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

Comparing AnyEvent/Changes (file contents):
Revision 1.152 by root, Thu Jun 5 18:30:08 2008 UTC vs.
Revision 1.536 by root, Fri Mar 30 02:26:07 2012 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2
3TODO: document TCP_*** constants
4TODO: add debug/trace logging to some modules
5TODO: AnyEvent::IO
6tODO: maxim souffle handle storable
7TODO: document import tags aeio, also ae_ naming
8TODO: t/io
9lib/AnyEvent/DNS.pm: open my $fh, "<", $resolv_conf
10lib/AnyEvent/DNS.pm: if (open my $fh, "ipconfig /all |") {
11lib/AnyEvent/Log.pm: open my $fh, ">>", $path
12lib/AnyEvent/Log.pm: open my $fh, ">>", $path
13lib/AnyEvent/TLS.pm: or croak "$dh_file: failed to open DH parameter file: $!";
14
156.2
16 - documented Rocco fabricating statements about AnyEvent and me.
17 - new module AnyEvent::IO, that is a frontend to either a pure-perl
18 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
19 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
20 - load /etc/hosts only when dns fails has no answer.
21 - stat /etc/hosts on every access and reload it if it changed.
22 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
23 - the default logging level was not properly documented in a variety of
24 places, this has hopefully been rectified.
25 - updated uts46data.pl for unicode 6.1.0.
26 - made log messages genearted by AnyEvent submodules not
27 include the package name anymore, as it will be logged
28 by default already.
29 - upgrade to the trick used by common-sense 3.5 to work
30 around extra warning torture/breakage under perl 5.15.x.
31 - log messages by anyevent are now ucfirst, are usually full
32 sentences and do no longer include the package name.
33
346.14 Tue Jan 31 20:00:24 CET 2012
35 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
36 probably an editing mistake (reported by Darin McBride).
37
386.13 Thu Jan 12 07:27:01 CET 2012
39 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
40 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
41 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
42 regex-matching inside the callback would change the parameter, despite
43 $1 being dynamically scoped per-block. this perl bug is now being
44 worked around (testcase by Cindy Wang).
45
466.12 Mon Dec 12 13:21:10 CET 2011
47 - $! was clobbered by subsequent calls in tcp_connect, due to
48 the postpone not saving/restoring it (analyzed by Richard Garnish).
49
506.11 Tue Nov 22 10:36:05 CET 2011
51 - Tk cannot create windows when tainted unless you set a title, so
52 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
53 - escape any nonprintable/nonascii characters when stringifying
54 backtraces.
55 - log the reason loading an interface module fails to load at level debug.
56
576.1 Tue Oct 4 19:44:30 CEST 2011
58 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
59 and above), and some messages inside AnyEvent have been elevated
60 to higher log levels to print by default.
61 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
62 - work around yet another signal race bug, newly introduced by POE. My
63 how that module sucks.
64 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
65 support ipv6-only operation even with defaults (also use ::1 instead
66 of 127.0.0.1).
67 - give better guidelines on how to chose a logging level.
68 - AnyEvent::Log can now cap (limit) log message priority by context.
69 - AE::log will now save and restore $! and $@, so callers don't have to.
70 - AE::Strict now checks that registered fds are still valid at regular
71 intervals.
72 - skip frequently called 'require' calls if the module appears loaded.
73 - increase signal latency test timer considerably for poor little osx,
74 which can't be bothered with basic posix functionality such as signal
75 delivery.
76 - mention some debugging aids in the tutorial.
77 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
78 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
79 of the context as title.
80 - AnyEvent::Strict would not restore read-write access to $_ when a callback
81 died - this is a user bug, but since AE::Strict exists to help dbeugging...
82 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
83 to work with it - the author is well aware of the issues but refuses
84 to fix them. This extreme step has been taken because the module actively
85 attacks the goals of AnyEvent and makes it harder for module authors to use
86 AnyEvent.
87
886.02 Fri Aug 26 20:08:31 CEST 2011
89 - log_to_syslog now accepts facility strings.
90 - log_to_syslog would not allow facilities - "$facility|$level"
91 *literally* means that string. sheesh.
92 - the CPAN parser got confused and thought we provide Sys::Syslog, try
93 to work around.
94 - renamed AnyEvent::Impl::FLTK2 to FLTK.
95
966.01 Fri Aug 26 07:04:11 CEST 2011
97 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
98 the TTL as fourth element - this affects ->resolve, ->request
99 and ->dns_unpack indirectly. Convenience functions such as
100 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
101 AnyEvent::Socket.
102 - INCOMPATIBLE CHANGE: backend authors now should not implement
103 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
104 methods.
105 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
106 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
107 - DEPRECIATION: please change
108 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
109 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
110 in your programs, for future compatibility.
111 - *blush* idle watcher emulation was borked.
112 - AnyEvent::Strict now checks the AE::xxx functions as well.
113 - querying too long domain names would barf deep inside AnyEvent::DNS - since
114 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
115 for too long domain names explicitly and passes an undef to the callback
116 instead (reported by James Bromberger).
117 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
118 - new functions: AnyEvent::postpone, AnyEvent::log.
119 - new module: AnyEvent::Log, for simple logging needs.
120 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
121 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
122 a debug shell.
123 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
124 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
125 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
126 /etc/hosts can be found.
127 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
128 configuration.
129 - env variables can now be specified using AE_xyz instead of
130 PERL_ANYEVENT_xyz, subject to some rules.
131 - lots of new AnyEvent::Debug functionality.
132 - AnyEvent::Strict now checks for modifications of $_ by the callback
133 (by setting it to readonly while executing the callback, if possible).
134 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
135 Invest extra overhead to work around this peculiar behaviour.
136 - split the pure perl backend into a loop module and an impl
137 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
138 - some read types (regex, netstring, json, storable) could
139 errornously keep an AE::Handle object alive on their own
140 (reported by Mohammad Toossi).
141 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
142 supports /etc/hosts overrides.
143 - add more workarounds around child watcher bugs in IO::Async
144 and (newly introduced) in POE.
145 - use glib child watchers and try to work around its limitations,
146 if possible (based on a patch by Kevin Ryde).
147 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
148 - many read types in AnyEvent::Handle would malfunction during
149 transitions to TLS because of a "delete $self->{rbuf}".
150 (reported by Gerald Galster).
151 - improve condvar blocking wait performance (probably only noticable
152 with faster backends), also make it easier for event loops that
153 do not support blocking waits.
154 - PERL_ANYEVENT_MODEL now also allows full module names, although
155 the usefulness of this feature is not yet clear.
156 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
157 - add t/80_ssltext.t.
158 - no longer use AUTOLOAD for the initial loading (for no very good reason).
159 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
160 - do not run event loop detection multiple times when programs call their cached
161 copies of AnyEvent::detect.
162 - parse_hostport now accepts absolute paths als unix domain sockets.
163 - suppress the idiotic warning inside POE by patching the POE::Kernel
164 run flags instead of asking the user to jump through contortions.
165 - add Task::Weaken as dependency (as recommends only at this time).
166 - replace PF_UNSPEC by 0, for good old minix.
167
1686.0 (not officially released, but was on CPAN for about 15 minutes :/)
169
1705.34 Thu May 12 09:59:41 CEST 2011
171 - AE::TLS didn't support both cert_file and key at the same time,
172 due to some copy&paste bug (reported and analyzed by vti).
173 - improve AE::Handle timeout documentation slightly.
174
1755.33 Mon Feb 7 21:51:41 CET 2011
176 - never appeared on CPAN, for unknown reasons.
177 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
178 object when it's already gone (analyzed by Ryan Bullock).
179 - update idna algorithm(s) to UTS#46 version 6.0.0.
180 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
181 uts46 says it's an error, and testsuite makes more sense.
182 - remove leading dots in idn_to_ascii - testsuite says to remove,
183 uts46 says it's an error, and testsuite makes more sense.
184 - fixed a bug in idn ignored character processing.
185 - passes the uts46 6.0.0 testsuite except for three cases which
186 apparently are bugs in the testsuite itself.
187 - passes the uts46 6.0.1 testsuite except for eleven cases which
188 apparently are bugs in the testsuite itself.
189
1905.31 Sun Jan 23 11:49:19 CET 2011
191 - work around a POE bug where POE would first poll for new events and
192 invoke handlers BEFORE handling the message queue.
193 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
194 - stop_read now does just that in TLS mode. see its
195 documentation for caveats.
196 - update to common::sense 3.4 warning set (some warnings in perl impose
197 an arbitrary 32kb limit on regex matches, which affects
198 AnyEvent::Handle).
199 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
200 - new experimental AE::Handle::resettls method.
201
2025.3 Fri Dec 31 05:48:13 CET 2010
203 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
204 issues with IO::Async have been solved in that version, including
205 the availabilty of a default loop!
206 - implement new wbuf_max parameter for AnyEvent::Handle.
207 - added Cocoa::EventLoop backend by Daisuke Murase.
208 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
209 - clarify handle on_prepare documentation.
210 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
211 to avoid leaking them to any child processes.
212 - try to force files into the arch-dependend lib directory, for the
213 half-"turly obsessive compulsive person"(s) that depend on it.
214
2155.29 Sun Dec 5 10:49:21 CET 2010
216 - convert EV backend to EV 4.00 API (so better upgrade EV too).
217 - AE::Handle: implement dynamic read_size adjustment, add
218 max_read_size parameter, reduce default read_size to 2048.
219 - add some FAQ entries for servers and/or on_eof vs. on_error.
220 - work around OS bugs (cygwin again) when lingering in AE::Handle.
221 - reduce memory usage slightly when connect was used in AE::Handle.
222 - wrok around more segfaults in Tk.
223 - document tls_ctx => undef as valid.
224 - detect WNOHANG value at compiletime.
225 - include some probably linux-only support for building AnyEvent
226 as part of the perl core.
227 - improve compatibility of t/08_idna.t to perl 5.8.
228 - make AnyEvent::Strict truly optional (in case somebody builds a
229 minimal perl...).
230
2315.28 Wed Oct 13 04:14:23 CEST 2010
232 - due to a glitch, AnyEvent's internal getprotobyname wasn't
233 actually used in tcp_server.
234 - implement AnyEvent::CondVar->new method, which was documented but
235 missing.
236 - support autoloading of read types in unshift_read.
237 - AnyEvent::Strict: do the test for invalid fh arguments differently,
238 resulting in more informative error messages.
239 - new AnyEvent::FAQ manpage.
240 - clarify that you can actually call push_read etc. in on_read.
241 - put keys and certs in /tmp, not into the current directory,
242 and document this in the SECURITY CONSIDERATIONS sections
243 of AnyEvent::TLS.
244 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
245 requested by Ron Isaacson.
246 - document $AnyEvent::DNS::RESOLVER variable.
247 - incorporated major typo patches by Abhijit Menon-Sen.
248 - AnyEvent::Handle now croaks when negative timeouts are passed.
249 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
250 undocumented, but who knows when they might come in handy.
251
2525.271 Tue Jun 8 12:05:46 CEST 2010
253 - backport to perl 5.8.x.
254
2555.27 Sun Jun 6 12:12:05 CEST 2010
256 - postpone differently in AnyEvent::Socket now, as
257 when not, canceling the connection attempt might fail
258 (found by Felix Antonius Wilhelm Ostmann).
259 - explicitly check for non-stream sockets in AE::Handle, too many
260 clueless people fell into the trap of this somehow working.
261 - simplified and reworked the "OTHER MODULES" section.
262 - better/more condvar examples.
263
2645.261 Wed Apr 28 16:13:36 CEST 2010
265 - AF_INET6 was not properly used from Socket6 during configuration
266 time (found by Andy Grundman).
267
2685.26 Mon Apr 12 04:49:35 CEST 2010
269 - don't generate spurious readyness notifications when select
270 returns EINTR in the pure perl backend (analysed and testcase
271 by Bas Denissen).
272 - give same examples in SYNOPSIS sections of AnyEvent and AE.
273 - provider faster implementations for AE::time/now/now_update for
274 EV and Perl backends.
275 - provide faster AE::cv and AE::time implementations for all
276 backends.
277 - no longer support lower/mixed-case signal names (this was
278 never documented nor universally supported).
279 - some more memory µ-optimisations, and somewhat less messy
280 code to cope with both AnyEvent and AE APIs.
281
2825.251 Sat Mar 13 00:58:21 CET 2010
283 - make test failed if it was run before make install on perl <=
284 5.8.8, and older versions of perl make our live very difficult,
285 so write constants.pl during Makefile.PL time.
286
2875.25 Sat Mar 13 00:23:14 CET 2010
288 - fix a race condition in AnyEvent::Handle that would cause
289 a "bio_read: unsupported method" error (in _another_
290 ssl connection) after user code threw an exception.
291 - added AnyEvent::Handle->destroyed method.
292 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
293 - the AnyEvent::Util::fh_nonblocking function worked only
294 by lucky accident on win32.
295 - smaller and faster AnyEvent::Util::fh_nonblocking.
296 - when the (required!) Time::HiRes module is missing, AnyEvent
297 did not fall back to built-in time properly.
298 - do not load Fcntl at runtime, saving memory and loading time.
299 - precompile a number of constants and use them instead of runtime
300 detection and eval.
301 - free detection code after detection and similar memory optimisations.
302 - Perl backend timer interval best effort drift has been improved
303 (same algorithm as EV).
304 - update unicode idna mapping table.
305
3065.24 Tue Jan 5 11:39:43 CET 2010
307 - cygwin never reports errors from failed connects. 1.5 just gives
308 you continous readyness and EAGAIN, 1.7 is even more broken and
309 just hangs. work around both issues in a cygwin-specific hack.
310 - improve idle watcher documentation slightly.
311
3125.23 Sun Dec 20 23:48:00 CET 2009
313 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
314 - implement punycode_encode/decode, idn_nameprep,
315 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
316 - provide $AE::VERSION.
317 - removed traces of "no strict 'refs'".
318
3195.22 Sat Dec 5 03:51:13 CET 2009
320 - downgrade-or-fail in AnyEvent::Handle::push_write, to
321 diagnose encoding failures earlier and more succinctly.
322 (this works around bugs in perl, throwing away encoding info
323 when passing scalar data to extensions).
324 - add more examples to AnyEvent::Socket manpage.
325 - upgrade internal warning set to the same as common::sense 2.03.
326 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
327 (John Beppu).
328 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
329 - add google public dns servers to fallback server set and make sure
330 we load-balance properly between the three sets. also add all
331 fallback dns servers, not just a random one, to each dns config.
332
3335.21 Thu Nov 19 02:48:47 CET 2009
334 - fix a problem where socket constants were called with parameters
335 (spotted by David Friedland).
336 - fork_call never use'd POSIX (reported by Daisuke Maki).
337 - improve perl 5.6 compatibility further (but it still won't work
338 unless you rip out everything but the core).
339 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
340 uses 10MB of RAM, but doesn't open a console window. *sigh*.
341
3425.202 Wed Oct 14 22:35:44 CEST 2009
343 - AnyEvent::DNS would unexpectedly clobber $_ under windows
344 (analysed by Matthias Waldorf).
345 - AnyEvent::Handle::run_cmd can now pass the PID of the
346 newly-created process, which is much less useful than it might
347 sound (based on patch by Yann Kerherve).
348
3495.201 Tue Sep 29 12:09:25 CEST 2009
350 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
351 (reported by Torsten Foertsch).
352 - common::sense 2.0 could cause tcp_server to throw an exception
353 (analysed by elmex).
354
3555.2 Mon Sep 14 07:04:49 CEST 2009
356 - INCOMPATIBLE CHANGE: do no longer support register_read_type
357 and register_write_type in AnyEvent::Handle, instead support
358 package names (the facility was mostly abused).
359 - implement "packagename-as-read/write type" support in
360 AnyEvent::Handle.
361 - AnyEvent::Handle: new options "keepalive" and "oobinline".
362 - oobinline set by default to avoid security issues.
363 - the pure-perl event loop backend wrongly detected times() fallback
364 support (spotted by Pavel Boldin).
365
3665.12 Tue Sep 1 20:26:50 CEST 2009
367 - be more lenient when parsing resolv.conf files, as some people
368 use hashmarks as comment indicator inside directives (reported
369 by Michael S. Fischer).
370 - use same set of warnings as common::sense 2.0.
371 - fix a potential 32 bit overflow issue due to perl having problems
372 with large hex constants in the Perl backend.
373 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
374 by default.
375 - fix some prototypes.
376 - (optionally) require Async::Interrupt 1.02.
377 - replace opendns fallbacks by verizon/level3 public dns servers.
378
3795.112 Fri Aug 21 13:59:04 CEST 2009
380 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
381 using starttls with an illegal/uncreatable tls context.
382 - port to broken win32 perls.
383
3845.111 Sun Aug 16 18:44:36 CEST 2009
385 - perl errornously requires sockaddr_un structures returned by the
386 kernel to have a specific length. work around this bug by padding
387 sockaddr_un structures (found to fail on OS X, as perl only uses
388 the correct code for GNU/Linux, but likely to fail on other
389 OSes as well).
390
3915.11 Wed Aug 12 17:49:37 CEST 2009
392 - fixed a bug in run_cmd with input from filename.
393 - create a json coder object when none is given in write direction
394 to reduce dependency on JSON.
395
3965.1 Tue Aug 11 03:17:32 CEST 2009
397 - tcp_connect now reports the correct error in $!, not an unrelated
398 one (was broken since 4.91).
399 - AE::Handle did not properly clear rtimeout/wtimeout watchers
400 when retrying a tcp_connect.
401 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
402 - fixed wrong documentation in AE::Socket.
403
4045.01 Mon Aug 10 03:16:32 CEST 2009
405 - last release broke EV child watchers :/.
406
4075.0 Sun Aug 9 17:32:09 CEST 2009
408 - officially introduce and document the AE API.
409 - lowering the timeout at runtime did not have immediate effect
410 in AnyEvent::Handle.
411 - AE::Handle now has separate and independent read and write timeout
412 settings.
413 - probe for Guard module when AnyEvent::Util is loaded, not at
414 runtime (Event doesn't like eval "use" from callbacks, and it avoids
415 exporting the probe function).
416 - Impl::IOAsync's io watchers did not use the filehandle emulation
417 code to work around it's watcher limits.
418 - work around IO::Async's problems with overloaded objects.
419 - try to disable wearnings inside IO::Async, it's just too buggy.
420 - do no longer use Event by default, only EV or Perl, Event is too buggy.
421 - hackishly provide fast AE:: interface when EV or Perl is the backend.
422 - inherit VERSION from AnyEvent::VERSION in most modules.
423
4244.91 Thu Aug 6 15:42:45 CEST 2009
425 - AE::Handle::starttls could get out of sync when the read buffer
426 already contains some TLS handshake.
427 - AE::Handle did not properly free the TLS session, causing bigger
428 memory leaks in Net::SSLeay than expected (reported by toaster).
429 - AE::Socket::tcp_connect will now artificially delay invoking
430 the callback to avoid returning after invoking the callback.
431 - convert many internal watcher uses to AE API.
432
4334.9 Sat Aug 1 11:07:01 CEST 2009
434 - Glib silently fails when registering a timeout with
435 a negative value, so avoid that.
436 - call condvar callback immediately when it is set
437 after the condition is already signalled.
438 - check rbuf_max condition only after trying to consume
439 data.
440 - normalise signal numbers to names when using the ae's
441 signal handling, but do not document this (yet).
442 - pure perl signal emulation did not properly set nonblocking
443 mode on the signal pipe on win32 (this is unlikely to have caused
444 any issues).
445 - new module: AnyEvent::Debug.
446 - AnyEvent::Strict now checks that a signal was specified
447 by name (not name or number).
448 - reduce memory footprint in typical cases by ~50kb by
449 only compiling signal code when necessary.
450 - add AnyEvent::Handle->rbuf_max.
451 - grab the AE namespace for future new API, implement stubs for
452 said future/alternative API.
453 - new function: AnyEvent::Socket::format_hostport.
454
4554.881 Tue Jul 28 12:51:53 CEST 2009
456 - work around a bug in local in pre-5.10 perls, causing
457 AnyEvent::Handle to recurse when it should not
458 (analysed by elmex).
459
4604.88 Tue Jul 28 04:04:37 CEST 2009
461 - re-bless the handle into a dummy package after calling
462 AnyEvent::Handle::destroy, so the user does not need to check
463 for errors after every push_write etc.
464 - do not attempt to run t/02_signals.t on obviously broken
465 platforms.
466
4674.87 Sun Jul 26 02:06:16 CEST 2009
468 - do not attempt to linger when there is no longer a valid fh.
469 - enforce tls mode to be either accept or connect, do not simply
470 segfault in Net::SSLeay.
471 - AnyEvent::Handle can now call tcp_connect itself (new parameters
472 connect, on_prepare, on_connect and on_connect_error). Updated
473 tutorial accordingly.
474 - add AnyEvent::Impl::Irssi backend.
475
4764.86 Mon Jul 20 23:52:29 CEST 2009
477 - since the verbose warning is not enough, explicitly document
478 that versions before 1.33 of Net::SSLeay are not secure.
479 - work around signal handling races in Event and (...) Event::Lib.
480 - try to align signal-race timer to full-second boundaries.
481 - work around Tk not liking negative timeouts.
482 - don't complain of different grades of environmental unfriendlyness
483 in IO::Async.
484
4854.85 Sat Jul 18 06:16:14 CEST 2009
486 - nail the signal race problem in perl once and for all
487 (see $AnyEvent::MAX_SIGNAL_LATENCY).
488 - take advantage of Async::Interrupt if it is available.
489 - load Time::HiRes and Guard modules on demand only.
490 - add optional/recommended modules section to AnyEvent
491 documentation.
492 - reduce memory usage considerably (and reduce startup penalty)
493 by not using "strict", "warnings" and "overload" modules.
494 - work around buggy windows/openbsd perls and provide EBADMSG
495 and EPROTO ourselves when missing.
496 - improve perl 5.6 compatibility of the core event loop.
497 - made Net::SSLeay version 1.33 a soft requirement.
498
4994.83 Fri Jul 17 16:56:26 CEST 2009
500 - implement AnyEvent::Socket::getprotobyname.
501 - AnyEvent::CondVar's will now detect recursive blocking
502 waits and will croak, as too many people fall into
503 this trap.
504 - AnyEvent::Handle will now call ->destroy on itself after
505 executing the on_error callback, instead of doing some
506 half-baked internal shutdown, for fatal errors.
507 - clarify on_eof behaviour w.r.t. the read queue and
508 on_read callbacks.
509 - ignore some possible spurious wake-ups in tcp_connect.
510
5114.82 Sat Jul 11 00:34:55 CEST 2009
512 - POE and Event backends didn't accept some callable objects as
513 callbacks.
514 - use Config module instead of POSIX module to detect signal names
515 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
516 module doesn't even have all POSIX signals :/.
517 - use more workarounds around the many refcnt/corruption bugs in
518 Event::Lib.
519 - work around a race condition in perl's select, causing t/03_child.t
520 to rarely fail.
521
5224.81 Thu Jul 9 10:30:30 CEST 2009
523 - AnyEvent::Handle didn't properly diagnose write errors
524 (it expected -1 from syswrite, how lame... :).
525 - support file descriptors in addition to file handles
526 in AnyEvent->io.
527 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
528 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
529 and PERL_ANYEVENT_CA_PATH.
530 - provide a sensible synopsis section for AnyEvent::TLS.
531 - add a "supported backends" section to the manpage.
532 - added simple io watcher test to testsuite, using a
533 portable_socketpair.
534 - tried to improve the stability of the Event::Lib backend,
535 YMMV.
536
5374.8 Mon Jul 6 23:45:16 CEST 2009
538 - AnyEvent::DNS did not properly follow CNAME records with
539 uppercase targets.
540 - AnyEvent::DNS would errornously return AAAA records
541 with v4 mapped addresses (a faulty record) as ipv4 addresses,
542 causing AnyEvent::Socket to throw an exception.
543 - added new module AnyEvent::TLS for easier SSL/TLS context
544 creation, with many options including hostname verification,
545 secure default configuration, lots of documentation and,
546 predefined diffie-hellman keys for perfect forward security
547 and much more. get it while it's still fresh!
548 - use AnyEvent::TLS in AnyEvent::Handle for context management.
549 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
550 so AnyEvent::Socket users have smaller memory footprint
551 in the common case.
552 - add AnyEvent::Handle->push_shutdown method.
553 - add an additional error message parameter to AnyEvent::Handle's
554 on_error callback (for TLS, $! is still available).
555 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
556 - make AnyEvent::Handle more robust against early conenction
557 failures (during new), and return C<undef> in that case
558 from the constructor.
559 - AnyEvent::Handle will now try to load only JSON::XS first,
560 then fall back to JSON.
561 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
562 and because it was documented that way.
563
5644.451 Fri Jul 3 00:28:58 CEST 2009
565 - do not clear rbuf when shutting down an AnyEvent::Handle
566 object - doing so breaks AnyEvent::HTTP.
567
5684.45 Mon Jun 29 22:59:26 CEST 2009
569 - a write error could cause AnyEvent::Handle to create
570 an I/O watcher with an undefined $fh.
571 - special-case mapped ipv4 addresses in both
572 AnyEvent::Socket::format_address and parse_address,
573 to treat them just like ipv4 addresses.
574 - updated and overhauled the AnyEvent::Intro doc.
575 - implement AnyEvent::Socket::format_ipv4/ipv6.
576 - slightly speed up portable_pipe/socketpair.
577 - expand condvar begin/end documentation.
578
5794.42 Fri Jun 26 08:32:18 CEST 2009
580 - preliminary and neccesarily incomplete support for IO::Async.
581 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
582 case it was set to IGNORE, to ensure we can catch child statuses
583 even when the calling env acts stupidly.
584 - updated benchmarks with IO::Async, which performs very well.
585
5864.412 Wed Jun 24 01:35:57 CEST 2009
587 - support an "untaint" attribute for AnyEvent::DNS
588 and set it on the default resolver.
589 - implement and document AnyEvent::Impl::Perl::loop.
590 - remove all anyevent-env variables from %ENV
591 when running in tainted mode.
592 - mention and extend the IO::Lambda benchmark.
593
5944.411 Sun Jun 7 18:48:02 CEST 2009
595 - do not try to use F_SETFD on windows.
596
5974.41 Thu May 14 06:40:11 CEST 2009
598 - work around issues in older perls (5.8.0?) when
599 a signal handler is deleted from the %SIG hash.
600 - use POSIX::_exit in child test, to avoid running
601 destructors.
602 - speed up CHLD handling by relying on SIGCHLD being
603 synchronously delivered, even when we roll our own
604 implementation.
605 - AnyEvent::DNS: add the "dname" resource record name
606 for cosmetic reasons.
607
6084.4 Sun Apr 26 20:12:33 CEST 2009
609 - implemented idle watchers, where applicable.
610 - AnyEvent->time died when Event backend was in use.
611 - fix a memleak in the Tk backend.
612 - sped up Tk timer handling.
613 - clip negative "after" values to 0 in AnyEvent::Impl::Event
614 to avoid spamming out warning messages.
615 - fix Qt timers without interval.
616 - avoid Qt zero-timeout special case, in old libqt's.
617
6184.352 Mon Apr 20 16:31:11 CEST 2009
619 - fix AnyEvent::Strict error messages for child watchers.
620 - fix/update Makefile.PL "recommends" versions.
621 - add AnyEvent->now_update.
622
6234.351 Sat Apr 11 07:56:14 CEST 2009
624 - actually make the signal pipe work under win32.
625 - localise $! in signal handler to avoid changing $!, although
626 perl itself does not seem to save/restore errno either.
627 - set the cloexec flag on the signal pipe (normally set by perl too).
628
6294.35 Fri Mar 27 11:48:20 CET 2009
630 - event models relying on AnyEvent's signal watcher emulation did
631 invoke the callback asynchronously, contradicting documentation
632 and causing signals to get lost (this includes AnyEvent's own
633 event loop). AnyEvent now uses the standard pipe trick to make
634 callback execution synchronous to the event loop.
635 - AnyEvent::Handle didn't free TLS context data on DESTROY
636 (patch by Pavel Shaydo).
637 - work around the ever-incompatibly-changing API of MakeMaker.
638 - document that changing global variables without restoring them
639 is a bad idea in Perl (noted by Adam Rosenstein).
640 - AnyEvent::Strict now barfs if ->io is passed a file.
641
6424.34 Thu Feb 12 18:32:45 CET 2009
643 - separately buffer TLS read data, as otherwise the read queue could
644 deadlock as receiving data is not expected while draining the
645 read queue (which cna only happen with TLS).
646 - raise EBADMSG error on JSON decoding errors.
647 - fix some minor manpage bugs (reported by Maximilian Gaß).
648 - speed up select bitmask parsing quite a bit in the pure perl
649 backend.
650 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
651
6524.331 Tue Jan 6 21:07:25 CET 2009
653 - socketpair fails on many vista machines because vista has
654 completely broken accept/getpeername and getsockname functions,
655 so we provide our own socketpair emulation that kind of works
656 (AnyEvent::Util::portable_pipe).
657 - new function: AnyEvent::Util::portable_socketpair.
658 - take advantage of the Guard module if it exists.
659
6604.33 Fri Nov 21 02:35:40 CET 2008
661 - AnyEvent::Strict did errornously flag a fileno of 0 as
662 illegal.
663 - reduce memory usage and slightly speed up the pure perl backend
664 by only storing the file descriptor, not the file handle.
665 - add missing autocork method to AnyEvent::Handle (reported
666 by Adam Rosenstein).
667 - AnyEvent::DNS->resolve errornously documented an $rcode result
668 argument, but there isn't (spotted by Henrik Krohns).
669 - the naked truth about Tk - it's basically unmaintained.
670
6714.32 Mon Nov 3 22:46:32 CET 2008
672 - fix AnyEvent::Socket::resolve_sockaddr to properly support
673 unix sockets again and choose a proper default for
674 the socket type.
675 - fix call to resolve_sockaddr in tcp_connect.
676
6774.31 Thu Oct 30 04:41:48 CET 2008
678 - implemented AnyEvent::Handle->destroy method.
679 - hint about unexpected effects in TLS mode.
680 - speed up AnyEvent::Impl::Perl by using more arrays
681 instead of hashes without the slightest loss of readability :)
682 - work around a perl argument refcounting bug.
683 - with some perl patches applied, AnyEvent now seems leak-free.
684
6854.3 Fri Oct 3 09:18:43 CEST 2008
686 - AnyEvent will now install a no-op signal handler for SIGPIPE,
687 unless one has been installed already.
688 - warn about not loaidng AnyEvent::Impl::POE early enough
689 (patch by Adam Rosenstein).
690 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
691 - fix a bug in where in SSL connect mode, the
692 client would first wait for some data by the server and
693 otherwise hang. (reported and analysed in an absolutely
694 exemplary manner by Adam Rosenstein).
695 - fix a bug in where SSL EOF would not be treated
696 as stream EOF, putting the connection into a hung state.
697 (reported and analysed in an absolutely exemplary manner by Adam
698 Rosenstein).
699 - fix a potential segfault when the TLS context
700 would go missing in a read callback (e.g. due to stoptls) -
701 Net::SSLeay of course makes no type checking whatsoever.
702 - AnyEvent::Handle will not stop the read watcher in TLS mode.
703 - AnyEvent::Handle->stoptls will send a TLS close notify
704 instead of simply destroying the stream now.
705 - fix a bug where large blocks of data written in TLS mode
706 would not be sent unless triggered by receives.
707 - on_drain will now take the tls write buffer into account.
708 - SSL operations have been streamlined a bit and should be faster
709 now, more to come, though.
710 - added a FAQ section, will document nontrivial issues in there.
711
7124.234 Mon Sep 29 04:08:13 CEST 2008
713 - fix child watcher documentation: only child exits will be reported,
714 no trace events.
715 - mention SIGPIPE in AnyEvent::Handle.
716 - perl backend will now use POSIX::times as monotonic clock
717 when available and the monotonic clock option is not.
718 - run even without Time::HiRes in some cases now, but do not
719 provide sub-second accuracy (all *supported* perl versions have
720 Time::HiRes).
721
7224.233 Fri Aug 22 01:48:05 CEST 2008
723 - fix a bug in the testsuite, causing a hang.
724 - clarified lots of AnyEvent::Handle constructor arguments.
725
7264.232 Thu Aug 21 20:44:25 CEST 2008
727 - fix a bug in AnyEvent::Handle that could cause two fatal error
728 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
729 (testcase by Leon Brocard).
730 - clarified on_eof/on_error documentation a bit.
731
7324.231 Tue Jul 29 13:12:15 CEST 2008
733 - remove some debugging code left in AnyEvent::Util::fork_call
734 (and no, it's impossible to implement with the broken windows
735 perls without resource leaks or worse).
736
7374.23 Tue Jul 29 12:19:59 CEST 2008
738 - document the first parameter passed to condvar callbacks to be
739 the callback.
740 - add AnyEvent::Socket::{ntoa,aton} aliases.
741 - optimize the AE::Handle->push_read (line) for the default
742 eol marker.
743 - optimize push_read (packstring|storable) for small packets.
744 - invoke on_error callback when no on_eof callback is set.
745 - fix a bug in push_read (storable) of unknown impact.
746
7474.22 Sun Jul 20 16:34:13 CEST 2008
748 - new function AnyEvent::Socket::parse_hostport.
749 - as the bulkheads at microsoft can't even get getprotobyname reliably
750 working on their shitty fucking broken os we need to hardcode
751 some common protocol numbers in AnyEvent::Socket. How can
752 people even bother with such a pile of shit as windows.
753
7544.21 Thu Jul 17 14:40:05 CEST 2008
755 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
756 which is incompatible to previous invocation syntax.
757 - work around more windows perl bugs in fork_call. I cannot imagine
758 why anybody would bother implementing fork in such an obviously
759 unusable way.
760 - avoid calling $do_retry if the request has already finished in
761 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
762 K. Sharifulin).
763
7644.2 Sat Jul 12 22:42:11 CEST 2008
765 - support an "interval" argument for repeating timers.
766 - fix a bug in the pure perl event loop that caused timers to
767 be delayed under some circumstances.
768 - implement optional argument checking via PERL_ANYEVENT_STRICT.
769 - not importing anything from Event causes it to malfunction
770 and not use Time::HiRes, work around this bug.
771 - more bug workarounds against the endlessly broken Tk module,
772 seems to work now on 32 bit machines, but less so on 64 bit.
773 - minor optimisations applied to most backends.
774
7754.161 Fri Jul 4 14:24:48 CEST 2008
776 - fixed the t/handle/02_write.t, it read after write, with the
777 assumption that all data could be written before read
778 (reported by Jim Razmus).
779
7804.160 Thu Jul 3 04:02:21 CEST 2008
781 - re-registering signal watchers after unregistering one did
782 not work if a backend falls back on AnyEvent's default
783 implementation (also affected child watchers).
784 - new AnyEvent::Handle options: autocork and no_delay.
785
7864.152 Sun Jun 22 14:15:44 CEST 2008
787 - allow for 32-bit perls that implement shifts differently
788 on different architectures in parse_ipv4 (reported and
789 analysed by Keiichi DAIBA).
790
7914.151 Fri Jun 6 17:34:24 CEST 2008
792 - make sure specifying _only_ on_read and never pushing reads
793 works.
794
7954.15 Fri Jun 6 13:00:46 CEST 2008
796 - the pure perl backend would keep some watchers alive when more than
797 one watcher was registered for the same fd.
798 - new "packstring" and "storable" read and write types
799 for AnyEvent::Handle.
800 - allow on_eof handler to be called after on_error with EPIPE returns.
801 - do not immediately call on_read callback in handle constructor.
2 802
34.14 Thu Jun 5 20:29:31 CEST 2008 8034.14 Thu Jun 5 20:29:31 CEST 2008
4 - Fixed a bug in DNS SRV priority sorting. 804 - Fixed a bug in DNS SRV priority sorting.
5 - AnyEvent::Util::guard now reports runtime errors while 805 - AnyEvent::Util::guard now reports runtime errors while
6 executing the guard block as warnings. 806 executing the guard block as warnings.
… …
131 - support service names in tcp_connect and tcp_server. 931 - support service names in tcp_connect and tcp_server.
132 - work around netbsd/openbsd bug causing unix sockets not to 932 - work around netbsd/openbsd bug causing unix sockets not to
133 become writable until empty. 933 become writable until empty.
134 - only make a dummy read for error detection when getpeername 934 - only make a dummy read for error detection when getpeername
135 returns ENOTCONN. 935 returns ENOTCONN.
136 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 936 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
137 override. 937 override.
138 - new AnyEvent::Guard::cancel. 938 - new AnyEvent::Guard::cancel.
139 939
1403.5 Sat May 17 23:17:14 CEST 2008 9403.5 Sat May 17 23:17:14 CEST 2008
141 - gracefully deal with systems not defining CLOCK_MONOTONIC 941 - gracefully deal with systems not defining CLOCK_MONOTONIC
… …
155 955
1563.41 Sun May 11 19:53:13 CEST 2008 9563.41 Sun May 11 19:53:13 CEST 2008
157 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 957 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
158 958
1593.4 Sun May 11 00:29:25 CEST 2008 9593.4 Sun May 11 00:29:25 CEST 2008
160 - complete re-implementation of condvars with a hopefully much mroe useful 960 - complete re-implementation of condvars with a hopefully much more useful
161 API (backward compatible functions exist but are undocumented). 961 API (backward compatible functions exist but are undocumented).
162 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 962 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
163 (still subject to change). 963 (still subject to change).
164 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 964 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
165 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 965 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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