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

Comparing AnyEvent/Changes (file contents):
Revision 1.149 by elmex, Thu Jun 5 07:13:07 2008 UTC vs.
Revision 1.532 by root, Wed Mar 28 22:28:48 2012 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
34.14 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 - the default logging level was not properly documented in a variety of
19 places, this has hopefully been rectified.
20 - updated uts46data.pl for unicode 6.1.0.
21 - made log messages genearted by AnyEvent submodules not
22 include the package name anymore, as it will be logged
23 by default already.
24 - upgrade to the trick used by common-sense 3.5 to work
25 around extra warning torture/breakage under perl 5.15.x.
26 - log messages by anyevent are now ucfirst, are usually full
27 sentences and do no longer include the package name.
28
296.14 Tue Jan 31 20:00:24 CET 2012
30 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
31 probably an editing mistake (reported by Darin McBride).
32
336.13 Thu Jan 12 07:27:01 CET 2012
34 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
35 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
36 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
37 regex-matching inside the callback would change the parameter, despite
38 $1 being dynamically scoped per-block. this perl bug is now being
39 worked around (testcase by Cindy Wang).
40
416.12 Mon Dec 12 13:21:10 CET 2011
42 - $! was clobbered by subsequent calls in tcp_connect, due to
43 the postpone not saving/restoring it (analyzed by Richard Garnish).
44
456.11 Tue Nov 22 10:36:05 CET 2011
46 - Tk cannot create windows when tainted unless you set a title, so
47 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
48 - escape any nonprintable/nonascii characters when stringifying
49 backtraces.
50 - log the reason loading an interface module fails to load at level debug.
51
526.1 Tue Oct 4 19:44:30 CEST 2011
53 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
54 and above), and some messages inside AnyEvent have been elevated
55 to higher log levels to print by default.
56 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
57 - work around yet another signal race bug, newly introduced by POE. My
58 how that module sucks.
59 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
60 support ipv6-only operation even with defaults (also use ::1 instead
61 of 127.0.0.1).
62 - give better guidelines on how to chose a logging level.
63 - AnyEvent::Log can now cap (limit) log message priority by context.
64 - AE::log will now save and restore $! and $@, so callers don't have to.
65 - AE::Strict now checks that registered fds are still valid at regular
66 intervals.
67 - skip frequently called 'require' calls if the module appears loaded.
68 - increase signal latency test timer considerably for poor little osx,
69 which can't be bothered with basic posix functionality such as signal
70 delivery.
71 - mention some debugging aids in the tutorial.
72 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
73 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
74 of the context as title.
75 - AnyEvent::Strict would not restore read-write access to $_ when a callback
76 died - this is a user bug, but since AE::Strict exists to help dbeugging...
77 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
78 to work with it - the author is well aware of the issues but refuses
79 to fix them. This extreme step has been taken because the module actively
80 attacks the goals of AnyEvent and makes it harder for module authors to use
81 AnyEvent.
82
836.02 Fri Aug 26 20:08:31 CEST 2011
84 - log_to_syslog now accepts facility strings.
85 - log_to_syslog would not allow facilities - "$facility|$level"
86 *literally* means that string. sheesh.
87 - the CPAN parser got confused and thought we provide Sys::Syslog, try
88 to work around.
89 - renamed AnyEvent::Impl::FLTK2 to FLTK.
90
916.01 Fri Aug 26 07:04:11 CEST 2011
92 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
93 the TTL as fourth element - this affects ->resolve, ->request
94 and ->dns_unpack indirectly. Convenience functions such as
95 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
96 AnyEvent::Socket.
97 - INCOMPATIBLE CHANGE: backend authors now should not implement
98 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
99 methods.
100 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
101 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
102 - DEPRECIATION: please change
103 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
104 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
105 in your programs, for future compatibility.
106 - *blush* idle watcher emulation was borked.
107 - AnyEvent::Strict now checks the AE::xxx functions as well.
108 - querying too long domain names would barf deep inside AnyEvent::DNS - since
109 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
110 for too long domain names explicitly and passes an undef to the callback
111 instead (reported by James Bromberger).
112 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
113 - new functions: AnyEvent::postpone, AnyEvent::log.
114 - new module: AnyEvent::Log, for simple logging needs.
115 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
116 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
117 a debug shell.
118 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
119 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
120 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
121 /etc/hosts can be found.
122 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
123 configuration.
124 - env variables can now be specified using AE_xyz instead of
125 PERL_ANYEVENT_xyz, subject to some rules.
126 - lots of new AnyEvent::Debug functionality.
127 - AnyEvent::Strict now checks for modifications $_ by the callback
128 (by setting it to readonly while executing the callback, if possible).
129 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
130 Invest extra overhead to work around this peculiar behaviour.
131 - split the pure perl backend into a loop module and an impl
132 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
133 - some read types (regex, netstring, json, storable) could
134 errornously keep an AE::Handle object alive on their own
135 (reported by Mohammad Toossi).
136 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
137 supports /etc/hosts overrides.
138 - add more workarounds around child watcher bugs in IO::Async
139 and (newly introduced) in POE.
140 - use glib child watchers and try to work around its limitations,
141 if possible (based on a patch by Kevin Ryde).
142 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
143 - many read types in AnyEvent::Handle would malfunction during
144 transitions to TLS because of a "delete $self->{rbuf}".
145 (reported by Gerald Galster).
146 - improve condvar blocking wait performance (probably only noticable
147 with faster backends), also make it easier for event loops that
148 do not support blocking waits.
149 - PERL_ANYEVENT_MODEL now also allows full module names, although
150 the usefulness of this feature is not yet clear.
151 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
152 - add t/80_ssltext.t.
153 - no longer use AUTOLOAD for the initial loading (for no very good reason).
154 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
155 - do not run event loop detection multiple times when programs call their cached
156 copies of AnyEvent::detect.
157 - parse_hostport now accepts absolute paths als unix domain sockets.
158 - suppress the idiotic warning inside POE by patching the POE::Kernel
159 run flags instead of asking the user to jump through contortions.
160 - add Task::Weaken as dependency (as recommends only at this time).
161 - replace PF_UNSPEC by 0, for good old minix.
162
1636.0 (not officially released, but was on CPAN for about 15 minutes :/)
164
1655.34 Thu May 12 09:59:41 CEST 2011
166 - AE::TLS didn't support both cert_file and key at the same time,
167 due to some copy&paste bug (reported and analyzed by vti).
168 - improve AE::Handle timeout documentation slightly.
169
1705.33 Mon Feb 7 21:51:41 CET 2011
171 - never appeared on CPAN, for unknown reasons.
172 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
173 object when it's already gone (analyzed by Ryan Bullock).
174 - update idna algorithm(s) to UTS#46 version 6.0.0.
175 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
176 uts46 says it's an error, and testsuite makes more sense.
177 - remove leading dots in idn_to_ascii - testsuite says to remove,
178 uts46 says it's an error, and testsuite makes more sense.
179 - fixed a bug in idn ignored character processing.
180 - passes the uts46 6.0.0 testsuite except for three cases which
181 apparently are bugs in the testsuite itself.
182 - passes the uts46 6.0.1 testsuite except for eleven cases which
183 apparently are bugs in the testsuite itself.
184
1855.31 Sun Jan 23 11:49:19 CET 2011
186 - work around a POE bug where POE would first poll for new events and
187 invoke handlers BEFORE handling the message queue.
188 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
189 - stop_read now does just that in TLS mode. see its
190 documentation for caveats.
191 - update to common::sense 3.4 warning set (some warnings in perl impose
192 an arbitrary 32kb limit on regex matches, which affects
193 AnyEvent::Handle).
194 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
195 - new experimental AE::Handle::resettls method.
196
1975.3 Fri Dec 31 05:48:13 CET 2010
198 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
199 issues with IO::Async have been solved in that version, including
200 the availabilty of a default loop!
201 - implement new wbuf_max parameter for AnyEvent::Handle.
202 - added Cocoa::EventLoop backend by Daisuke Murase.
203 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
204 - clarify handle on_prepare documentation.
205 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
206 to avoid leaking them to any child processes.
207 - try to force files into the arch-dependend lib directory, for the
208 half-"turly obsessive compulsive person"(s) that depend on it.
209
2105.29 Sun Dec 5 10:49:21 CET 2010
211 - convert EV backend to EV 4.00 API (so better upgrade EV too).
212 - AE::Handle: implement dynamic read_size adjustment, add
213 max_read_size parameter, reduce default read_size to 2048.
214 - add some FAQ entries for servers and/or on_eof vs. on_error.
215 - work around OS bugs (cygwin again) when lingering in AE::Handle.
216 - reduce memory usage slightly when connect was used in AE::Handle.
217 - wrok around more segfaults in Tk.
218 - document tls_ctx => undef as valid.
219 - detect WNOHANG value at compiletime.
220 - include some probably linux-only support for building AnyEvent
221 as part of the perl core.
222 - improve compatibility of t/08_idna.t to perl 5.8.
223 - make AnyEvent::Strict truly optional (in case somebody builds a
224 minimal perl...).
225
2265.28 Wed Oct 13 04:14:23 CEST 2010
227 - due to a glitch, AnyEvent's internal getprotobyname wasn't
228 actually used in tcp_server.
229 - implement AnyEvent::CondVar->new method, which was documented but
230 missing.
231 - support autoloading of read types in unshift_read.
232 - AnyEvent::Strict: do the test for invalid fh arguments differently,
233 resulting in more informative error messages.
234 - new AnyEvent::FAQ manpage.
235 - clarify that you can actually call push_read etc. in on_read.
236 - put keys and certs in /tmp, not into the current directory,
237 and document this in the SECURITY CONSIDERATIONS sections
238 of AnyEvent::TLS.
239 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
240 requested by Ron Isaacson.
241 - document $AnyEvent::DNS::RESOLVER variable.
242 - incorporated major typo patches by Abhijit Menon-Sen.
243 - AnyEvent::Handle now croaks when negative timeouts are passed.
244 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
245 undocumented, but who knows when they might come in handy.
246
2475.271 Tue Jun 8 12:05:46 CEST 2010
248 - backport to perl 5.8.x.
249
2505.27 Sun Jun 6 12:12:05 CEST 2010
251 - postpone differently in AnyEvent::Socket now, as
252 when not, canceling the connection attempt might fail
253 (found by Felix Antonius Wilhelm Ostmann).
254 - explicitly check for non-stream sockets in AE::Handle, too many
255 clueless people fell into the trap of this somehow working.
256 - simplified and reworked the "OTHER MODULES" section.
257 - better/more condvar examples.
258
2595.261 Wed Apr 28 16:13:36 CEST 2010
260 - AF_INET6 was not properly used from Socket6 during configuration
261 time (found by Andy Grundman).
262
2635.26 Mon Apr 12 04:49:35 CEST 2010
264 - don't generate spurious readyness notifications when select
265 returns EINTR in the pure perl backend (analysed and testcase
266 by Bas Denissen).
267 - give same examples in SYNOPSIS sections of AnyEvent and AE.
268 - provider faster implementations for AE::time/now/now_update for
269 EV and Perl backends.
270 - provide faster AE::cv and AE::time implementations for all
271 backends.
272 - no longer support lower/mixed-case signal names (this was
273 never documented nor universally supported).
274 - some more memory µ-optimisations, and somewhat less messy
275 code to cope with both AnyEvent and AE APIs.
276
2775.251 Sat Mar 13 00:58:21 CET 2010
278 - make test failed if it was run before make install on perl <=
279 5.8.8, and older versions of perl make our live very difficult,
280 so write constants.pl during Makefile.PL time.
281
2825.25 Sat Mar 13 00:23:14 CET 2010
283 - fix a race condition in AnyEvent::Handle that would cause
284 a "bio_read: unsupported method" error (in _another_
285 ssl connection) after user code threw an exception.
286 - added AnyEvent::Handle->destroyed method.
287 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
288 - the AnyEvent::Util::fh_nonblocking function worked only
289 by lucky accident on win32.
290 - smaller and faster AnyEvent::Util::fh_nonblocking.
291 - when the (required!) Time::HiRes module is missing, AnyEvent
292 did not fall back to built-in time properly.
293 - do not load Fcntl at runtime, saving memory and loading time.
294 - precompile a number of constants and use them instead of runtime
295 detection and eval.
296 - free detection code after detection and similar memory optimisations.
297 - Perl backend timer interval best effort drift has been improved
298 (same algorithm as EV).
299 - update unicode idna mapping table.
300
3015.24 Tue Jan 5 11:39:43 CET 2010
302 - cygwin never reports errors from failed connects. 1.5 just gives
303 you continous readyness and EAGAIN, 1.7 is even more broken and
304 just hangs. work around both issues in a cygwin-specific hack.
305 - improve idle watcher documentation slightly.
306
3075.23 Sun Dec 20 23:48:00 CET 2009
308 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
309 - implement punycode_encode/decode, idn_nameprep,
310 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
311 - provide $AE::VERSION.
312 - removed traces of "no strict 'refs'".
313
3145.22 Sat Dec 5 03:51:13 CET 2009
315 - downgrade-or-fail in AnyEvent::Handle::push_write, to
316 diagnose encoding failures earlier and more succinctly.
317 (this works around bugs in perl, throwing away encoding info
318 when passing scalar data to extensions).
319 - add more examples to AnyEvent::Socket manpage.
320 - upgrade internal warning set to the same as common::sense 2.03.
321 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
322 (John Beppu).
323 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
324 - add google public dns servers to fallback server set and make sure
325 we load-balance properly between the three sets. also add all
326 fallback dns servers, not just a random one, to each dns config.
327
3285.21 Thu Nov 19 02:48:47 CET 2009
329 - fix a problem where socket constants were called with parameters
330 (spotted by David Friedland).
331 - fork_call never use'd POSIX (reported by Daisuke Maki).
332 - improve perl 5.6 compatibility further (but it still won't work
333 unless you rip out everything but the core).
334 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
335 uses 10MB of RAM, but doesn't open a console window. *sigh*.
336
3375.202 Wed Oct 14 22:35:44 CEST 2009
338 - AnyEvent::DNS would unexpectedly clobber $_ under windows
339 (analysed by Matthias Waldorf).
340 - AnyEvent::Handle::run_cmd can now pass the PID of the
341 newly-created process, which is much less useful than it might
342 sound (based on patch by Yann Kerherve).
343
3445.201 Tue Sep 29 12:09:25 CEST 2009
345 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
346 (reported by Torsten Foertsch).
347 - common::sense 2.0 could cause tcp_server to throw an exception
348 (analysed by elmex).
349
3505.2 Mon Sep 14 07:04:49 CEST 2009
351 - INCOMPATIBLE CHANGE: do no longer support register_read_type
352 and register_write_type in AnyEvent::Handle, instead support
353 package names (the facility was mostly abused).
354 - implement "packagename-as-read/write type" support in
355 AnyEvent::Handle.
356 - AnyEvent::Handle: new options "keepalive" and "oobinline".
357 - oobinline set by default to avoid security issues.
358 - the pure-perl event loop backend wrongly detected times() fallback
359 support (spotted by Pavel Boldin).
360
3615.12 Tue Sep 1 20:26:50 CEST 2009
362 - be more lenient when parsing resolv.conf files, as some people
363 use hashmarks as comment indicator inside directives (reported
364 by Michael S. Fischer).
365 - use same set of warnings as common::sense 2.0.
366 - fix a potential 32 bit overflow issue due to perl having problems
367 with large hex constants in the Perl backend.
368 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
369 by default.
370 - fix some prototypes.
371 - (optionally) require Async::Interrupt 1.02.
372 - replace opendns fallbacks by verizon/level3 public dns servers.
373
3745.112 Fri Aug 21 13:59:04 CEST 2009
375 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
376 using starttls with an illegal/uncreatable tls context.
377 - port to broken win32 perls.
378
3795.111 Sun Aug 16 18:44:36 CEST 2009
380 - perl errornously requires sockaddr_un structures returned by the
381 kernel to have a specific length. work around this bug by padding
382 sockaddr_un structures (found to fail on OS X, as perl only uses
383 the correct code for GNU/Linux, but likely to fail on other
384 OSes as well).
385
3865.11 Wed Aug 12 17:49:37 CEST 2009
387 - fixed a bug in run_cmd with input from filename.
388 - create a json coder object when none is given in write direction
389 to reduce dependency on JSON.
390
3915.1 Tue Aug 11 03:17:32 CEST 2009
392 - tcp_connect now reports the correct error in $!, not an unrelated
393 one (was broken since 4.91).
394 - AE::Handle did not properly clear rtimeout/wtimeout watchers
395 when retrying a tcp_connect.
396 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
397 - fixed wrong documentation in AE::Socket.
398
3995.01 Mon Aug 10 03:16:32 CEST 2009
400 - last release broke EV child watchers :/.
401
4025.0 Sun Aug 9 17:32:09 CEST 2009
403 - officially introduce and document the AE API.
404 - lowering the timeout at runtime did not have immediate effect
405 in AnyEvent::Handle.
406 - AE::Handle now has separate and independent read and write timeout
407 settings.
408 - probe for Guard module when AnyEvent::Util is loaded, not at
409 runtime (Event doesn't like eval "use" from callbacks, and it avoids
410 exporting the probe function).
411 - Impl::IOAsync's io watchers did not use the filehandle emulation
412 code to work around it's watcher limits.
413 - work around IO::Async's problems with overloaded objects.
414 - try to disable wearnings inside IO::Async, it's just too buggy.
415 - do no longer use Event by default, only EV or Perl, Event is too buggy.
416 - hackishly provide fast AE:: interface when EV or Perl is the backend.
417 - inherit VERSION from AnyEvent::VERSION in most modules.
418
4194.91 Thu Aug 6 15:42:45 CEST 2009
420 - AE::Handle::starttls could get out of sync when the read buffer
421 already contains some TLS handshake.
422 - AE::Handle did not properly free the TLS session, causing bigger
423 memory leaks in Net::SSLeay than expected (reported by toaster).
424 - AE::Socket::tcp_connect will now artificially delay invoking
425 the callback to avoid returning after invoking the callback.
426 - convert many internal watcher uses to AE API.
427
4284.9 Sat Aug 1 11:07:01 CEST 2009
429 - Glib silently fails when registering a timeout with
430 a negative value, so avoid that.
431 - call condvar callback immediately when it is set
432 after the condition is already signalled.
433 - check rbuf_max condition only after trying to consume
434 data.
435 - normalise signal numbers to names when using the ae's
436 signal handling, but do not document this (yet).
437 - pure perl signal emulation did not properly set nonblocking
438 mode on the signal pipe on win32 (this is unlikely to have caused
439 any issues).
440 - new module: AnyEvent::Debug.
441 - AnyEvent::Strict now checks that a signal was specified
442 by name (not name or number).
443 - reduce memory footprint in typical cases by ~50kb by
444 only compiling signal code when necessary.
445 - add AnyEvent::Handle->rbuf_max.
446 - grab the AE namespace for future new API, implement stubs for
447 said future/alternative API.
448 - new function: AnyEvent::Socket::format_hostport.
449
4504.881 Tue Jul 28 12:51:53 CEST 2009
451 - work around a bug in local in pre-5.10 perls, causing
452 AnyEvent::Handle to recurse when it should not
453 (analysed by elmex).
454
4554.88 Tue Jul 28 04:04:37 CEST 2009
456 - re-bless the handle into a dummy package after calling
457 AnyEvent::Handle::destroy, so the user does not need to check
458 for errors after every push_write etc.
459 - do not attempt to run t/02_signals.t on obviously broken
460 platforms.
461
4624.87 Sun Jul 26 02:06:16 CEST 2009
463 - do not attempt to linger when there is no longer a valid fh.
464 - enforce tls mode to be either accept or connect, do not simply
465 segfault in Net::SSLeay.
466 - AnyEvent::Handle can now call tcp_connect itself (new parameters
467 connect, on_prepare, on_connect and on_connect_error). Updated
468 tutorial accordingly.
469 - add AnyEvent::Impl::Irssi backend.
470
4714.86 Mon Jul 20 23:52:29 CEST 2009
472 - since the verbose warning is not enough, explicitly document
473 that versions before 1.33 of Net::SSLeay are not secure.
474 - work around signal handling races in Event and (...) Event::Lib.
475 - try to align signal-race timer to full-second boundaries.
476 - work around Tk not liking negative timeouts.
477 - don't complain of different grades of environmental unfriendlyness
478 in IO::Async.
479
4804.85 Sat Jul 18 06:16:14 CEST 2009
481 - nail the signal race problem in perl once and for all
482 (see $AnyEvent::MAX_SIGNAL_LATENCY).
483 - take advantage of Async::Interrupt if it is available.
484 - load Time::HiRes and Guard modules on demand only.
485 - add optional/recommended modules section to AnyEvent
486 documentation.
487 - reduce memory usage considerably (and reduce startup penalty)
488 by not using "strict", "warnings" and "overload" modules.
489 - work around buggy windows/openbsd perls and provide EBADMSG
490 and EPROTO ourselves when missing.
491 - improve perl 5.6 compatibility of the core event loop.
492 - made Net::SSLeay version 1.33 a soft requirement.
493
4944.83 Fri Jul 17 16:56:26 CEST 2009
495 - implement AnyEvent::Socket::getprotobyname.
496 - AnyEvent::CondVar's will now detect recursive blocking
497 waits and will croak, as too many people fall into
498 this trap.
499 - AnyEvent::Handle will now call ->destroy on itself after
500 executing the on_error callback, instead of doing some
501 half-baked internal shutdown, for fatal errors.
502 - clarify on_eof behaviour w.r.t. the read queue and
503 on_read callbacks.
504 - ignore some possible spurious wake-ups in tcp_connect.
505
5064.82 Sat Jul 11 00:34:55 CEST 2009
507 - POE and Event backends didn't accept some callable objects as
508 callbacks.
509 - use Config module instead of POSIX module to detect signal names
510 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
511 module doesn't even have all POSIX signals :/.
512 - use more workarounds around the many refcnt/corruption bugs in
513 Event::Lib.
514 - work around a race condition in perl's select, causing t/03_child.t
515 to rarely fail.
516
5174.81 Thu Jul 9 10:30:30 CEST 2009
518 - AnyEvent::Handle didn't properly diagnose write errors
519 (it expected -1 from syswrite, how lame... :).
520 - support file descriptors in addition to file handles
521 in AnyEvent->io.
522 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
523 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
524 and PERL_ANYEVENT_CA_PATH.
525 - provide a sensible synopsis section for AnyEvent::TLS.
526 - add a "supported backends" section to the manpage.
527 - added simple io watcher test to testsuite, using a
528 portable_socketpair.
529 - tried to improve the stability of the Event::Lib backend,
530 YMMV.
531
5324.8 Mon Jul 6 23:45:16 CEST 2009
533 - AnyEvent::DNS did not properly follow CNAME records with
534 uppercase targets.
535 - AnyEvent::DNS would errornously return AAAA records
536 with v4 mapped addresses (a faulty record) as ipv4 addresses,
537 causing AnyEvent::Socket to throw an exception.
538 - added new module AnyEvent::TLS for easier SSL/TLS context
539 creation, with many options including hostname verification,
540 secure default configuration, lots of documentation and,
541 predefined diffie-hellman keys for perfect forward security
542 and much more. get it while it's still fresh!
543 - use AnyEvent::TLS in AnyEvent::Handle for context management.
544 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
545 so AnyEvent::Socket users have smaller memory footprint
546 in the common case.
547 - add AnyEvent::Handle->push_shutdown method.
548 - add an additional error message parameter to AnyEvent::Handle's
549 on_error callback (for TLS, $! is still available).
550 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
551 - make AnyEvent::Handle more robust against early conenction
552 failures (during new), and return C<undef> in that case
553 from the constructor.
554 - AnyEvent::Handle will now try to load only JSON::XS first,
555 then fall back to JSON.
556 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
557 and because it was documented that way.
558
5594.451 Fri Jul 3 00:28:58 CEST 2009
560 - do not clear rbuf when shutting down an AnyEvent::Handle
561 object - doing so breaks AnyEvent::HTTP.
562
5634.45 Mon Jun 29 22:59:26 CEST 2009
564 - a write error could cause AnyEvent::Handle to create
565 an I/O watcher with an undefined $fh.
566 - special-case mapped ipv4 addresses in both
567 AnyEvent::Socket::format_address and parse_address,
568 to treat them just like ipv4 addresses.
569 - updated and overhauled the AnyEvent::Intro doc.
570 - implement AnyEvent::Socket::format_ipv4/ipv6.
571 - slightly speed up portable_pipe/socketpair.
572 - expand condvar begin/end documentation.
573
5744.42 Fri Jun 26 08:32:18 CEST 2009
575 - preliminary and neccesarily incomplete support for IO::Async.
576 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
577 case it was set to IGNORE, to ensure we can catch child statuses
578 even when the calling env acts stupidly.
579 - updated benchmarks with IO::Async, which performs very well.
580
5814.412 Wed Jun 24 01:35:57 CEST 2009
582 - support an "untaint" attribute for AnyEvent::DNS
583 and set it on the default resolver.
584 - implement and document AnyEvent::Impl::Perl::loop.
585 - remove all anyevent-env variables from %ENV
586 when running in tainted mode.
587 - mention and extend the IO::Lambda benchmark.
588
5894.411 Sun Jun 7 18:48:02 CEST 2009
590 - do not try to use F_SETFD on windows.
591
5924.41 Thu May 14 06:40:11 CEST 2009
593 - work around issues in older perls (5.8.0?) when
594 a signal handler is deleted from the %SIG hash.
595 - use POSIX::_exit in child test, to avoid running
596 destructors.
597 - speed up CHLD handling by relying on SIGCHLD being
598 synchronously delivered, even when we roll our own
599 implementation.
600 - AnyEvent::DNS: add the "dname" resource record name
601 for cosmetic reasons.
602
6034.4 Sun Apr 26 20:12:33 CEST 2009
604 - implemented idle watchers, where applicable.
605 - AnyEvent->time died when Event backend was in use.
606 - fix a memleak in the Tk backend.
607 - sped up Tk timer handling.
608 - clip negative "after" values to 0 in AnyEvent::Impl::Event
609 to avoid spamming out warning messages.
610 - fix Qt timers without interval.
611 - avoid Qt zero-timeout special case, in old libqt's.
612
6134.352 Mon Apr 20 16:31:11 CEST 2009
614 - fix AnyEvent::Strict error messages for child watchers.
615 - fix/update Makefile.PL "recommends" versions.
616 - add AnyEvent->now_update.
617
6184.351 Sat Apr 11 07:56:14 CEST 2009
619 - actually make the signal pipe work under win32.
620 - localise $! in signal handler to avoid changing $!, although
621 perl itself does not seem to save/restore errno either.
622 - set the cloexec flag on the signal pipe (normally set by perl too).
623
6244.35 Fri Mar 27 11:48:20 CET 2009
625 - event models relying on AnyEvent's signal watcher emulation did
626 invoke the callback asynchronously, contradicting documentation
627 and causing signals to get lost (this includes AnyEvent's own
628 event loop). AnyEvent now uses the standard pipe trick to make
629 callback execution synchronous to the event loop.
630 - AnyEvent::Handle didn't free TLS context data on DESTROY
631 (patch by Pavel Shaydo).
632 - work around the ever-incompatibly-changing API of MakeMaker.
633 - document that changing global variables without restoring them
634 is a bad idea in Perl (noted by Adam Rosenstein).
635 - AnyEvent::Strict now barfs if ->io is passed a file.
636
6374.34 Thu Feb 12 18:32:45 CET 2009
638 - separately buffer TLS read data, as otherwise the read queue could
639 deadlock as receiving data is not expected while draining the
640 read queue (which cna only happen with TLS).
641 - raise EBADMSG error on JSON decoding errors.
642 - fix some minor manpage bugs (reported by Maximilian Gaß).
643 - speed up select bitmask parsing quite a bit in the pure perl
644 backend.
645 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
646
6474.331 Tue Jan 6 21:07:25 CET 2009
648 - socketpair fails on many vista machines because vista has
649 completely broken accept/getpeername and getsockname functions,
650 so we provide our own socketpair emulation that kind of works
651 (AnyEvent::Util::portable_pipe).
652 - new function: AnyEvent::Util::portable_socketpair.
653 - take advantage of the Guard module if it exists.
654
6554.33 Fri Nov 21 02:35:40 CET 2008
656 - AnyEvent::Strict did errornously flag a fileno of 0 as
657 illegal.
658 - reduce memory usage and slightly speed up the pure perl backend
659 by only storing the file descriptor, not the file handle.
660 - add missing autocork method to AnyEvent::Handle (reported
661 by Adam Rosenstein).
662 - AnyEvent::DNS->resolve errornously documented an $rcode result
663 argument, but there isn't (spotted by Henrik Krohns).
664 - the naked truth about Tk - it's basically unmaintained.
665
6664.32 Mon Nov 3 22:46:32 CET 2008
667 - fix AnyEvent::Socket::resolve_sockaddr to properly support
668 unix sockets again and choose a proper default for
669 the socket type.
670 - fix call to resolve_sockaddr in tcp_connect.
671
6724.31 Thu Oct 30 04:41:48 CET 2008
673 - implemented AnyEvent::Handle->destroy method.
674 - hint about unexpected effects in TLS mode.
675 - speed up AnyEvent::Impl::Perl by using more arrays
676 instead of hashes without the slightest loss of readability :)
677 - work around a perl argument refcounting bug.
678 - with some perl patches applied, AnyEvent now seems leak-free.
679
6804.3 Fri Oct 3 09:18:43 CEST 2008
681 - AnyEvent will now install a no-op signal handler for SIGPIPE,
682 unless one has been installed already.
683 - warn about not loaidng AnyEvent::Impl::POE early enough
684 (patch by Adam Rosenstein).
685 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
686 - fix a bug in where in SSL connect mode, the
687 client would first wait for some data by the server and
688 otherwise hang. (reported and analysed in an absolutely
689 exemplary manner by Adam Rosenstein).
690 - fix a bug in where SSL EOF would not be treated
691 as stream EOF, putting the connection into a hung state.
692 (reported and analysed in an absolutely exemplary manner by Adam
693 Rosenstein).
694 - fix a potential segfault when the TLS context
695 would go missing in a read callback (e.g. due to stoptls) -
696 Net::SSLeay of course makes no type checking whatsoever.
697 - AnyEvent::Handle will not stop the read watcher in TLS mode.
698 - AnyEvent::Handle->stoptls will send a TLS close notify
699 instead of simply destroying the stream now.
700 - fix a bug where large blocks of data written in TLS mode
701 would not be sent unless triggered by receives.
702 - on_drain will now take the tls write buffer into account.
703 - SSL operations have been streamlined a bit and should be faster
704 now, more to come, though.
705 - added a FAQ section, will document nontrivial issues in there.
706
7074.234 Mon Sep 29 04:08:13 CEST 2008
708 - fix child watcher documentation: only child exits will be reported,
709 no trace events.
710 - mention SIGPIPE in AnyEvent::Handle.
711 - perl backend will now use POSIX::times as monotonic clock
712 when available and the monotonic clock option is not.
713 - run even without Time::HiRes in some cases now, but do not
714 provide sub-second accuracy (all *supported* perl versions have
715 Time::HiRes).
716
7174.233 Fri Aug 22 01:48:05 CEST 2008
718 - fix a bug in the testsuite, causing a hang.
719 - clarified lots of AnyEvent::Handle constructor arguments.
720
7214.232 Thu Aug 21 20:44:25 CEST 2008
722 - fix a bug in AnyEvent::Handle that could cause two fatal error
723 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
724 (testcase by Leon Brocard).
725 - clarified on_eof/on_error documentation a bit.
726
7274.231 Tue Jul 29 13:12:15 CEST 2008
728 - remove some debugging code left in AnyEvent::Util::fork_call
729 (and no, it's impossible to implement with the broken windows
730 perls without resource leaks or worse).
731
7324.23 Tue Jul 29 12:19:59 CEST 2008
733 - document the first parameter passed to condvar callbacks to be
734 the callback.
735 - add AnyEvent::Socket::{ntoa,aton} aliases.
736 - optimize the AE::Handle->push_read (line) for the default
737 eol marker.
738 - optimize push_read (packstring|storable) for small packets.
739 - invoke on_error callback when no on_eof callback is set.
740 - fix a bug in push_read (storable) of unknown impact.
741
7424.22 Sun Jul 20 16:34:13 CEST 2008
743 - new function AnyEvent::Socket::parse_hostport.
744 - as the bulkheads at microsoft can't even get getprotobyname reliably
745 working on their shitty fucking broken os we need to hardcode
746 some common protocol numbers in AnyEvent::Socket. How can
747 people even bother with such a pile of shit as windows.
748
7494.21 Thu Jul 17 14:40:05 CEST 2008
750 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
751 which is incompatible to previous invocation syntax.
752 - work around more windows perl bugs in fork_call. I cannot imagine
753 why anybody would bother implementing fork in such an obviously
754 unusable way.
755 - avoid calling $do_retry if the request has already finished in
756 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
757 K. Sharifulin).
758
7594.2 Sat Jul 12 22:42:11 CEST 2008
760 - support an "interval" argument for repeating timers.
761 - fix a bug in the pure perl event loop that caused timers to
762 be delayed under some circumstances.
763 - implement optional argument checking via PERL_ANYEVENT_STRICT.
764 - not importing anything from Event causes it to malfunction
765 and not use Time::HiRes, work around this bug.
766 - more bug workarounds against the endlessly broken Tk module,
767 seems to work now on 32 bit machines, but less so on 64 bit.
768 - minor optimisations applied to most backends.
769
7704.161 Fri Jul 4 14:24:48 CEST 2008
771 - fixed the t/handle/02_write.t, it read after write, with the
772 assumption that all data could be written before read
773 (reported by Jim Razmus).
774
7754.160 Thu Jul 3 04:02:21 CEST 2008
776 - re-registering signal watchers after unregistering one did
777 not work if a backend falls back on AnyEvent's default
778 implementation (also affected child watchers).
779 - new AnyEvent::Handle options: autocork and no_delay.
780
7814.152 Sun Jun 22 14:15:44 CEST 2008
782 - allow for 32-bit perls that implement shifts differently
783 on different architectures in parse_ipv4 (reported and
784 analysed by Keiichi DAIBA).
785
7864.151 Fri Jun 6 17:34:24 CEST 2008
787 - make sure specifying _only_ on_read and never pushing reads
788 works.
789
7904.15 Fri Jun 6 13:00:46 CEST 2008
791 - the pure perl backend would keep some watchers alive when more than
792 one watcher was registered for the same fd.
793 - new "packstring" and "storable" read and write types
794 for AnyEvent::Handle.
795 - allow on_eof handler to be called after on_error with EPIPE returns.
796 - do not immediately call on_read callback in handle constructor.
797
7984.14 Thu Jun 5 20:29:31 CEST 2008
4 - Fixed a bug in DNS SRV priority sorting. 799 - Fixed a bug in DNS SRV priority sorting.
800 - AnyEvent::Util::guard now reports runtime errors while
801 executing the guard block as warnings.
802 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
5 803
64.13 Thu Jun 5 00:47:59 CEST 2008 8044.13 Thu Jun 5 00:47:59 CEST 2008
7 - AnyEvent::DNS only followed cname chains with length 2, 805 - AnyEvent::DNS only followed cname chains with length 2,
8 contrary to documentation. bumped it up to 10, thanks to 806 contrary to documentation. bumped it up to 10, thanks to
9 microsoft, the current broken-dns-config-king. 807 microsoft, the current broken-dns-config-king.
… …
128 - support service names in tcp_connect and tcp_server. 926 - support service names in tcp_connect and tcp_server.
129 - work around netbsd/openbsd bug causing unix sockets not to 927 - work around netbsd/openbsd bug causing unix sockets not to
130 become writable until empty. 928 become writable until empty.
131 - only make a dummy read for error detection when getpeername 929 - only make a dummy read for error detection when getpeername
132 returns ENOTCONN. 930 returns ENOTCONN.
133 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 931 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
134 override. 932 override.
135 - new AnyEvent::Guard::cancel. 933 - new AnyEvent::Guard::cancel.
136 934
1373.5 Sat May 17 23:17:14 CEST 2008 9353.5 Sat May 17 23:17:14 CEST 2008
138 - gracefully deal with systems not defining CLOCK_MONOTONIC 936 - gracefully deal with systems not defining CLOCK_MONOTONIC
… …
152 950
1533.41 Sun May 11 19:53:13 CEST 2008 9513.41 Sun May 11 19:53:13 CEST 2008
154 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 952 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
155 953
1563.4 Sun May 11 00:29:25 CEST 2008 9543.4 Sun May 11 00:29:25 CEST 2008
157 - complete re-implementation of condvars with a hopefully much mroe useful 955 - complete re-implementation of condvars with a hopefully much more useful
158 API (backward compatible functions exist but are undocumented). 956 API (backward compatible functions exist but are undocumented).
159 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 957 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
160 (still subject to change). 958 (still subject to change).
161 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 959 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
162 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 960 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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