ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent/Changes
Revision: 1.572
Committed: Tue Jul 30 23:31:56 2013 UTC (10 years, 9 months ago) by root
Branch: MAIN
Changes since 1.571: +1 -0 lines
Log Message:
*** empty log message ***

File Contents

# Content
1 Revision history for Perl extension AnyEvent.
2
3 TODO: document TCP_*** constants
4 TODO: add debug/trace logging to some modules
5 TODO: use the eval-with-cb-call trick also to speed up JSON?
6 TODO: FAQ: common pitfalls? invoke-before-return
7 TODO: FAQ: scope-ids? hosts file order?
8 TODO: AnyEvent::Handle with AnyEvent::IO?
9 TODO: invalid. and localhost. specialcasing inside AEDNS and not AESocket (rfc6761)
10
11 TODO: expose default formatting implementation (ft &c) in AE::log, also allow overridign fatal exit
12 TODO: José Micó (3.0K) 1 inbox Error connecting to hosts resolved from /etc/hosts
13 TODO: Daniel Carosone (5.2K) 1 inbox AnyEvent::IO vs AnyEvent::AIO (doc patch)
14 TODO: maybe implement env variable to give hosts precedence
15 - uts46data.pl couldn't be found due to wrong naming of the file
16 (reported by Fulko Hew).
17 - AnyEvent::Debug shell can now run coro shell commands, if available.
18 - t/63* tests were wrongly in MANIFEST.
19 - mention IO::AIO as optional dependency.
20 - kernel.org's finger server went MIA, switch to freebsd.org and
21 icculus.org.
22 - handle lone \015's properly in AE::Handle's default line read
23 (reported by various people).
24 - hard-disable $^W in most tests, it generates too much garbage output.
25 - use a (hopefully) more future-proof method to emulate common::sense.
26 - upgrade to UTS-46:6.2.0.
27
28 7.04 Thu Nov 15 02:23:58 CET 2012
29 - debugging code left in.
30
31 7.03 Thu Nov 15 02:16:11 CET 2012
32 - AnyEvent::Socket::inet_aton did not work when dns resolution was
33 used to find the addresses (analyzed and patch by Maxime Soulé).
34 - fix a memory leak in the /etc/hosts lookup code when hosts don't
35 resolve and are not in hosts (analyzed by sten).
36
37 7.02 Tue Aug 14 04:27:58 CEST 2012
38 - AnyEvent::Util::run_cmd could block indefinitely (analyzed and test
39 program by Yorhel).
40 - verified that AnyEvent::Socket follows RFC5952.
41 - try to parse "ADDR#PORT" in addition to "ADDR PORT".
42
43 7.01 Sun May 13 01:03:17 CEST 2012
44 - fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but
45 not available, instead of throwing an exception.
46 - use File::Spec to get the tmpdir in t/*, to avoid needless
47 failures on (most, not mine :) windows boxes.
48 - new handle read types: tls_detect and tls_autostart.
49
50 7.0 Fri Apr 13 06:33:30 CEST 2012
51 - child watchers are broken in POE 1.352 (also many earlier
52 versions) and there seems to be no way to work around it, as POE
53 itself is inherently racy. Document this regression and add a delay
54 in t/68_poe_03_child.t for the time being.
55 - new module AnyEvent::IO, that is a frontend to either a pure-perl
56 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
57 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
58 - load /etc/hosts only when DNS has no answer.
59 - stat /etc/hosts on every access and reload it if it changed.
60 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
61 - fix a buggy croak in the dh parameter reading in AnyEvent::TLS.
62 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
63 a side effect, they now use true appending as opposed to libc
64 appending, and the file might not have been opened when the function
65 returns.
66 - the default logging level was not properly documented in a variety of
67 places, this has hopefully been rectified.
68 - updated uts46data.pl for unicode 6.1.0.
69 - made log messages generated by AnyEvent submodules not
70 include the package name anymore, as it will be logged
71 by default already.
72 - upgrade to the trick used by common-sense 3.5 to work
73 around extra warning torture/breakage under perl 5.15.x.
74 - log messages by anyevent are now ucfirst, are usually full
75 sentences and do no longer include the package name.
76 - the storable read type would sometimes throw an exception instead
77 of causing EBADMSG (analyzed by Maxime Soulé).
78 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
79 keysize (reported by Paul Howarth).
80 - documented Rocco fabricating statements about AnyEvent and me.
81
82 6.14 Tue Jan 31 20:00:24 CET 2012
83 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
84 probably an editing mistake (reported by Darin McBride).
85
86 6.13 Thu Jan 12 07:27:01 CET 2012
87 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
88 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
89 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
90 regex-matching inside the callback would change the parameter, despite
91 $1 being dynamically scoped per-block. this perl bug is now being
92 worked around (testcase by Cindy Wang).
93
94 6.12 Mon Dec 12 13:21:10 CET 2011
95 - $! was clobbered by subsequent calls in tcp_connect, due to
96 the postpone not saving/restoring it (analyzed by Richard Garnish).
97
98 6.11 Tue Nov 22 10:36:05 CET 2011
99 - Tk cannot create windows when tainted unless you set a title, so
100 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
101 - escape any nonprintable/nonascii characters when stringifying
102 backtraces.
103 - log the reason loading an interface module fails to load at level debug.
104
105 6.1 Tue Oct 4 19:44:30 CEST 2011
106 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
107 and above), and some messages inside AnyEvent have been elevated
108 to higher log levels to print by default.
109 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
110 - work around yet another signal race bug, newly introduced by POE. My
111 how that module sucks.
112 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
113 support ipv6-only operation even with defaults (also use ::1 instead
114 of 127.0.0.1).
115 - give better guidelines on how to chose a logging level.
116 - AnyEvent::Log can now cap (limit) log message priority by context.
117 - AE::log will now save and restore $! and $@, so callers don't have to.
118 - AE::Strict now checks that registered fds are still valid at regular
119 intervals.
120 - skip frequently called 'require' calls if the module appears loaded.
121 - increase signal latency test timer considerably for poor little osx,
122 which can't be bothered with basic posix functionality such as signal
123 delivery.
124 - mention some debugging aids in the tutorial.
125 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
126 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
127 of the context as title.
128 - AnyEvent::Strict would not restore read-write access to $_ when a callback
129 died - this is a user bug, but since AE::Strict exists to help dbeugging...
130 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
131 to work with it - the author is well aware of the issues but refuses
132 to fix them. This extreme step has been taken because the module actively
133 attacks the goals of AnyEvent and makes it harder for module authors to use
134 AnyEvent.
135
136 6.02 Fri Aug 26 20:08:31 CEST 2011
137 - log_to_syslog now accepts facility strings.
138 - log_to_syslog would not allow facilities - "$facility|$level"
139 *literally* means that string. sheesh.
140 - the CPAN parser got confused and thought we provide Sys::Syslog, try
141 to work around.
142 - renamed AnyEvent::Impl::FLTK2 to FLTK.
143
144 6.01 Fri Aug 26 07:04:11 CEST 2011
145 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
146 the TTL as fourth element - this affects ->resolve, ->request
147 and ->dns_unpack indirectly. Convenience functions such as
148 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
149 AnyEvent::Socket.
150 - INCOMPATIBLE CHANGE: backend authors now should not implement
151 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
152 methods.
153 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
154 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
155 - DEPRECIATION: please change
156 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
157 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
158 in your programs, for future compatibility.
159 - *blush* idle watcher emulation was borked.
160 - AnyEvent::Strict now checks the AE::xxx functions as well.
161 - querying too long domain names would barf deep inside AnyEvent::DNS - since
162 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
163 for too long domain names explicitly and passes an undef to the callback
164 instead (reported by James Bromberger).
165 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
166 - new functions: AnyEvent::postpone, AnyEvent::log.
167 - new module: AnyEvent::Log, for simple logging needs.
168 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
169 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
170 a debug shell.
171 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
172 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
173 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
174 /etc/hosts can be found.
175 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
176 configuration.
177 - env variables can now be specified using AE_xyz instead of
178 PERL_ANYEVENT_xyz, subject to some rules.
179 - lots of new AnyEvent::Debug functionality.
180 - AnyEvent::Strict now checks for modifications of $_ by the callback
181 (by setting it to readonly while executing the callback, if possible).
182 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
183 Invest extra overhead to work around this peculiar behaviour.
184 - split the pure perl backend into a loop module and an impl
185 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
186 - some read types (regex, netstring, json, storable) could
187 errornously keep an AE::Handle object alive on their own
188 (reported by Mohammad Toossi).
189 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
190 supports /etc/hosts overrides.
191 - add more workarounds around child watcher bugs in IO::Async
192 and (newly introduced) in POE.
193 - use glib child watchers and try to work around its limitations,
194 if possible (based on a patch by Kevin Ryde).
195 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
196 - many read types in AnyEvent::Handle would malfunction during
197 transitions to TLS because of a "delete $self->{rbuf}".
198 (reported by Gerald Galster).
199 - improve condvar blocking wait performance (probably only noticable
200 with faster backends), also make it easier for event loops that
201 do not support blocking waits.
202 - PERL_ANYEVENT_MODEL now also allows full module names, although
203 the usefulness of this feature is not yet clear.
204 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
205 - add t/80_ssltext.t.
206 - no longer use AUTOLOAD for the initial loading (for no very good reason).
207 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
208 - do not run event loop detection multiple times when programs call their cached
209 copies of AnyEvent::detect.
210 - parse_hostport now accepts absolute paths als unix domain sockets.
211 - suppress the idiotic warning inside POE by patching the POE::Kernel
212 run flags instead of asking the user to jump through contortions.
213 - add Task::Weaken as dependency (as recommends only at this time).
214 - replace PF_UNSPEC by 0, for good old minix.
215
216 6.0 (not officially released, but was on CPAN for about 15 minutes :/)
217
218 5.34 Thu May 12 09:59:41 CEST 2011
219 - AE::TLS didn't support both cert_file and key at the same time,
220 due to some copy&paste bug (reported and analyzed by vti).
221 - improve AE::Handle timeout documentation slightly.
222
223 5.33 Mon Feb 7 21:51:41 CET 2011
224 - never appeared on CPAN, for unknown reasons.
225 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
226 object when it's already gone (analyzed by Ryan Bullock).
227 - update idna algorithm(s) to UTS#46 version 6.0.0.
228 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
229 uts46 says it's an error, and testsuite makes more sense.
230 - remove leading dots in idn_to_ascii - testsuite says to remove,
231 uts46 says it's an error, and testsuite makes more sense.
232 - fixed a bug in idn ignored character processing.
233 - passes the uts46 6.0.0 testsuite except for three cases which
234 apparently are bugs in the testsuite itself.
235 - passes the uts46 6.0.1 testsuite except for eleven cases which
236 apparently are bugs in the testsuite itself.
237
238 5.31 Sun Jan 23 11:49:19 CET 2011
239 - work around a POE bug where POE would first poll for new events and
240 invoke handlers BEFORE handling the message queue.
241 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
242 - stop_read now does just that in TLS mode. see its
243 documentation for caveats.
244 - update to common::sense 3.4 warning set (some warnings in perl impose
245 an arbitrary 32kb limit on regex matches, which affects
246 AnyEvent::Handle).
247 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
248 - new experimental AE::Handle::resettls method.
249
250 5.3 Fri Dec 31 05:48:13 CET 2010
251 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
252 issues with IO::Async have been solved in that version, including
253 the availabilty of a default loop!
254 - implement new wbuf_max parameter for AnyEvent::Handle.
255 - added Cocoa::EventLoop backend by Daisuke Murase.
256 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
257 - clarify handle on_prepare documentation.
258 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
259 to avoid leaking them to any child processes.
260 - try to force files into the arch-dependend lib directory, for the
261 half-"turly obsessive compulsive person"(s) that depend on it.
262
263 5.29 Sun Dec 5 10:49:21 CET 2010
264 - convert EV backend to EV 4.00 API (so better upgrade EV too).
265 - AE::Handle: implement dynamic read_size adjustment, add
266 max_read_size parameter, reduce default read_size to 2048.
267 - add some FAQ entries for servers and/or on_eof vs. on_error.
268 - work around OS bugs (cygwin again) when lingering in AE::Handle.
269 - reduce memory usage slightly when connect was used in AE::Handle.
270 - wrok around more segfaults in Tk.
271 - document tls_ctx => undef as valid.
272 - detect WNOHANG value at compiletime.
273 - include some probably linux-only support for building AnyEvent
274 as part of the perl core.
275 - improve compatibility of t/08_idna.t to perl 5.8.
276 - make AnyEvent::Strict truly optional (in case somebody builds a
277 minimal perl...).
278
279 5.28 Wed Oct 13 04:14:23 CEST 2010
280 - due to a glitch, AnyEvent's internal getprotobyname wasn't
281 actually used in tcp_server.
282 - implement AnyEvent::CondVar->new method, which was documented but
283 missing.
284 - support autoloading of read types in unshift_read.
285 - AnyEvent::Strict: do the test for invalid fh arguments differently,
286 resulting in more informative error messages.
287 - new AnyEvent::FAQ manpage.
288 - clarify that you can actually call push_read etc. in on_read.
289 - put keys and certs in /tmp, not into the current directory,
290 and document this in the SECURITY CONSIDERATIONS sections
291 of AnyEvent::TLS.
292 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
293 requested by Ron Isaacson.
294 - document $AnyEvent::DNS::RESOLVER variable.
295 - incorporated major typo patches by Abhijit Menon-Sen.
296 - AnyEvent::Handle now croaks when negative timeouts are passed.
297 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
298 undocumented, but who knows when they might come in handy.
299
300 5.271 Tue Jun 8 12:05:46 CEST 2010
301 - backport to perl 5.8.x.
302
303 5.27 Sun Jun 6 12:12:05 CEST 2010
304 - postpone differently in AnyEvent::Socket now, as
305 when not, canceling the connection attempt might fail
306 (found by Felix Antonius Wilhelm Ostmann).
307 - explicitly check for non-stream sockets in AE::Handle, too many
308 clueless people fell into the trap of this somehow working.
309 - simplified and reworked the "OTHER MODULES" section.
310 - better/more condvar examples.
311
312 5.261 Wed Apr 28 16:13:36 CEST 2010
313 - AF_INET6 was not properly used from Socket6 during configuration
314 time (found by Andy Grundman).
315
316 5.26 Mon Apr 12 04:49:35 CEST 2010
317 - don't generate spurious readyness notifications when select
318 returns EINTR in the pure perl backend (analysed and testcase
319 by Bas Denissen).
320 - give same examples in SYNOPSIS sections of AnyEvent and AE.
321 - provider faster implementations for AE::time/now/now_update for
322 EV and Perl backends.
323 - provide faster AE::cv and AE::time implementations for all
324 backends.
325 - no longer support lower/mixed-case signal names (this was
326 never documented nor universally supported).
327 - some more memory µ-optimisations, and somewhat less messy
328 code to cope with both AnyEvent and AE APIs.
329
330 5.251 Sat Mar 13 00:58:21 CET 2010
331 - make test failed if it was run before make install on perl <=
332 5.8.8, and older versions of perl make our live very difficult,
333 so write constants.pl during Makefile.PL time.
334
335 5.25 Sat Mar 13 00:23:14 CET 2010
336 - fix a race condition in AnyEvent::Handle that would cause
337 a "bio_read: unsupported method" error (in _another_
338 ssl connection) after user code threw an exception.
339 - added AnyEvent::Handle->destroyed method.
340 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
341 - the AnyEvent::Util::fh_nonblocking function worked only
342 by lucky accident on win32.
343 - smaller and faster AnyEvent::Util::fh_nonblocking.
344 - when the (required!) Time::HiRes module is missing, AnyEvent
345 did not fall back to built-in time properly.
346 - do not load Fcntl at runtime, saving memory and loading time.
347 - precompile a number of constants and use them instead of runtime
348 detection and eval.
349 - free detection code after detection and similar memory optimisations.
350 - Perl backend timer interval best effort drift has been improved
351 (same algorithm as EV).
352 - update unicode idna mapping table.
353
354 5.24 Tue Jan 5 11:39:43 CET 2010
355 - cygwin never reports errors from failed connects. 1.5 just gives
356 you continous readyness and EAGAIN, 1.7 is even more broken and
357 just hangs. work around both issues in a cygwin-specific hack.
358 - improve idle watcher documentation slightly.
359
360 5.23 Sun Dec 20 23:48:00 CET 2009
361 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
362 - implement punycode_encode/decode, idn_nameprep,
363 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
364 - provide $AE::VERSION.
365 - removed traces of "no strict 'refs'".
366
367 5.22 Sat Dec 5 03:51:13 CET 2009
368 - downgrade-or-fail in AnyEvent::Handle::push_write, to
369 diagnose encoding failures earlier and more succinctly.
370 (this works around bugs in perl, throwing away encoding info
371 when passing scalar data to extensions).
372 - add more examples to AnyEvent::Socket manpage.
373 - upgrade internal warning set to the same as common::sense 2.03.
374 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
375 (John Beppu).
376 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
377 - add google public dns servers to fallback server set and make sure
378 we load-balance properly between the three sets. also add all
379 fallback dns servers, not just a random one, to each dns config.
380
381 5.21 Thu Nov 19 02:48:47 CET 2009
382 - fix a problem where socket constants were called with parameters
383 (spotted by David Friedland).
384 - fork_call never use'd POSIX (reported by Daisuke Maki).
385 - improve perl 5.6 compatibility further (but it still won't work
386 unless you rip out everything but the core).
387 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
388 uses 10MB of RAM, but doesn't open a console window. *sigh*.
389
390 5.202 Wed Oct 14 22:35:44 CEST 2009
391 - AnyEvent::DNS would unexpectedly clobber $_ under windows
392 (analysed by Matthias Waldorf).
393 - AnyEvent::Handle::run_cmd can now pass the PID of the
394 newly-created process, which is much less useful than it might
395 sound (based on patch by Yann Kerherve).
396
397 5.201 Tue Sep 29 12:09:25 CEST 2009
398 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
399 (reported by Torsten Foertsch).
400 - common::sense 2.0 could cause tcp_server to throw an exception
401 (analysed by elmex).
402
403 5.2 Mon Sep 14 07:04:49 CEST 2009
404 - INCOMPATIBLE CHANGE: do no longer support register_read_type
405 and register_write_type in AnyEvent::Handle, instead support
406 package names (the facility was mostly abused).
407 - implement "packagename-as-read/write type" support in
408 AnyEvent::Handle.
409 - AnyEvent::Handle: new options "keepalive" and "oobinline".
410 - oobinline set by default to avoid security issues.
411 - the pure-perl event loop backend wrongly detected times() fallback
412 support (spotted by Pavel Boldin).
413
414 5.12 Tue Sep 1 20:26:50 CEST 2009
415 - be more lenient when parsing resolv.conf files, as some people
416 use hashmarks as comment indicator inside directives (reported
417 by Michael S. Fischer).
418 - use same set of warnings as common::sense 2.0.
419 - fix a potential 32 bit overflow issue due to perl having problems
420 with large hex constants in the Perl backend.
421 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
422 by default.
423 - fix some prototypes.
424 - (optionally) require Async::Interrupt 1.02.
425 - replace opendns fallbacks by verizon/level3 public dns servers.
426
427 5.112 Fri Aug 21 13:59:04 CEST 2009
428 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
429 using starttls with an illegal/uncreatable tls context.
430 - port to broken win32 perls.
431
432 5.111 Sun Aug 16 18:44:36 CEST 2009
433 - perl errornously requires sockaddr_un structures returned by the
434 kernel to have a specific length. work around this bug by padding
435 sockaddr_un structures (found to fail on OS X, as perl only uses
436 the correct code for GNU/Linux, but likely to fail on other
437 OSes as well).
438
439 5.11 Wed Aug 12 17:49:37 CEST 2009
440 - fixed a bug in run_cmd with input from filename.
441 - create a json coder object when none is given in write direction
442 to reduce dependency on JSON.
443
444 5.1 Tue Aug 11 03:17:32 CEST 2009
445 - tcp_connect now reports the correct error in $!, not an unrelated
446 one (was broken since 4.91).
447 - AE::Handle did not properly clear rtimeout/wtimeout watchers
448 when retrying a tcp_connect.
449 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
450 - fixed wrong documentation in AE::Socket.
451
452 5.01 Mon Aug 10 03:16:32 CEST 2009
453 - last release broke EV child watchers :/.
454
455 5.0 Sun Aug 9 17:32:09 CEST 2009
456 - officially introduce and document the AE API.
457 - lowering the timeout at runtime did not have immediate effect
458 in AnyEvent::Handle.
459 - AE::Handle now has separate and independent read and write timeout
460 settings.
461 - probe for Guard module when AnyEvent::Util is loaded, not at
462 runtime (Event doesn't like eval "use" from callbacks, and it avoids
463 exporting the probe function).
464 - Impl::IOAsync's io watchers did not use the filehandle emulation
465 code to work around it's watcher limits.
466 - work around IO::Async's problems with overloaded objects.
467 - try to disable wearnings inside IO::Async, it's just too buggy.
468 - do no longer use Event by default, only EV or Perl, Event is too buggy.
469 - hackishly provide fast AE:: interface when EV or Perl is the backend.
470 - inherit VERSION from AnyEvent::VERSION in most modules.
471
472 4.91 Thu Aug 6 15:42:45 CEST 2009
473 - AE::Handle::starttls could get out of sync when the read buffer
474 already contains some TLS handshake.
475 - AE::Handle did not properly free the TLS session, causing bigger
476 memory leaks in Net::SSLeay than expected (reported by toaster).
477 - AE::Socket::tcp_connect will now artificially delay invoking
478 the callback to avoid returning after invoking the callback.
479 - convert many internal watcher uses to AE API.
480
481 4.9 Sat Aug 1 11:07:01 CEST 2009
482 - Glib silently fails when registering a timeout with
483 a negative value, so avoid that.
484 - call condvar callback immediately when it is set
485 after the condition is already signalled.
486 - check rbuf_max condition only after trying to consume
487 data.
488 - normalise signal numbers to names when using the ae's
489 signal handling, but do not document this (yet).
490 - pure perl signal emulation did not properly set nonblocking
491 mode on the signal pipe on win32 (this is unlikely to have caused
492 any issues).
493 - new module: AnyEvent::Debug.
494 - AnyEvent::Strict now checks that a signal was specified
495 by name (not name or number).
496 - reduce memory footprint in typical cases by ~50kb by
497 only compiling signal code when necessary.
498 - add AnyEvent::Handle->rbuf_max.
499 - grab the AE namespace for future new API, implement stubs for
500 said future/alternative API.
501 - new function: AnyEvent::Socket::format_hostport.
502
503 4.881 Tue Jul 28 12:51:53 CEST 2009
504 - work around a bug in local in pre-5.10 perls, causing
505 AnyEvent::Handle to recurse when it should not
506 (analysed by elmex).
507
508 4.88 Tue Jul 28 04:04:37 CEST 2009
509 - re-bless the handle into a dummy package after calling
510 AnyEvent::Handle::destroy, so the user does not need to check
511 for errors after every push_write etc.
512 - do not attempt to run t/02_signals.t on obviously broken
513 platforms.
514
515 4.87 Sun Jul 26 02:06:16 CEST 2009
516 - do not attempt to linger when there is no longer a valid fh.
517 - enforce tls mode to be either accept or connect, do not simply
518 segfault in Net::SSLeay.
519 - AnyEvent::Handle can now call tcp_connect itself (new parameters
520 connect, on_prepare, on_connect and on_connect_error). Updated
521 tutorial accordingly.
522 - add AnyEvent::Impl::Irssi backend.
523
524 4.86 Mon Jul 20 23:52:29 CEST 2009
525 - since the verbose warning is not enough, explicitly document
526 that versions before 1.33 of Net::SSLeay are not secure.
527 - work around signal handling races in Event and (...) Event::Lib.
528 - try to align signal-race timer to full-second boundaries.
529 - work around Tk not liking negative timeouts.
530 - don't complain of different grades of environmental unfriendlyness
531 in IO::Async.
532
533 4.85 Sat Jul 18 06:16:14 CEST 2009
534 - nail the signal race problem in perl once and for all
535 (see $AnyEvent::MAX_SIGNAL_LATENCY).
536 - take advantage of Async::Interrupt if it is available.
537 - load Time::HiRes and Guard modules on demand only.
538 - add optional/recommended modules section to AnyEvent
539 documentation.
540 - reduce memory usage considerably (and reduce startup penalty)
541 by not using "strict", "warnings" and "overload" modules.
542 - work around buggy windows/openbsd perls and provide EBADMSG
543 and EPROTO ourselves when missing.
544 - improve perl 5.6 compatibility of the core event loop.
545 - made Net::SSLeay version 1.33 a soft requirement.
546
547 4.83 Fri Jul 17 16:56:26 CEST 2009
548 - implement AnyEvent::Socket::getprotobyname.
549 - AnyEvent::CondVar's will now detect recursive blocking
550 waits and will croak, as too many people fall into
551 this trap.
552 - AnyEvent::Handle will now call ->destroy on itself after
553 executing the on_error callback, instead of doing some
554 half-baked internal shutdown, for fatal errors.
555 - clarify on_eof behaviour w.r.t. the read queue and
556 on_read callbacks.
557 - ignore some possible spurious wake-ups in tcp_connect.
558
559 4.82 Sat Jul 11 00:34:55 CEST 2009
560 - POE and Event backends didn't accept some callable objects as
561 callbacks.
562 - use Config module instead of POSIX module to detect signal names
563 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
564 module doesn't even have all POSIX signals :/.
565 - use more workarounds around the many refcnt/corruption bugs in
566 Event::Lib.
567 - work around a race condition in perl's select, causing t/03_child.t
568 to rarely fail.
569
570 4.81 Thu Jul 9 10:30:30 CEST 2009
571 - AnyEvent::Handle didn't properly diagnose write errors
572 (it expected -1 from syswrite, how lame... :).
573 - support file descriptors in addition to file handles
574 in AnyEvent->io.
575 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
576 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
577 and PERL_ANYEVENT_CA_PATH.
578 - provide a sensible synopsis section for AnyEvent::TLS.
579 - add a "supported backends" section to the manpage.
580 - added simple io watcher test to testsuite, using a
581 portable_socketpair.
582 - tried to improve the stability of the Event::Lib backend,
583 YMMV.
584
585 4.8 Mon Jul 6 23:45:16 CEST 2009
586 - AnyEvent::DNS did not properly follow CNAME records with
587 uppercase targets.
588 - AnyEvent::DNS would errornously return AAAA records
589 with v4 mapped addresses (a faulty record) as ipv4 addresses,
590 causing AnyEvent::Socket to throw an exception.
591 - added new module AnyEvent::TLS for easier SSL/TLS context
592 creation, with many options including hostname verification,
593 secure default configuration, lots of documentation and,
594 predefined diffie-hellman keys for perfect forward security
595 and much more. get it while it's still fresh!
596 - use AnyEvent::TLS in AnyEvent::Handle for context management.
597 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
598 so AnyEvent::Socket users have smaller memory footprint
599 in the common case.
600 - add AnyEvent::Handle->push_shutdown method.
601 - add an additional error message parameter to AnyEvent::Handle's
602 on_error callback (for TLS, $! is still available).
603 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
604 - make AnyEvent::Handle more robust against early conenction
605 failures (during new), and return C<undef> in that case
606 from the constructor.
607 - AnyEvent::Handle will now try to load only JSON::XS first,
608 then fall back to JSON.
609 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
610 and because it was documented that way.
611
612 4.451 Fri Jul 3 00:28:58 CEST 2009
613 - do not clear rbuf when shutting down an AnyEvent::Handle
614 object - doing so breaks AnyEvent::HTTP.
615
616 4.45 Mon Jun 29 22:59:26 CEST 2009
617 - a write error could cause AnyEvent::Handle to create
618 an I/O watcher with an undefined $fh.
619 - special-case mapped ipv4 addresses in both
620 AnyEvent::Socket::format_address and parse_address,
621 to treat them just like ipv4 addresses.
622 - updated and overhauled the AnyEvent::Intro doc.
623 - implement AnyEvent::Socket::format_ipv4/ipv6.
624 - slightly speed up portable_pipe/socketpair.
625 - expand condvar begin/end documentation.
626
627 4.42 Fri Jun 26 08:32:18 CEST 2009
628 - preliminary and neccesarily incomplete support for IO::Async.
629 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
630 case it was set to IGNORE, to ensure we can catch child statuses
631 even when the calling env acts stupidly.
632 - updated benchmarks with IO::Async, which performs very well.
633
634 4.412 Wed Jun 24 01:35:57 CEST 2009
635 - support an "untaint" attribute for AnyEvent::DNS
636 and set it on the default resolver.
637 - implement and document AnyEvent::Impl::Perl::loop.
638 - remove all anyevent-env variables from %ENV
639 when running in tainted mode.
640 - mention and extend the IO::Lambda benchmark.
641
642 4.411 Sun Jun 7 18:48:02 CEST 2009
643 - do not try to use F_SETFD on windows.
644
645 4.41 Thu May 14 06:40:11 CEST 2009
646 - work around issues in older perls (5.8.0?) when
647 a signal handler is deleted from the %SIG hash.
648 - use POSIX::_exit in child test, to avoid running
649 destructors.
650 - speed up CHLD handling by relying on SIGCHLD being
651 synchronously delivered, even when we roll our own
652 implementation.
653 - AnyEvent::DNS: add the "dname" resource record name
654 for cosmetic reasons.
655
656 4.4 Sun Apr 26 20:12:33 CEST 2009
657 - implemented idle watchers, where applicable.
658 - AnyEvent->time died when Event backend was in use.
659 - fix a memleak in the Tk backend.
660 - sped up Tk timer handling.
661 - clip negative "after" values to 0 in AnyEvent::Impl::Event
662 to avoid spamming out warning messages.
663 - fix Qt timers without interval.
664 - avoid Qt zero-timeout special case, in old libqt's.
665
666 4.352 Mon Apr 20 16:31:11 CEST 2009
667 - fix AnyEvent::Strict error messages for child watchers.
668 - fix/update Makefile.PL "recommends" versions.
669 - add AnyEvent->now_update.
670
671 4.351 Sat Apr 11 07:56:14 CEST 2009
672 - actually make the signal pipe work under win32.
673 - localise $! in signal handler to avoid changing $!, although
674 perl itself does not seem to save/restore errno either.
675 - set the cloexec flag on the signal pipe (normally set by perl too).
676
677 4.35 Fri Mar 27 11:48:20 CET 2009
678 - event models relying on AnyEvent's signal watcher emulation did
679 invoke the callback asynchronously, contradicting documentation
680 and causing signals to get lost (this includes AnyEvent's own
681 event loop). AnyEvent now uses the standard pipe trick to make
682 callback execution synchronous to the event loop.
683 - AnyEvent::Handle didn't free TLS context data on DESTROY
684 (patch by Pavel Shaydo).
685 - work around the ever-incompatibly-changing API of MakeMaker.
686 - document that changing global variables without restoring them
687 is a bad idea in Perl (noted by Adam Rosenstein).
688 - AnyEvent::Strict now barfs if ->io is passed a file.
689
690 4.34 Thu Feb 12 18:32:45 CET 2009
691 - separately buffer TLS read data, as otherwise the read queue could
692 deadlock as receiving data is not expected while draining the
693 read queue (which cna only happen with TLS).
694 - raise EBADMSG error on JSON decoding errors.
695 - fix some minor manpage bugs (reported by Maximilian Gaß).
696 - speed up select bitmask parsing quite a bit in the pure perl
697 backend.
698 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
699
700 4.331 Tue Jan 6 21:07:25 CET 2009
701 - socketpair fails on many vista machines because vista has
702 completely broken accept/getpeername and getsockname functions,
703 so we provide our own socketpair emulation that kind of works
704 (AnyEvent::Util::portable_pipe).
705 - new function: AnyEvent::Util::portable_socketpair.
706 - take advantage of the Guard module if it exists.
707
708 4.33 Fri Nov 21 02:35:40 CET 2008
709 - AnyEvent::Strict did errornously flag a fileno of 0 as
710 illegal.
711 - reduce memory usage and slightly speed up the pure perl backend
712 by only storing the file descriptor, not the file handle.
713 - add missing autocork method to AnyEvent::Handle (reported
714 by Adam Rosenstein).
715 - AnyEvent::DNS->resolve errornously documented an $rcode result
716 argument, but there isn't (spotted by Henrik Krohns).
717 - the naked truth about Tk - it's basically unmaintained.
718
719 4.32 Mon Nov 3 22:46:32 CET 2008
720 - fix AnyEvent::Socket::resolve_sockaddr to properly support
721 unix sockets again and choose a proper default for
722 the socket type.
723 - fix call to resolve_sockaddr in tcp_connect.
724
725 4.31 Thu Oct 30 04:41:48 CET 2008
726 - implemented AnyEvent::Handle->destroy method.
727 - hint about unexpected effects in TLS mode.
728 - speed up AnyEvent::Impl::Perl by using more arrays
729 instead of hashes without the slightest loss of readability :)
730 - work around a perl argument refcounting bug.
731 - with some perl patches applied, AnyEvent now seems leak-free.
732
733 4.3 Fri Oct 3 09:18:43 CEST 2008
734 - AnyEvent will now install a no-op signal handler for SIGPIPE,
735 unless one has been installed already.
736 - warn about not loaidng AnyEvent::Impl::POE early enough
737 (patch by Adam Rosenstein).
738 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
739 - fix a bug in where in SSL connect mode, the
740 client would first wait for some data by the server and
741 otherwise hang. (reported and analysed in an absolutely
742 exemplary manner by Adam Rosenstein).
743 - fix a bug in where SSL EOF would not be treated
744 as stream EOF, putting the connection into a hung state.
745 (reported and analysed in an absolutely exemplary manner by Adam
746 Rosenstein).
747 - fix a potential segfault when the TLS context
748 would go missing in a read callback (e.g. due to stoptls) -
749 Net::SSLeay of course makes no type checking whatsoever.
750 - AnyEvent::Handle will not stop the read watcher in TLS mode.
751 - AnyEvent::Handle->stoptls will send a TLS close notify
752 instead of simply destroying the stream now.
753 - fix a bug where large blocks of data written in TLS mode
754 would not be sent unless triggered by receives.
755 - on_drain will now take the tls write buffer into account.
756 - SSL operations have been streamlined a bit and should be faster
757 now, more to come, though.
758 - added a FAQ section, will document nontrivial issues in there.
759
760 4.234 Mon Sep 29 04:08:13 CEST 2008
761 - fix child watcher documentation: only child exits will be reported,
762 no trace events.
763 - mention SIGPIPE in AnyEvent::Handle.
764 - perl backend will now use POSIX::times as monotonic clock
765 when available and the monotonic clock option is not.
766 - run even without Time::HiRes in some cases now, but do not
767 provide sub-second accuracy (all *supported* perl versions have
768 Time::HiRes).
769
770 4.233 Fri Aug 22 01:48:05 CEST 2008
771 - fix a bug in the testsuite, causing a hang.
772 - clarified lots of AnyEvent::Handle constructor arguments.
773
774 4.232 Thu Aug 21 20:44:25 CEST 2008
775 - fix a bug in AnyEvent::Handle that could cause two fatal error
776 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
777 (testcase by Leon Brocard).
778 - clarified on_eof/on_error documentation a bit.
779
780 4.231 Tue Jul 29 13:12:15 CEST 2008
781 - remove some debugging code left in AnyEvent::Util::fork_call
782 (and no, it's impossible to implement with the broken windows
783 perls without resource leaks or worse).
784
785 4.23 Tue Jul 29 12:19:59 CEST 2008
786 - document the first parameter passed to condvar callbacks to be
787 the callback.
788 - add AnyEvent::Socket::{ntoa,aton} aliases.
789 - optimize the AE::Handle->push_read (line) for the default
790 eol marker.
791 - optimize push_read (packstring|storable) for small packets.
792 - invoke on_error callback when no on_eof callback is set.
793 - fix a bug in push_read (storable) of unknown impact.
794
795 4.22 Sun Jul 20 16:34:13 CEST 2008
796 - new function AnyEvent::Socket::parse_hostport.
797 - as the bulkheads at microsoft can't even get getprotobyname reliably
798 working on their shitty fucking broken os we need to hardcode
799 some common protocol numbers in AnyEvent::Socket. How can
800 people even bother with such a pile of shit as windows.
801
802 4.21 Thu Jul 17 14:40:05 CEST 2008
803 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
804 which is incompatible to previous invocation syntax.
805 - work around more windows perl bugs in fork_call. I cannot imagine
806 why anybody would bother implementing fork in such an obviously
807 unusable way.
808 - avoid calling $do_retry if the request has already finished in
809 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
810 K. Sharifulin).
811
812 4.2 Sat Jul 12 22:42:11 CEST 2008
813 - support an "interval" argument for repeating timers.
814 - fix a bug in the pure perl event loop that caused timers to
815 be delayed under some circumstances.
816 - implement optional argument checking via PERL_ANYEVENT_STRICT.
817 - not importing anything from Event causes it to malfunction
818 and not use Time::HiRes, work around this bug.
819 - more bug workarounds against the endlessly broken Tk module,
820 seems to work now on 32 bit machines, but less so on 64 bit.
821 - minor optimisations applied to most backends.
822
823 4.161 Fri Jul 4 14:24:48 CEST 2008
824 - fixed the t/handle/02_write.t, it read after write, with the
825 assumption that all data could be written before read
826 (reported by Jim Razmus).
827
828 4.160 Thu Jul 3 04:02:21 CEST 2008
829 - re-registering signal watchers after unregistering one did
830 not work if a backend falls back on AnyEvent's default
831 implementation (also affected child watchers).
832 - new AnyEvent::Handle options: autocork and no_delay.
833
834 4.152 Sun Jun 22 14:15:44 CEST 2008
835 - allow for 32-bit perls that implement shifts differently
836 on different architectures in parse_ipv4 (reported and
837 analysed by Keiichi DAIBA).
838
839 4.151 Fri Jun 6 17:34:24 CEST 2008
840 - make sure specifying _only_ on_read and never pushing reads
841 works.
842
843 4.15 Fri Jun 6 13:00:46 CEST 2008
844 - the pure perl backend would keep some watchers alive when more than
845 one watcher was registered for the same fd.
846 - new "packstring" and "storable" read and write types
847 for AnyEvent::Handle.
848 - allow on_eof handler to be called after on_error with EPIPE returns.
849 - do not immediately call on_read callback in handle constructor.
850
851 4.14 Thu Jun 5 20:29:31 CEST 2008
852 - Fixed a bug in DNS SRV priority sorting.
853 - AnyEvent::Util::guard now reports runtime errors while
854 executing the guard block as warnings.
855 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
856
857 4.13 Thu Jun 5 00:47:59 CEST 2008
858 - AnyEvent::DNS only followed cname chains with length 2,
859 contrary to documentation. bumped it up to 10, thanks to
860 microsoft, the current broken-dns-config-king.
861 - AnyEvent::DNS didn't check for socket return status
862 but instead relied on perl not creating filehandles in that
863 case - too bad it gives you a bogus file handle (reported
864 and analysed by Vladimir Timofeev).
865 - fix queue management logic in AnyEvent::Handle:
866 when on_read was registered, the queue was empty
867 and no progress could be made AnyEvent::Handle would
868 enter an endless loop.
869 - correctly start reading again when the handle became
870 busy again after idling.
871 - correctly treat tls shut-downs as EOF condition
872 (for the time being, would be nice to have a callback for that).
873 - correctly call eof callback under all conditions
874 (hopefully).
875 - the timeout callback did not expect that $self can go
876 away any time.
877
878 4.12 Tue Jun 3 10:58:04 CEST 2008
879 - include AnyEvent::Intro, a tutorial for anyevent,
880 anyevent::socket and anyevent::handle.
881 - allow more options in on_error.
882
883 4.11 Fri May 30 23:42:25 CEST 2008
884 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
885 reverse_lookup and reverse_verify to replace it, support
886 v4mapped and v4compat addresses.
887 - provide more documentation for the resolver class.
888 - really replace longest run of :0: by :: in format_address,
889 also properly convert :: and ::1 again.
890 - support NAPTR record name and decode it.
891 - implement random weight sampling for SRV records, as per
892 rfc 2782.
893 - correctly abort on srv-record targets of ".".
894 - added AnyEvent::DNS::wait_for_slot.
895 - in the unlikely event of a virtual circuit connection
896 being invalidated by a delayed udp reply, AnyEvent::DNS
897 could die. this has been fixed.
898 - plug a probable memleak in the DNS vc code.
899 - use configured timeout also for VC request phase.
900 - implement timeout and max_outstanding methods
901 for resolver class.
902 - update version numbers in all modules.
903
904 4.1 Thu May 29 05:45:40 CEST 2008
905 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
906 which mirrors their purpose better (old names still available
907 for a while). Also moved AnyEvent::DNS::addr to
908 AnyEvent::Socket::resolve_sockaddr.
909 - implement AnyEvent->time and AnyEvent->now.
910 - fix IPv6 support in pack_sockaddr.
911 - officially un-experimentalise all newly introduced code.
912 - support unix domain sockets everywhere by specifying
913 a host of "unix/" and the pathname as service.
914 - implement an activity timeout in AnyEvent::Handle.
915 - added a regex read type to AnyEvent::Handle.
916 - added a json read/write type to AnyEvent::Handle.
917 - always croak in on_error in AnyEvent::Handle.
918 - document how to subclass AnyEvent::Handle.
919 - implement AnyEvent::Util::fork_call.
920 - add support for IPv6 nameservers and nameserver statements.
921 - work around _yet_ _another_ windows perl bug in where empty
922 select masks cause errors in the pure perl backend (this
923 workaround cannot be implemented for the other event loops,
924 but EV also works around this).
925 - supply AnyEvent::Socket::address_type and make good use of it.
926 - clarify and fix the Handle documentation w.r.t $self vs. $handle
927 vs. callback arguments.
928 - add some recommends to the META.yml.
929
930 4.05 Mon May 26 19:44:06 CEST 2008
931 - some platforms ignore the rfc and prepend an extra sa_len member to
932 the sockaddr structure, cater for those. also use sockaddr_family,
933 didn't know it was there, but comes in handy for the workaround.
934 - undo the import WIN32 hack.
935
936 4.04 Mon May 26 08:03:31 CEST 2008
937 - try to work around yet another windows bug: failed connects are reported
938 as if it were out-of-band data. windows users: you suck. Only
939 EV, Glib, Event and the pure perl backend can handle this condition.
940 - optimize write algorithm in AnyEvent::Handle.
941 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
942 - tcp_connect now properly iterates through all targets.
943 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
944 support entirely. also hardcode some AF_INET6 constants for the
945 sake of perl 5.8.
946 - vastly improved nameserver/suffix detection algorithm for windows
947 (still a hack...)
948 - try to use a fallback dns server if no dns servers could be found.
949 - splatter around some local $SIG{__DIE__}'s to work around bad
950 die handlers.
951
952 4.03 Sun May 25 05:05:57 CEST 2008
953 - work around perl bugs on windows, where perl returns undocumented
954 error codes for sysread, syswrite etc.
955 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
956 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
957 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
958 - updated documentation.
959 - lots of workarounds for perl 5.8 quirks.
960 - implement a more flexible type system for both reads and writes
961 in AnyEvent::Handle.
962 - support IPv6 in tcp_server.
963 - pass local hostname and port to the prepare callback in tcp_server.
964 - make types extendable in AnyEvent::Handle.
965 - croak, not die, in AnyEvent::Handle;
966
967 4.0 Sat May 24 19:58:08 CEST 2008
968 - added (experimental) AnyEvent::DNS module.
969 - added (experimental) AnyEvent::Socket module.
970 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
971 - make condvar->end call ->send as documented.
972 - make condvar constructor support the documented cb => argument.
973 - overload condvars so you can use them instead of code references.
974 - support more of the POSIX forms instead of a pure dotted quad for
975 verbatim ip addresses.
976 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
977 - try very hard not to freeze on broken windows perls.
978 - make on_eof optional in AnyEvent::Handle.
979 - support service names in tcp_connect and tcp_server.
980 - work around netbsd/openbsd bug causing unix sockets not to
981 become writable until empty.
982 - only make a dummy read for error detection when getpeername
983 returns ENOTCONN.
984 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
985 override.
986 - new AnyEvent::Guard::cancel.
987
988 3.5 Sat May 17 23:17:14 CEST 2008
989 - gracefully deal with systems not defining CLOCK_MONOTONIC
990 (as opposed to not supporting the clock itself, which worked in 3.41).
991 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
992 helpers, to replace the ill-designed AnyEvent::Socket class.
993 - added some examples (eg/connect, eg/handle).
994 - added AnyEvent::Util::guard.
995 - AnyEvent::Util now exports most of its functions.
996 - fixed a bug in *_read_chunk, where the callback was called with $self as
997 first and second argument.
998 - simplified and fixed regex handling in AnyEvent::Handle::*_read_line.
999 - fix manpage condvars examples to use new syntax.
1000 - considerably improve inet_aton with EV::ADNS backend to support
1001 cname chains (common dns configuration bug, e.g. with www.google.de)
1002 and aaaa records if no a record could be found.
1003
1004 3.41 Sun May 11 19:53:13 CEST 2008
1005 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
1006
1007 3.4 Sun May 11 00:29:25 CEST 2008
1008 - complete re-implementation of condvars with a hopefully much more useful
1009 API (backward compatible functions exist but are undocumented).
1010 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
1011 (still subject to change).
1012 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
1013 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which
1014 allows module authors to avoid forcing event loop detection.
1015 (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example).
1016 - remove coro backends: Coro now provides generic support for AnyEvent
1017 via Coro::AnyEvent.
1018
1019 3.3 Mon Apr 28 09:51:06 CEST 2008
1020 - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules.
1021 - fix a bug in the pure perl backend that kept watchers alive
1022 when multiple watchers were registered for the same fh/poll combo.
1023 - add a benchmark section showing AnyEvent overhead and comparing
1024 the different event loops with each other.
1025 - prefer pure perl over tk when autoprobing, as it's about as fast,
1026 but doesn't crash with many watchers.
1027 - declare Qt support non-experimental.
1028 - clarify the confusing section about the file descriptor being
1029 kept alive.
1030 - document the race between loading of an event module and
1031 child processes exiting.
1032 - support POE as "backend" (with some caveats, POE is not generic
1033 enough, and darn slow).
1034 - support Wx and Prima through POE.
1035 - optimise perl backend to use 20% less memory and take advantage
1036 of typical timeout behaviour. It can now compete with
1037 select/poll-based C event loops in most cases (it is usually
1038 faster than Event and Glib :).
1039 - roughly cut EV memory use in half and increase its speed by 30%,
1040 by removing undocumented functionality. Did something similar
1041 to other event loops where possible.
1042
1043 3.2 Thu Apr 24 10:10:40 CEST 2008
1044 - do not die when anyevent watchers are destroyed while running callbacks
1045 in the pureperl backend (could only happen when two watchers are registered
1046 for the same fh/poll combo).
1047 - support autoloading for child watchers, was broken in all versions.
1048 - implement PERL_ANYEVENT_MODEL env variable.
1049 - (experimental) implement interface to the (very crashy/buggy) Event::Lib module.
1050 - (experimental) implement interface to the Qt module (cannot be autoprobed).
1051 - this release is IO:AnyEvent-proof.
1052
1053 3.12 Tue Apr 22 07:11:46 CEST 2008
1054 - reinstate AUTHOR section that got lost somehow.
1055 - do not hang in the testsuite with badly broken perls (activestate,
1056 strawberry...), but instead diagnose the problem and continue.
1057 - use INT instead of CHLD in an attempt to support broken windows perls
1058 better (this decreases the test quality, unfortunately).
1059 - do not send the signal to the process group
1060 (no problem for CHLD, bad for INT :).
1061
1062 3.11 Sat Apr 19 06:57:31 CEST 2008
1063 - major documentation rework.
1064 - document the fact that child watchers only watch for zombies.
1065 - fix the child watcher example.
1066
1067 3.1 Wed Apr 16 17:09:01 CEST 2008
1068 - work around recurring bugs in Tk by dup'ing filehandles, the
1069 only method with good success chances on Tk (the bugs apparently
1070 don't get fixed anytime soon).
1071 - lift the restriction of only one watcher per fh direction
1072 (as the Tk bug workaround also lifts it and only Tk imposed
1073 such strong limits).
1074 - changed probe order to prefer coro adaptors.
1075 - explain why recursion into the event loop is not supported
1076 unless the backend supports it (only Coro::EV does without
1077 any restrictions...).
1078 - add simple manpages for all backend modules.
1079
1080 3.0 Mon Apr 7 21:30:23 CEST 2008
1081 - Coro::Signal changed semantics, roll our own, also cleaning
1082 up the Coro implementation in general.
1083 - rename Coro backend to CoroEvent.
1084 - add some decision helping paragraph to the manpage that should
1085 help people to decide whether AnyEvent is the right thing for them.
1086
1087 2.9 Mon Jan 28 13:31:54 CET 2008
1088 - update for EV 3.0 API changes.
1089
1090 2.8 Sun Nov 25 15:06:03 CET 2007
1091 - waitpid can validly return 0. accept this fact of life
1092 instead of reporting it to any watchers.
1093
1094 2.7 Fri Nov 23 11:41:14 CET 2007
1095 - force use of AnyEvent::Impl::Perl in testsuite, there is
1096 too much breakage outside AnyEvent.
1097 - deliver signals synchronously in AnyEvent::Impl::Perl
1098 or any other event loop that relies on AnyEvents child
1099 watcher emulation. *Could* help with hanging testsuite
1100 (except when Event or EV are installed).
1101
1102 2.6 Fri Nov 9 20:36:35 CET 2007
1103 - fix bug in testsuite.
1104 - move EV adaptor modules to AnyEvent.
1105 - add Coro+EV adaptor module.
1106
1107 2.55 Tue Nov 6 17:41:32 CET 2007
1108 - add EV to the list of supported event models.
1109 - do not auto-reset pid watchers, pass pid and status to them.
1110 - allow a pid of zero to watch for all children in child watchers.
1111
1112 2.54 Wed Jul 18 17:36:23 CEST 2007
1113 - work around a perl bug that results in
1114 BEGIN not safe after errors--compilation aborted without
1115 any discernible reason or error message by once
1116 again not use'ing strict.
1117
1118 2.53 Sun Jul 8 10:51:53 CEST 2007
1119 - make Glib interface work again (spotted by elmex).
1120
1121 2.52 Wed Mar 7 18:36:16 CET 2007
1122 - child watchers stopped working permanently when
1123 all current child watchers were destroyed.
1124
1125 2.51 Mon Dec 11 21:33:24 CET 2006
1126 - work around bugs in perl where eval "require Module" returns
1127 true even if the module couldn't be loaded :/.
1128
1129 2.5 Mon Dec 11 02:15:28 CET 2006
1130 - avoid $AUTOLOAD because many perls corrupt it.
1131 - AnyEvent::detect forces autodetection.
1132 - implement signal watchers (experimental)
1133 (Tk does not support async signals (<= 804.027 at least),
1134 so they might get delayed indefinitely on Tk).
1135 - implement child watchers (experimental).
1136 - moved default condvar implementation into base module, simplifying
1137 most implementation modules.
1138
1139 2.1 Fri Nov 24 15:50:48 CET 2006
1140 - better docs.
1141 - simple testsuite.
1142 - added AnyEvent->one_event method for special purposes.
1143
1144 2.0 Wed Nov 1 02:21:30 CET 2006
1145 - INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both.
1146 - INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore.
1147 - EXPERIMENTAL:
1148 new pure-perl model added, AnyEvent now always finds a suitable
1149 event model.
1150 - improved documentation.
1151 - reduced cpu overhead.
1152
1153 1.02 Fri Jan 13 14:15:40 CET 2006
1154 - add COPYING to clarify license.
1155
1156 1.01 Sun Jan 8 05:48:09 CET 2006
1157 - correctly cancel Event timers, avoid memory leaks.
1158
1159 1.0 Sun Jan 8 05:40:58 CET 2006
1160 - allow third-party interfaces to be added at runtime.
1161 - fix version number.
1162
1163 0.9 Fri Dec 30 02:25:41 CET 2005
1164 - fixed errors in the documentation.
1165 - going back to using $VERSION as "load check",
1166 as suggested by Jörn Reder. Let's see why I stopped using
1167 this method.
1168 - add 'hup' to the list of events to receive for glib
1169 and generate 'rw' events for it, as some versions of
1170 glib seem to require this, while this breaks other
1171 (older) versions of glib.
1172 - implemented PERL_ANYEVENT_VERBOSE env variable.
1173
1174 0.3 Sun Dec 4 10:43:33 CET 2005
1175 - no changes to the code.
1176 - fixed documentation.
1177 - clarified some corner cases.
1178 - vastly improved documentation with a long example.
1179
1180 0.2 Thu Dec 1 22:18:49 CET 2005
1181 - check whether a valid method was called, to avoid endless loops.
1182 - move condvars into their own class (cleaner, safer).
1183 - don't die when some modules are not available.
1184 - Tk is too broken to allow for words. Will work once Tk works.
1185
1186 0.1 Thu Dec 1 19:44:57 CET 2005
1187 - spiced up and cleaned up for release.
1188
1189 0.01 Tue Sep 21 18:12:09 CEST 2004
1190 - original version; copied from Convert::Scalar.
1191