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

Comparing AnyEvent/Changes (file contents):
Revision 1.63 by root, Thu May 1 11:04:24 2008 UTC vs.
Revision 1.547 by root, Fri Apr 13 04:38:13 2012 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: common pitfalls?
7
87.0 Fri Apr 13 06:33:30 CEST 2012
9 - child watchers are broken in POE 1.352 (also many earlier
10 versions) and there seems to be no way to work around it, as POE
11 itself is inherently racy. Document this regression and add a delay
12 in t/68_poe_03_child.t for the time being.
13 - new module AnyEvent::IO, that is a frontend to either a pure-perl
14 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
15 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
16 - load /etc/hosts only when DNS has no answer.
17 - stat /etc/hosts on every access and reload it if it changed.
18 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
19 - fix a buggy croak in the dh parameter rsading in AnyEvent::TLS.
20 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
21 a side effect, they now use true appending as opposed to libc
22 appending, and the file might not have been opened when the function
23 returns.
24 - the default logging level was not properly documented in a variety of
25 places, this has hopefully been rectified.
26 - updated uts46data.pl for unicode 6.1.0.
27 - made log messages genearted by AnyEvent submodules not
28 include the package name anymore, as it will be logged
29 by default already.
30 - upgrade to the trick used by common-sense 3.5 to work
31 around extra warning torture/breakage under perl 5.15.x.
32 - log messages by anyevent are now ucfirst, are usually full
33 sentences and do no longer include the package name.
34 - the storable read type would sometimes throw an exception instead
35 of causing EBADMSG (analyzed by Maxime Soulé).
36 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
37 keysize (reported by Paul Howarth).
38 - documented Rocco fabricating statements about AnyEvent and me.
39
406.14 Tue Jan 31 20:00:24 CET 2012
41 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
42 probably an editing mistake (reported by Darin McBride).
43
446.13 Thu Jan 12 07:27:01 CET 2012
45 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
46 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
47 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
48 regex-matching inside the callback would change the parameter, despite
49 $1 being dynamically scoped per-block. this perl bug is now being
50 worked around (testcase by Cindy Wang).
51
526.12 Mon Dec 12 13:21:10 CET 2011
53 - $! was clobbered by subsequent calls in tcp_connect, due to
54 the postpone not saving/restoring it (analyzed by Richard Garnish).
55
566.11 Tue Nov 22 10:36:05 CET 2011
57 - Tk cannot create windows when tainted unless you set a title, so
58 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
59 - escape any nonprintable/nonascii characters when stringifying
60 backtraces.
61 - log the reason loading an interface module fails to load at level debug.
62
636.1 Tue Oct 4 19:44:30 CEST 2011
64 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
65 and above), and some messages inside AnyEvent have been elevated
66 to higher log levels to print by default.
67 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
68 - work around yet another signal race bug, newly introduced by POE. My
69 how that module sucks.
70 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
71 support ipv6-only operation even with defaults (also use ::1 instead
72 of 127.0.0.1).
73 - give better guidelines on how to chose a logging level.
74 - AnyEvent::Log can now cap (limit) log message priority by context.
75 - AE::log will now save and restore $! and $@, so callers don't have to.
76 - AE::Strict now checks that registered fds are still valid at regular
77 intervals.
78 - skip frequently called 'require' calls if the module appears loaded.
79 - increase signal latency test timer considerably for poor little osx,
80 which can't be bothered with basic posix functionality such as signal
81 delivery.
82 - mention some debugging aids in the tutorial.
83 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
84 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
85 of the context as title.
86 - AnyEvent::Strict would not restore read-write access to $_ when a callback
87 died - this is a user bug, but since AE::Strict exists to help dbeugging...
88 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
89 to work with it - the author is well aware of the issues but refuses
90 to fix them. This extreme step has been taken because the module actively
91 attacks the goals of AnyEvent and makes it harder for module authors to use
92 AnyEvent.
93
946.02 Fri Aug 26 20:08:31 CEST 2011
95 - log_to_syslog now accepts facility strings.
96 - log_to_syslog would not allow facilities - "$facility|$level"
97 *literally* means that string. sheesh.
98 - the CPAN parser got confused and thought we provide Sys::Syslog, try
99 to work around.
100 - renamed AnyEvent::Impl::FLTK2 to FLTK.
101
1026.01 Fri Aug 26 07:04:11 CEST 2011
103 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
104 the TTL as fourth element - this affects ->resolve, ->request
105 and ->dns_unpack indirectly. Convenience functions such as
106 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
107 AnyEvent::Socket.
108 - INCOMPATIBLE CHANGE: backend authors now should not implement
109 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
110 methods.
111 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
112 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
113 - DEPRECIATION: please change
114 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
115 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
116 in your programs, for future compatibility.
117 - *blush* idle watcher emulation was borked.
118 - AnyEvent::Strict now checks the AE::xxx functions as well.
119 - querying too long domain names would barf deep inside AnyEvent::DNS - since
120 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
121 for too long domain names explicitly and passes an undef to the callback
122 instead (reported by James Bromberger).
123 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
124 - new functions: AnyEvent::postpone, AnyEvent::log.
125 - new module: AnyEvent::Log, for simple logging needs.
126 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
127 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
128 a debug shell.
129 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
130 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
131 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
132 /etc/hosts can be found.
133 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
134 configuration.
135 - env variables can now be specified using AE_xyz instead of
136 PERL_ANYEVENT_xyz, subject to some rules.
137 - lots of new AnyEvent::Debug functionality.
138 - AnyEvent::Strict now checks for modifications of $_ by the callback
139 (by setting it to readonly while executing the callback, if possible).
140 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
141 Invest extra overhead to work around this peculiar behaviour.
142 - split the pure perl backend into a loop module and an impl
143 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
144 - some read types (regex, netstring, json, storable) could
145 errornously keep an AE::Handle object alive on their own
146 (reported by Mohammad Toossi).
147 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
148 supports /etc/hosts overrides.
149 - add more workarounds around child watcher bugs in IO::Async
150 and (newly introduced) in POE.
151 - use glib child watchers and try to work around its limitations,
152 if possible (based on a patch by Kevin Ryde).
153 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
154 - many read types in AnyEvent::Handle would malfunction during
155 transitions to TLS because of a "delete $self->{rbuf}".
156 (reported by Gerald Galster).
157 - improve condvar blocking wait performance (probably only noticable
158 with faster backends), also make it easier for event loops that
159 do not support blocking waits.
160 - PERL_ANYEVENT_MODEL now also allows full module names, although
161 the usefulness of this feature is not yet clear.
162 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
163 - add t/80_ssltext.t.
164 - no longer use AUTOLOAD for the initial loading (for no very good reason).
165 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
166 - do not run event loop detection multiple times when programs call their cached
167 copies of AnyEvent::detect.
168 - parse_hostport now accepts absolute paths als unix domain sockets.
169 - suppress the idiotic warning inside POE by patching the POE::Kernel
170 run flags instead of asking the user to jump through contortions.
171 - add Task::Weaken as dependency (as recommends only at this time).
172 - replace PF_UNSPEC by 0, for good old minix.
173
1746.0 (not officially released, but was on CPAN for about 15 minutes :/)
175
1765.34 Thu May 12 09:59:41 CEST 2011
177 - AE::TLS didn't support both cert_file and key at the same time,
178 due to some copy&paste bug (reported and analyzed by vti).
179 - improve AE::Handle timeout documentation slightly.
180
1815.33 Mon Feb 7 21:51:41 CET 2011
182 - never appeared on CPAN, for unknown reasons.
183 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
184 object when it's already gone (analyzed by Ryan Bullock).
185 - update idna algorithm(s) to UTS#46 version 6.0.0.
186 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
187 uts46 says it's an error, and testsuite makes more sense.
188 - remove leading dots in idn_to_ascii - testsuite says to remove,
189 uts46 says it's an error, and testsuite makes more sense.
190 - fixed a bug in idn ignored character processing.
191 - passes the uts46 6.0.0 testsuite except for three cases which
192 apparently are bugs in the testsuite itself.
193 - passes the uts46 6.0.1 testsuite except for eleven cases which
194 apparently are bugs in the testsuite itself.
195
1965.31 Sun Jan 23 11:49:19 CET 2011
197 - work around a POE bug where POE would first poll for new events and
198 invoke handlers BEFORE handling the message queue.
199 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
200 - stop_read now does just that in TLS mode. see its
201 documentation for caveats.
202 - update to common::sense 3.4 warning set (some warnings in perl impose
203 an arbitrary 32kb limit on regex matches, which affects
204 AnyEvent::Handle).
205 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
206 - new experimental AE::Handle::resettls method.
207
2085.3 Fri Dec 31 05:48:13 CET 2010
209 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
210 issues with IO::Async have been solved in that version, including
211 the availabilty of a default loop!
212 - implement new wbuf_max parameter for AnyEvent::Handle.
213 - added Cocoa::EventLoop backend by Daisuke Murase.
214 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
215 - clarify handle on_prepare documentation.
216 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
217 to avoid leaking them to any child processes.
218 - try to force files into the arch-dependend lib directory, for the
219 half-"turly obsessive compulsive person"(s) that depend on it.
220
2215.29 Sun Dec 5 10:49:21 CET 2010
222 - convert EV backend to EV 4.00 API (so better upgrade EV too).
223 - AE::Handle: implement dynamic read_size adjustment, add
224 max_read_size parameter, reduce default read_size to 2048.
225 - add some FAQ entries for servers and/or on_eof vs. on_error.
226 - work around OS bugs (cygwin again) when lingering in AE::Handle.
227 - reduce memory usage slightly when connect was used in AE::Handle.
228 - wrok around more segfaults in Tk.
229 - document tls_ctx => undef as valid.
230 - detect WNOHANG value at compiletime.
231 - include some probably linux-only support for building AnyEvent
232 as part of the perl core.
233 - improve compatibility of t/08_idna.t to perl 5.8.
234 - make AnyEvent::Strict truly optional (in case somebody builds a
235 minimal perl...).
236
2375.28 Wed Oct 13 04:14:23 CEST 2010
238 - due to a glitch, AnyEvent's internal getprotobyname wasn't
239 actually used in tcp_server.
240 - implement AnyEvent::CondVar->new method, which was documented but
241 missing.
242 - support autoloading of read types in unshift_read.
243 - AnyEvent::Strict: do the test for invalid fh arguments differently,
244 resulting in more informative error messages.
245 - new AnyEvent::FAQ manpage.
246 - clarify that you can actually call push_read etc. in on_read.
247 - put keys and certs in /tmp, not into the current directory,
248 and document this in the SECURITY CONSIDERATIONS sections
249 of AnyEvent::TLS.
250 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
251 requested by Ron Isaacson.
252 - document $AnyEvent::DNS::RESOLVER variable.
253 - incorporated major typo patches by Abhijit Menon-Sen.
254 - AnyEvent::Handle now croaks when negative timeouts are passed.
255 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
256 undocumented, but who knows when they might come in handy.
257
2585.271 Tue Jun 8 12:05:46 CEST 2010
259 - backport to perl 5.8.x.
260
2615.27 Sun Jun 6 12:12:05 CEST 2010
262 - postpone differently in AnyEvent::Socket now, as
263 when not, canceling the connection attempt might fail
264 (found by Felix Antonius Wilhelm Ostmann).
265 - explicitly check for non-stream sockets in AE::Handle, too many
266 clueless people fell into the trap of this somehow working.
267 - simplified and reworked the "OTHER MODULES" section.
268 - better/more condvar examples.
269
2705.261 Wed Apr 28 16:13:36 CEST 2010
271 - AF_INET6 was not properly used from Socket6 during configuration
272 time (found by Andy Grundman).
273
2745.26 Mon Apr 12 04:49:35 CEST 2010
275 - don't generate spurious readyness notifications when select
276 returns EINTR in the pure perl backend (analysed and testcase
277 by Bas Denissen).
278 - give same examples in SYNOPSIS sections of AnyEvent and AE.
279 - provider faster implementations for AE::time/now/now_update for
280 EV and Perl backends.
281 - provide faster AE::cv and AE::time implementations for all
282 backends.
283 - no longer support lower/mixed-case signal names (this was
284 never documented nor universally supported).
285 - some more memory µ-optimisations, and somewhat less messy
286 code to cope with both AnyEvent and AE APIs.
287
2885.251 Sat Mar 13 00:58:21 CET 2010
289 - make test failed if it was run before make install on perl <=
290 5.8.8, and older versions of perl make our live very difficult,
291 so write constants.pl during Makefile.PL time.
292
2935.25 Sat Mar 13 00:23:14 CET 2010
294 - fix a race condition in AnyEvent::Handle that would cause
295 a "bio_read: unsupported method" error (in _another_
296 ssl connection) after user code threw an exception.
297 - added AnyEvent::Handle->destroyed method.
298 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
299 - the AnyEvent::Util::fh_nonblocking function worked only
300 by lucky accident on win32.
301 - smaller and faster AnyEvent::Util::fh_nonblocking.
302 - when the (required!) Time::HiRes module is missing, AnyEvent
303 did not fall back to built-in time properly.
304 - do not load Fcntl at runtime, saving memory and loading time.
305 - precompile a number of constants and use them instead of runtime
306 detection and eval.
307 - free detection code after detection and similar memory optimisations.
308 - Perl backend timer interval best effort drift has been improved
309 (same algorithm as EV).
310 - update unicode idna mapping table.
311
3125.24 Tue Jan 5 11:39:43 CET 2010
313 - cygwin never reports errors from failed connects. 1.5 just gives
314 you continous readyness and EAGAIN, 1.7 is even more broken and
315 just hangs. work around both issues in a cygwin-specific hack.
316 - improve idle watcher documentation slightly.
317
3185.23 Sun Dec 20 23:48:00 CET 2009
319 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
320 - implement punycode_encode/decode, idn_nameprep,
321 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
322 - provide $AE::VERSION.
323 - removed traces of "no strict 'refs'".
324
3255.22 Sat Dec 5 03:51:13 CET 2009
326 - downgrade-or-fail in AnyEvent::Handle::push_write, to
327 diagnose encoding failures earlier and more succinctly.
328 (this works around bugs in perl, throwing away encoding info
329 when passing scalar data to extensions).
330 - add more examples to AnyEvent::Socket manpage.
331 - upgrade internal warning set to the same as common::sense 2.03.
332 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
333 (John Beppu).
334 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
335 - add google public dns servers to fallback server set and make sure
336 we load-balance properly between the three sets. also add all
337 fallback dns servers, not just a random one, to each dns config.
338
3395.21 Thu Nov 19 02:48:47 CET 2009
340 - fix a problem where socket constants were called with parameters
341 (spotted by David Friedland).
342 - fork_call never use'd POSIX (reported by Daisuke Maki).
343 - improve perl 5.6 compatibility further (but it still won't work
344 unless you rip out everything but the core).
345 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
346 uses 10MB of RAM, but doesn't open a console window. *sigh*.
347
3485.202 Wed Oct 14 22:35:44 CEST 2009
349 - AnyEvent::DNS would unexpectedly clobber $_ under windows
350 (analysed by Matthias Waldorf).
351 - AnyEvent::Handle::run_cmd can now pass the PID of the
352 newly-created process, which is much less useful than it might
353 sound (based on patch by Yann Kerherve).
354
3555.201 Tue Sep 29 12:09:25 CEST 2009
356 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
357 (reported by Torsten Foertsch).
358 - common::sense 2.0 could cause tcp_server to throw an exception
359 (analysed by elmex).
360
3615.2 Mon Sep 14 07:04:49 CEST 2009
362 - INCOMPATIBLE CHANGE: do no longer support register_read_type
363 and register_write_type in AnyEvent::Handle, instead support
364 package names (the facility was mostly abused).
365 - implement "packagename-as-read/write type" support in
366 AnyEvent::Handle.
367 - AnyEvent::Handle: new options "keepalive" and "oobinline".
368 - oobinline set by default to avoid security issues.
369 - the pure-perl event loop backend wrongly detected times() fallback
370 support (spotted by Pavel Boldin).
371
3725.12 Tue Sep 1 20:26:50 CEST 2009
373 - be more lenient when parsing resolv.conf files, as some people
374 use hashmarks as comment indicator inside directives (reported
375 by Michael S. Fischer).
376 - use same set of warnings as common::sense 2.0.
377 - fix a potential 32 bit overflow issue due to perl having problems
378 with large hex constants in the Perl backend.
379 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
380 by default.
381 - fix some prototypes.
382 - (optionally) require Async::Interrupt 1.02.
383 - replace opendns fallbacks by verizon/level3 public dns servers.
384
3855.112 Fri Aug 21 13:59:04 CEST 2009
386 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
387 using starttls with an illegal/uncreatable tls context.
388 - port to broken win32 perls.
389
3905.111 Sun Aug 16 18:44:36 CEST 2009
391 - perl errornously requires sockaddr_un structures returned by the
392 kernel to have a specific length. work around this bug by padding
393 sockaddr_un structures (found to fail on OS X, as perl only uses
394 the correct code for GNU/Linux, but likely to fail on other
395 OSes as well).
396
3975.11 Wed Aug 12 17:49:37 CEST 2009
398 - fixed a bug in run_cmd with input from filename.
399 - create a json coder object when none is given in write direction
400 to reduce dependency on JSON.
401
4025.1 Tue Aug 11 03:17:32 CEST 2009
403 - tcp_connect now reports the correct error in $!, not an unrelated
404 one (was broken since 4.91).
405 - AE::Handle did not properly clear rtimeout/wtimeout watchers
406 when retrying a tcp_connect.
407 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
408 - fixed wrong documentation in AE::Socket.
409
4105.01 Mon Aug 10 03:16:32 CEST 2009
411 - last release broke EV child watchers :/.
412
4135.0 Sun Aug 9 17:32:09 CEST 2009
414 - officially introduce and document the AE API.
415 - lowering the timeout at runtime did not have immediate effect
416 in AnyEvent::Handle.
417 - AE::Handle now has separate and independent read and write timeout
418 settings.
419 - probe for Guard module when AnyEvent::Util is loaded, not at
420 runtime (Event doesn't like eval "use" from callbacks, and it avoids
421 exporting the probe function).
422 - Impl::IOAsync's io watchers did not use the filehandle emulation
423 code to work around it's watcher limits.
424 - work around IO::Async's problems with overloaded objects.
425 - try to disable wearnings inside IO::Async, it's just too buggy.
426 - do no longer use Event by default, only EV or Perl, Event is too buggy.
427 - hackishly provide fast AE:: interface when EV or Perl is the backend.
428 - inherit VERSION from AnyEvent::VERSION in most modules.
429
4304.91 Thu Aug 6 15:42:45 CEST 2009
431 - AE::Handle::starttls could get out of sync when the read buffer
432 already contains some TLS handshake.
433 - AE::Handle did not properly free the TLS session, causing bigger
434 memory leaks in Net::SSLeay than expected (reported by toaster).
435 - AE::Socket::tcp_connect will now artificially delay invoking
436 the callback to avoid returning after invoking the callback.
437 - convert many internal watcher uses to AE API.
438
4394.9 Sat Aug 1 11:07:01 CEST 2009
440 - Glib silently fails when registering a timeout with
441 a negative value, so avoid that.
442 - call condvar callback immediately when it is set
443 after the condition is already signalled.
444 - check rbuf_max condition only after trying to consume
445 data.
446 - normalise signal numbers to names when using the ae's
447 signal handling, but do not document this (yet).
448 - pure perl signal emulation did not properly set nonblocking
449 mode on the signal pipe on win32 (this is unlikely to have caused
450 any issues).
451 - new module: AnyEvent::Debug.
452 - AnyEvent::Strict now checks that a signal was specified
453 by name (not name or number).
454 - reduce memory footprint in typical cases by ~50kb by
455 only compiling signal code when necessary.
456 - add AnyEvent::Handle->rbuf_max.
457 - grab the AE namespace for future new API, implement stubs for
458 said future/alternative API.
459 - new function: AnyEvent::Socket::format_hostport.
460
4614.881 Tue Jul 28 12:51:53 CEST 2009
462 - work around a bug in local in pre-5.10 perls, causing
463 AnyEvent::Handle to recurse when it should not
464 (analysed by elmex).
465
4664.88 Tue Jul 28 04:04:37 CEST 2009
467 - re-bless the handle into a dummy package after calling
468 AnyEvent::Handle::destroy, so the user does not need to check
469 for errors after every push_write etc.
470 - do not attempt to run t/02_signals.t on obviously broken
471 platforms.
472
4734.87 Sun Jul 26 02:06:16 CEST 2009
474 - do not attempt to linger when there is no longer a valid fh.
475 - enforce tls mode to be either accept or connect, do not simply
476 segfault in Net::SSLeay.
477 - AnyEvent::Handle can now call tcp_connect itself (new parameters
478 connect, on_prepare, on_connect and on_connect_error). Updated
479 tutorial accordingly.
480 - add AnyEvent::Impl::Irssi backend.
481
4824.86 Mon Jul 20 23:52:29 CEST 2009
483 - since the verbose warning is not enough, explicitly document
484 that versions before 1.33 of Net::SSLeay are not secure.
485 - work around signal handling races in Event and (...) Event::Lib.
486 - try to align signal-race timer to full-second boundaries.
487 - work around Tk not liking negative timeouts.
488 - don't complain of different grades of environmental unfriendlyness
489 in IO::Async.
490
4914.85 Sat Jul 18 06:16:14 CEST 2009
492 - nail the signal race problem in perl once and for all
493 (see $AnyEvent::MAX_SIGNAL_LATENCY).
494 - take advantage of Async::Interrupt if it is available.
495 - load Time::HiRes and Guard modules on demand only.
496 - add optional/recommended modules section to AnyEvent
497 documentation.
498 - reduce memory usage considerably (and reduce startup penalty)
499 by not using "strict", "warnings" and "overload" modules.
500 - work around buggy windows/openbsd perls and provide EBADMSG
501 and EPROTO ourselves when missing.
502 - improve perl 5.6 compatibility of the core event loop.
503 - made Net::SSLeay version 1.33 a soft requirement.
504
5054.83 Fri Jul 17 16:56:26 CEST 2009
506 - implement AnyEvent::Socket::getprotobyname.
507 - AnyEvent::CondVar's will now detect recursive blocking
508 waits and will croak, as too many people fall into
509 this trap.
510 - AnyEvent::Handle will now call ->destroy on itself after
511 executing the on_error callback, instead of doing some
512 half-baked internal shutdown, for fatal errors.
513 - clarify on_eof behaviour w.r.t. the read queue and
514 on_read callbacks.
515 - ignore some possible spurious wake-ups in tcp_connect.
516
5174.82 Sat Jul 11 00:34:55 CEST 2009
518 - POE and Event backends didn't accept some callable objects as
519 callbacks.
520 - use Config module instead of POSIX module to detect signal names
521 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
522 module doesn't even have all POSIX signals :/.
523 - use more workarounds around the many refcnt/corruption bugs in
524 Event::Lib.
525 - work around a race condition in perl's select, causing t/03_child.t
526 to rarely fail.
527
5284.81 Thu Jul 9 10:30:30 CEST 2009
529 - AnyEvent::Handle didn't properly diagnose write errors
530 (it expected -1 from syswrite, how lame... :).
531 - support file descriptors in addition to file handles
532 in AnyEvent->io.
533 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
534 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
535 and PERL_ANYEVENT_CA_PATH.
536 - provide a sensible synopsis section for AnyEvent::TLS.
537 - add a "supported backends" section to the manpage.
538 - added simple io watcher test to testsuite, using a
539 portable_socketpair.
540 - tried to improve the stability of the Event::Lib backend,
541 YMMV.
542
5434.8 Mon Jul 6 23:45:16 CEST 2009
544 - AnyEvent::DNS did not properly follow CNAME records with
545 uppercase targets.
546 - AnyEvent::DNS would errornously return AAAA records
547 with v4 mapped addresses (a faulty record) as ipv4 addresses,
548 causing AnyEvent::Socket to throw an exception.
549 - added new module AnyEvent::TLS for easier SSL/TLS context
550 creation, with many options including hostname verification,
551 secure default configuration, lots of documentation and,
552 predefined diffie-hellman keys for perfect forward security
553 and much more. get it while it's still fresh!
554 - use AnyEvent::TLS in AnyEvent::Handle for context management.
555 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
556 so AnyEvent::Socket users have smaller memory footprint
557 in the common case.
558 - add AnyEvent::Handle->push_shutdown method.
559 - add an additional error message parameter to AnyEvent::Handle's
560 on_error callback (for TLS, $! is still available).
561 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
562 - make AnyEvent::Handle more robust against early conenction
563 failures (during new), and return C<undef> in that case
564 from the constructor.
565 - AnyEvent::Handle will now try to load only JSON::XS first,
566 then fall back to JSON.
567 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
568 and because it was documented that way.
569
5704.451 Fri Jul 3 00:28:58 CEST 2009
571 - do not clear rbuf when shutting down an AnyEvent::Handle
572 object - doing so breaks AnyEvent::HTTP.
573
5744.45 Mon Jun 29 22:59:26 CEST 2009
575 - a write error could cause AnyEvent::Handle to create
576 an I/O watcher with an undefined $fh.
577 - special-case mapped ipv4 addresses in both
578 AnyEvent::Socket::format_address and parse_address,
579 to treat them just like ipv4 addresses.
580 - updated and overhauled the AnyEvent::Intro doc.
581 - implement AnyEvent::Socket::format_ipv4/ipv6.
582 - slightly speed up portable_pipe/socketpair.
583 - expand condvar begin/end documentation.
584
5854.42 Fri Jun 26 08:32:18 CEST 2009
586 - preliminary and neccesarily incomplete support for IO::Async.
587 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
588 case it was set to IGNORE, to ensure we can catch child statuses
589 even when the calling env acts stupidly.
590 - updated benchmarks with IO::Async, which performs very well.
591
5924.412 Wed Jun 24 01:35:57 CEST 2009
593 - support an "untaint" attribute for AnyEvent::DNS
594 and set it on the default resolver.
595 - implement and document AnyEvent::Impl::Perl::loop.
596 - remove all anyevent-env variables from %ENV
597 when running in tainted mode.
598 - mention and extend the IO::Lambda benchmark.
599
6004.411 Sun Jun 7 18:48:02 CEST 2009
601 - do not try to use F_SETFD on windows.
602
6034.41 Thu May 14 06:40:11 CEST 2009
604 - work around issues in older perls (5.8.0?) when
605 a signal handler is deleted from the %SIG hash.
606 - use POSIX::_exit in child test, to avoid running
607 destructors.
608 - speed up CHLD handling by relying on SIGCHLD being
609 synchronously delivered, even when we roll our own
610 implementation.
611 - AnyEvent::DNS: add the "dname" resource record name
612 for cosmetic reasons.
613
6144.4 Sun Apr 26 20:12:33 CEST 2009
615 - implemented idle watchers, where applicable.
616 - AnyEvent->time died when Event backend was in use.
617 - fix a memleak in the Tk backend.
618 - sped up Tk timer handling.
619 - clip negative "after" values to 0 in AnyEvent::Impl::Event
620 to avoid spamming out warning messages.
621 - fix Qt timers without interval.
622 - avoid Qt zero-timeout special case, in old libqt's.
623
6244.352 Mon Apr 20 16:31:11 CEST 2009
625 - fix AnyEvent::Strict error messages for child watchers.
626 - fix/update Makefile.PL "recommends" versions.
627 - add AnyEvent->now_update.
628
6294.351 Sat Apr 11 07:56:14 CEST 2009
630 - actually make the signal pipe work under win32.
631 - localise $! in signal handler to avoid changing $!, although
632 perl itself does not seem to save/restore errno either.
633 - set the cloexec flag on the signal pipe (normally set by perl too).
634
6354.35 Fri Mar 27 11:48:20 CET 2009
636 - event models relying on AnyEvent's signal watcher emulation did
637 invoke the callback asynchronously, contradicting documentation
638 and causing signals to get lost (this includes AnyEvent's own
639 event loop). AnyEvent now uses the standard pipe trick to make
640 callback execution synchronous to the event loop.
641 - AnyEvent::Handle didn't free TLS context data on DESTROY
642 (patch by Pavel Shaydo).
643 - work around the ever-incompatibly-changing API of MakeMaker.
644 - document that changing global variables without restoring them
645 is a bad idea in Perl (noted by Adam Rosenstein).
646 - AnyEvent::Strict now barfs if ->io is passed a file.
647
6484.34 Thu Feb 12 18:32:45 CET 2009
649 - separately buffer TLS read data, as otherwise the read queue could
650 deadlock as receiving data is not expected while draining the
651 read queue (which cna only happen with TLS).
652 - raise EBADMSG error on JSON decoding errors.
653 - fix some minor manpage bugs (reported by Maximilian Gaß).
654 - speed up select bitmask parsing quite a bit in the pure perl
655 backend.
656 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
657
6584.331 Tue Jan 6 21:07:25 CET 2009
659 - socketpair fails on many vista machines because vista has
660 completely broken accept/getpeername and getsockname functions,
661 so we provide our own socketpair emulation that kind of works
662 (AnyEvent::Util::portable_pipe).
663 - new function: AnyEvent::Util::portable_socketpair.
664 - take advantage of the Guard module if it exists.
665
6664.33 Fri Nov 21 02:35:40 CET 2008
667 - AnyEvent::Strict did errornously flag a fileno of 0 as
668 illegal.
669 - reduce memory usage and slightly speed up the pure perl backend
670 by only storing the file descriptor, not the file handle.
671 - add missing autocork method to AnyEvent::Handle (reported
672 by Adam Rosenstein).
673 - AnyEvent::DNS->resolve errornously documented an $rcode result
674 argument, but there isn't (spotted by Henrik Krohns).
675 - the naked truth about Tk - it's basically unmaintained.
676
6774.32 Mon Nov 3 22:46:32 CET 2008
678 - fix AnyEvent::Socket::resolve_sockaddr to properly support
679 unix sockets again and choose a proper default for
680 the socket type.
681 - fix call to resolve_sockaddr in tcp_connect.
682
6834.31 Thu Oct 30 04:41:48 CET 2008
684 - implemented AnyEvent::Handle->destroy method.
685 - hint about unexpected effects in TLS mode.
686 - speed up AnyEvent::Impl::Perl by using more arrays
687 instead of hashes without the slightest loss of readability :)
688 - work around a perl argument refcounting bug.
689 - with some perl patches applied, AnyEvent now seems leak-free.
690
6914.3 Fri Oct 3 09:18:43 CEST 2008
692 - AnyEvent will now install a no-op signal handler for SIGPIPE,
693 unless one has been installed already.
694 - warn about not loaidng AnyEvent::Impl::POE early enough
695 (patch by Adam Rosenstein).
696 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
697 - fix a bug in where in SSL connect mode, the
698 client would first wait for some data by the server and
699 otherwise hang. (reported and analysed in an absolutely
700 exemplary manner by Adam Rosenstein).
701 - fix a bug in where SSL EOF would not be treated
702 as stream EOF, putting the connection into a hung state.
703 (reported and analysed in an absolutely exemplary manner by Adam
704 Rosenstein).
705 - fix a potential segfault when the TLS context
706 would go missing in a read callback (e.g. due to stoptls) -
707 Net::SSLeay of course makes no type checking whatsoever.
708 - AnyEvent::Handle will not stop the read watcher in TLS mode.
709 - AnyEvent::Handle->stoptls will send a TLS close notify
710 instead of simply destroying the stream now.
711 - fix a bug where large blocks of data written in TLS mode
712 would not be sent unless triggered by receives.
713 - on_drain will now take the tls write buffer into account.
714 - SSL operations have been streamlined a bit and should be faster
715 now, more to come, though.
716 - added a FAQ section, will document nontrivial issues in there.
717
7184.234 Mon Sep 29 04:08:13 CEST 2008
719 - fix child watcher documentation: only child exits will be reported,
720 no trace events.
721 - mention SIGPIPE in AnyEvent::Handle.
722 - perl backend will now use POSIX::times as monotonic clock
723 when available and the monotonic clock option is not.
724 - run even without Time::HiRes in some cases now, but do not
725 provide sub-second accuracy (all *supported* perl versions have
726 Time::HiRes).
727
7284.233 Fri Aug 22 01:48:05 CEST 2008
729 - fix a bug in the testsuite, causing a hang.
730 - clarified lots of AnyEvent::Handle constructor arguments.
731
7324.232 Thu Aug 21 20:44:25 CEST 2008
733 - fix a bug in AnyEvent::Handle that could cause two fatal error
734 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
735 (testcase by Leon Brocard).
736 - clarified on_eof/on_error documentation a bit.
737
7384.231 Tue Jul 29 13:12:15 CEST 2008
739 - remove some debugging code left in AnyEvent::Util::fork_call
740 (and no, it's impossible to implement with the broken windows
741 perls without resource leaks or worse).
742
7434.23 Tue Jul 29 12:19:59 CEST 2008
744 - document the first parameter passed to condvar callbacks to be
745 the callback.
746 - add AnyEvent::Socket::{ntoa,aton} aliases.
747 - optimize the AE::Handle->push_read (line) for the default
748 eol marker.
749 - optimize push_read (packstring|storable) for small packets.
750 - invoke on_error callback when no on_eof callback is set.
751 - fix a bug in push_read (storable) of unknown impact.
752
7534.22 Sun Jul 20 16:34:13 CEST 2008
754 - new function AnyEvent::Socket::parse_hostport.
755 - as the bulkheads at microsoft can't even get getprotobyname reliably
756 working on their shitty fucking broken os we need to hardcode
757 some common protocol numbers in AnyEvent::Socket. How can
758 people even bother with such a pile of shit as windows.
759
7604.21 Thu Jul 17 14:40:05 CEST 2008
761 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
762 which is incompatible to previous invocation syntax.
763 - work around more windows perl bugs in fork_call. I cannot imagine
764 why anybody would bother implementing fork in such an obviously
765 unusable way.
766 - avoid calling $do_retry if the request has already finished in
767 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
768 K. Sharifulin).
769
7704.2 Sat Jul 12 22:42:11 CEST 2008
771 - support an "interval" argument for repeating timers.
772 - fix a bug in the pure perl event loop that caused timers to
773 be delayed under some circumstances.
774 - implement optional argument checking via PERL_ANYEVENT_STRICT.
775 - not importing anything from Event causes it to malfunction
776 and not use Time::HiRes, work around this bug.
777 - more bug workarounds against the endlessly broken Tk module,
778 seems to work now on 32 bit machines, but less so on 64 bit.
779 - minor optimisations applied to most backends.
780
7814.161 Fri Jul 4 14:24:48 CEST 2008
782 - fixed the t/handle/02_write.t, it read after write, with the
783 assumption that all data could be written before read
784 (reported by Jim Razmus).
785
7864.160 Thu Jul 3 04:02:21 CEST 2008
787 - re-registering signal watchers after unregistering one did
788 not work if a backend falls back on AnyEvent's default
789 implementation (also affected child watchers).
790 - new AnyEvent::Handle options: autocork and no_delay.
791
7924.152 Sun Jun 22 14:15:44 CEST 2008
793 - allow for 32-bit perls that implement shifts differently
794 on different architectures in parse_ipv4 (reported and
795 analysed by Keiichi DAIBA).
796
7974.151 Fri Jun 6 17:34:24 CEST 2008
798 - make sure specifying _only_ on_read and never pushing reads
799 works.
800
8014.15 Fri Jun 6 13:00:46 CEST 2008
802 - the pure perl backend would keep some watchers alive when more than
803 one watcher was registered for the same fd.
804 - new "packstring" and "storable" read and write types
805 for AnyEvent::Handle.
806 - allow on_eof handler to be called after on_error with EPIPE returns.
807 - do not immediately call on_read callback in handle constructor.
808
8094.14 Thu Jun 5 20:29:31 CEST 2008
810 - Fixed a bug in DNS SRV priority sorting.
811 - AnyEvent::Util::guard now reports runtime errors while
812 executing the guard block as warnings.
813 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
814
8154.13 Thu Jun 5 00:47:59 CEST 2008
816 - AnyEvent::DNS only followed cname chains with length 2,
817 contrary to documentation. bumped it up to 10, thanks to
818 microsoft, the current broken-dns-config-king.
819 - AnyEvent::DNS didn't check for socket return status
820 but instead relied on perl not creating filehandles in that
821 case - too bad it gives you a bogus file handle (reported
822 and analysed by Vladimir Timofeev).
823 - fix queue management logic in AnyEvent::Handle:
824 when on_read was registered, the queue was empty
825 and no progress could be made AnyEvent::Handle would
826 enter an endless loop.
827 - correctly start reading again when the handle became
828 busy again after idling.
829 - correctly treat tls shut-downs as EOF condition
830 (for the time being, would be nice to have a callback for that).
831 - correctly call eof callback under all conditions
832 (hopefully).
833 - the timeout callback did not expect that $self can go
834 away any time.
835
8364.12 Tue Jun 3 10:58:04 CEST 2008
837 - include AnyEvent::Intro, a tutorial for anyevent,
838 anyevent::socket and anyevent::handle.
839 - allow more options in on_error.
840
8414.11 Fri May 30 23:42:25 CEST 2008
842 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
843 reverse_lookup and reverse_verify to replace it, support
844 v4mapped and v4compat addresses.
845 - provide more documentation for the resolver class.
846 - really replace longest run of :0: by :: in format_address,
847 also properly convert :: and ::1 again.
848 - support NAPTR record name and decode it.
849 - implement random weight sampling for SRV records, as per
850 rfc 2782.
851 - correctly abort on srv-record targets of ".".
852 - added AnyEvent::DNS::wait_for_slot.
853 - in the unlikely event of a virtual circuit connection
854 being invalidated by a delayed udp reply, AnyEvent::DNS
855 could die. this has been fixed.
856 - plug a probable memleak in the DNS vc code.
857 - use configured timeout also for VC request phase.
858 - implement timeout and max_outstanding methods
859 for resolver class.
860 - update version numbers in all modules.
861
8624.1 Thu May 29 05:45:40 CEST 2008
863 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
864 which mirrors their purpose better (old names still available
865 for a while). Also moved AnyEvent::DNS::addr to
866 AnyEvent::Socket::resolve_sockaddr.
867 - implement AnyEvent->time and AnyEvent->now.
868 - fix IPv6 support in pack_sockaddr.
869 - officially un-experimentalise all newly introduced code.
870 - support unix domain sockets everywhere by specifying
871 a host of "unix/" and the pathname as service.
872 - implement an activity timeout in AnyEvent::Handle.
873 - added a regex read type to AnyEvent::Handle.
874 - added a json read/write type to AnyEvent::Handle.
875 - always croak in on_error in AnyEvent::Handle.
876 - document how to subclass AnyEvent::Handle.
877 - implement AnyEvent::Util::fork_call.
878 - add support for IPv6 nameservers and nameserver statements.
879 - work around _yet_ _another_ windows perl bug in where empty
880 select masks cause errors in the pure perl backend (this
881 workaround cannot be implemented for the other event loops,
882 but EV also works around this).
883 - supply AnyEvent::Socket::address_type and make good use of it.
884 - clarify and fix the Handle documentation w.r.t $self vs. $handle
885 vs. callback arguments.
886 - add some recommends to the META.yml.
887
8884.05 Mon May 26 19:44:06 CEST 2008
889 - some platforms ignore the rfc and prepend an extra sa_len member to
890 the sockaddr structure, cater for those. also use sockaddr_family,
891 didn't know it was there, but comes in handy for the workaround.
892 - undo the import WIN32 hack.
893
8944.04 Mon May 26 08:03:31 CEST 2008
895 - try to work around yet another windows bug: failed connects are reported
896 as if it were out-of-band data. windows users: you suck. Only
897 EV, Glib, Event and the pure perl backend can handle this condition.
898 - optimize write algorithm in AnyEvent::Handle.
899 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
900 - tcp_connect now properly iterates through all targets.
901 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
902 support entirely. also hardcode some AF_INET6 constants for the
903 sake of perl 5.8.
904 - vastly improved nameserver/suffix detection algorithm for windows
905 (still a hack...)
906 - try to use a fallback dns server if no dns servers could be found.
907 - splatter around some local $SIG{__DIE__}'s to work around bad
908 die handlers.
909
9104.03 Sun May 25 05:05:57 CEST 2008
911 - work around perl bugs on windows, where perl returns undocumented
912 error codes for sysread, syswrite etc.
913 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
914 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
915 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
916 - updated documentation.
917 - lots of workarounds for perl 5.8 quirks.
918 - implement a more flexible type system for both reads and writes
919 in AnyEvent::Handle.
920 - support IPv6 in tcp_server.
921 - pass local hostname and port to the prepare callback in tcp_server.
922 - make types extendable in AnyEvent::Handle.
923 - croak, not die, in AnyEvent::Handle;
924
9254.0 Sat May 24 19:58:08 CEST 2008
926 - added (experimental) AnyEvent::DNS module.
927 - added (experimental) AnyEvent::Socket module.
928 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
929 - make condvar->end call ->send as documented.
930 - make condvar constructor support the documented cb => argument.
931 - overload condvars so you can use them instead of code references.
932 - support more of the POSIX forms instead of a pure dotted quad for
933 verbatim ip addresses.
934 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
935 - try very hard not to freeze on broken windows perls.
936 - make on_eof optional in AnyEvent::Handle.
937 - support service names in tcp_connect and tcp_server.
938 - work around netbsd/openbsd bug causing unix sockets not to
939 become writable until empty.
940 - only make a dummy read for error detection when getpeername
941 returns ENOTCONN.
942 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
943 override.
944 - new AnyEvent::Guard::cancel.
945
9463.5 Sat May 17 23:17:14 CEST 2008
947 - gracefully deal with systems not defining CLOCK_MONOTONIC
948 (as opposed to not supporting the clock itself, which worked in 3.41).
949 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
950 helpers, to replace the ill-designed AnyEvent::Socket class.
951 - added some examples (eg/connect, eg/handle).
952 - added AnyEvent::Util::guard.
953 - AnyEvent::Util now exports most of its functions.
954 - fixed a bug in *_read_chunk, where the callback was called with $self as
955 first and second argument.
956 - simplified and fixed regex handling in AnyEvent::Handle::*_read_line.
957 - fix manpage condvars examples to use new syntax.
958 - considerably improve inet_aton with EV::ADNS backend to support
959 cname chains (common dns configuration bug, e.g. with www.google.de)
960 and aaaa records if no a record could be found.
961
9623.41 Sun May 11 19:53:13 CEST 2008
963 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
964
9653.4 Sun May 11 00:29:25 CEST 2008
966 - complete re-implementation of condvars with a hopefully much more useful
967 API (backward compatible functions exist but are undocumented).
968 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
969 (still subject to change).
3 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 970 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
971 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which
972 allows module authors to avoid forcing event loop detection.
973 (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example).
974 - remove coro backends: Coro now provides generic support for AnyEvent
975 via Coro::AnyEvent.
4 976
53.3 Mon Apr 28 09:51:06 CEST 2008 9773.3 Mon Apr 28 09:51:06 CEST 2008
6 - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. 978 - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules.
7 - fix a bug in the pure perl backend that kept watchers alive 979 - fix a bug in the pure perl backend that kept watchers alive
8 when multiple watchers were registered for the same fh/poll combo. 980 when multiple watchers were registered for the same fh/poll combo.

Diff Legend

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