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

Comparing AnyEvent/Changes (file contents):
Revision 1.87 by root, Fri May 23 16:36:02 2008 UTC vs.
Revision 1.523 by root, Thu Jan 12 06:27:44 2012 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
33.51 3TODO: document TCP_*** constants
4TODO: add debug/trace logging to some modules
5
66.13 Thu Jan 12 07:27:01 CET 2012
7 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
8 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
9 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
10 regex-matching inside the callback would change the parameter, despite
11 $1 being dynamically scoped per-block. this perl bug is now being
12 worked around (testcase by Cindy Wang).
13
146.12 Mon Dec 12 13:21:10 CET 2011
15 - $! was clobbered by subsequent calls in tcp_connect, due to
16 the postpone not saving/restoring it (analyzed by Richard Garnish).
17
186.11 Tue Nov 22 10:36:05 CET 2011
19 - Tk cannot create windows when tainted unless you set a title, so
20 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
21 - escape any nonprintable/nonascii characters when stringifying
22 backtraces.
23 - log the reason loading an interface module fails to load at level debug.
24
256.1 Tue Oct 4 19:44:30 CEST 2011
26 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
27 and above), and some messages inside AnyEvent have been elevated
28 to higher log levels to print by default.
29 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
30 - work around yet another signal race bug, newly introduced by POE. My
31 how that module sucks.
32 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
33 support ipv6-only operation even with defaults (also use ::1 instead
34 of 127.0.0.1).
35 - give better guidelines on how to chose a logging level.
36 - AnyEvent::Log can now cap (limit) log message priority by context.
37 - AE::log will now save and restore $! and $@, so callers don't have to.
38 - AE::Strict now checks that registered fds are still valid at regular
39 intervals.
40 - skip frequently called 'require' calls if the module appears loaded.
41 - increase signal latency test timer considerably for poor little osx,
42 which can't be bothered with basic posix functionality such as signal
43 delivery.
44 - mention some debugging aids in the tutorial.
45 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
46 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
47 of the context as title.
48 - AnyEvent::Strict would not restore read-write access to $_ when a callback
49 died - this is a user bug, but since AE::Strict exists to help dbeugging...
50 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
51 to work with it - the author is well aware of the issues but refuses
52 to fix them. This extreme step has been taken because the module actively
53 attacks the goals of AnyEvent and makes it harder for module authors to use
54 AnyEvent.
55
566.02 Fri Aug 26 20:08:31 CEST 2011
57 - log_to_syslog now accepts facility strings.
58 - log_to_syslog would not allow facilities - "$facility|$level"
59 *literally* means that string. sheesh.
60 - the CPAN parser got confused and thought we provide Sys::Syslog, try
61 to work around.
62 - renamed AnyEvent::Impl::FLTK2 to FLTK.
63
646.01 Fri Aug 26 07:04:11 CEST 2011
65 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
66 the TTL as fourth element - this affects ->resolve, ->request
67 and ->dns_unpack indirectly. Convenience functions such as
68 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
69 AnyEvent::Socket.
70 - INCOMPATIBLE CHANGE: backend authors now should not implement
71 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
72 methods.
73 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
74 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
75 - DEPRECIATION: please change
76 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
77 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
78 in your programs, for future compatibility.
79 - *blush* idle watcher emulation was borked.
80 - AnyEvent::Strict now checks the AE::xxx functions as well.
81 - querying too long domain names would barf deep inside AnyEvent::DNS - since
82 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
83 for too long domain names explicitly and passes an undef to the callback
84 instead (reported by James Bromberger).
85 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
86 - new functions: AnyEvent::postpone, AnyEvent::log.
87 - new module: AnyEvent::Log, for simple logging needs.
88 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
89 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
90 a debug shell.
91 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
92 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
93 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
94 /etc/hosts can be found.
95 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
96 configuration.
97 - env variables can now be specified using AE_xyz instead of
98 PERL_ANYEVENT_xyz, subject to some rules.
99 - lots of new AnyEvent::Debug functionality.
100 - AnyEvent::Strict now checks for modifications $_ by the callback
101 (by setting it to readonly while executing the callback, if possible).
102 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
103 Invest extra overhead to work around this peculiar behaviour.
104 - split the pure perl backend into a loop module and an impl
105 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
106 - some read types (regex, netstring, json, storable) could
107 errornously keep an AE::Handle object alive on their own
108 (reported by Mohammad Toossi).
109 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
110 supports /etc/hosts overrides.
111 - add more workarounds around child watcher bugs in IO::Async
112 and (newly introduced) in POE.
113 - use glib child watchers and try to work around its limitations,
114 if possible (based on a patch by Kevin Ryde).
115 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
116 - many read types in AnyEvent::Handle would malfunction during
117 transitions to TLS because of a "delete $self->{rbuf}".
118 (reported by Gerald Galster).
119 - improve condvar blocking wait performance (probably only noticable
120 with faster backends), also make it easier for event loops that
121 do not support blocking waits.
122 - PERL_ANYEVENT_MODEL now also allows full module names, although
123 the usefulness of this feature is not yet clear.
124 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
125 - add t/80_ssltext.t.
126 - no longer use AUTOLOAD for the initial loading (for no very good reason).
127 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
128 - do not run event loop detection multiple times when programs call their cached
129 copies of AnyEvent::detect.
130 - parse_hostport now accepts absolute paths als unix domain sockets.
131 - suppress the idiotic warning inside POE by patching the POE::Kernel
132 run flags instead of asking the user to jump through contortions.
133 - add Task::Weaken as dependency (as recommends only at this time).
134 - replace PF_UNSPEC by 0, for good old minix.
135
1366.0 (not officially released, but was on CPAN for about 15 minutes :/)
137
1385.34 Thu May 12 09:59:41 CEST 2011
139 - AE::TLS didn't support both cert_file and key at the same time,
140 due to some copy&paste bug (reported and analyzed by vti).
141 - improve AE::Handle timeout documentation slightly.
142
1435.33 Mon Feb 7 21:51:41 CET 2011
144 - never appeared on CPAN, for unknown reasons.
145 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
146 object when it's already gone (analyzed by Ryan Bullock).
147 - update idna algorithm(s) to UTS#46 version 6.0.0.
148 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
149 uts46 says it's an error, and testsuite makes more sense.
150 - remove leading dots in idn_to_ascii - testsuite says to remove,
151 uts46 says it's an error, and testsuite makes more sense.
152 - fixed a bug in idn ignored character processing.
153 - passes the uts46 6.0.0 testsuite except for three cases which
154 apparently are bugs in the testsuite itself.
155 - passes the uts46 6.0.1 testsuite except for eleven cases which
156 apparently are bugs in the testsuite itself.
157
1585.31 Sun Jan 23 11:49:19 CET 2011
159 - work around a POE bug where POE would first poll for new events and
160 invoke handlers BEFORE handling the message queue.
161 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
162 - stop_read now does just that in TLS mode. see its
163 documentation for caveats.
164 - update to common::sense 3.4 warning set (some warnings in perl impose
165 an arbitrary 32kb limit on regex matches, which affects
166 AnyEvent::Handle).
167 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
168 - new experimental AE::Handle::resettls method.
169
1705.3 Fri Dec 31 05:48:13 CET 2010
171 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
172 issues with IO::Async have been solved in that version, including
173 the availabilty of a default loop!
174 - implement new wbuf_max parameter for AnyEvent::Handle.
175 - added Cocoa::EventLoop backend by Daisuke Murase.
176 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
177 - clarify handle on_prepare documentation.
178 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
179 to avoid leaking them to any child processes.
180 - try to force files into the arch-dependend lib directory, for the
181 half-"turly obsessive compulsive person"(s) that depend on it.
182
1835.29 Sun Dec 5 10:49:21 CET 2010
184 - convert EV backend to EV 4.00 API (so better upgrade EV too).
185 - AE::Handle: implement dynamic read_size adjustment, add
186 max_read_size parameter, reduce default read_size to 2048.
187 - add some FAQ entries for servers and/or on_eof vs. on_error.
188 - work around OS bugs (cygwin again) when lingering in AE::Handle.
189 - reduce memory usage slightly when connect was used in AE::Handle.
190 - wrok around more segfaults in Tk.
191 - document tls_ctx => undef as valid.
192 - detect WNOHANG value at compiletime.
193 - include some probably linux-only support for building AnyEvent
194 as part of the perl core.
195 - improve compatibility of t/08_idna.t to perl 5.8.
196 - make AnyEvent::Strict truly optional (in case somebody builds a
197 minimal perl...).
198
1995.28 Wed Oct 13 04:14:23 CEST 2010
200 - due to a glitch, AnyEvent's internal getprotobyname wasn't
201 actually used in tcp_server.
202 - implement AnyEvent::CondVar->new method, which was documented but
203 missing.
204 - support autoloading of read types in unshift_read.
205 - AnyEvent::Strict: do the test for invalid fh arguments differently,
206 resulting in more informative error messages.
207 - new AnyEvent::FAQ manpage.
208 - clarify that you can actually call push_read etc. in on_read.
209 - put keys and certs in /tmp, not into the current directory,
210 and document this in the SECURITY CONSIDERATIONS sections
211 of AnyEvent::TLS.
212 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
213 requested by Ron Isaacson.
214 - document $AnyEvent::DNS::RESOLVER variable.
215 - incorporated major typo patches by Abhijit Menon-Sen.
216 - AnyEvent::Handle now croaks when negative timeouts are passed.
217 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
218 undocumented, but who knows when they might come in handy.
219
2205.271 Tue Jun 8 12:05:46 CEST 2010
221 - backport to perl 5.8.x.
222
2235.27 Sun Jun 6 12:12:05 CEST 2010
224 - postpone differently in AnyEvent::Socket now, as
225 when not, canceling the connection attempt might fail
226 (found by Felix Antonius Wilhelm Ostmann).
227 - explicitly check for non-stream sockets in AE::Handle, too many
228 clueless people fell into the trap of this somehow working.
229 - simplified and reworked the "OTHER MODULES" section.
230 - better/more condvar examples.
231
2325.261 Wed Apr 28 16:13:36 CEST 2010
233 - AF_INET6 was not properly used from Socket6 during configuration
234 time (found by Andy Grundman).
235
2365.26 Mon Apr 12 04:49:35 CEST 2010
237 - don't generate spurious readyness notifications when select
238 returns EINTR in the pure perl backend (analysed and testcase
239 by Bas Denissen).
240 - give same examples in SYNOPSIS sections of AnyEvent and AE.
241 - provider faster implementations for AE::time/now/now_update for
242 EV and Perl backends.
243 - provide faster AE::cv and AE::time implementations for all
244 backends.
245 - no longer support lower/mixed-case signal names (this was
246 never documented nor universally supported).
247 - some more memory µ-optimisations, and somewhat less messy
248 code to cope with both AnyEvent and AE APIs.
249
2505.251 Sat Mar 13 00:58:21 CET 2010
251 - make test failed if it was run before make install on perl <=
252 5.8.8, and older versions of perl make our live very difficult,
253 so write constants.pl during Makefile.PL time.
254
2555.25 Sat Mar 13 00:23:14 CET 2010
256 - fix a race condition in AnyEvent::Handle that would cause
257 a "bio_read: unsupported method" error (in _another_
258 ssl connection) after user code threw an exception.
259 - added AnyEvent::Handle->destroyed method.
260 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
261 - the AnyEvent::Util::fh_nonblocking function worked only
262 by lucky accident on win32.
263 - smaller and faster AnyEvent::Util::fh_nonblocking.
264 - when the (required!) Time::HiRes module is missing, AnyEvent
265 did not fall back to built-in time properly.
266 - do not load Fcntl at runtime, saving memory and loading time.
267 - precompile a number of constants and use them instead of runtime
268 detection and eval.
269 - free detection code after detection and similar memory optimisations.
270 - Perl backend timer interval best effort drift has been improved
271 (same algorithm as EV).
272 - update unicode idna mapping table.
273
2745.24 Tue Jan 5 11:39:43 CET 2010
275 - cygwin never reports errors from failed connects. 1.5 just gives
276 you continous readyness and EAGAIN, 1.7 is even more broken and
277 just hangs. work around both issues in a cygwin-specific hack.
278 - improve idle watcher documentation slightly.
279
2805.23 Sun Dec 20 23:48:00 CET 2009
281 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
282 - implement punycode_encode/decode, idn_nameprep,
283 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
284 - provide $AE::VERSION.
285 - removed traces of "no strict 'refs'".
286
2875.22 Sat Dec 5 03:51:13 CET 2009
288 - downgrade-or-fail in AnyEvent::Handle::push_write, to
289 diagnose encoding failures earlier and more succinctly.
290 (this works around bugs in perl, throwing away encoding info
291 when passing scalar data to extensions).
292 - add more examples to AnyEvent::Socket manpage.
293 - upgrade internal warning set to the same as common::sense 2.03.
294 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
295 (John Beppu).
296 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
297 - add google public dns servers to fallback server set and make sure
298 we load-balance properly between the three sets. also add all
299 fallback dns servers, not just a random one, to each dns config.
300
3015.21 Thu Nov 19 02:48:47 CET 2009
302 - fix a problem where socket constants were called with parameters
303 (spotted by David Friedland).
304 - fork_call never use'd POSIX (reported by Daisuke Maki).
305 - improve perl 5.6 compatibility further (but it still won't work
306 unless you rip out everything but the core).
307 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
308 uses 10MB of RAM, but doesn't open a console window. *sigh*.
309
3105.202 Wed Oct 14 22:35:44 CEST 2009
311 - AnyEvent::DNS would unexpectedly clobber $_ under windows
312 (analysed by Matthias Waldorf).
313 - AnyEvent::Handle::run_cmd can now pass the PID of the
314 newly-created process, which is much less useful than it might
315 sound (based on patch by Yann Kerherve).
316
3175.201 Tue Sep 29 12:09:25 CEST 2009
318 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
319 (reported by Torsten Foertsch).
320 - common::sense 2.0 could cause tcp_server to throw an exception
321 (analysed by elmex).
322
3235.2 Mon Sep 14 07:04:49 CEST 2009
324 - INCOMPATIBLE CHANGE: do no longer support register_read_type
325 and register_write_type in AnyEvent::Handle, instead support
326 package names (the facility was mostly abused).
327 - implement "packagename-as-read/write type" support in
328 AnyEvent::Handle.
329 - AnyEvent::Handle: new options "keepalive" and "oobinline".
330 - oobinline set by default to avoid security issues.
331 - the pure-perl event loop backend wrongly detected times() fallback
332 support (spotted by Pavel Boldin).
333
3345.12 Tue Sep 1 20:26:50 CEST 2009
335 - be more lenient when parsing resolv.conf files, as some people
336 use hashmarks as comment indicator inside directives (reported
337 by Michael S. Fischer).
338 - use same set of warnings as common::sense 2.0.
339 - fix a potential 32 bit overflow issue due to perl having problems
340 with large hex constants in the Perl backend.
341 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
342 by default.
343 - fix some prototypes.
344 - (optionally) require Async::Interrupt 1.02.
345 - replace opendns fallbacks by verizon/level3 public dns servers.
346
3475.112 Fri Aug 21 13:59:04 CEST 2009
348 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
349 using starttls with an illegal/uncreatable tls context.
350 - port to broken win32 perls.
351
3525.111 Sun Aug 16 18:44:36 CEST 2009
353 - perl errornously requires sockaddr_un structures returned by the
354 kernel to have a specific length. work around this bug by padding
355 sockaddr_un structures (found to fail on OS X, as perl only uses
356 the correct code for GNU/Linux, but likely to fail on other
357 OSes as well).
358
3595.11 Wed Aug 12 17:49:37 CEST 2009
360 - fixed a bug in run_cmd with input from filename.
361 - create a json coder object when none is given in write direction
362 to reduce dependency on JSON.
363
3645.1 Tue Aug 11 03:17:32 CEST 2009
365 - tcp_connect now reports the correct error in $!, not an unrelated
366 one (was broken since 4.91).
367 - AE::Handle did not properly clear rtimeout/wtimeout watchers
368 when retrying a tcp_connect.
369 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
370 - fixed wrong documentation in AE::Socket.
371
3725.01 Mon Aug 10 03:16:32 CEST 2009
373 - last release broke EV child watchers :/.
374
3755.0 Sun Aug 9 17:32:09 CEST 2009
376 - officially introduce and document the AE API.
377 - lowering the timeout at runtime did not have immediate effect
378 in AnyEvent::Handle.
379 - AE::Handle now has separate and independent read and write timeout
380 settings.
381 - probe for Guard module when AnyEvent::Util is loaded, not at
382 runtime (Event doesn't like eval "use" from callbacks, and it avoids
383 exporting the probe function).
384 - Impl::IOAsync's io watchers did not use the filehandle emulation
385 code to work around it's watcher limits.
386 - work around IO::Async's problems with overloaded objects.
387 - try to disable wearnings inside IO::Async, it's just too buggy.
388 - do no longer use Event by default, only EV or Perl, Event is too buggy.
389 - hackishly provide fast AE:: interface when EV or Perl is the backend.
390 - inherit VERSION from AnyEvent::VERSION in most modules.
391
3924.91 Thu Aug 6 15:42:45 CEST 2009
393 - AE::Handle::starttls could get out of sync when the read buffer
394 already contains some TLS handshake.
395 - AE::Handle did not properly free the TLS session, causing bigger
396 memory leaks in Net::SSLeay than expected (reported by toaster).
397 - AE::Socket::tcp_connect will now artificially delay invoking
398 the callback to avoid returning after invoking the callback.
399 - convert many internal watcher uses to AE API.
400
4014.9 Sat Aug 1 11:07:01 CEST 2009
402 - Glib silently fails when registering a timeout with
403 a negative value, so avoid that.
404 - call condvar callback immediately when it is set
405 after the condition is already signalled.
406 - check rbuf_max condition only after trying to consume
407 data.
408 - normalise signal numbers to names when using the ae's
409 signal handling, but do not document this (yet).
410 - pure perl signal emulation did not properly set nonblocking
411 mode on the signal pipe on win32 (this is unlikely to have caused
412 any issues).
413 - new module: AnyEvent::Debug.
414 - AnyEvent::Strict now checks that a signal was specified
415 by name (not name or number).
416 - reduce memory footprint in typical cases by ~50kb by
417 only compiling signal code when necessary.
418 - add AnyEvent::Handle->rbuf_max.
419 - grab the AE namespace for future new API, implement stubs for
420 said future/alternative API.
421 - new function: AnyEvent::Socket::format_hostport.
422
4234.881 Tue Jul 28 12:51:53 CEST 2009
424 - work around a bug in local in pre-5.10 perls, causing
425 AnyEvent::Handle to recurse when it should not
426 (analysed by elmex).
427
4284.88 Tue Jul 28 04:04:37 CEST 2009
429 - re-bless the handle into a dummy package after calling
430 AnyEvent::Handle::destroy, so the user does not need to check
431 for errors after every push_write etc.
432 - do not attempt to run t/02_signals.t on obviously broken
433 platforms.
434
4354.87 Sun Jul 26 02:06:16 CEST 2009
436 - do not attempt to linger when there is no longer a valid fh.
437 - enforce tls mode to be either accept or connect, do not simply
438 segfault in Net::SSLeay.
439 - AnyEvent::Handle can now call tcp_connect itself (new parameters
440 connect, on_prepare, on_connect and on_connect_error). Updated
441 tutorial accordingly.
442 - add AnyEvent::Impl::Irssi backend.
443
4444.86 Mon Jul 20 23:52:29 CEST 2009
445 - since the verbose warning is not enough, explicitly document
446 that versions before 1.33 of Net::SSLeay are not secure.
447 - work around signal handling races in Event and (...) Event::Lib.
448 - try to align signal-race timer to full-second boundaries.
449 - work around Tk not liking negative timeouts.
450 - don't complain of different grades of environmental unfriendlyness
451 in IO::Async.
452
4534.85 Sat Jul 18 06:16:14 CEST 2009
454 - nail the signal race problem in perl once and for all
455 (see $AnyEvent::MAX_SIGNAL_LATENCY).
456 - take advantage of Async::Interrupt if it is available.
457 - load Time::HiRes and Guard modules on demand only.
458 - add optional/recommended modules section to AnyEvent
459 documentation.
460 - reduce memory usage considerably (and reduce startup penalty)
461 by not using "strict", "warnings" and "overload" modules.
462 - work around buggy windows/openbsd perls and provide EBADMSG
463 and EPROTO ourselves when missing.
464 - improve perl 5.6 compatibility of the core event loop.
465 - made Net::SSLeay version 1.33 a soft requirement.
466
4674.83 Fri Jul 17 16:56:26 CEST 2009
468 - implement AnyEvent::Socket::getprotobyname.
469 - AnyEvent::CondVar's will now detect recursive blocking
470 waits and will croak, as too many people fall into
471 this trap.
472 - AnyEvent::Handle will now call ->destroy on itself after
473 executing the on_error callback, instead of doing some
474 half-baked internal shutdown, for fatal errors.
475 - clarify on_eof behaviour w.r.t. the read queue and
476 on_read callbacks.
477 - ignore some possible spurious wake-ups in tcp_connect.
478
4794.82 Sat Jul 11 00:34:55 CEST 2009
480 - POE and Event backends didn't accept some callable objects as
481 callbacks.
482 - use Config module instead of POSIX module to detect signal names
483 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
484 module doesn't even have all POSIX signals :/.
485 - use more workarounds around the many refcnt/corruption bugs in
486 Event::Lib.
487 - work around a race condition in perl's select, causing t/03_child.t
488 to rarely fail.
489
4904.81 Thu Jul 9 10:30:30 CEST 2009
491 - AnyEvent::Handle didn't properly diagnose write errors
492 (it expected -1 from syswrite, how lame... :).
493 - support file descriptors in addition to file handles
494 in AnyEvent->io.
495 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
496 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
497 and PERL_ANYEVENT_CA_PATH.
498 - provide a sensible synopsis section for AnyEvent::TLS.
499 - add a "supported backends" section to the manpage.
500 - added simple io watcher test to testsuite, using a
501 portable_socketpair.
502 - tried to improve the stability of the Event::Lib backend,
503 YMMV.
504
5054.8 Mon Jul 6 23:45:16 CEST 2009
506 - AnyEvent::DNS did not properly follow CNAME records with
507 uppercase targets.
508 - AnyEvent::DNS would errornously return AAAA records
509 with v4 mapped addresses (a faulty record) as ipv4 addresses,
510 causing AnyEvent::Socket to throw an exception.
511 - added new module AnyEvent::TLS for easier SSL/TLS context
512 creation, with many options including hostname verification,
513 secure default configuration, lots of documentation and,
514 predefined diffie-hellman keys for perfect forward security
515 and much more. get it while it's still fresh!
516 - use AnyEvent::TLS in AnyEvent::Handle for context management.
517 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
518 so AnyEvent::Socket users have smaller memory footprint
519 in the common case.
520 - add AnyEvent::Handle->push_shutdown method.
521 - add an additional error message parameter to AnyEvent::Handle's
522 on_error callback (for TLS, $! is still available).
523 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
524 - make AnyEvent::Handle more robust against early conenction
525 failures (during new), and return C<undef> in that case
526 from the constructor.
527 - AnyEvent::Handle will now try to load only JSON::XS first,
528 then fall back to JSON.
529 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
530 and because it was documented that way.
531
5324.451 Fri Jul 3 00:28:58 CEST 2009
533 - do not clear rbuf when shutting down an AnyEvent::Handle
534 object - doing so breaks AnyEvent::HTTP.
535
5364.45 Mon Jun 29 22:59:26 CEST 2009
537 - a write error could cause AnyEvent::Handle to create
538 an I/O watcher with an undefined $fh.
539 - special-case mapped ipv4 addresses in both
540 AnyEvent::Socket::format_address and parse_address,
541 to treat them just like ipv4 addresses.
542 - updated and overhauled the AnyEvent::Intro doc.
543 - implement AnyEvent::Socket::format_ipv4/ipv6.
544 - slightly speed up portable_pipe/socketpair.
545 - expand condvar begin/end documentation.
546
5474.42 Fri Jun 26 08:32:18 CEST 2009
548 - preliminary and neccesarily incomplete support for IO::Async.
549 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
550 case it was set to IGNORE, to ensure we can catch child statuses
551 even when the calling env acts stupidly.
552 - updated benchmarks with IO::Async, which performs very well.
553
5544.412 Wed Jun 24 01:35:57 CEST 2009
555 - support an "untaint" attribute for AnyEvent::DNS
556 and set it on the default resolver.
557 - implement and document AnyEvent::Impl::Perl::loop.
558 - remove all anyevent-env variables from %ENV
559 when running in tainted mode.
560 - mention and extend the IO::Lambda benchmark.
561
5624.411 Sun Jun 7 18:48:02 CEST 2009
563 - do not try to use F_SETFD on windows.
564
5654.41 Thu May 14 06:40:11 CEST 2009
566 - work around issues in older perls (5.8.0?) when
567 a signal handler is deleted from the %SIG hash.
568 - use POSIX::_exit in child test, to avoid running
569 destructors.
570 - speed up CHLD handling by relying on SIGCHLD being
571 synchronously delivered, even when we roll our own
572 implementation.
573 - AnyEvent::DNS: add the "dname" resource record name
574 for cosmetic reasons.
575
5764.4 Sun Apr 26 20:12:33 CEST 2009
577 - implemented idle watchers, where applicable.
578 - AnyEvent->time died when Event backend was in use.
579 - fix a memleak in the Tk backend.
580 - sped up Tk timer handling.
581 - clip negative "after" values to 0 in AnyEvent::Impl::Event
582 to avoid spamming out warning messages.
583 - fix Qt timers without interval.
584 - avoid Qt zero-timeout special case, in old libqt's.
585
5864.352 Mon Apr 20 16:31:11 CEST 2009
587 - fix AnyEvent::Strict error messages for child watchers.
588 - fix/update Makefile.PL "recommends" versions.
589 - add AnyEvent->now_update.
590
5914.351 Sat Apr 11 07:56:14 CEST 2009
592 - actually make the signal pipe work under win32.
593 - localise $! in signal handler to avoid changing $!, although
594 perl itself does not seem to save/restore errno either.
595 - set the cloexec flag on the signal pipe (normally set by perl too).
596
5974.35 Fri Mar 27 11:48:20 CET 2009
598 - event models relying on AnyEvent's signal watcher emulation did
599 invoke the callback asynchronously, contradicting documentation
600 and causing signals to get lost (this includes AnyEvent's own
601 event loop). AnyEvent now uses the standard pipe trick to make
602 callback execution synchronous to the event loop.
603 - AnyEvent::Handle didn't free TLS context data on DESTROY
604 (patch by Pavel Shaydo).
605 - work around the ever-incompatibly-changing API of MakeMaker.
606 - document that changing global variables without restoring them
607 is a bad idea in Perl (noted by Adam Rosenstein).
608 - AnyEvent::Strict now barfs if ->io is passed a file.
609
6104.34 Thu Feb 12 18:32:45 CET 2009
611 - separately buffer TLS read data, as otherwise the read queue could
612 deadlock as receiving data is not expected while draining the
613 read queue (which cna only happen with TLS).
614 - raise EBADMSG error on JSON decoding errors.
615 - fix some minor manpage bugs (reported by Maximilian Gaß).
616 - speed up select bitmask parsing quite a bit in the pure perl
617 backend.
618 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
619
6204.331 Tue Jan 6 21:07:25 CET 2009
621 - socketpair fails on many vista machines because vista has
622 completely broken accept/getpeername and getsockname functions,
623 so we provide our own socketpair emulation that kind of works
624 (AnyEvent::Util::portable_pipe).
625 - new function: AnyEvent::Util::portable_socketpair.
626 - take advantage of the Guard module if it exists.
627
6284.33 Fri Nov 21 02:35:40 CET 2008
629 - AnyEvent::Strict did errornously flag a fileno of 0 as
630 illegal.
631 - reduce memory usage and slightly speed up the pure perl backend
632 by only storing the file descriptor, not the file handle.
633 - add missing autocork method to AnyEvent::Handle (reported
634 by Adam Rosenstein).
635 - AnyEvent::DNS->resolve errornously documented an $rcode result
636 argument, but there isn't (spotted by Henrik Krohns).
637 - the naked truth about Tk - it's basically unmaintained.
638
6394.32 Mon Nov 3 22:46:32 CET 2008
640 - fix AnyEvent::Socket::resolve_sockaddr to properly support
641 unix sockets again and choose a proper default for
642 the socket type.
643 - fix call to resolve_sockaddr in tcp_connect.
644
6454.31 Thu Oct 30 04:41:48 CET 2008
646 - implemented AnyEvent::Handle->destroy method.
647 - hint about unexpected effects in TLS mode.
648 - speed up AnyEvent::Impl::Perl by using more arrays
649 instead of hashes without the slightest loss of readability :)
650 - work around a perl argument refcounting bug.
651 - with some perl patches applied, AnyEvent now seems leak-free.
652
6534.3 Fri Oct 3 09:18:43 CEST 2008
654 - AnyEvent will now install a no-op signal handler for SIGPIPE,
655 unless one has been installed already.
656 - warn about not loaidng AnyEvent::Impl::POE early enough
657 (patch by Adam Rosenstein).
658 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
659 - fix a bug in where in SSL connect mode, the
660 client would first wait for some data by the server and
661 otherwise hang. (reported and analysed in an absolutely
662 exemplary manner by Adam Rosenstein).
663 - fix a bug in where SSL EOF would not be treated
664 as stream EOF, putting the connection into a hung state.
665 (reported and analysed in an absolutely exemplary manner by Adam
666 Rosenstein).
667 - fix a potential segfault when the TLS context
668 would go missing in a read callback (e.g. due to stoptls) -
669 Net::SSLeay of course makes no type checking whatsoever.
670 - AnyEvent::Handle will not stop the read watcher in TLS mode.
671 - AnyEvent::Handle->stoptls will send a TLS close notify
672 instead of simply destroying the stream now.
673 - fix a bug where large blocks of data written in TLS mode
674 would not be sent unless triggered by receives.
675 - on_drain will now take the tls write buffer into account.
676 - SSL operations have been streamlined a bit and should be faster
677 now, more to come, though.
678 - added a FAQ section, will document nontrivial issues in there.
679
6804.234 Mon Sep 29 04:08:13 CEST 2008
681 - fix child watcher documentation: only child exits will be reported,
682 no trace events.
683 - mention SIGPIPE in AnyEvent::Handle.
684 - perl backend will now use POSIX::times as monotonic clock
685 when available and the monotonic clock option is not.
686 - run even without Time::HiRes in some cases now, but do not
687 provide sub-second accuracy (all *supported* perl versions have
688 Time::HiRes).
689
6904.233 Fri Aug 22 01:48:05 CEST 2008
691 - fix a bug in the testsuite, causing a hang.
692 - clarified lots of AnyEvent::Handle constructor arguments.
693
6944.232 Thu Aug 21 20:44:25 CEST 2008
695 - fix a bug in AnyEvent::Handle that could cause two fatal error
696 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
697 (testcase by Leon Brocard).
698 - clarified on_eof/on_error documentation a bit.
699
7004.231 Tue Jul 29 13:12:15 CEST 2008
701 - remove some debugging code left in AnyEvent::Util::fork_call
702 (and no, it's impossible to implement with the broken windows
703 perls without resource leaks or worse).
704
7054.23 Tue Jul 29 12:19:59 CEST 2008
706 - document the first parameter passed to condvar callbacks to be
707 the callback.
708 - add AnyEvent::Socket::{ntoa,aton} aliases.
709 - optimize the AE::Handle->push_read (line) for the default
710 eol marker.
711 - optimize push_read (packstring|storable) for small packets.
712 - invoke on_error callback when no on_eof callback is set.
713 - fix a bug in push_read (storable) of unknown impact.
714
7154.22 Sun Jul 20 16:34:13 CEST 2008
716 - new function AnyEvent::Socket::parse_hostport.
717 - as the bulkheads at microsoft can't even get getprotobyname reliably
718 working on their shitty fucking broken os we need to hardcode
719 some common protocol numbers in AnyEvent::Socket. How can
720 people even bother with such a pile of shit as windows.
721
7224.21 Thu Jul 17 14:40:05 CEST 2008
723 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
724 which is incompatible to previous invocation syntax.
725 - work around more windows perl bugs in fork_call. I cannot imagine
726 why anybody would bother implementing fork in such an obviously
727 unusable way.
728 - avoid calling $do_retry if the request has already finished in
729 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
730 K. Sharifulin).
731
7324.2 Sat Jul 12 22:42:11 CEST 2008
733 - support an "interval" argument for repeating timers.
734 - fix a bug in the pure perl event loop that caused timers to
735 be delayed under some circumstances.
736 - implement optional argument checking via PERL_ANYEVENT_STRICT.
737 - not importing anything from Event causes it to malfunction
738 and not use Time::HiRes, work around this bug.
739 - more bug workarounds against the endlessly broken Tk module,
740 seems to work now on 32 bit machines, but less so on 64 bit.
741 - minor optimisations applied to most backends.
742
7434.161 Fri Jul 4 14:24:48 CEST 2008
744 - fixed the t/handle/02_write.t, it read after write, with the
745 assumption that all data could be written before read
746 (reported by Jim Razmus).
747
7484.160 Thu Jul 3 04:02:21 CEST 2008
749 - re-registering signal watchers after unregistering one did
750 not work if a backend falls back on AnyEvent's default
751 implementation (also affected child watchers).
752 - new AnyEvent::Handle options: autocork and no_delay.
753
7544.152 Sun Jun 22 14:15:44 CEST 2008
755 - allow for 32-bit perls that implement shifts differently
756 on different architectures in parse_ipv4 (reported and
757 analysed by Keiichi DAIBA).
758
7594.151 Fri Jun 6 17:34:24 CEST 2008
760 - make sure specifying _only_ on_read and never pushing reads
761 works.
762
7634.15 Fri Jun 6 13:00:46 CEST 2008
764 - the pure perl backend would keep some watchers alive when more than
765 one watcher was registered for the same fd.
766 - new "packstring" and "storable" read and write types
767 for AnyEvent::Handle.
768 - allow on_eof handler to be called after on_error with EPIPE returns.
769 - do not immediately call on_read callback in handle constructor.
770
7714.14 Thu Jun 5 20:29:31 CEST 2008
772 - Fixed a bug in DNS SRV priority sorting.
773 - AnyEvent::Util::guard now reports runtime errors while
774 executing the guard block as warnings.
775 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
776
7774.13 Thu Jun 5 00:47:59 CEST 2008
778 - AnyEvent::DNS only followed cname chains with length 2,
779 contrary to documentation. bumped it up to 10, thanks to
780 microsoft, the current broken-dns-config-king.
781 - AnyEvent::DNS didn't check for socket return status
782 but instead relied on perl not creating filehandles in that
783 case - too bad it gives you a bogus file handle (reported
784 and analysed by Vladimir Timofeev).
785 - fix queue management logic in AnyEvent::Handle:
786 when on_read was registered, the queue was empty
787 and no progress could be made AnyEvent::Handle would
788 enter an endless loop.
789 - correctly start reading again when the handle became
790 busy again after idling.
791 - correctly treat tls shut-downs as EOF condition
792 (for the time being, would be nice to have a callback for that).
793 - correctly call eof callback under all conditions
794 (hopefully).
795 - the timeout callback did not expect that $self can go
796 away any time.
797
7984.12 Tue Jun 3 10:58:04 CEST 2008
799 - include AnyEvent::Intro, a tutorial for anyevent,
800 anyevent::socket and anyevent::handle.
801 - allow more options in on_error.
802
8034.11 Fri May 30 23:42:25 CEST 2008
804 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
805 reverse_lookup and reverse_verify to replace it, support
806 v4mapped and v4compat addresses.
807 - provide more documentation for the resolver class.
808 - really replace longest run of :0: by :: in format_address,
809 also properly convert :: and ::1 again.
810 - support NAPTR record name and decode it.
811 - implement random weight sampling for SRV records, as per
812 rfc 2782.
813 - correctly abort on srv-record targets of ".".
814 - added AnyEvent::DNS::wait_for_slot.
815 - in the unlikely event of a virtual circuit connection
816 being invalidated by a delayed udp reply, AnyEvent::DNS
817 could die. this has been fixed.
818 - plug a probable memleak in the DNS vc code.
819 - use configured timeout also for VC request phase.
820 - implement timeout and max_outstanding methods
821 for resolver class.
822 - update version numbers in all modules.
823
8244.1 Thu May 29 05:45:40 CEST 2008
825 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
826 which mirrors their purpose better (old names still available
827 for a while). Also moved AnyEvent::DNS::addr to
828 AnyEvent::Socket::resolve_sockaddr.
829 - implement AnyEvent->time and AnyEvent->now.
830 - fix IPv6 support in pack_sockaddr.
831 - officially un-experimentalise all newly introduced code.
832 - support unix domain sockets everywhere by specifying
833 a host of "unix/" and the pathname as service.
834 - implement an activity timeout in AnyEvent::Handle.
835 - added a regex read type to AnyEvent::Handle.
836 - added a json read/write type to AnyEvent::Handle.
837 - always croak in on_error in AnyEvent::Handle.
838 - document how to subclass AnyEvent::Handle.
839 - implement AnyEvent::Util::fork_call.
840 - add support for IPv6 nameservers and nameserver statements.
841 - work around _yet_ _another_ windows perl bug in where empty
842 select masks cause errors in the pure perl backend (this
843 workaround cannot be implemented for the other event loops,
844 but EV also works around this).
845 - supply AnyEvent::Socket::address_type and make good use of it.
846 - clarify and fix the Handle documentation w.r.t $self vs. $handle
847 vs. callback arguments.
848 - add some recommends to the META.yml.
849
8504.05 Mon May 26 19:44:06 CEST 2008
851 - some platforms ignore the rfc and prepend an extra sa_len member to
852 the sockaddr structure, cater for those. also use sockaddr_family,
853 didn't know it was there, but comes in handy for the workaround.
854 - undo the import WIN32 hack.
855
8564.04 Mon May 26 08:03:31 CEST 2008
857 - try to work around yet another windows bug: failed connects are reported
858 as if it were out-of-band data. windows users: you suck. Only
859 EV, Glib, Event and the pure perl backend can handle this condition.
860 - optimize write algorithm in AnyEvent::Handle.
861 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
862 - tcp_connect now properly iterates through all targets.
863 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
864 support entirely. also hardcode some AF_INET6 constants for the
865 sake of perl 5.8.
866 - vastly improved nameserver/suffix detection algorithm for windows
867 (still a hack...)
868 - try to use a fallback dns server if no dns servers could be found.
869 - splatter around some local $SIG{__DIE__}'s to work around bad
870 die handlers.
871
8724.03 Sun May 25 05:05:57 CEST 2008
873 - work around perl bugs on windows, where perl returns undocumented
874 error codes for sysread, syswrite etc.
875 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
876 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
877 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
878 - updated documentation.
879 - lots of workarounds for perl 5.8 quirks.
880 - implement a more flexible type system for both reads and writes
881 in AnyEvent::Handle.
882 - support IPv6 in tcp_server.
883 - pass local hostname and port to the prepare callback in tcp_server.
884 - make types extendable in AnyEvent::Handle.
885 - croak, not die, in AnyEvent::Handle;
886
8874.0 Sat May 24 19:58:08 CEST 2008
888 - added (experimental) AnyEvent::DNS module.
889 - added (experimental) AnyEvent::Socket module.
890 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
891 - make condvar->end call ->send as documented.
892 - make condvar constructor support the documented cb => argument.
893 - overload condvars so you can use them instead of code references.
894 - support more of the POSIX forms instead of a pure dotted quad for
895 verbatim ip addresses.
896 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
4 - try very hard not to freeze on broken windows perls. 897 - try very hard not to freeze on broken windows perls.
5 - added (experimental) AnyEvent::DNS module.
6 - make on_eof optional in AnyEvent::Handle. 898 - make on_eof optional in AnyEvent::Handle.
7 - support service names in tcp_connect and tcp_server. 899 - support service names in tcp_connect and tcp_server.
8 - work around netbsd/openbsd bug causing unix sockets not to 900 - work around netbsd/openbsd bug causing unix sockets not to
9 become writable untill empty. 901 become writable until empty.
10 - only make a dummy read for error detection when getpeername 902 - only make a dummy read for error detection when getpeername
11 returns ENOTCONN. 903 returns ENOTCONN.
12 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 904 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
13 override. 905 override.
906 - new AnyEvent::Guard::cancel.
14 907
153.5 Sat May 17 23:17:14 CEST 2008 9083.5 Sat May 17 23:17:14 CEST 2008
16 - gracefully deal with systems not defining CLOCK_MONOTONIC 909 - gracefully deal with systems not defining CLOCK_MONOTONIC
17 (as opposed to not supporting the clock itself, which worked in 3.41). 910 (as opposed to not supporting the clock itself, which worked in 3.41).
18 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server 911 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
… …
30 923
313.41 Sun May 11 19:53:13 CEST 2008 9243.41 Sun May 11 19:53:13 CEST 2008
32 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 925 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
33 926
343.4 Sun May 11 00:29:25 CEST 2008 9273.4 Sun May 11 00:29:25 CEST 2008
35 - complete re-implementation of condvars with a hopefully much mroe useful 928 - complete re-implementation of condvars with a hopefully much more useful
36 API (backward compatible functions exist but are undocumented). 929 API (backward compatible functions exist but are undocumented).
37 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 930 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
38 (still subject to change). 931 (still subject to change).
39 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 932 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
40 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 933 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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