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

Comparing AnyEvent/Changes (file contents):
Revision 1.206 by root, Wed Nov 5 02:21:27 2008 UTC vs.
Revision 1.567 by root, Thu Mar 21 15:40:09 2013 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: document TCP_*** constants
4TODO: add debug/trace logging to some modules
5TODO: use the eval-with-cb-call trick also to speed up JSON?
6TODO: FAQ: common pitfalls? invoke-before-return
7TODO: FAQ: scope-ids? hosts file order?
8TODO: AnyEvent::Handle with AnyEvent::IO?
9
10TODO: invalid. and localhost. specialcasing inside AEDNS and not AESocket (rfc6761)
11TODO: maybe always try to downgrade in push_write.
12 - AnyEvent::Debug shell can now run coro shell commands, if
13 available.
14 - t/63* tests were wrongly in MANIFEST.
15 - mention IO::AIO as optional dependency.
16 - kernel.org's finger server went MIA, switch to freebsd.org.
17
187.04 Thu Nov 15 02:23:58 CET 2012
19 - debugging code left in.
20
217.03 Thu Nov 15 02:16:11 CET 2012
22 - AnyEvent::Socket::inet_aton did not work when dns resolution was
23 used to find the addresses (analyzed and patch by Maxime Soulé).
24 - fix a memory leak in the /etc/hosts lookup code when hosts don't
25 resolve and are not in hosts (analyzed by sten).
26
277.02 Tue Aug 14 04:27:58 CEST 2012
28 - AnyEvent::Util::run_cmd could block indefinitely (analyzed and test
29 program by Yorhel).
30 - verified that AnyEvent::Socket follows RFC5952.
31 - try to parse "ADDR#PORT" in addition to "ADDR PORT".
32
337.01 Sun May 13 01:03:17 CEST 2012
34 - fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but
35 not available, instead of throwing an exception.
36 - use File::Spec to get the tmpdir in t/*, to avoid needless
37 failures on (most, not mine :) windows boxes.
38 - new handle read types: tls_detect and tls_autostart.
39
407.0 Fri Apr 13 06:33:30 CEST 2012
41 - child watchers are broken in POE 1.352 (also many earlier
42 versions) and there seems to be no way to work around it, as POE
43 itself is inherently racy. Document this regression and add a delay
44 in t/68_poe_03_child.t for the time being.
45 - new module AnyEvent::IO, that is a frontend to either a pure-perl
46 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
47 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
48 - load /etc/hosts only when DNS has no answer.
49 - stat /etc/hosts on every access and reload it if it changed.
50 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
51 - fix a buggy croak in the dh parameter reading in AnyEvent::TLS.
52 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
53 a side effect, they now use true appending as opposed to libc
54 appending, and the file might not have been opened when the function
55 returns.
56 - the default logging level was not properly documented in a variety of
57 places, this has hopefully been rectified.
58 - updated uts46data.pl for unicode 6.1.0.
59 - made log messages generated by AnyEvent submodules not
60 include the package name anymore, as it will be logged
61 by default already.
62 - upgrade to the trick used by common-sense 3.5 to work
63 around extra warning torture/breakage under perl 5.15.x.
64 - log messages by anyevent are now ucfirst, are usually full
65 sentences and do no longer include the package name.
66 - the storable read type would sometimes throw an exception instead
67 of causing EBADMSG (analyzed by Maxime Soulé).
68 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
69 keysize (reported by Paul Howarth).
70 - documented Rocco fabricating statements about AnyEvent and me.
71
726.14 Tue Jan 31 20:00:24 CET 2012
73 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
74 probably an editing mistake (reported by Darin McBride).
75
766.13 Thu Jan 12 07:27:01 CET 2012
77 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
78 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
79 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
80 regex-matching inside the callback would change the parameter, despite
81 $1 being dynamically scoped per-block. this perl bug is now being
82 worked around (testcase by Cindy Wang).
83
846.12 Mon Dec 12 13:21:10 CET 2011
85 - $! was clobbered by subsequent calls in tcp_connect, due to
86 the postpone not saving/restoring it (analyzed by Richard Garnish).
87
886.11 Tue Nov 22 10:36:05 CET 2011
89 - Tk cannot create windows when tainted unless you set a title, so
90 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
91 - escape any nonprintable/nonascii characters when stringifying
92 backtraces.
93 - log the reason loading an interface module fails to load at level debug.
94
956.1 Tue Oct 4 19:44:30 CEST 2011
96 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
97 and above), and some messages inside AnyEvent have been elevated
98 to higher log levels to print by default.
99 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
100 - work around yet another signal race bug, newly introduced by POE. My
101 how that module sucks.
102 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
103 support ipv6-only operation even with defaults (also use ::1 instead
104 of 127.0.0.1).
105 - give better guidelines on how to chose a logging level.
106 - AnyEvent::Log can now cap (limit) log message priority by context.
107 - AE::log will now save and restore $! and $@, so callers don't have to.
108 - AE::Strict now checks that registered fds are still valid at regular
109 intervals.
110 - skip frequently called 'require' calls if the module appears loaded.
111 - increase signal latency test timer considerably for poor little osx,
112 which can't be bothered with basic posix functionality such as signal
113 delivery.
114 - mention some debugging aids in the tutorial.
115 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
116 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
117 of the context as title.
118 - AnyEvent::Strict would not restore read-write access to $_ when a callback
119 died - this is a user bug, but since AE::Strict exists to help dbeugging...
120 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
121 to work with it - the author is well aware of the issues but refuses
122 to fix them. This extreme step has been taken because the module actively
123 attacks the goals of AnyEvent and makes it harder for module authors to use
124 AnyEvent.
125
1266.02 Fri Aug 26 20:08:31 CEST 2011
127 - log_to_syslog now accepts facility strings.
128 - log_to_syslog would not allow facilities - "$facility|$level"
129 *literally* means that string. sheesh.
130 - the CPAN parser got confused and thought we provide Sys::Syslog, try
131 to work around.
132 - renamed AnyEvent::Impl::FLTK2 to FLTK.
133
1346.01 Fri Aug 26 07:04:11 CEST 2011
135 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
136 the TTL as fourth element - this affects ->resolve, ->request
137 and ->dns_unpack indirectly. Convenience functions such as
138 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
139 AnyEvent::Socket.
140 - INCOMPATIBLE CHANGE: backend authors now should not implement
141 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
142 methods.
143 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
144 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
145 - DEPRECIATION: please change
146 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
147 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
148 in your programs, for future compatibility.
149 - *blush* idle watcher emulation was borked.
150 - AnyEvent::Strict now checks the AE::xxx functions as well.
151 - querying too long domain names would barf deep inside AnyEvent::DNS - since
152 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
153 for too long domain names explicitly and passes an undef to the callback
154 instead (reported by James Bromberger).
155 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
156 - new functions: AnyEvent::postpone, AnyEvent::log.
157 - new module: AnyEvent::Log, for simple logging needs.
158 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
159 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
160 a debug shell.
161 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
162 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
163 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
164 /etc/hosts can be found.
165 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
166 configuration.
167 - env variables can now be specified using AE_xyz instead of
168 PERL_ANYEVENT_xyz, subject to some rules.
169 - lots of new AnyEvent::Debug functionality.
170 - AnyEvent::Strict now checks for modifications of $_ by the callback
171 (by setting it to readonly while executing the callback, if possible).
172 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
173 Invest extra overhead to work around this peculiar behaviour.
174 - split the pure perl backend into a loop module and an impl
175 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
176 - some read types (regex, netstring, json, storable) could
177 errornously keep an AE::Handle object alive on their own
178 (reported by Mohammad Toossi).
179 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
180 supports /etc/hosts overrides.
181 - add more workarounds around child watcher bugs in IO::Async
182 and (newly introduced) in POE.
183 - use glib child watchers and try to work around its limitations,
184 if possible (based on a patch by Kevin Ryde).
185 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
186 - many read types in AnyEvent::Handle would malfunction during
187 transitions to TLS because of a "delete $self->{rbuf}".
188 (reported by Gerald Galster).
189 - improve condvar blocking wait performance (probably only noticable
190 with faster backends), also make it easier for event loops that
191 do not support blocking waits.
192 - PERL_ANYEVENT_MODEL now also allows full module names, although
193 the usefulness of this feature is not yet clear.
194 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
195 - add t/80_ssltext.t.
196 - no longer use AUTOLOAD for the initial loading (for no very good reason).
197 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
198 - do not run event loop detection multiple times when programs call their cached
199 copies of AnyEvent::detect.
200 - parse_hostport now accepts absolute paths als unix domain sockets.
201 - suppress the idiotic warning inside POE by patching the POE::Kernel
202 run flags instead of asking the user to jump through contortions.
203 - add Task::Weaken as dependency (as recommends only at this time).
204 - replace PF_UNSPEC by 0, for good old minix.
205
2066.0 (not officially released, but was on CPAN for about 15 minutes :/)
207
2085.34 Thu May 12 09:59:41 CEST 2011
209 - AE::TLS didn't support both cert_file and key at the same time,
210 due to some copy&paste bug (reported and analyzed by vti).
211 - improve AE::Handle timeout documentation slightly.
212
2135.33 Mon Feb 7 21:51:41 CET 2011
214 - never appeared on CPAN, for unknown reasons.
215 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
216 object when it's already gone (analyzed by Ryan Bullock).
217 - update idna algorithm(s) to UTS#46 version 6.0.0.
218 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
219 uts46 says it's an error, and testsuite makes more sense.
220 - remove leading dots in idn_to_ascii - testsuite says to remove,
221 uts46 says it's an error, and testsuite makes more sense.
222 - fixed a bug in idn ignored character processing.
223 - passes the uts46 6.0.0 testsuite except for three cases which
224 apparently are bugs in the testsuite itself.
225 - passes the uts46 6.0.1 testsuite except for eleven cases which
226 apparently are bugs in the testsuite itself.
227
2285.31 Sun Jan 23 11:49:19 CET 2011
229 - work around a POE bug where POE would first poll for new events and
230 invoke handlers BEFORE handling the message queue.
231 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
232 - stop_read now does just that in TLS mode. see its
233 documentation for caveats.
234 - update to common::sense 3.4 warning set (some warnings in perl impose
235 an arbitrary 32kb limit on regex matches, which affects
236 AnyEvent::Handle).
237 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
238 - new experimental AE::Handle::resettls method.
239
2405.3 Fri Dec 31 05:48:13 CET 2010
241 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
242 issues with IO::Async have been solved in that version, including
243 the availabilty of a default loop!
244 - implement new wbuf_max parameter for AnyEvent::Handle.
245 - added Cocoa::EventLoop backend by Daisuke Murase.
246 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
247 - clarify handle on_prepare documentation.
248 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
249 to avoid leaking them to any child processes.
250 - try to force files into the arch-dependend lib directory, for the
251 half-"turly obsessive compulsive person"(s) that depend on it.
252
2535.29 Sun Dec 5 10:49:21 CET 2010
254 - convert EV backend to EV 4.00 API (so better upgrade EV too).
255 - AE::Handle: implement dynamic read_size adjustment, add
256 max_read_size parameter, reduce default read_size to 2048.
257 - add some FAQ entries for servers and/or on_eof vs. on_error.
258 - work around OS bugs (cygwin again) when lingering in AE::Handle.
259 - reduce memory usage slightly when connect was used in AE::Handle.
260 - wrok around more segfaults in Tk.
261 - document tls_ctx => undef as valid.
262 - detect WNOHANG value at compiletime.
263 - include some probably linux-only support for building AnyEvent
264 as part of the perl core.
265 - improve compatibility of t/08_idna.t to perl 5.8.
266 - make AnyEvent::Strict truly optional (in case somebody builds a
267 minimal perl...).
268
2695.28 Wed Oct 13 04:14:23 CEST 2010
270 - due to a glitch, AnyEvent's internal getprotobyname wasn't
271 actually used in tcp_server.
272 - implement AnyEvent::CondVar->new method, which was documented but
273 missing.
274 - support autoloading of read types in unshift_read.
275 - AnyEvent::Strict: do the test for invalid fh arguments differently,
276 resulting in more informative error messages.
277 - new AnyEvent::FAQ manpage.
278 - clarify that you can actually call push_read etc. in on_read.
279 - put keys and certs in /tmp, not into the current directory,
280 and document this in the SECURITY CONSIDERATIONS sections
281 of AnyEvent::TLS.
282 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
283 requested by Ron Isaacson.
284 - document $AnyEvent::DNS::RESOLVER variable.
285 - incorporated major typo patches by Abhijit Menon-Sen.
286 - AnyEvent::Handle now croaks when negative timeouts are passed.
287 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
288 undocumented, but who knows when they might come in handy.
289
2905.271 Tue Jun 8 12:05:46 CEST 2010
291 - backport to perl 5.8.x.
292
2935.27 Sun Jun 6 12:12:05 CEST 2010
294 - postpone differently in AnyEvent::Socket now, as
295 when not, canceling the connection attempt might fail
296 (found by Felix Antonius Wilhelm Ostmann).
297 - explicitly check for non-stream sockets in AE::Handle, too many
298 clueless people fell into the trap of this somehow working.
299 - simplified and reworked the "OTHER MODULES" section.
300 - better/more condvar examples.
301
3025.261 Wed Apr 28 16:13:36 CEST 2010
303 - AF_INET6 was not properly used from Socket6 during configuration
304 time (found by Andy Grundman).
305
3065.26 Mon Apr 12 04:49:35 CEST 2010
307 - don't generate spurious readyness notifications when select
308 returns EINTR in the pure perl backend (analysed and testcase
309 by Bas Denissen).
310 - give same examples in SYNOPSIS sections of AnyEvent and AE.
311 - provider faster implementations for AE::time/now/now_update for
312 EV and Perl backends.
313 - provide faster AE::cv and AE::time implementations for all
314 backends.
315 - no longer support lower/mixed-case signal names (this was
316 never documented nor universally supported).
317 - some more memory µ-optimisations, and somewhat less messy
318 code to cope with both AnyEvent and AE APIs.
319
3205.251 Sat Mar 13 00:58:21 CET 2010
321 - make test failed if it was run before make install on perl <=
322 5.8.8, and older versions of perl make our live very difficult,
323 so write constants.pl during Makefile.PL time.
324
3255.25 Sat Mar 13 00:23:14 CET 2010
326 - fix a race condition in AnyEvent::Handle that would cause
327 a "bio_read: unsupported method" error (in _another_
328 ssl connection) after user code threw an exception.
329 - added AnyEvent::Handle->destroyed method.
330 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
331 - the AnyEvent::Util::fh_nonblocking function worked only
332 by lucky accident on win32.
333 - smaller and faster AnyEvent::Util::fh_nonblocking.
334 - when the (required!) Time::HiRes module is missing, AnyEvent
335 did not fall back to built-in time properly.
336 - do not load Fcntl at runtime, saving memory and loading time.
337 - precompile a number of constants and use them instead of runtime
338 detection and eval.
339 - free detection code after detection and similar memory optimisations.
340 - Perl backend timer interval best effort drift has been improved
341 (same algorithm as EV).
342 - update unicode idna mapping table.
343
3445.24 Tue Jan 5 11:39:43 CET 2010
345 - cygwin never reports errors from failed connects. 1.5 just gives
346 you continous readyness and EAGAIN, 1.7 is even more broken and
347 just hangs. work around both issues in a cygwin-specific hack.
348 - improve idle watcher documentation slightly.
349
3505.23 Sun Dec 20 23:48:00 CET 2009
351 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
352 - implement punycode_encode/decode, idn_nameprep,
353 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
354 - provide $AE::VERSION.
355 - removed traces of "no strict 'refs'".
356
3575.22 Sat Dec 5 03:51:13 CET 2009
358 - downgrade-or-fail in AnyEvent::Handle::push_write, to
359 diagnose encoding failures earlier and more succinctly.
360 (this works around bugs in perl, throwing away encoding info
361 when passing scalar data to extensions).
362 - add more examples to AnyEvent::Socket manpage.
363 - upgrade internal warning set to the same as common::sense 2.03.
364 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
365 (John Beppu).
366 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
367 - add google public dns servers to fallback server set and make sure
368 we load-balance properly between the three sets. also add all
369 fallback dns servers, not just a random one, to each dns config.
370
3715.21 Thu Nov 19 02:48:47 CET 2009
372 - fix a problem where socket constants were called with parameters
373 (spotted by David Friedland).
374 - fork_call never use'd POSIX (reported by Daisuke Maki).
375 - improve perl 5.6 compatibility further (but it still won't work
376 unless you rip out everything but the core).
377 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
378 uses 10MB of RAM, but doesn't open a console window. *sigh*.
379
3805.202 Wed Oct 14 22:35:44 CEST 2009
381 - AnyEvent::DNS would unexpectedly clobber $_ under windows
382 (analysed by Matthias Waldorf).
383 - AnyEvent::Handle::run_cmd can now pass the PID of the
384 newly-created process, which is much less useful than it might
385 sound (based on patch by Yann Kerherve).
386
3875.201 Tue Sep 29 12:09:25 CEST 2009
388 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
389 (reported by Torsten Foertsch).
390 - common::sense 2.0 could cause tcp_server to throw an exception
391 (analysed by elmex).
392
3935.2 Mon Sep 14 07:04:49 CEST 2009
394 - INCOMPATIBLE CHANGE: do no longer support register_read_type
395 and register_write_type in AnyEvent::Handle, instead support
396 package names (the facility was mostly abused).
397 - implement "packagename-as-read/write type" support in
398 AnyEvent::Handle.
399 - AnyEvent::Handle: new options "keepalive" and "oobinline".
400 - oobinline set by default to avoid security issues.
401 - the pure-perl event loop backend wrongly detected times() fallback
402 support (spotted by Pavel Boldin).
403
4045.12 Tue Sep 1 20:26:50 CEST 2009
405 - be more lenient when parsing resolv.conf files, as some people
406 use hashmarks as comment indicator inside directives (reported
407 by Michael S. Fischer).
408 - use same set of warnings as common::sense 2.0.
409 - fix a potential 32 bit overflow issue due to perl having problems
410 with large hex constants in the Perl backend.
411 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
412 by default.
413 - fix some prototypes.
414 - (optionally) require Async::Interrupt 1.02.
415 - replace opendns fallbacks by verizon/level3 public dns servers.
416
4175.112 Fri Aug 21 13:59:04 CEST 2009
418 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
419 using starttls with an illegal/uncreatable tls context.
420 - port to broken win32 perls.
421
4225.111 Sun Aug 16 18:44:36 CEST 2009
423 - perl errornously requires sockaddr_un structures returned by the
424 kernel to have a specific length. work around this bug by padding
425 sockaddr_un structures (found to fail on OS X, as perl only uses
426 the correct code for GNU/Linux, but likely to fail on other
427 OSes as well).
428
4295.11 Wed Aug 12 17:49:37 CEST 2009
430 - fixed a bug in run_cmd with input from filename.
431 - create a json coder object when none is given in write direction
432 to reduce dependency on JSON.
433
4345.1 Tue Aug 11 03:17:32 CEST 2009
435 - tcp_connect now reports the correct error in $!, not an unrelated
436 one (was broken since 4.91).
437 - AE::Handle did not properly clear rtimeout/wtimeout watchers
438 when retrying a tcp_connect.
439 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
440 - fixed wrong documentation in AE::Socket.
441
4425.01 Mon Aug 10 03:16:32 CEST 2009
443 - last release broke EV child watchers :/.
444
4455.0 Sun Aug 9 17:32:09 CEST 2009
446 - officially introduce and document the AE API.
447 - lowering the timeout at runtime did not have immediate effect
448 in AnyEvent::Handle.
449 - AE::Handle now has separate and independent read and write timeout
450 settings.
451 - probe for Guard module when AnyEvent::Util is loaded, not at
452 runtime (Event doesn't like eval "use" from callbacks, and it avoids
453 exporting the probe function).
454 - Impl::IOAsync's io watchers did not use the filehandle emulation
455 code to work around it's watcher limits.
456 - work around IO::Async's problems with overloaded objects.
457 - try to disable wearnings inside IO::Async, it's just too buggy.
458 - do no longer use Event by default, only EV or Perl, Event is too buggy.
459 - hackishly provide fast AE:: interface when EV or Perl is the backend.
460 - inherit VERSION from AnyEvent::VERSION in most modules.
461
4624.91 Thu Aug 6 15:42:45 CEST 2009
463 - AE::Handle::starttls could get out of sync when the read buffer
464 already contains some TLS handshake.
465 - AE::Handle did not properly free the TLS session, causing bigger
466 memory leaks in Net::SSLeay than expected (reported by toaster).
467 - AE::Socket::tcp_connect will now artificially delay invoking
468 the callback to avoid returning after invoking the callback.
469 - convert many internal watcher uses to AE API.
470
4714.9 Sat Aug 1 11:07:01 CEST 2009
472 - Glib silently fails when registering a timeout with
473 a negative value, so avoid that.
474 - call condvar callback immediately when it is set
475 after the condition is already signalled.
476 - check rbuf_max condition only after trying to consume
477 data.
478 - normalise signal numbers to names when using the ae's
479 signal handling, but do not document this (yet).
480 - pure perl signal emulation did not properly set nonblocking
481 mode on the signal pipe on win32 (this is unlikely to have caused
482 any issues).
483 - new module: AnyEvent::Debug.
484 - AnyEvent::Strict now checks that a signal was specified
485 by name (not name or number).
486 - reduce memory footprint in typical cases by ~50kb by
487 only compiling signal code when necessary.
488 - add AnyEvent::Handle->rbuf_max.
489 - grab the AE namespace for future new API, implement stubs for
490 said future/alternative API.
491 - new function: AnyEvent::Socket::format_hostport.
492
4934.881 Tue Jul 28 12:51:53 CEST 2009
494 - work around a bug in local in pre-5.10 perls, causing
495 AnyEvent::Handle to recurse when it should not
496 (analysed by elmex).
497
4984.88 Tue Jul 28 04:04:37 CEST 2009
499 - re-bless the handle into a dummy package after calling
500 AnyEvent::Handle::destroy, so the user does not need to check
501 for errors after every push_write etc.
502 - do not attempt to run t/02_signals.t on obviously broken
503 platforms.
504
5054.87 Sun Jul 26 02:06:16 CEST 2009
506 - do not attempt to linger when there is no longer a valid fh.
507 - enforce tls mode to be either accept or connect, do not simply
508 segfault in Net::SSLeay.
509 - AnyEvent::Handle can now call tcp_connect itself (new parameters
510 connect, on_prepare, on_connect and on_connect_error). Updated
511 tutorial accordingly.
512 - add AnyEvent::Impl::Irssi backend.
513
5144.86 Mon Jul 20 23:52:29 CEST 2009
515 - since the verbose warning is not enough, explicitly document
516 that versions before 1.33 of Net::SSLeay are not secure.
517 - work around signal handling races in Event and (...) Event::Lib.
518 - try to align signal-race timer to full-second boundaries.
519 - work around Tk not liking negative timeouts.
520 - don't complain of different grades of environmental unfriendlyness
521 in IO::Async.
522
5234.85 Sat Jul 18 06:16:14 CEST 2009
524 - nail the signal race problem in perl once and for all
525 (see $AnyEvent::MAX_SIGNAL_LATENCY).
526 - take advantage of Async::Interrupt if it is available.
527 - load Time::HiRes and Guard modules on demand only.
528 - add optional/recommended modules section to AnyEvent
529 documentation.
530 - reduce memory usage considerably (and reduce startup penalty)
531 by not using "strict", "warnings" and "overload" modules.
532 - work around buggy windows/openbsd perls and provide EBADMSG
533 and EPROTO ourselves when missing.
534 - improve perl 5.6 compatibility of the core event loop.
535 - made Net::SSLeay version 1.33 a soft requirement.
536
5374.83 Fri Jul 17 16:56:26 CEST 2009
538 - implement AnyEvent::Socket::getprotobyname.
539 - AnyEvent::CondVar's will now detect recursive blocking
540 waits and will croak, as too many people fall into
541 this trap.
542 - AnyEvent::Handle will now call ->destroy on itself after
543 executing the on_error callback, instead of doing some
544 half-baked internal shutdown, for fatal errors.
545 - clarify on_eof behaviour w.r.t. the read queue and
546 on_read callbacks.
547 - ignore some possible spurious wake-ups in tcp_connect.
548
5494.82 Sat Jul 11 00:34:55 CEST 2009
550 - POE and Event backends didn't accept some callable objects as
551 callbacks.
552 - use Config module instead of POSIX module to detect signal names
553 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
554 module doesn't even have all POSIX signals :/.
555 - use more workarounds around the many refcnt/corruption bugs in
556 Event::Lib.
557 - work around a race condition in perl's select, causing t/03_child.t
558 to rarely fail.
559
5604.81 Thu Jul 9 10:30:30 CEST 2009
561 - AnyEvent::Handle didn't properly diagnose write errors
562 (it expected -1 from syswrite, how lame... :).
563 - support file descriptors in addition to file handles
564 in AnyEvent->io.
565 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
566 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
567 and PERL_ANYEVENT_CA_PATH.
568 - provide a sensible synopsis section for AnyEvent::TLS.
569 - add a "supported backends" section to the manpage.
570 - added simple io watcher test to testsuite, using a
571 portable_socketpair.
572 - tried to improve the stability of the Event::Lib backend,
573 YMMV.
574
5754.8 Mon Jul 6 23:45:16 CEST 2009
576 - AnyEvent::DNS did not properly follow CNAME records with
577 uppercase targets.
578 - AnyEvent::DNS would errornously return AAAA records
579 with v4 mapped addresses (a faulty record) as ipv4 addresses,
580 causing AnyEvent::Socket to throw an exception.
581 - added new module AnyEvent::TLS for easier SSL/TLS context
582 creation, with many options including hostname verification,
583 secure default configuration, lots of documentation and,
584 predefined diffie-hellman keys for perfect forward security
585 and much more. get it while it's still fresh!
586 - use AnyEvent::TLS in AnyEvent::Handle for context management.
587 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
588 so AnyEvent::Socket users have smaller memory footprint
589 in the common case.
590 - add AnyEvent::Handle->push_shutdown method.
591 - add an additional error message parameter to AnyEvent::Handle's
592 on_error callback (for TLS, $! is still available).
593 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
594 - make AnyEvent::Handle more robust against early conenction
595 failures (during new), and return C<undef> in that case
596 from the constructor.
597 - AnyEvent::Handle will now try to load only JSON::XS first,
598 then fall back to JSON.
599 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
600 and because it was documented that way.
601
6024.451 Fri Jul 3 00:28:58 CEST 2009
603 - do not clear rbuf when shutting down an AnyEvent::Handle
604 object - doing so breaks AnyEvent::HTTP.
605
6064.45 Mon Jun 29 22:59:26 CEST 2009
607 - a write error could cause AnyEvent::Handle to create
608 an I/O watcher with an undefined $fh.
609 - special-case mapped ipv4 addresses in both
610 AnyEvent::Socket::format_address and parse_address,
611 to treat them just like ipv4 addresses.
612 - updated and overhauled the AnyEvent::Intro doc.
613 - implement AnyEvent::Socket::format_ipv4/ipv6.
614 - slightly speed up portable_pipe/socketpair.
615 - expand condvar begin/end documentation.
616
6174.42 Fri Jun 26 08:32:18 CEST 2009
618 - preliminary and neccesarily incomplete support for IO::Async.
619 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
620 case it was set to IGNORE, to ensure we can catch child statuses
621 even when the calling env acts stupidly.
622 - updated benchmarks with IO::Async, which performs very well.
623
6244.412 Wed Jun 24 01:35:57 CEST 2009
625 - support an "untaint" attribute for AnyEvent::DNS
626 and set it on the default resolver.
627 - implement and document AnyEvent::Impl::Perl::loop.
628 - remove all anyevent-env variables from %ENV
629 when running in tainted mode.
630 - mention and extend the IO::Lambda benchmark.
631
6324.411 Sun Jun 7 18:48:02 CEST 2009
633 - do not try to use F_SETFD on windows.
634
6354.41 Thu May 14 06:40:11 CEST 2009
636 - work around issues in older perls (5.8.0?) when
637 a signal handler is deleted from the %SIG hash.
638 - use POSIX::_exit in child test, to avoid running
639 destructors.
640 - speed up CHLD handling by relying on SIGCHLD being
641 synchronously delivered, even when we roll our own
642 implementation.
643 - AnyEvent::DNS: add the "dname" resource record name
644 for cosmetic reasons.
645
6464.4 Sun Apr 26 20:12:33 CEST 2009
647 - implemented idle watchers, where applicable.
648 - AnyEvent->time died when Event backend was in use.
649 - fix a memleak in the Tk backend.
650 - sped up Tk timer handling.
651 - clip negative "after" values to 0 in AnyEvent::Impl::Event
652 to avoid spamming out warning messages.
653 - fix Qt timers without interval.
654 - avoid Qt zero-timeout special case, in old libqt's.
655
6564.352 Mon Apr 20 16:31:11 CEST 2009
657 - fix AnyEvent::Strict error messages for child watchers.
658 - fix/update Makefile.PL "recommends" versions.
659 - add AnyEvent->now_update.
660
6614.351 Sat Apr 11 07:56:14 CEST 2009
662 - actually make the signal pipe work under win32.
663 - localise $! in signal handler to avoid changing $!, although
664 perl itself does not seem to save/restore errno either.
665 - set the cloexec flag on the signal pipe (normally set by perl too).
666
6674.35 Fri Mar 27 11:48:20 CET 2009
668 - event models relying on AnyEvent's signal watcher emulation did
669 invoke the callback asynchronously, contradicting documentation
670 and causing signals to get lost (this includes AnyEvent's own
671 event loop). AnyEvent now uses the standard pipe trick to make
672 callback execution synchronous to the event loop.
673 - AnyEvent::Handle didn't free TLS context data on DESTROY
674 (patch by Pavel Shaydo).
675 - work around the ever-incompatibly-changing API of MakeMaker.
676 - document that changing global variables without restoring them
677 is a bad idea in Perl (noted by Adam Rosenstein).
678 - AnyEvent::Strict now barfs if ->io is passed a file.
679
6804.34 Thu Feb 12 18:32:45 CET 2009
681 - separately buffer TLS read data, as otherwise the read queue could
682 deadlock as receiving data is not expected while draining the
683 read queue (which cna only happen with TLS).
684 - raise EBADMSG error on JSON decoding errors.
685 - fix some minor manpage bugs (reported by Maximilian Gaß).
686 - speed up select bitmask parsing quite a bit in the pure perl
687 backend.
688 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
689
6904.331 Tue Jan 6 21:07:25 CET 2009
691 - socketpair fails on many vista machines because vista has
692 completely broken accept/getpeername and getsockname functions,
693 so we provide our own socketpair emulation that kind of works
694 (AnyEvent::Util::portable_pipe).
695 - new function: AnyEvent::Util::portable_socketpair.
696 - take advantage of the Guard module if it exists.
697
6984.33 Fri Nov 21 02:35:40 CET 2008
699 - AnyEvent::Strict did errornously flag a fileno of 0 as
700 illegal.
3 - reduce memory usage and slightly speed up the pure perl backend 701 - reduce memory usage and slightly speed up the pure perl backend
4 by only storing the file descriptor, not the file handle. 702 by only storing the file descriptor, not the file handle.
703 - add missing autocork method to AnyEvent::Handle (reported
704 by Adam Rosenstein).
705 - AnyEvent::DNS->resolve errornously documented an $rcode result
706 argument, but there isn't (spotted by Henrik Krohns).
707 - the naked truth about Tk - it's basically unmaintained.
5 708
64.32 Mon Nov 3 22:46:32 CET 2008 7094.32 Mon Nov 3 22:46:32 CET 2008
7 - fix AnyEvent::Socket::resolve_sockaddr to properly support 710 - fix AnyEvent::Socket::resolve_sockaddr to properly support
8 unix sockets again and choose a proper default for 711 unix sockets again and choose a proper default for
9 the socket type. 712 the socket type.
… …
204.3 Fri Oct 3 09:18:43 CEST 2008 7234.3 Fri Oct 3 09:18:43 CEST 2008
21 - AnyEvent will now install a no-op signal handler for SIGPIPE, 724 - AnyEvent will now install a no-op signal handler for SIGPIPE,
22 unless one has been installed already. 725 unless one has been installed already.
23 - warn about not loaidng AnyEvent::Impl::POE early enough 726 - warn about not loaidng AnyEvent::Impl::POE early enough
24 (patch by Adam Rosenstein). 727 (patch by Adam Rosenstein).
25 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 728 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
26 - fix a bug in where in SSL connect mode, the 729 - fix a bug in where in SSL connect mode, the
27 client would first wait for some data by the server and 730 client would first wait for some data by the server and
28 otherwise hang. (reported and analyzed in an absolutely 731 otherwise hang. (reported and analysed in an absolutely
29 exemplary manner by Adam Rosenstein). 732 exemplary manner by Adam Rosenstein).
30 - fix a bug in where SSL EOF would not be treated 733 - fix a bug in where SSL EOF would not be treated
31 as stream EOF, putting the connection into a hung state. 734 as stream EOF, putting the connection into a hung state.
32 (reported and analyzed in an absolutely exemplary manner by Adam 735 (reported and analysed in an absolutely exemplary manner by Adam
33 Rosenstein). 736 Rosenstein).
34 - fix a potential segfault when the TLS context 737 - fix a potential segfault when the TLS context
35 would go missing in a read callback (e.g. due to stoptls) - 738 would go missing in a read callback (e.g. due to stoptls) -
36 Net::SSLeay of course makes no type checking whatsoever. 739 Net::SSLeay of course makes no type checking whatsoever.
37 - AnyEvent::Handle will not stop the read watcher in TLS mode. 740 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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