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

Comparing AnyEvent/Changes (file contents):
Revision 1.109 by root, Mon May 26 04:30:30 2008 UTC vs.
Revision 1.526 by root, Fri Feb 3 01:41:35 2012 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: exceptbit 3TODO: document TCP_*** constants
4TODO: add debug/trace logging to some modules
5
6 - documented Rocco fabricating statements about AnyEvent and me.
7 - updated uts46data.pl for unicode 6.1.0.
8
96.14 Tue Jan 31 20:00:24 CET 2012
10 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
11 probably an editing mistake (reported by Darin McBride).
12
136.13 Thu Jan 12 07:27:01 CET 2012
14 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
15 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
16 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
17 regex-matching inside the callback would change the parameter, despite
18 $1 being dynamically scoped per-block. this perl bug is now being
19 worked around (testcase by Cindy Wang).
20
216.12 Mon Dec 12 13:21:10 CET 2011
22 - $! was clobbered by subsequent calls in tcp_connect, due to
23 the postpone not saving/restoring it (analyzed by Richard Garnish).
24
256.11 Tue Nov 22 10:36:05 CET 2011
26 - Tk cannot create windows when tainted unless you set a title, so
27 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
28 - escape any nonprintable/nonascii characters when stringifying
29 backtraces.
30 - log the reason loading an interface module fails to load at level debug.
31
326.1 Tue Oct 4 19:44:30 CEST 2011
33 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
34 and above), and some messages inside AnyEvent have been elevated
35 to higher log levels to print by default.
36 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
37 - work around yet another signal race bug, newly introduced by POE. My
38 how that module sucks.
39 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
40 support ipv6-only operation even with defaults (also use ::1 instead
41 of 127.0.0.1).
42 - give better guidelines on how to chose a logging level.
43 - AnyEvent::Log can now cap (limit) log message priority by context.
44 - AE::log will now save and restore $! and $@, so callers don't have to.
45 - AE::Strict now checks that registered fds are still valid at regular
46 intervals.
47 - skip frequently called 'require' calls if the module appears loaded.
48 - increase signal latency test timer considerably for poor little osx,
49 which can't be bothered with basic posix functionality such as signal
50 delivery.
51 - mention some debugging aids in the tutorial.
52 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
53 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
54 of the context as title.
55 - AnyEvent::Strict would not restore read-write access to $_ when a callback
56 died - this is a user bug, but since AE::Strict exists to help dbeugging...
57 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
58 to work with it - the author is well aware of the issues but refuses
59 to fix them. This extreme step has been taken because the module actively
60 attacks the goals of AnyEvent and makes it harder for module authors to use
61 AnyEvent.
62
636.02 Fri Aug 26 20:08:31 CEST 2011
64 - log_to_syslog now accepts facility strings.
65 - log_to_syslog would not allow facilities - "$facility|$level"
66 *literally* means that string. sheesh.
67 - the CPAN parser got confused and thought we provide Sys::Syslog, try
68 to work around.
69 - renamed AnyEvent::Impl::FLTK2 to FLTK.
70
716.01 Fri Aug 26 07:04:11 CEST 2011
72 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
73 the TTL as fourth element - this affects ->resolve, ->request
74 and ->dns_unpack indirectly. Convenience functions such as
75 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
76 AnyEvent::Socket.
77 - INCOMPATIBLE CHANGE: backend authors now should not implement
78 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
79 methods.
80 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
81 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
82 - DEPRECIATION: please change
83 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
84 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
85 in your programs, for future compatibility.
86 - *blush* idle watcher emulation was borked.
87 - AnyEvent::Strict now checks the AE::xxx functions as well.
88 - querying too long domain names would barf deep inside AnyEvent::DNS - since
89 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
90 for too long domain names explicitly and passes an undef to the callback
91 instead (reported by James Bromberger).
92 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
93 - new functions: AnyEvent::postpone, AnyEvent::log.
94 - new module: AnyEvent::Log, for simple logging needs.
95 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
96 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
97 a debug shell.
98 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
99 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
100 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
101 /etc/hosts can be found.
102 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
103 configuration.
104 - env variables can now be specified using AE_xyz instead of
105 PERL_ANYEVENT_xyz, subject to some rules.
106 - lots of new AnyEvent::Debug functionality.
107 - AnyEvent::Strict now checks for modifications $_ by the callback
108 (by setting it to readonly while executing the callback, if possible).
109 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
110 Invest extra overhead to work around this peculiar behaviour.
111 - split the pure perl backend into a loop module and an impl
112 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
113 - some read types (regex, netstring, json, storable) could
114 errornously keep an AE::Handle object alive on their own
115 (reported by Mohammad Toossi).
116 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
117 supports /etc/hosts overrides.
118 - add more workarounds around child watcher bugs in IO::Async
119 and (newly introduced) in POE.
120 - use glib child watchers and try to work around its limitations,
121 if possible (based on a patch by Kevin Ryde).
122 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
123 - many read types in AnyEvent::Handle would malfunction during
124 transitions to TLS because of a "delete $self->{rbuf}".
125 (reported by Gerald Galster).
126 - improve condvar blocking wait performance (probably only noticable
127 with faster backends), also make it easier for event loops that
128 do not support blocking waits.
129 - PERL_ANYEVENT_MODEL now also allows full module names, although
130 the usefulness of this feature is not yet clear.
131 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
132 - add t/80_ssltext.t.
133 - no longer use AUTOLOAD for the initial loading (for no very good reason).
134 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
135 - do not run event loop detection multiple times when programs call their cached
136 copies of AnyEvent::detect.
137 - parse_hostport now accepts absolute paths als unix domain sockets.
138 - suppress the idiotic warning inside POE by patching the POE::Kernel
139 run flags instead of asking the user to jump through contortions.
140 - add Task::Weaken as dependency (as recommends only at this time).
141 - replace PF_UNSPEC by 0, for good old minix.
142
1436.0 (not officially released, but was on CPAN for about 15 minutes :/)
144
1455.34 Thu May 12 09:59:41 CEST 2011
146 - AE::TLS didn't support both cert_file and key at the same time,
147 due to some copy&paste bug (reported and analyzed by vti).
148 - improve AE::Handle timeout documentation slightly.
149
1505.33 Mon Feb 7 21:51:41 CET 2011
151 - never appeared on CPAN, for unknown reasons.
152 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
153 object when it's already gone (analyzed by Ryan Bullock).
154 - update idna algorithm(s) to UTS#46 version 6.0.0.
155 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
156 uts46 says it's an error, and testsuite makes more sense.
157 - remove leading dots in idn_to_ascii - testsuite says to remove,
158 uts46 says it's an error, and testsuite makes more sense.
159 - fixed a bug in idn ignored character processing.
160 - passes the uts46 6.0.0 testsuite except for three cases which
161 apparently are bugs in the testsuite itself.
162 - passes the uts46 6.0.1 testsuite except for eleven cases which
163 apparently are bugs in the testsuite itself.
164
1655.31 Sun Jan 23 11:49:19 CET 2011
166 - work around a POE bug where POE would first poll for new events and
167 invoke handlers BEFORE handling the message queue.
168 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
169 - stop_read now does just that in TLS mode. see its
170 documentation for caveats.
171 - update to common::sense 3.4 warning set (some warnings in perl impose
172 an arbitrary 32kb limit on regex matches, which affects
173 AnyEvent::Handle).
174 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
175 - new experimental AE::Handle::resettls method.
176
1775.3 Fri Dec 31 05:48:13 CET 2010
178 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
179 issues with IO::Async have been solved in that version, including
180 the availabilty of a default loop!
181 - implement new wbuf_max parameter for AnyEvent::Handle.
182 - added Cocoa::EventLoop backend by Daisuke Murase.
183 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
184 - clarify handle on_prepare documentation.
185 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
186 to avoid leaking them to any child processes.
187 - try to force files into the arch-dependend lib directory, for the
188 half-"turly obsessive compulsive person"(s) that depend on it.
189
1905.29 Sun Dec 5 10:49:21 CET 2010
191 - convert EV backend to EV 4.00 API (so better upgrade EV too).
192 - AE::Handle: implement dynamic read_size adjustment, add
193 max_read_size parameter, reduce default read_size to 2048.
194 - add some FAQ entries for servers and/or on_eof vs. on_error.
195 - work around OS bugs (cygwin again) when lingering in AE::Handle.
196 - reduce memory usage slightly when connect was used in AE::Handle.
197 - wrok around more segfaults in Tk.
198 - document tls_ctx => undef as valid.
199 - detect WNOHANG value at compiletime.
200 - include some probably linux-only support for building AnyEvent
201 as part of the perl core.
202 - improve compatibility of t/08_idna.t to perl 5.8.
203 - make AnyEvent::Strict truly optional (in case somebody builds a
204 minimal perl...).
205
2065.28 Wed Oct 13 04:14:23 CEST 2010
207 - due to a glitch, AnyEvent's internal getprotobyname wasn't
208 actually used in tcp_server.
209 - implement AnyEvent::CondVar->new method, which was documented but
210 missing.
211 - support autoloading of read types in unshift_read.
212 - AnyEvent::Strict: do the test for invalid fh arguments differently,
213 resulting in more informative error messages.
214 - new AnyEvent::FAQ manpage.
215 - clarify that you can actually call push_read etc. in on_read.
216 - put keys and certs in /tmp, not into the current directory,
217 and document this in the SECURITY CONSIDERATIONS sections
218 of AnyEvent::TLS.
219 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
220 requested by Ron Isaacson.
221 - document $AnyEvent::DNS::RESOLVER variable.
222 - incorporated major typo patches by Abhijit Menon-Sen.
223 - AnyEvent::Handle now croaks when negative timeouts are passed.
224 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
225 undocumented, but who knows when they might come in handy.
226
2275.271 Tue Jun 8 12:05:46 CEST 2010
228 - backport to perl 5.8.x.
229
2305.27 Sun Jun 6 12:12:05 CEST 2010
231 - postpone differently in AnyEvent::Socket now, as
232 when not, canceling the connection attempt might fail
233 (found by Felix Antonius Wilhelm Ostmann).
234 - explicitly check for non-stream sockets in AE::Handle, too many
235 clueless people fell into the trap of this somehow working.
236 - simplified and reworked the "OTHER MODULES" section.
237 - better/more condvar examples.
238
2395.261 Wed Apr 28 16:13:36 CEST 2010
240 - AF_INET6 was not properly used from Socket6 during configuration
241 time (found by Andy Grundman).
242
2435.26 Mon Apr 12 04:49:35 CEST 2010
244 - don't generate spurious readyness notifications when select
245 returns EINTR in the pure perl backend (analysed and testcase
246 by Bas Denissen).
247 - give same examples in SYNOPSIS sections of AnyEvent and AE.
248 - provider faster implementations for AE::time/now/now_update for
249 EV and Perl backends.
250 - provide faster AE::cv and AE::time implementations for all
251 backends.
252 - no longer support lower/mixed-case signal names (this was
253 never documented nor universally supported).
254 - some more memory µ-optimisations, and somewhat less messy
255 code to cope with both AnyEvent and AE APIs.
256
2575.251 Sat Mar 13 00:58:21 CET 2010
258 - make test failed if it was run before make install on perl <=
259 5.8.8, and older versions of perl make our live very difficult,
260 so write constants.pl during Makefile.PL time.
261
2625.25 Sat Mar 13 00:23:14 CET 2010
263 - fix a race condition in AnyEvent::Handle that would cause
264 a "bio_read: unsupported method" error (in _another_
265 ssl connection) after user code threw an exception.
266 - added AnyEvent::Handle->destroyed method.
267 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
268 - the AnyEvent::Util::fh_nonblocking function worked only
269 by lucky accident on win32.
270 - smaller and faster AnyEvent::Util::fh_nonblocking.
271 - when the (required!) Time::HiRes module is missing, AnyEvent
272 did not fall back to built-in time properly.
273 - do not load Fcntl at runtime, saving memory and loading time.
274 - precompile a number of constants and use them instead of runtime
275 detection and eval.
276 - free detection code after detection and similar memory optimisations.
277 - Perl backend timer interval best effort drift has been improved
278 (same algorithm as EV).
279 - update unicode idna mapping table.
280
2815.24 Tue Jan 5 11:39:43 CET 2010
282 - cygwin never reports errors from failed connects. 1.5 just gives
283 you continous readyness and EAGAIN, 1.7 is even more broken and
284 just hangs. work around both issues in a cygwin-specific hack.
285 - improve idle watcher documentation slightly.
286
2875.23 Sun Dec 20 23:48:00 CET 2009
288 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
289 - implement punycode_encode/decode, idn_nameprep,
290 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
291 - provide $AE::VERSION.
292 - removed traces of "no strict 'refs'".
293
2945.22 Sat Dec 5 03:51:13 CET 2009
295 - downgrade-or-fail in AnyEvent::Handle::push_write, to
296 diagnose encoding failures earlier and more succinctly.
297 (this works around bugs in perl, throwing away encoding info
298 when passing scalar data to extensions).
299 - add more examples to AnyEvent::Socket manpage.
300 - upgrade internal warning set to the same as common::sense 2.03.
301 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
302 (John Beppu).
303 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
304 - add google public dns servers to fallback server set and make sure
305 we load-balance properly between the three sets. also add all
306 fallback dns servers, not just a random one, to each dns config.
307
3085.21 Thu Nov 19 02:48:47 CET 2009
309 - fix a problem where socket constants were called with parameters
310 (spotted by David Friedland).
311 - fork_call never use'd POSIX (reported by Daisuke Maki).
312 - improve perl 5.6 compatibility further (but it still won't work
313 unless you rip out everything but the core).
314 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
315 uses 10MB of RAM, but doesn't open a console window. *sigh*.
316
3175.202 Wed Oct 14 22:35:44 CEST 2009
318 - AnyEvent::DNS would unexpectedly clobber $_ under windows
319 (analysed by Matthias Waldorf).
320 - AnyEvent::Handle::run_cmd can now pass the PID of the
321 newly-created process, which is much less useful than it might
322 sound (based on patch by Yann Kerherve).
323
3245.201 Tue Sep 29 12:09:25 CEST 2009
325 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
326 (reported by Torsten Foertsch).
327 - common::sense 2.0 could cause tcp_server to throw an exception
328 (analysed by elmex).
329
3305.2 Mon Sep 14 07:04:49 CEST 2009
331 - INCOMPATIBLE CHANGE: do no longer support register_read_type
332 and register_write_type in AnyEvent::Handle, instead support
333 package names (the facility was mostly abused).
334 - implement "packagename-as-read/write type" support in
335 AnyEvent::Handle.
336 - AnyEvent::Handle: new options "keepalive" and "oobinline".
337 - oobinline set by default to avoid security issues.
338 - the pure-perl event loop backend wrongly detected times() fallback
339 support (spotted by Pavel Boldin).
340
3415.12 Tue Sep 1 20:26:50 CEST 2009
342 - be more lenient when parsing resolv.conf files, as some people
343 use hashmarks as comment indicator inside directives (reported
344 by Michael S. Fischer).
345 - use same set of warnings as common::sense 2.0.
346 - fix a potential 32 bit overflow issue due to perl having problems
347 with large hex constants in the Perl backend.
348 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
349 by default.
350 - fix some prototypes.
351 - (optionally) require Async::Interrupt 1.02.
352 - replace opendns fallbacks by verizon/level3 public dns servers.
353
3545.112 Fri Aug 21 13:59:04 CEST 2009
355 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
356 using starttls with an illegal/uncreatable tls context.
357 - port to broken win32 perls.
358
3595.111 Sun Aug 16 18:44:36 CEST 2009
360 - perl errornously requires sockaddr_un structures returned by the
361 kernel to have a specific length. work around this bug by padding
362 sockaddr_un structures (found to fail on OS X, as perl only uses
363 the correct code for GNU/Linux, but likely to fail on other
364 OSes as well).
365
3665.11 Wed Aug 12 17:49:37 CEST 2009
367 - fixed a bug in run_cmd with input from filename.
368 - create a json coder object when none is given in write direction
369 to reduce dependency on JSON.
370
3715.1 Tue Aug 11 03:17:32 CEST 2009
372 - tcp_connect now reports the correct error in $!, not an unrelated
373 one (was broken since 4.91).
374 - AE::Handle did not properly clear rtimeout/wtimeout watchers
375 when retrying a tcp_connect.
376 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
377 - fixed wrong documentation in AE::Socket.
378
3795.01 Mon Aug 10 03:16:32 CEST 2009
380 - last release broke EV child watchers :/.
381
3825.0 Sun Aug 9 17:32:09 CEST 2009
383 - officially introduce and document the AE API.
384 - lowering the timeout at runtime did not have immediate effect
385 in AnyEvent::Handle.
386 - AE::Handle now has separate and independent read and write timeout
387 settings.
388 - probe for Guard module when AnyEvent::Util is loaded, not at
389 runtime (Event doesn't like eval "use" from callbacks, and it avoids
390 exporting the probe function).
391 - Impl::IOAsync's io watchers did not use the filehandle emulation
392 code to work around it's watcher limits.
393 - work around IO::Async's problems with overloaded objects.
394 - try to disable wearnings inside IO::Async, it's just too buggy.
395 - do no longer use Event by default, only EV or Perl, Event is too buggy.
396 - hackishly provide fast AE:: interface when EV or Perl is the backend.
397 - inherit VERSION from AnyEvent::VERSION in most modules.
398
3994.91 Thu Aug 6 15:42:45 CEST 2009
400 - AE::Handle::starttls could get out of sync when the read buffer
401 already contains some TLS handshake.
402 - AE::Handle did not properly free the TLS session, causing bigger
403 memory leaks in Net::SSLeay than expected (reported by toaster).
404 - AE::Socket::tcp_connect will now artificially delay invoking
405 the callback to avoid returning after invoking the callback.
406 - convert many internal watcher uses to AE API.
407
4084.9 Sat Aug 1 11:07:01 CEST 2009
409 - Glib silently fails when registering a timeout with
410 a negative value, so avoid that.
411 - call condvar callback immediately when it is set
412 after the condition is already signalled.
413 - check rbuf_max condition only after trying to consume
414 data.
415 - normalise signal numbers to names when using the ae's
416 signal handling, but do not document this (yet).
417 - pure perl signal emulation did not properly set nonblocking
418 mode on the signal pipe on win32 (this is unlikely to have caused
419 any issues).
420 - new module: AnyEvent::Debug.
421 - AnyEvent::Strict now checks that a signal was specified
422 by name (not name or number).
423 - reduce memory footprint in typical cases by ~50kb by
424 only compiling signal code when necessary.
425 - add AnyEvent::Handle->rbuf_max.
426 - grab the AE namespace for future new API, implement stubs for
427 said future/alternative API.
428 - new function: AnyEvent::Socket::format_hostport.
429
4304.881 Tue Jul 28 12:51:53 CEST 2009
431 - work around a bug in local in pre-5.10 perls, causing
432 AnyEvent::Handle to recurse when it should not
433 (analysed by elmex).
434
4354.88 Tue Jul 28 04:04:37 CEST 2009
436 - re-bless the handle into a dummy package after calling
437 AnyEvent::Handle::destroy, so the user does not need to check
438 for errors after every push_write etc.
439 - do not attempt to run t/02_signals.t on obviously broken
440 platforms.
441
4424.87 Sun Jul 26 02:06:16 CEST 2009
443 - do not attempt to linger when there is no longer a valid fh.
444 - enforce tls mode to be either accept or connect, do not simply
445 segfault in Net::SSLeay.
446 - AnyEvent::Handle can now call tcp_connect itself (new parameters
447 connect, on_prepare, on_connect and on_connect_error). Updated
448 tutorial accordingly.
449 - add AnyEvent::Impl::Irssi backend.
450
4514.86 Mon Jul 20 23:52:29 CEST 2009
452 - since the verbose warning is not enough, explicitly document
453 that versions before 1.33 of Net::SSLeay are not secure.
454 - work around signal handling races in Event and (...) Event::Lib.
455 - try to align signal-race timer to full-second boundaries.
456 - work around Tk not liking negative timeouts.
457 - don't complain of different grades of environmental unfriendlyness
458 in IO::Async.
459
4604.85 Sat Jul 18 06:16:14 CEST 2009
461 - nail the signal race problem in perl once and for all
462 (see $AnyEvent::MAX_SIGNAL_LATENCY).
463 - take advantage of Async::Interrupt if it is available.
464 - load Time::HiRes and Guard modules on demand only.
465 - add optional/recommended modules section to AnyEvent
466 documentation.
467 - reduce memory usage considerably (and reduce startup penalty)
468 by not using "strict", "warnings" and "overload" modules.
469 - work around buggy windows/openbsd perls and provide EBADMSG
470 and EPROTO ourselves when missing.
471 - improve perl 5.6 compatibility of the core event loop.
472 - made Net::SSLeay version 1.33 a soft requirement.
473
4744.83 Fri Jul 17 16:56:26 CEST 2009
475 - implement AnyEvent::Socket::getprotobyname.
476 - AnyEvent::CondVar's will now detect recursive blocking
477 waits and will croak, as too many people fall into
478 this trap.
479 - AnyEvent::Handle will now call ->destroy on itself after
480 executing the on_error callback, instead of doing some
481 half-baked internal shutdown, for fatal errors.
482 - clarify on_eof behaviour w.r.t. the read queue and
483 on_read callbacks.
484 - ignore some possible spurious wake-ups in tcp_connect.
485
4864.82 Sat Jul 11 00:34:55 CEST 2009
487 - POE and Event backends didn't accept some callable objects as
488 callbacks.
489 - use Config module instead of POSIX module to detect signal names
490 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
491 module doesn't even have all POSIX signals :/.
492 - use more workarounds around the many refcnt/corruption bugs in
493 Event::Lib.
494 - work around a race condition in perl's select, causing t/03_child.t
495 to rarely fail.
496
4974.81 Thu Jul 9 10:30:30 CEST 2009
498 - AnyEvent::Handle didn't properly diagnose write errors
499 (it expected -1 from syswrite, how lame... :).
500 - support file descriptors in addition to file handles
501 in AnyEvent->io.
502 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
503 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
504 and PERL_ANYEVENT_CA_PATH.
505 - provide a sensible synopsis section for AnyEvent::TLS.
506 - add a "supported backends" section to the manpage.
507 - added simple io watcher test to testsuite, using a
508 portable_socketpair.
509 - tried to improve the stability of the Event::Lib backend,
510 YMMV.
511
5124.8 Mon Jul 6 23:45:16 CEST 2009
513 - AnyEvent::DNS did not properly follow CNAME records with
514 uppercase targets.
515 - AnyEvent::DNS would errornously return AAAA records
516 with v4 mapped addresses (a faulty record) as ipv4 addresses,
517 causing AnyEvent::Socket to throw an exception.
518 - added new module AnyEvent::TLS for easier SSL/TLS context
519 creation, with many options including hostname verification,
520 secure default configuration, lots of documentation and,
521 predefined diffie-hellman keys for perfect forward security
522 and much more. get it while it's still fresh!
523 - use AnyEvent::TLS in AnyEvent::Handle for context management.
524 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
525 so AnyEvent::Socket users have smaller memory footprint
526 in the common case.
527 - add AnyEvent::Handle->push_shutdown method.
528 - add an additional error message parameter to AnyEvent::Handle's
529 on_error callback (for TLS, $! is still available).
530 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
531 - make AnyEvent::Handle more robust against early conenction
532 failures (during new), and return C<undef> in that case
533 from the constructor.
534 - AnyEvent::Handle will now try to load only JSON::XS first,
535 then fall back to JSON.
536 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
537 and because it was documented that way.
538
5394.451 Fri Jul 3 00:28:58 CEST 2009
540 - do not clear rbuf when shutting down an AnyEvent::Handle
541 object - doing so breaks AnyEvent::HTTP.
542
5434.45 Mon Jun 29 22:59:26 CEST 2009
544 - a write error could cause AnyEvent::Handle to create
545 an I/O watcher with an undefined $fh.
546 - special-case mapped ipv4 addresses in both
547 AnyEvent::Socket::format_address and parse_address,
548 to treat them just like ipv4 addresses.
549 - updated and overhauled the AnyEvent::Intro doc.
550 - implement AnyEvent::Socket::format_ipv4/ipv6.
551 - slightly speed up portable_pipe/socketpair.
552 - expand condvar begin/end documentation.
553
5544.42 Fri Jun 26 08:32:18 CEST 2009
555 - preliminary and neccesarily incomplete support for IO::Async.
556 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
557 case it was set to IGNORE, to ensure we can catch child statuses
558 even when the calling env acts stupidly.
559 - updated benchmarks with IO::Async, which performs very well.
560
5614.412 Wed Jun 24 01:35:57 CEST 2009
562 - support an "untaint" attribute for AnyEvent::DNS
563 and set it on the default resolver.
564 - implement and document AnyEvent::Impl::Perl::loop.
565 - remove all anyevent-env variables from %ENV
566 when running in tainted mode.
567 - mention and extend the IO::Lambda benchmark.
568
5694.411 Sun Jun 7 18:48:02 CEST 2009
570 - do not try to use F_SETFD on windows.
571
5724.41 Thu May 14 06:40:11 CEST 2009
573 - work around issues in older perls (5.8.0?) when
574 a signal handler is deleted from the %SIG hash.
575 - use POSIX::_exit in child test, to avoid running
576 destructors.
577 - speed up CHLD handling by relying on SIGCHLD being
578 synchronously delivered, even when we roll our own
579 implementation.
580 - AnyEvent::DNS: add the "dname" resource record name
581 for cosmetic reasons.
582
5834.4 Sun Apr 26 20:12:33 CEST 2009
584 - implemented idle watchers, where applicable.
585 - AnyEvent->time died when Event backend was in use.
586 - fix a memleak in the Tk backend.
587 - sped up Tk timer handling.
588 - clip negative "after" values to 0 in AnyEvent::Impl::Event
589 to avoid spamming out warning messages.
590 - fix Qt timers without interval.
591 - avoid Qt zero-timeout special case, in old libqt's.
592
5934.352 Mon Apr 20 16:31:11 CEST 2009
594 - fix AnyEvent::Strict error messages for child watchers.
595 - fix/update Makefile.PL "recommends" versions.
596 - add AnyEvent->now_update.
597
5984.351 Sat Apr 11 07:56:14 CEST 2009
599 - actually make the signal pipe work under win32.
600 - localise $! in signal handler to avoid changing $!, although
601 perl itself does not seem to save/restore errno either.
602 - set the cloexec flag on the signal pipe (normally set by perl too).
603
6044.35 Fri Mar 27 11:48:20 CET 2009
605 - event models relying on AnyEvent's signal watcher emulation did
606 invoke the callback asynchronously, contradicting documentation
607 and causing signals to get lost (this includes AnyEvent's own
608 event loop). AnyEvent now uses the standard pipe trick to make
609 callback execution synchronous to the event loop.
610 - AnyEvent::Handle didn't free TLS context data on DESTROY
611 (patch by Pavel Shaydo).
612 - work around the ever-incompatibly-changing API of MakeMaker.
613 - document that changing global variables without restoring them
614 is a bad idea in Perl (noted by Adam Rosenstein).
615 - AnyEvent::Strict now barfs if ->io is passed a file.
616
6174.34 Thu Feb 12 18:32:45 CET 2009
618 - separately buffer TLS read data, as otherwise the read queue could
619 deadlock as receiving data is not expected while draining the
620 read queue (which cna only happen with TLS).
621 - raise EBADMSG error on JSON decoding errors.
622 - fix some minor manpage bugs (reported by Maximilian Gaß).
623 - speed up select bitmask parsing quite a bit in the pure perl
624 backend.
625 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
626
6274.331 Tue Jan 6 21:07:25 CET 2009
628 - socketpair fails on many vista machines because vista has
629 completely broken accept/getpeername and getsockname functions,
630 so we provide our own socketpair emulation that kind of works
631 (AnyEvent::Util::portable_pipe).
632 - new function: AnyEvent::Util::portable_socketpair.
633 - take advantage of the Guard module if it exists.
634
6354.33 Fri Nov 21 02:35:40 CET 2008
636 - AnyEvent::Strict did errornously flag a fileno of 0 as
637 illegal.
638 - reduce memory usage and slightly speed up the pure perl backend
639 by only storing the file descriptor, not the file handle.
640 - add missing autocork method to AnyEvent::Handle (reported
641 by Adam Rosenstein).
642 - AnyEvent::DNS->resolve errornously documented an $rcode result
643 argument, but there isn't (spotted by Henrik Krohns).
644 - the naked truth about Tk - it's basically unmaintained.
645
6464.32 Mon Nov 3 22:46:32 CET 2008
647 - fix AnyEvent::Socket::resolve_sockaddr to properly support
648 unix sockets again and choose a proper default for
649 the socket type.
650 - fix call to resolve_sockaddr in tcp_connect.
651
6524.31 Thu Oct 30 04:41:48 CET 2008
653 - implemented AnyEvent::Handle->destroy method.
654 - hint about unexpected effects in TLS mode.
655 - speed up AnyEvent::Impl::Perl by using more arrays
656 instead of hashes without the slightest loss of readability :)
657 - work around a perl argument refcounting bug.
658 - with some perl patches applied, AnyEvent now seems leak-free.
659
6604.3 Fri Oct 3 09:18:43 CEST 2008
661 - AnyEvent will now install a no-op signal handler for SIGPIPE,
662 unless one has been installed already.
663 - warn about not loaidng AnyEvent::Impl::POE early enough
664 (patch by Adam Rosenstein).
665 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
666 - fix a bug in where in SSL connect mode, the
667 client would first wait for some data by the server and
668 otherwise hang. (reported and analysed in an absolutely
669 exemplary manner by Adam Rosenstein).
670 - fix a bug in where SSL EOF would not be treated
671 as stream EOF, putting the connection into a hung state.
672 (reported and analysed in an absolutely exemplary manner by Adam
673 Rosenstein).
674 - fix a potential segfault when the TLS context
675 would go missing in a read callback (e.g. due to stoptls) -
676 Net::SSLeay of course makes no type checking whatsoever.
677 - AnyEvent::Handle will not stop the read watcher in TLS mode.
678 - AnyEvent::Handle->stoptls will send a TLS close notify
679 instead of simply destroying the stream now.
680 - fix a bug where large blocks of data written in TLS mode
681 would not be sent unless triggered by receives.
682 - on_drain will now take the tls write buffer into account.
683 - SSL operations have been streamlined a bit and should be faster
684 now, more to come, though.
685 - added a FAQ section, will document nontrivial issues in there.
686
6874.234 Mon Sep 29 04:08:13 CEST 2008
688 - fix child watcher documentation: only child exits will be reported,
689 no trace events.
690 - mention SIGPIPE in AnyEvent::Handle.
691 - perl backend will now use POSIX::times as monotonic clock
692 when available and the monotonic clock option is not.
693 - run even without Time::HiRes in some cases now, but do not
694 provide sub-second accuracy (all *supported* perl versions have
695 Time::HiRes).
696
6974.233 Fri Aug 22 01:48:05 CEST 2008
698 - fix a bug in the testsuite, causing a hang.
699 - clarified lots of AnyEvent::Handle constructor arguments.
700
7014.232 Thu Aug 21 20:44:25 CEST 2008
702 - fix a bug in AnyEvent::Handle that could cause two fatal error
703 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
704 (testcase by Leon Brocard).
705 - clarified on_eof/on_error documentation a bit.
706
7074.231 Tue Jul 29 13:12:15 CEST 2008
708 - remove some debugging code left in AnyEvent::Util::fork_call
709 (and no, it's impossible to implement with the broken windows
710 perls without resource leaks or worse).
711
7124.23 Tue Jul 29 12:19:59 CEST 2008
713 - document the first parameter passed to condvar callbacks to be
714 the callback.
715 - add AnyEvent::Socket::{ntoa,aton} aliases.
716 - optimize the AE::Handle->push_read (line) for the default
717 eol marker.
718 - optimize push_read (packstring|storable) for small packets.
719 - invoke on_error callback when no on_eof callback is set.
720 - fix a bug in push_read (storable) of unknown impact.
721
7224.22 Sun Jul 20 16:34:13 CEST 2008
723 - new function AnyEvent::Socket::parse_hostport.
724 - as the bulkheads at microsoft can't even get getprotobyname reliably
725 working on their shitty fucking broken os we need to hardcode
726 some common protocol numbers in AnyEvent::Socket. How can
727 people even bother with such a pile of shit as windows.
728
7294.21 Thu Jul 17 14:40:05 CEST 2008
730 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
731 which is incompatible to previous invocation syntax.
732 - work around more windows perl bugs in fork_call. I cannot imagine
733 why anybody would bother implementing fork in such an obviously
734 unusable way.
735 - avoid calling $do_retry if the request has already finished in
736 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
737 K. Sharifulin).
738
7394.2 Sat Jul 12 22:42:11 CEST 2008
740 - support an "interval" argument for repeating timers.
741 - fix a bug in the pure perl event loop that caused timers to
742 be delayed under some circumstances.
743 - implement optional argument checking via PERL_ANYEVENT_STRICT.
744 - not importing anything from Event causes it to malfunction
745 and not use Time::HiRes, work around this bug.
746 - more bug workarounds against the endlessly broken Tk module,
747 seems to work now on 32 bit machines, but less so on 64 bit.
748 - minor optimisations applied to most backends.
749
7504.161 Fri Jul 4 14:24:48 CEST 2008
751 - fixed the t/handle/02_write.t, it read after write, with the
752 assumption that all data could be written before read
753 (reported by Jim Razmus).
754
7554.160 Thu Jul 3 04:02:21 CEST 2008
756 - re-registering signal watchers after unregistering one did
757 not work if a backend falls back on AnyEvent's default
758 implementation (also affected child watchers).
759 - new AnyEvent::Handle options: autocork and no_delay.
760
7614.152 Sun Jun 22 14:15:44 CEST 2008
762 - allow for 32-bit perls that implement shifts differently
763 on different architectures in parse_ipv4 (reported and
764 analysed by Keiichi DAIBA).
765
7664.151 Fri Jun 6 17:34:24 CEST 2008
767 - make sure specifying _only_ on_read and never pushing reads
768 works.
769
7704.15 Fri Jun 6 13:00:46 CEST 2008
771 - the pure perl backend would keep some watchers alive when more than
772 one watcher was registered for the same fd.
773 - new "packstring" and "storable" read and write types
774 for AnyEvent::Handle.
775 - allow on_eof handler to be called after on_error with EPIPE returns.
776 - do not immediately call on_read callback in handle constructor.
777
7784.14 Thu Jun 5 20:29:31 CEST 2008
779 - Fixed a bug in DNS SRV priority sorting.
780 - AnyEvent::Util::guard now reports runtime errors while
781 executing the guard block as warnings.
782 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
783
7844.13 Thu Jun 5 00:47:59 CEST 2008
785 - AnyEvent::DNS only followed cname chains with length 2,
786 contrary to documentation. bumped it up to 10, thanks to
787 microsoft, the current broken-dns-config-king.
788 - AnyEvent::DNS didn't check for socket return status
789 but instead relied on perl not creating filehandles in that
790 case - too bad it gives you a bogus file handle (reported
791 and analysed by Vladimir Timofeev).
792 - fix queue management logic in AnyEvent::Handle:
793 when on_read was registered, the queue was empty
794 and no progress could be made AnyEvent::Handle would
795 enter an endless loop.
796 - correctly start reading again when the handle became
797 busy again after idling.
798 - correctly treat tls shut-downs as EOF condition
799 (for the time being, would be nice to have a callback for that).
800 - correctly call eof callback under all conditions
801 (hopefully).
802 - the timeout callback did not expect that $self can go
803 away any time.
804
8054.12 Tue Jun 3 10:58:04 CEST 2008
806 - include AnyEvent::Intro, a tutorial for anyevent,
807 anyevent::socket and anyevent::handle.
808 - allow more options in on_error.
809
8104.11 Fri May 30 23:42:25 CEST 2008
811 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
812 reverse_lookup and reverse_verify to replace it, support
813 v4mapped and v4compat addresses.
814 - provide more documentation for the resolver class.
815 - really replace longest run of :0: by :: in format_address,
816 also properly convert :: and ::1 again.
817 - support NAPTR record name and decode it.
818 - implement random weight sampling for SRV records, as per
819 rfc 2782.
820 - correctly abort on srv-record targets of ".".
821 - added AnyEvent::DNS::wait_for_slot.
822 - in the unlikely event of a virtual circuit connection
823 being invalidated by a delayed udp reply, AnyEvent::DNS
824 could die. this has been fixed.
825 - plug a probable memleak in the DNS vc code.
826 - use configured timeout also for VC request phase.
827 - implement timeout and max_outstanding methods
828 for resolver class.
829 - update version numbers in all modules.
830
8314.1 Thu May 29 05:45:40 CEST 2008
832 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
833 which mirrors their purpose better (old names still available
834 for a while). Also moved AnyEvent::DNS::addr to
835 AnyEvent::Socket::resolve_sockaddr.
836 - implement AnyEvent->time and AnyEvent->now.
837 - fix IPv6 support in pack_sockaddr.
838 - officially un-experimentalise all newly introduced code.
839 - support unix domain sockets everywhere by specifying
840 a host of "unix/" and the pathname as service.
841 - implement an activity timeout in AnyEvent::Handle.
842 - added a regex read type to AnyEvent::Handle.
843 - added a json read/write type to AnyEvent::Handle.
844 - always croak in on_error in AnyEvent::Handle.
845 - document how to subclass AnyEvent::Handle.
846 - implement AnyEvent::Util::fork_call.
847 - add support for IPv6 nameservers and nameserver statements.
848 - work around _yet_ _another_ windows perl bug in where empty
849 select masks cause errors in the pure perl backend (this
850 workaround cannot be implemented for the other event loops,
851 but EV also works around this).
852 - supply AnyEvent::Socket::address_type and make good use of it.
853 - clarify and fix the Handle documentation w.r.t $self vs. $handle
854 vs. callback arguments.
855 - add some recommends to the META.yml.
856
8574.05 Mon May 26 19:44:06 CEST 2008
858 - some platforms ignore the rfc and prepend an extra sa_len member to
859 the sockaddr structure, cater for those. also use sockaddr_family,
860 didn't know it was there, but comes in handy for the workaround.
861 - undo the import WIN32 hack.
862
8634.04 Mon May 26 08:03:31 CEST 2008
4 - try to work around yet another windows bug: failed connects are reported 864 - try to work around yet another windows bug: failed connects are reported
5 as if it were out-of-band data. windows users: you suck. 865 as if it were out-of-band data. windows users: you suck. Only
866 EV, Glib, Event and the pure perl backend can handle this condition.
867 - optimize write algorithm in AnyEvent::Handle.
6 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly. 868 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
869 - tcp_connect now properly iterates through all targets.
7 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6 870 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
8 support entirely. also hardcode some AF_INEt6 constants for the 871 support entirely. also hardcode some AF_INET6 constants for the
9 sake of perl 5.8. 872 sake of perl 5.8.
10 - vastly improved nameserver/suffix detection algorithm for windows 873 - vastly improved nameserver/suffix detection algorithm for windows
11 (still a hack...) 874 (still a hack...)
12 - try to use a fallback dns server if no dns servers could be found. 875 - try to use a fallback dns server if no dns servers could be found.
13 - splatter around some local $SIG{__DIE__}'s to work around bad 876 - splatter around some local $SIG{__DIE__}'s to work around bad
… …
40 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. 903 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
41 - try very hard not to freeze on broken windows perls. 904 - try very hard not to freeze on broken windows perls.
42 - make on_eof optional in AnyEvent::Handle. 905 - make on_eof optional in AnyEvent::Handle.
43 - support service names in tcp_connect and tcp_server. 906 - support service names in tcp_connect and tcp_server.
44 - work around netbsd/openbsd bug causing unix sockets not to 907 - work around netbsd/openbsd bug causing unix sockets not to
45 become writable untill empty. 908 become writable until empty.
46 - only make a dummy read for error detection when getpeername 909 - only make a dummy read for error detection when getpeername
47 returns ENOTCONN. 910 returns ENOTCONN.
48 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 911 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
49 override. 912 override.
50 - new AnyEvent::Guard::cancel. 913 - new AnyEvent::Guard::cancel.
51 914
523.5 Sat May 17 23:17:14 CEST 2008 9153.5 Sat May 17 23:17:14 CEST 2008
53 - gracefully deal with systems not defining CLOCK_MONOTONIC 916 - gracefully deal with systems not defining CLOCK_MONOTONIC
… …
67 930
683.41 Sun May 11 19:53:13 CEST 2008 9313.41 Sun May 11 19:53:13 CEST 2008
69 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 932 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
70 933
713.4 Sun May 11 00:29:25 CEST 2008 9343.4 Sun May 11 00:29:25 CEST 2008
72 - complete re-implementation of condvars with a hopefully much mroe useful 935 - complete re-implementation of condvars with a hopefully much more useful
73 API (backward compatible functions exist but are undocumented). 936 API (backward compatible functions exist but are undocumented).
74 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 937 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
75 (still subject to change). 938 (still subject to change).
76 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 939 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
77 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 940 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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