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

Comparing AnyEvent/Changes (file contents):
Revision 1.212 by root, Fri Jan 2 22:09:12 2009 UTC vs.
Revision 1.506 by root, Fri Sep 2 04:41:16 2011 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: document TCP_*** constants
4
5TODO: log message level redistribute?
6TOOD: stop internals abusing
7 - INCOMPATIBLE CHANGE: the default log level is now "3" (critical
8 and above), and some emssages inside AnyEvent have bene elevated
9 to critical, so.
10 - add google public ipv6 dns as flalback servers, so AE::DNS can finally
11 support ipv6-only operation even with defaults (also use ::1 instead
12 of 127.0.0.1).
13 - give guidelines on how to chose a logging level.
14 - AE::Strict now checks that registered fds are still valid at regular
15 intervals.
16 - increase signal latency test timer considerably for poor little osx.
17 - mention some debugging aids in the tutorial.
18
196.02 Fri Aug 26 20:08:31 CEST 2011
20 - log_to_syslog now accepts facility strings.
21 - log_to_syslog would not allow facilities - "$facility|$level"
22 *literally* means that string. sheesh.
23 - the CPAN parser got confused and thought we provide Sys::Syslog, try
24 to work around.
25 - renamed AnyEvent::Impl::FLTK2 to FLTK.
26
276.01 Fri Aug 26 07:04:11 CEST 2011
28 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
29 the TTL as fourth element - this affects ->resolve, ->request
30 and ->dns_unpack indirectly. Convenience functions such as
31 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
32 AnyEvent::Socket.
33 - INCOMPATIBLE CHANGE: backend authors now should not implement
34 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
35 methods.
36 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
37 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
38 - DEPRECIATION: please change
39 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
40 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
41 in your programs, for future compatibility.
42 - *blush* idle watcher emulation was borked.
43 - AnyEvent::Strict now checks the AE::xxx functions as well.
44 - querying too long domain names would barf deep inside AnyEvent::DNS - since
45 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
46 for too long domain names explicitly and passes an undef to the callback
47 instead (reported by James Bromberger).
48 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
49 - new functions: AnyEvent::postpone, AnyEvent::log.
50 - new module: AnyEvent::Log, for simple logging needs.
51 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
52 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
53 a debug shell.
54 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
55 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
56 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
57 F</etc/hosts> can be found.
58 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
59 configuration.
60 - env variables can now be specified using AE_xyz instead of
61 PERL_ANYEVENT_xyz, subject to some rules.
62 - lots of new AnyEvent::Debug functionality.
63 - AnyEvent::Strict now checks for modifications $_ by the callback
64 (by setting it to readonly while executing the callback, if possible).
65 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
66 Invest extra overhead to work around this peculiar behaviour.
67 - split the pure perl backend into a loop module and an impl
68 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
69 - some read types (regex, netstring, json, storable) could
70 errornously keep an AE::Handle object alive on their own
71 (reported by Mohammad Toossi).
72 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
73 supports /etc/hosts overrides.
74 - add more workarounds around child watcher bugs in IO::Async
75 and (newly introduced) in POE.
76 - use glib child watchers and try to work around its limitations,
77 if possible (based on a patch by Kevin Ryde).
78 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
79 - many read types in AnyEvent::Handle would malfunction during
80 transitions to TLS because of a "delete $self->{rbuf}".
81 (reported by Gerald Galster).
82 - improve condvar blocking wait performance (probably only noticable
83 with faster backends), also make it easier for event loops that
84 do not support blocking waits.
85 - PERL_ANYEVENT_MODEL now also allows full module names, although
86 the usefulness of this feature is not yet clear.
87 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
88 - add t/80_ssltext.t.
89 - no longer use AUTOLOAD for the initial loading (for no very good reason).
90 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
91 - do not run event loop detection multiple times when programs call their cached
92 copies of AnyEvent::detect.
93 - parse_hostport now accepts absolute paths als unix domain sockets.
94 - suppress the idiotic warning inside POE by patching the POE::Kernel
95 run flags instead of asking the user to jump through contortions.
96 - add Task::Weaken as dependency (as recommends only at this time).
97 - replace PF_UNSPEC by 0, for good old minix.
98
996.0 (not officially released, but was on CPAN for about 15 minutes :/)
100
1015.34 Thu May 12 09:59:41 CEST 2011
102 - AE::TLS didn't support both cert_file and key at the same time,
103 due to some copy&paste bug (reported and analyzed by vti).
104 - improve AE::Handle timeout documentation slightly.
105
1065.33 Mon Feb 7 21:51:41 CET 2011
107 - never appeared on CPAN, for unknown reasons.
108 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
109 object when it's already gone (analyzed by Ryan Bullock).
110 - update idna algorithm(s) to UTS#46 version 6.0.0.
111 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
112 uts46 says it's an error, and testsuite makes more sense.
113 - remove leading dots in idn_to_ascii - testsuite says to remove,
114 uts46 says it's an error, and testsuite makes more sense.
115 - fixed a bug in idn ignored character processing.
116 - passes the uts46 6.0.0 testsuite except for three cases which
117 apparently are bugs in the testsuite itself.
118 - passes the uts46 6.0.1 testsuite except for eleven cases which
119 apparently are bugs in the testsuite itself.
120
1215.31 Sun Jan 23 11:49:19 CET 2011
122 - work around a POE bug where POE would first poll for new events and
123 invoke handlers BEFORE handling the message queue.
124 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
125 - stop_read now does just that in TLS mode. see its
126 documentation for caveats.
127 - update to common::sense 3.4 warning set (some warnings in perl impose
128 an arbitrary 32kb limit on regex matches, which affects
129 AnyEvent::Handle).
130 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
131 - new experimental AE::Handle::resettls method.
132
1335.3 Fri Dec 31 05:48:13 CET 2010
134 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
135 issues with IO::Async have been solved in that version, including
136 the availabilty of a default loop!
137 - implement new wbuf_max parameter for AnyEvent::Handle.
138 - added Cocoa::EventLoop backend by Daisuke Murase.
139 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
140 - clarify handle on_prepare documentation.
141 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
142 to avoid leaking them to any child processes.
143 - try to force files into the arch-dependend lib directory, for the
144 half-"turly obsessive compulsive person"(s) that depend on it.
145
1465.29 Sun Dec 5 10:49:21 CET 2010
147 - convert EV backend to EV 4.00 API (so better upgrade EV too).
148 - AE::Handle: implement dynamic read_size adjustment, add
149 max_read_size parameter, reduce default read_size to 2048.
150 - add some FAQ entries for servers and/or on_eof vs. on_error.
151 - work around OS bugs (cygwin again) when lingering in AE::Handle.
152 - reduce memory usage slightly when connect was used in AE::Handle.
153 - wrok around more segfaults in Tk.
154 - document tls_ctx => undef as valid.
155 - detect WNOHANG value at compiletime.
156 - include some probably linux-only support for building AnyEvent
157 as part of the perl core.
158 - improve compatibility of t/08_idna.t to perl 5.8.
159 - make AnyEvent::Strict truly optional (in case somebody builds a
160 minimal perl...).
161
1625.28 Wed Oct 13 04:14:23 CEST 2010
163 - due to a glitch, AnyEvent's internal getprotobyname wasn't
164 actually used in tcp_server.
165 - implement AnyEvent::CondVar->new method, which was documented but
166 missing.
167 - support autoloading of read types in unshift_read.
168 - AnyEvent::Strict: do the test for invalid fh arguments differently,
169 resulting in more informative error messages.
170 - new AnyEvent::FAQ manpage.
171 - clarify that you can actually call push_read etc. in on_read.
172 - put keys and certs in /tmp, not into the current directory,
173 and document this in the SECURITY CONSIDERATIONS sections
174 of AnyEvent::TLS.
175 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
176 requested by Ron Isaacson.
177 - document $AnyEvent::DNS::RESOLVER variable.
178 - incorporated major typo patches by Abhijit Menon-Sen.
179 - AnyEvent::Handle now croaks when negative timeouts are passed.
180 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
181 undocumented, but who knows when they might come in handy.
182
1835.271 Tue Jun 8 12:05:46 CEST 2010
184 - backport to perl 5.8.x.
185
1865.27 Sun Jun 6 12:12:05 CEST 2010
187 - postpone differently in AnyEvent::Socket now, as
188 when not, canceling the connection attempt might fail
189 (found by Felix Antonius Wilhelm Ostmann).
190 - explicitly check for non-stream sockets in AE::Handle, too many
191 clueless people fell into the trap of this somehow working.
192 - simplified and reworked the "OTHER MODULES" section.
193 - better/more condvar examples.
194
1955.261 Wed Apr 28 16:13:36 CEST 2010
196 - AF_INET6 was not properly used from Socket6 during configuration
197 time (found by Andy Grundman).
198
1995.26 Mon Apr 12 04:49:35 CEST 2010
200 - don't generate spurious readyness notifications when select
201 returns EINTR in the pure perl backend (analysed and testcase
202 by Bas Denissen).
203 - give same examples in SYNOPSIS sections of AnyEvent and AE.
204 - provider faster implementations for AE::time/now/now_update for
205 EV and Perl backends.
206 - provide faster AE::cv and AE::time implementations for all
207 backends.
208 - no longer support lower/mixed-case signal names (this was
209 never documented nor universally supported).
210 - some more memory µ-optimisations, and somewhat less messy
211 code to cope with both AnyEvent and AE APIs.
212
2135.251 Sat Mar 13 00:58:21 CET 2010
214 - make test failed if it was run before make install on perl <=
215 5.8.8, and older versions of perl make our live very difficult,
216 so write constants.pl during Makefile.PL time.
217
2185.25 Sat Mar 13 00:23:14 CET 2010
219 - fix a race condition in AnyEvent::Handle that would cause
220 a "bio_read: unsupported method" error (in _another_
221 ssl connection) after user code threw an exception.
222 - added AnyEvent::Handle->destroyed method.
223 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
224 - the AnyEvent::Util::fh_nonblocking function worked only
225 by lucky accident on win32.
226 - smaller and faster AnyEvent::Util::fh_nonblocking.
227 - when the (required!) Time::HiRes module is missing, AnyEvent
228 did not fall back to built-in time properly.
229 - do not load Fcntl at runtime, saving memory and loading time.
230 - precompile a number of constants and use them instead of runtime
231 detection and eval.
232 - free detection code after detection and similar memory optimisations.
233 - Perl backend timer interval best effort drift has been improved
234 (same algorithm as EV).
235 - update unicode idna mapping table.
236
2375.24 Tue Jan 5 11:39:43 CET 2010
238 - cygwin never reports errors from failed connects. 1.5 just gives
239 you continous readyness and EAGAIN, 1.7 is even more broken and
240 just hangs. work around both issues in a cygwin-specific hack.
241 - improve idle watcher documentation slightly.
242
2435.23 Sun Dec 20 23:48:00 CET 2009
244 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
245 - implement punycode_encode/decode, idn_nameprep,
246 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
247 - provide $AE::VERSION.
248 - removed traces of "no strict 'refs'".
249
2505.22 Sat Dec 5 03:51:13 CET 2009
251 - downgrade-or-fail in AnyEvent::Handle::push_write, to
252 diagnose encoding failures earlier and more succinctly.
253 (this works around bugs in perl, throwing away encoding info
254 when passing scalar data to extensions).
255 - add more examples to AnyEvent::Socket manpage.
256 - upgrade internal warning set to the same as common::sense 2.03.
257 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
258 (John Beppu).
259 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
260 - add google public dns servers to fallback server set and make sure
261 we load-balance properly between the three sets. also add all
262 fallback dns servers, not just a random one, to each dns config.
263
2645.21 Thu Nov 19 02:48:47 CET 2009
265 - fix a problem where socket constants were called with parameters
266 (spotted by David Friedland).
267 - fork_call never use'd POSIX (reported by Daisuke Maki).
268 - improve perl 5.6 compatibility further (but it still won't work
269 unless you rip out everything but the core).
270 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
271 uses 10MB of RAM, but doesn't open a console window. *sigh*.
272
2735.202 Wed Oct 14 22:35:44 CEST 2009
274 - AnyEvent::DNS would unexpectedly clobber $_ under windows
275 (analysed by Matthias Waldorf).
276 - AnyEvent::Handle::run_cmd can now pass the PID of the
277 newly-created process, which is much less useful than it might
278 sound (based on patch by Yann Kerherve).
279
2805.201 Tue Sep 29 12:09:25 CEST 2009
281 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
282 (reported by Torsten Foertsch).
283 - common::sense 2.0 could cause tcp_server to throw an exception
284 (analysed by elmex).
285
2865.2 Mon Sep 14 07:04:49 CEST 2009
287 - INCOMPATIBLE CHANGE: do no longer support register_read_type
288 and register_write_type in AnyEvent::Handle, instead support
289 package names (the facility was mostly abused).
290 - implement "packagename-as-read/write type" support in
291 AnyEvent::Handle.
292 - AnyEvent::Handle: new options "keepalive" and "oobinline".
293 - oobinline set by default to avoid security issues.
294 - the pure-perl event loop backend wrongly detected times() fallback
295 support (spotted by Pavel Boldin).
296
2975.12 Tue Sep 1 20:26:50 CEST 2009
298 - be more lenient when parsing resolv.conf files, as some people
299 use hashmarks as comment indicator inside directives (reported
300 by Michael S. Fischer).
301 - use same set of warnings as common::sense 2.0.
302 - fix a potential 32 bit overflow issue due to perl having problems
303 with large hex constants in the Perl backend.
304 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
305 by default.
306 - fix some prototypes.
307 - (optionally) require Async::Interrupt 1.02.
308 - replace opendns fallbacks by verizon/level3 public dns servers.
309
3105.112 Fri Aug 21 13:59:04 CEST 2009
311 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
312 using starttls with an illegal/uncreatable tls context.
313 - port to broken win32 perls.
314
3155.111 Sun Aug 16 18:44:36 CEST 2009
316 - perl errornously requires sockaddr_un structures returned by the
317 kernel to have a specific length. work around this bug by padding
318 sockaddr_un structures (found to fail on OS X, as perl only uses
319 the correct code for GNU/Linux, but likely to fail on other
320 OSes as well).
321
3225.11 Wed Aug 12 17:49:37 CEST 2009
323 - fixed a bug in run_cmd with input from filename.
324 - create a json coder object when none is given in write direction
325 to reduce dependency on JSON.
326
3275.1 Tue Aug 11 03:17:32 CEST 2009
328 - tcp_connect now reports the correct error in $!, not an unrelated
329 one (was broken since 4.91).
330 - AE::Handle did not properly clear rtimeout/wtimeout watchers
331 when retrying a tcp_connect.
332 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
333 - fixed wrong documentation in AE::Socket.
334
3355.01 Mon Aug 10 03:16:32 CEST 2009
336 - last release broke EV child watchers :/.
337
3385.0 Sun Aug 9 17:32:09 CEST 2009
339 - officially introduce and document the AE API.
340 - lowering the timeout at runtime did not have immediate effect
341 in AnyEvent::Handle.
342 - AE::Handle now has separate and independent read and write timeout
343 settings.
344 - probe for Guard module when AnyEvent::Util is loaded, not at
345 runtime (Event doesn't like eval "use" from callbacks, and it avoids
346 exporting the probe function).
347 - Impl::IOAsync's io watchers did not use the filehandle emulation
348 code to work around it's watcher limits.
349 - work around IO::Async's problems with overloaded objects.
350 - try to disable wearnings inside IO::Async, it's just too buggy.
351 - do no longer use Event by default, only EV or Perl, Event is too buggy.
352 - hackishly provide fast AE:: interface when EV or Perl is the backend.
353 - inherit VERSION from AnyEvent::VERSION in most modules.
354
3554.91 Thu Aug 6 15:42:45 CEST 2009
356 - AE::Handle::starttls could get out of sync when the read buffer
357 already contains some TLS handshake.
358 - AE::Handle did not properly free the TLS session, causing bigger
359 memory leaks in Net::SSLeay than expected (reported by toaster).
360 - AE::Socket::tcp_connect will now artificially delay invoking
361 the callback to avoid returning after invoking the callback.
362 - convert many internal watcher uses to AE API.
363
3644.9 Sat Aug 1 11:07:01 CEST 2009
365 - Glib silently fails when registering a timeout with
366 a negative value, so avoid that.
367 - call condvar callback immediately when it is set
368 after the condition is already signalled.
369 - check rbuf_max condition only after trying to consume
370 data.
371 - normalise signal numbers to names when using the ae's
372 signal handling, but do not document this (yet).
373 - pure perl signal emulation did not properly set nonblocking
374 mode on the signal pipe on win32 (this is unlikely to have caused
375 any issues).
376 - new module: AnyEvent::Debug.
377 - AnyEvent::Strict now checks that a signal was specified
378 by name (not name or number).
379 - reduce memory footprint in typical cases by ~50kb by
380 only compiling signal code when necessary.
381 - add AnyEvent::Handle->rbuf_max.
382 - grab the AE namespace for future new API, implement stubs for
383 said future/alternative API.
384 - new function: AnyEvent::Socket::format_hostport.
385
3864.881 Tue Jul 28 12:51:53 CEST 2009
387 - work around a bug in local in pre-5.10 perls, causing
388 AnyEvent::Handle to recurse when it should not
389 (analysed by elmex).
390
3914.88 Tue Jul 28 04:04:37 CEST 2009
392 - re-bless the handle into a dummy package after calling
393 AnyEvent::Handle::destroy, so the user does not need to check
394 for errors after every push_write etc.
395 - do not attempt to run t/02_signals.t on obviously broken
396 platforms.
397
3984.87 Sun Jul 26 02:06:16 CEST 2009
399 - do not attempt to linger when there is no longer a valid fh.
400 - enforce tls mode to be either accept or connect, do not simply
401 segfault in Net::SSLeay.
402 - AnyEvent::Handle can now call tcp_connect itself (new parameters
403 connect, on_prepare, on_connect and on_connect_error). Updated
404 tutorial accordingly.
405 - add AnyEvent::Impl::Irssi backend.
406
4074.86 Mon Jul 20 23:52:29 CEST 2009
408 - since the verbose warning is not enough, explicitly document
409 that versions before 1.33 of Net::SSLeay are not secure.
410 - work around signal handling races in Event and (...) Event::Lib.
411 - try to align signal-race timer to full-second boundaries.
412 - work around Tk not liking negative timeouts.
413 - don't complain of different grades of environmental unfriendlyness
414 in IO::Async.
415
4164.85 Sat Jul 18 06:16:14 CEST 2009
417 - nail the signal race problem in perl once and for all
418 (see $AnyEvent::MAX_SIGNAL_LATENCY).
419 - take advantage of Async::Interrupt if it is available.
420 - load Time::HiRes and Guard modules on demand only.
421 - add optional/recommended modules section to AnyEvent
422 documentation.
423 - reduce memory usage considerably (and reduce startup penalty)
424 by not using "strict", "warnings" and "overload" modules.
425 - work around buggy windows/openbsd perls and provide EBADMSG
426 and EPROTO ourselves when missing.
427 - improve perl 5.6 compatibility of the core event loop.
428 - made Net::SSLeay version 1.33 a soft requirement.
429
4304.83 Fri Jul 17 16:56:26 CEST 2009
431 - implement AnyEvent::Socket::getprotobyname.
432 - AnyEvent::CondVar's will now detect recursive blocking
433 waits and will croak, as too many people fall into
434 this trap.
435 - AnyEvent::Handle will now call ->destroy on itself after
436 executing the on_error callback, instead of doing some
437 half-baked internal shutdown, for fatal errors.
438 - clarify on_eof behaviour w.r.t. the read queue and
439 on_read callbacks.
440 - ignore some possible spurious wake-ups in tcp_connect.
441
4424.82 Sat Jul 11 00:34:55 CEST 2009
443 - POE and Event backends didn't accept some callable objects as
444 callbacks.
445 - use Config module instead of POSIX module to detect signal names
446 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
447 module doesn't even have all POSIX signals :/.
448 - use more workarounds around the many refcnt/corruption bugs in
449 Event::Lib.
450 - work around a race condition in perl's select, causing t/03_child.t
451 to rarely fail.
452
4534.81 Thu Jul 9 10:30:30 CEST 2009
454 - AnyEvent::Handle didn't properly diagnose write errors
455 (it expected -1 from syswrite, how lame... :).
456 - support file descriptors in addition to file handles
457 in AnyEvent->io.
458 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
459 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
460 and PERL_ANYEVENT_CA_PATH.
461 - provide a sensible synopsis section for AnyEvent::TLS.
462 - add a "supported backends" section to the manpage.
463 - added simple io watcher test to testsuite, using a
464 portable_socketpair.
465 - tried to improve the stability of the Event::Lib backend,
466 YMMV.
467
4684.8 Mon Jul 6 23:45:16 CEST 2009
469 - AnyEvent::DNS did not properly follow CNAME records with
470 uppercase targets.
471 - AnyEvent::DNS would errornously return AAAA records
472 with v4 mapped addresses (a faulty record) as ipv4 addresses,
473 causing AnyEvent::Socket to throw an exception.
474 - added new module AnyEvent::TLS for easier SSL/TLS context
475 creation, with many options including hostname verification,
476 secure default configuration, lots of documentation and,
477 predefined diffie-hellman keys for perfect forward security
478 and much more. get it while it's still fresh!
479 - use AnyEvent::TLS in AnyEvent::Handle for context management.
480 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
481 so AnyEvent::Socket users have smaller memory footprint
482 in the common case.
483 - add AnyEvent::Handle->push_shutdown method.
484 - add an additional error message parameter to AnyEvent::Handle's
485 on_error callback (for TLS, $! is still available).
486 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
487 - make AnyEvent::Handle more robust against early conenction
488 failures (during new), and return C<undef> in that case
489 from the constructor.
490 - AnyEvent::Handle will now try to load only JSON::XS first,
491 then fall back to JSON.
492 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
493 and because it was documented that way.
494
4954.451 Fri Jul 3 00:28:58 CEST 2009
496 - do not clear rbuf when shutting down an AnyEvent::Handle
497 object - doing so breaks AnyEvent::HTTP.
498
4994.45 Mon Jun 29 22:59:26 CEST 2009
500 - a write error could cause AnyEvent::Handle to create
501 an I/O watcher with an undefined $fh.
502 - special-case mapped ipv4 addresses in both
503 AnyEvent::Socket::format_address and parse_address,
504 to treat them just like ipv4 addresses.
505 - updated and overhauled the AnyEvent::Intro doc.
506 - implement AnyEvent::Socket::format_ipv4/ipv6.
507 - slightly speed up portable_pipe/socketpair.
508 - expand condvar begin/end documentation.
509
5104.42 Fri Jun 26 08:32:18 CEST 2009
511 - preliminary and neccesarily incomplete support for IO::Async.
512 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
513 case it was set to IGNORE, to ensure we can catch child statuses
514 even when the calling env acts stupidly.
515 - updated benchmarks with IO::Async, which performs very well.
516
5174.412 Wed Jun 24 01:35:57 CEST 2009
518 - support an "untaint" attribute for AnyEvent::DNS
519 and set it on the default resolver.
520 - implement and document AnyEvent::Impl::Perl::loop.
521 - remove all anyevent-env variables from %ENV
522 when running in tainted mode.
523 - mention and extend the IO::Lambda benchmark.
524
5254.411 Sun Jun 7 18:48:02 CEST 2009
526 - do not try to use F_SETFD on windows.
527
5284.41 Thu May 14 06:40:11 CEST 2009
529 - work around issues in older perls (5.8.0?) when
530 a signal handler is deleted from the %SIG hash.
531 - use POSIX::_exit in child test, to avoid running
532 destructors.
533 - speed up CHLD handling by relying on SIGCHLD being
534 synchronously delivered, even when we roll our own
535 implementation.
536 - AnyEvent::DNS: add the "dname" resource record name
537 for cosmetic reasons.
538
5394.4 Sun Apr 26 20:12:33 CEST 2009
540 - implemented idle watchers, where applicable.
541 - AnyEvent->time died when Event backend was in use.
542 - fix a memleak in the Tk backend.
543 - sped up Tk timer handling.
544 - clip negative "after" values to 0 in AnyEvent::Impl::Event
545 to avoid spamming out warning messages.
546 - fix Qt timers without interval.
547 - avoid Qt zero-timeout special case, in old libqt's.
548
5494.352 Mon Apr 20 16:31:11 CEST 2009
550 - fix AnyEvent::Strict error messages for child watchers.
551 - fix/update Makefile.PL "recommends" versions.
552 - add AnyEvent->now_update.
553
5544.351 Sat Apr 11 07:56:14 CEST 2009
555 - actually make the signal pipe work under win32.
556 - localise $! in signal handler to avoid changing $!, although
557 perl itself does not seem to save/restore errno either.
558 - set the cloexec flag on the signal pipe (normally set by perl too).
559
5604.35 Fri Mar 27 11:48:20 CET 2009
561 - event models relying on AnyEvent's signal watcher emulation did
562 invoke the callback asynchronously, contradicting documentation
563 and causing signals to get lost (this includes AnyEvent's own
564 event loop). AnyEvent now uses the standard pipe trick to make
565 callback execution synchronous to the event loop.
566 - AnyEvent::Handle didn't free TLS context data on DESTROY
567 (patch by Pavel Shaydo).
568 - work around the ever-incompatibly-changing API of MakeMaker.
569 - document that changing global variables without restoring them
570 is a bad idea in Perl (noted by Adam Rosenstein).
571 - AnyEvent::Strict now barfs if ->io is passed a file.
572
5734.34 Thu Feb 12 18:32:45 CET 2009
574 - separately buffer TLS read data, as otherwise the read queue could
575 deadlock as receiving data is not expected while draining the
576 read queue (which cna only happen with TLS).
577 - raise EBADMSG error on JSON decoding errors.
578 - fix some minor manpage bugs (reported by Maximilian Gaß).
579 - speed up select bitmask parsing quite a bit in the pure perl
580 backend.
581 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
582
5834.331 Tue Jan 6 21:07:25 CET 2009
3 - socketpair fails on many vista machines because vista has 584 - socketpair fails on many vista machines because vista has
4 completely broken accept/getpeername and getsockname functions, 585 completely broken accept/getpeername and getsockname functions,
5 so we provide our own socketpair emulation that kind of works 586 so we provide our own socketpair emulation that kind of works
6 (AnyEvent::Util::portable_pipe). 587 (AnyEvent::Util::portable_pipe).
7 - new function: AnyEvent::Util::portable_socketpair. 588 - new function: AnyEvent::Util::portable_socketpair.
… …
354.3 Fri Oct 3 09:18:43 CEST 2008 6164.3 Fri Oct 3 09:18:43 CEST 2008
36 - AnyEvent will now install a no-op signal handler for SIGPIPE, 617 - AnyEvent will now install a no-op signal handler for SIGPIPE,
37 unless one has been installed already. 618 unless one has been installed already.
38 - warn about not loaidng AnyEvent::Impl::POE early enough 619 - warn about not loaidng AnyEvent::Impl::POE early enough
39 (patch by Adam Rosenstein). 620 (patch by Adam Rosenstein).
40 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 621 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
41 - fix a bug in where in SSL connect mode, the 622 - fix a bug in where in SSL connect mode, the
42 client would first wait for some data by the server and 623 client would first wait for some data by the server and
43 otherwise hang. (reported and analyzed in an absolutely 624 otherwise hang. (reported and analysed in an absolutely
44 exemplary manner by Adam Rosenstein). 625 exemplary manner by Adam Rosenstein).
45 - fix a bug in where SSL EOF would not be treated 626 - fix a bug in where SSL EOF would not be treated
46 as stream EOF, putting the connection into a hung state. 627 as stream EOF, putting the connection into a hung state.
47 (reported and analyzed in an absolutely exemplary manner by Adam 628 (reported and analysed in an absolutely exemplary manner by Adam
48 Rosenstein). 629 Rosenstein).
49 - fix a potential segfault when the TLS context 630 - fix a potential segfault when the TLS context
50 would go missing in a read callback (e.g. due to stoptls) - 631 would go missing in a read callback (e.g. due to stoptls) -
51 Net::SSLeay of course makes no type checking whatsoever. 632 Net::SSLeay of course makes no type checking whatsoever.
52 - AnyEvent::Handle will not stop the read watcher in TLS mode. 633 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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