small packets.
742 - invoke on_error callback when no on_eof callback is set.
9 - fix a bug in push_read (storable) of unknown impact. 743 - fix a bug in push_read (storable) of unknown impact.
10 744
114.22 Sun Jul 20 16:34:13 CEST 2008 7454.22 Sun Jul 20 16:34:13 CEST 2008
12 - new function AnyEvent::Socket::parse_hostport. 746 - new function AnyEvent::Socket::parse_hostport.
13 - as the bulkheads at microsoft can't even get getprotobyname reliably 747 - as the bulkheads at microsoft can't even get getprotobyname reliably
… …
195 - support service names in tcp_connect and tcp_server. 929 - support service names in tcp_connect and tcp_server.
196 - work around netbsd/openbsd bug causing unix sockets not to 930 - work around netbsd/openbsd bug causing unix sockets not to
197 become writable until empty. 931 become writable until empty.
198 - only make a dummy read for error detection when getpeername 932 - only make a dummy read for error detection when getpeername
199 returns ENOTCONN. 933 returns ENOTCONN.
200 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 934 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
201 override. 935 override.
202 - new AnyEvent::Guard::cancel. 936 - new AnyEvent::Guard::cancel.
203 937
2043.5 Sat May 17 23:17:14 CEST 2008 9383.5 Sat May 17 23:17:14 CEST 2008
205 - gracefully deal with systems not defining CLOCK_MONOTONIC 939 - gracefully deal with systems not defining CLOCK_MONOTONIC
… …
219 953
2203.41 Sun May 11 19:53:13 CEST 2008 9543.41 Sun May 11 19:53:13 CEST 2008
221 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 955 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
222 956
2233.4 Sun May 11 00:29:25 CEST 2008 9573.4 Sun May 11 00:29:25 CEST 2008
224 - complete re-implementation of condvars with a hopefully much mroe useful 958 - complete re-implementation of condvars with a hopefully much more useful
225 API (backward compatible functions exist but are undocumented). 959 API (backward compatible functions exist but are undocumented).
226 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 960 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
227 (still subject to change). 961 (still subject to change).
228 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 962 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
229 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 963 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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