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

Comparing AnyEvent/Changes (file contents):
Revision 1.135 by root, Thu May 29 23:15:38 2008 UTC vs.
Revision 1.564 by root, Thu Dec 13 02:33:26 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: FAQ: common pitfalls? invoke-before-return
7TODO: FAQ: scope-ids? hosts file order?
8TODO: AnyEvent::Handle with AnyEvent::IO?
9
10 - AnyEvent::Debug shell can now run coro shell commands, if
11 available.
12 - t/63* tests were wrongly in MANIFEST.
13
147.04 Thu Nov 15 02:23:58 CET 2012
15 - debugging code left in.
16
177.03 Thu Nov 15 02:16:11 CET 2012
18 - AnyEvent::Socket::inet_aton did not work when dns resolution was
19 used to find the addresses (analyzed and patch by Maxime Soulé).
20 - fix a memory leak in the /etc/hosts lookup code when hosts don't
21 resolve and are not in hosts (analyzed by sten).
22
237.02 Tue Aug 14 04:27:58 CEST 2012
24 - AnyEvent::Util::run_cmd could block indefinitely (analyzed and test
25 program by Yorhel).
26 - verified that AnyEvent::Socket follows RFC5952.
27 - try to parse "ADDR#PORT" in addition to "ADDR PORT".
28
297.01 Sun May 13 01:03:17 CEST 2012
30 - fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but
31 not available, instead of throwing an exception.
32 - use File::Spec to get the tmpdir in t/*, to avoid needless
33 failures on (most, not mine :) windows boxes.
34 - new handle read types: tls_detect and tls_autostart.
35
367.0 Fri Apr 13 06:33:30 CEST 2012
37 - child watchers are broken in POE 1.352 (also many earlier
38 versions) and there seems to be no way to work around it, as POE
39 itself is inherently racy. Document this regression and add a delay
40 in t/68_poe_03_child.t for the time being.
41 - new module AnyEvent::IO, that is a frontend to either a pure-perl
42 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
43 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
44 - load /etc/hosts only when DNS has no answer.
45 - stat /etc/hosts on every access and reload it if it changed.
46 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
47 - fix a buggy croak in the dh parameter reading in AnyEvent::TLS.
48 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
49 a side effect, they now use true appending as opposed to libc
50 appending, and the file might not have been opened when the function
51 returns.
52 - the default logging level was not properly documented in a variety of
53 places, this has hopefully been rectified.
54 - updated uts46data.pl for unicode 6.1.0.
55 - made log messages generated by AnyEvent submodules not
56 include the package name anymore, as it will be logged
57 by default already.
58 - upgrade to the trick used by common-sense 3.5 to work
59 around extra warning torture/breakage under perl 5.15.x.
60 - log messages by anyevent are now ucfirst, are usually full
61 sentences and do no longer include the package name.
62 - the storable read type would sometimes throw an exception instead
63 of causing EBADMSG (analyzed by Maxime Soulé).
64 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
65 keysize (reported by Paul Howarth).
66 - documented Rocco fabricating statements about AnyEvent and me.
67
686.14 Tue Jan 31 20:00:24 CET 2012
69 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
70 probably an editing mistake (reported by Darin McBride).
71
726.13 Thu Jan 12 07:27:01 CET 2012
73 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
74 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
75 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
76 regex-matching inside the callback would change the parameter, despite
77 $1 being dynamically scoped per-block. this perl bug is now being
78 worked around (testcase by Cindy Wang).
79
806.12 Mon Dec 12 13:21:10 CET 2011
81 - $! was clobbered by subsequent calls in tcp_connect, due to
82 the postpone not saving/restoring it (analyzed by Richard Garnish).
83
846.11 Tue Nov 22 10:36:05 CET 2011
85 - Tk cannot create windows when tainted unless you set a title, so
86 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
87 - escape any nonprintable/nonascii characters when stringifying
88 backtraces.
89 - log the reason loading an interface module fails to load at level debug.
90
916.1 Tue Oct 4 19:44:30 CEST 2011
92 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
93 and above), and some messages inside AnyEvent have been elevated
94 to higher log levels to print by default.
95 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
96 - work around yet another signal race bug, newly introduced by POE. My
97 how that module sucks.
98 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
99 support ipv6-only operation even with defaults (also use ::1 instead
100 of 127.0.0.1).
101 - give better guidelines on how to chose a logging level.
102 - AnyEvent::Log can now cap (limit) log message priority by context.
103 - AE::log will now save and restore $! and $@, so callers don't have to.
104 - AE::Strict now checks that registered fds are still valid at regular
105 intervals.
106 - skip frequently called 'require' calls if the module appears loaded.
107 - increase signal latency test timer considerably for poor little osx,
108 which can't be bothered with basic posix functionality such as signal
109 delivery.
110 - mention some debugging aids in the tutorial.
111 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
112 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
113 of the context as title.
114 - AnyEvent::Strict would not restore read-write access to $_ when a callback
115 died - this is a user bug, but since AE::Strict exists to help dbeugging...
116 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
117 to work with it - the author is well aware of the issues but refuses
118 to fix them. This extreme step has been taken because the module actively
119 attacks the goals of AnyEvent and makes it harder for module authors to use
120 AnyEvent.
121
1226.02 Fri Aug 26 20:08:31 CEST 2011
123 - log_to_syslog now accepts facility strings.
124 - log_to_syslog would not allow facilities - "$facility|$level"
125 *literally* means that string. sheesh.
126 - the CPAN parser got confused and thought we provide Sys::Syslog, try
127 to work around.
128 - renamed AnyEvent::Impl::FLTK2 to FLTK.
129
1306.01 Fri Aug 26 07:04:11 CEST 2011
131 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
132 the TTL as fourth element - this affects ->resolve, ->request
133 and ->dns_unpack indirectly. Convenience functions such as
134 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
135 AnyEvent::Socket.
136 - INCOMPATIBLE CHANGE: backend authors now should not implement
137 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
138 methods.
139 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
140 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
141 - DEPRECIATION: please change
142 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
143 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
144 in your programs, for future compatibility.
145 - *blush* idle watcher emulation was borked.
146 - AnyEvent::Strict now checks the AE::xxx functions as well.
147 - querying too long domain names would barf deep inside AnyEvent::DNS - since
148 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
149 for too long domain names explicitly and passes an undef to the callback
150 instead (reported by James Bromberger).
151 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
152 - new functions: AnyEvent::postpone, AnyEvent::log.
153 - new module: AnyEvent::Log, for simple logging needs.
154 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
155 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
156 a debug shell.
157 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
158 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
159 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
160 /etc/hosts can be found.
161 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
162 configuration.
163 - env variables can now be specified using AE_xyz instead of
164 PERL_ANYEVENT_xyz, subject to some rules.
165 - lots of new AnyEvent::Debug functionality.
166 - AnyEvent::Strict now checks for modifications of $_ by the callback
167 (by setting it to readonly while executing the callback, if possible).
168 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
169 Invest extra overhead to work around this peculiar behaviour.
170 - split the pure perl backend into a loop module and an impl
171 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
172 - some read types (regex, netstring, json, storable) could
173 errornously keep an AE::Handle object alive on their own
174 (reported by Mohammad Toossi).
175 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
176 supports /etc/hosts overrides.
177 - add more workarounds around child watcher bugs in IO::Async
178 and (newly introduced) in POE.
179 - use glib child watchers and try to work around its limitations,
180 if possible (based on a patch by Kevin Ryde).
181 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
182 - many read types in AnyEvent::Handle would malfunction during
183 transitions to TLS because of a "delete $self->{rbuf}".
184 (reported by Gerald Galster).
185 - improve condvar blocking wait performance (probably only noticable
186 with faster backends), also make it easier for event loops that
187 do not support blocking waits.
188 - PERL_ANYEVENT_MODEL now also allows full module names, although
189 the usefulness of this feature is not yet clear.
190 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
191 - add t/80_ssltext.t.
192 - no longer use AUTOLOAD for the initial loading (for no very good reason).
193 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
194 - do not run event loop detection multiple times when programs call their cached
195 copies of AnyEvent::detect.
196 - parse_hostport now accepts absolute paths als unix domain sockets.
197 - suppress the idiotic warning inside POE by patching the POE::Kernel
198 run flags instead of asking the user to jump through contortions.
199 - add Task::Weaken as dependency (as recommends only at this time).
200 - replace PF_UNSPEC by 0, for good old minix.
201
2026.0 (not officially released, but was on CPAN for about 15 minutes :/)
203
2045.34 Thu May 12 09:59:41 CEST 2011
205 - AE::TLS didn't support both cert_file and key at the same time,
206 due to some copy&paste bug (reported and analyzed by vti).
207 - improve AE::Handle timeout documentation slightly.
208
2095.33 Mon Feb 7 21:51:41 CET 2011
210 - never appeared on CPAN, for unknown reasons.
211 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
212 object when it's already gone (analyzed by Ryan Bullock).
213 - update idna algorithm(s) to UTS#46 version 6.0.0.
214 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
215 uts46 says it's an error, and testsuite makes more sense.
216 - remove leading dots in idn_to_ascii - testsuite says to remove,
217 uts46 says it's an error, and testsuite makes more sense.
218 - fixed a bug in idn ignored character processing.
219 - passes the uts46 6.0.0 testsuite except for three cases which
220 apparently are bugs in the testsuite itself.
221 - passes the uts46 6.0.1 testsuite except for eleven cases which
222 apparently are bugs in the testsuite itself.
223
2245.31 Sun Jan 23 11:49:19 CET 2011
225 - work around a POE bug where POE would first poll for new events and
226 invoke handlers BEFORE handling the message queue.
227 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
228 - stop_read now does just that in TLS mode. see its
229 documentation for caveats.
230 - update to common::sense 3.4 warning set (some warnings in perl impose
231 an arbitrary 32kb limit on regex matches, which affects
232 AnyEvent::Handle).
233 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
234 - new experimental AE::Handle::resettls method.
235
2365.3 Fri Dec 31 05:48:13 CET 2010
237 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
238 issues with IO::Async have been solved in that version, including
239 the availabilty of a default loop!
240 - implement new wbuf_max parameter for AnyEvent::Handle.
241 - added Cocoa::EventLoop backend by Daisuke Murase.
242 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
243 - clarify handle on_prepare documentation.
244 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
245 to avoid leaking them to any child processes.
246 - try to force files into the arch-dependend lib directory, for the
247 half-"turly obsessive compulsive person"(s) that depend on it.
248
2495.29 Sun Dec 5 10:49:21 CET 2010
250 - convert EV backend to EV 4.00 API (so better upgrade EV too).
251 - AE::Handle: implement dynamic read_size adjustment, add
252 max_read_size parameter, reduce default read_size to 2048.
253 - add some FAQ entries for servers and/or on_eof vs. on_error.
254 - work around OS bugs (cygwin again) when lingering in AE::Handle.
255 - reduce memory usage slightly when connect was used in AE::Handle.
256 - wrok around more segfaults in Tk.
257 - document tls_ctx => undef as valid.
258 - detect WNOHANG value at compiletime.
259 - include some probably linux-only support for building AnyEvent
260 as part of the perl core.
261 - improve compatibility of t/08_idna.t to perl 5.8.
262 - make AnyEvent::Strict truly optional (in case somebody builds a
263 minimal perl...).
264
2655.28 Wed Oct 13 04:14:23 CEST 2010
266 - due to a glitch, AnyEvent's internal getprotobyname wasn't
267 actually used in tcp_server.
268 - implement AnyEvent::CondVar->new method, which was documented but
269 missing.
270 - support autoloading of read types in unshift_read.
271 - AnyEvent::Strict: do the test for invalid fh arguments differently,
272 resulting in more informative error messages.
273 - new AnyEvent::FAQ manpage.
274 - clarify that you can actually call push_read etc. in on_read.
275 - put keys and certs in /tmp, not into the current directory,
276 and document this in the SECURITY CONSIDERATIONS sections
277 of AnyEvent::TLS.
278 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
279 requested by Ron Isaacson.
280 - document $AnyEvent::DNS::RESOLVER variable.
281 - incorporated major typo patches by Abhijit Menon-Sen.
282 - AnyEvent::Handle now croaks when negative timeouts are passed.
283 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
284 undocumented, but who knows when they might come in handy.
285
2865.271 Tue Jun 8 12:05:46 CEST 2010
287 - backport to perl 5.8.x.
288
2895.27 Sun Jun 6 12:12:05 CEST 2010
290 - postpone differently in AnyEvent::Socket now, as
291 when not, canceling the connection attempt might fail
292 (found by Felix Antonius Wilhelm Ostmann).
293 - explicitly check for non-stream sockets in AE::Handle, too many
294 clueless people fell into the trap of this somehow working.
295 - simplified and reworked the "OTHER MODULES" section.
296 - better/more condvar examples.
297
2985.261 Wed Apr 28 16:13:36 CEST 2010
299 - AF_INET6 was not properly used from Socket6 during configuration
300 time (found by Andy Grundman).
301
3025.26 Mon Apr 12 04:49:35 CEST 2010
303 - don't generate spurious readyness notifications when select
304 returns EINTR in the pure perl backend (analysed and testcase
305 by Bas Denissen).
306 - give same examples in SYNOPSIS sections of AnyEvent and AE.
307 - provider faster implementations for AE::time/now/now_update for
308 EV and Perl backends.
309 - provide faster AE::cv and AE::time implementations for all
310 backends.
311 - no longer support lower/mixed-case signal names (this was
312 never documented nor universally supported).
313 - some more memory µ-optimisations, and somewhat less messy
314 code to cope with both AnyEvent and AE APIs.
315
3165.251 Sat Mar 13 00:58:21 CET 2010
317 - make test failed if it was run before make install on perl <=
318 5.8.8, and older versions of perl make our live very difficult,
319 so write constants.pl during Makefile.PL time.
320
3215.25 Sat Mar 13 00:23:14 CET 2010
322 - fix a race condition in AnyEvent::Handle that would cause
323 a "bio_read: unsupported method" error (in _another_
324 ssl connection) after user code threw an exception.
325 - added AnyEvent::Handle->destroyed method.
326 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
327 - the AnyEvent::Util::fh_nonblocking function worked only
328 by lucky accident on win32.
329 - smaller and faster AnyEvent::Util::fh_nonblocking.
330 - when the (required!) Time::HiRes module is missing, AnyEvent
331 did not fall back to built-in time properly.
332 - do not load Fcntl at runtime, saving memory and loading time.
333 - precompile a number of constants and use them instead of runtime
334 detection and eval.
335 - free detection code after detection and similar memory optimisations.
336 - Perl backend timer interval best effort drift has been improved
337 (same algorithm as EV).
338 - update unicode idna mapping table.
339
3405.24 Tue Jan 5 11:39:43 CET 2010
341 - cygwin never reports errors from failed connects. 1.5 just gives
342 you continous readyness and EAGAIN, 1.7 is even more broken and
343 just hangs. work around both issues in a cygwin-specific hack.
344 - improve idle watcher documentation slightly.
345
3465.23 Sun Dec 20 23:48:00 CET 2009
347 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
348 - implement punycode_encode/decode, idn_nameprep,
349 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
350 - provide $AE::VERSION.
351 - removed traces of "no strict 'refs'".
352
3535.22 Sat Dec 5 03:51:13 CET 2009
354 - downgrade-or-fail in AnyEvent::Handle::push_write, to
355 diagnose encoding failures earlier and more succinctly.
356 (this works around bugs in perl, throwing away encoding info
357 when passing scalar data to extensions).
358 - add more examples to AnyEvent::Socket manpage.
359 - upgrade internal warning set to the same as common::sense 2.03.
360 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
361 (John Beppu).
362 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
363 - add google public dns servers to fallback server set and make sure
364 we load-balance properly between the three sets. also add all
365 fallback dns servers, not just a random one, to each dns config.
366
3675.21 Thu Nov 19 02:48:47 CET 2009
368 - fix a problem where socket constants were called with parameters
369 (spotted by David Friedland).
370 - fork_call never use'd POSIX (reported by Daisuke Maki).
371 - improve perl 5.6 compatibility further (but it still won't work
372 unless you rip out everything but the core).
373 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
374 uses 10MB of RAM, but doesn't open a console window. *sigh*.
375
3765.202 Wed Oct 14 22:35:44 CEST 2009
377 - AnyEvent::DNS would unexpectedly clobber $_ under windows
378 (analysed by Matthias Waldorf).
379 - AnyEvent::Handle::run_cmd can now pass the PID of the
380 newly-created process, which is much less useful than it might
381 sound (based on patch by Yann Kerherve).
382
3835.201 Tue Sep 29 12:09:25 CEST 2009
384 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
385 (reported by Torsten Foertsch).
386 - common::sense 2.0 could cause tcp_server to throw an exception
387 (analysed by elmex).
388
3895.2 Mon Sep 14 07:04:49 CEST 2009
390 - INCOMPATIBLE CHANGE: do no longer support register_read_type
391 and register_write_type in AnyEvent::Handle, instead support
392 package names (the facility was mostly abused).
393 - implement "packagename-as-read/write type" support in
394 AnyEvent::Handle.
395 - AnyEvent::Handle: new options "keepalive" and "oobinline".
396 - oobinline set by default to avoid security issues.
397 - the pure-perl event loop backend wrongly detected times() fallback
398 support (spotted by Pavel Boldin).
399
4005.12 Tue Sep 1 20:26:50 CEST 2009
401 - be more lenient when parsing resolv.conf files, as some people
402 use hashmarks as comment indicator inside directives (reported
403 by Michael S. Fischer).
404 - use same set of warnings as common::sense 2.0.
405 - fix a potential 32 bit overflow issue due to perl having problems
406 with large hex constants in the Perl backend.
407 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
408 by default.
409 - fix some prototypes.
410 - (optionally) require Async::Interrupt 1.02.
411 - replace opendns fallbacks by verizon/level3 public dns servers.
412
4135.112 Fri Aug 21 13:59:04 CEST 2009
414 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
415 using starttls with an illegal/uncreatable tls context.
416 - port to broken win32 perls.
417
4185.111 Sun Aug 16 18:44:36 CEST 2009
419 - perl errornously requires sockaddr_un structures returned by the
420 kernel to have a specific length. work around this bug by padding
421 sockaddr_un structures (found to fail on OS X, as perl only uses
422 the correct code for GNU/Linux, but likely to fail on other
423 OSes as well).
424
4255.11 Wed Aug 12 17:49:37 CEST 2009
426 - fixed a bug in run_cmd with input from filename.
427 - create a json coder object when none is given in write direction
428 to reduce dependency on JSON.
429
4305.1 Tue Aug 11 03:17:32 CEST 2009
431 - tcp_connect now reports the correct error in $!, not an unrelated
432 one (was broken since 4.91).
433 - AE::Handle did not properly clear rtimeout/wtimeout watchers
434 when retrying a tcp_connect.
435 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
436 - fixed wrong documentation in AE::Socket.
437
4385.01 Mon Aug 10 03:16:32 CEST 2009
439 - last release broke EV child watchers :/.
440
4415.0 Sun Aug 9 17:32:09 CEST 2009
442 - officially introduce and document the AE API.
443 - lowering the timeout at runtime did not have immediate effect
444 in AnyEvent::Handle.
445 - AE::Handle now has separate and independent read and write timeout
446 settings.
447 - probe for Guard module when AnyEvent::Util is loaded, not at
448 runtime (Event doesn't like eval "use" from callbacks, and it avoids
449 exporting the probe function).
450 - Impl::IOAsync's io watchers did not use the filehandle emulation
451 code to work around it's watcher limits.
452 - work around IO::Async's problems with overloaded objects.
453 - try to disable wearnings inside IO::Async, it's just too buggy.
454 - do no longer use Event by default, only EV or Perl, Event is too buggy.
455 - hackishly provide fast AE:: interface when EV or Perl is the backend.
456 - inherit VERSION from AnyEvent::VERSION in most modules.
457
4584.91 Thu Aug 6 15:42:45 CEST 2009
459 - AE::Handle::starttls could get out of sync when the read buffer
460 already contains some TLS handshake.
461 - AE::Handle did not properly free the TLS session, causing bigger
462 memory leaks in Net::SSLeay than expected (reported by toaster).
463 - AE::Socket::tcp_connect will now artificially delay invoking
464 the callback to avoid returning after invoking the callback.
465 - convert many internal watcher uses to AE API.
466
4674.9 Sat Aug 1 11:07:01 CEST 2009
468 - Glib silently fails when registering a timeout with
469 a negative value, so avoid that.
470 - call condvar callback immediately when it is set
471 after the condition is already signalled.
472 - check rbuf_max condition only after trying to consume
473 data.
474 - normalise signal numbers to names when using the ae's
475 signal handling, but do not document this (yet).
476 - pure perl signal emulation did not properly set nonblocking
477 mode on the signal pipe on win32 (this is unlikely to have caused
478 any issues).
479 - new module: AnyEvent::Debug.
480 - AnyEvent::Strict now checks that a signal was specified
481 by name (not name or number).
482 - reduce memory footprint in typical cases by ~50kb by
483 only compiling signal code when necessary.
484 - add AnyEvent::Handle->rbuf_max.
485 - grab the AE namespace for future new API, implement stubs for
486 said future/alternative API.
487 - new function: AnyEvent::Socket::format_hostport.
488
4894.881 Tue Jul 28 12:51:53 CEST 2009
490 - work around a bug in local in pre-5.10 perls, causing
491 AnyEvent::Handle to recurse when it should not
492 (analysed by elmex).
493
4944.88 Tue Jul 28 04:04:37 CEST 2009
495 - re-bless the handle into a dummy package after calling
496 AnyEvent::Handle::destroy, so the user does not need to check
497 for errors after every push_write etc.
498 - do not attempt to run t/02_signals.t on obviously broken
499 platforms.
500
5014.87 Sun Jul 26 02:06:16 CEST 2009
502 - do not attempt to linger when there is no longer a valid fh.
503 - enforce tls mode to be either accept or connect, do not simply
504 segfault in Net::SSLeay.
505 - AnyEvent::Handle can now call tcp_connect itself (new parameters
506 connect, on_prepare, on_connect and on_connect_error). Updated
507 tutorial accordingly.
508 - add AnyEvent::Impl::Irssi backend.
509
5104.86 Mon Jul 20 23:52:29 CEST 2009
511 - since the verbose warning is not enough, explicitly document
512 that versions before 1.33 of Net::SSLeay are not secure.
513 - work around signal handling races in Event and (...) Event::Lib.
514 - try to align signal-race timer to full-second boundaries.
515 - work around Tk not liking negative timeouts.
516 - don't complain of different grades of environmental unfriendlyness
517 in IO::Async.
518
5194.85 Sat Jul 18 06:16:14 CEST 2009
520 - nail the signal race problem in perl once and for all
521 (see $AnyEvent::MAX_SIGNAL_LATENCY).
522 - take advantage of Async::Interrupt if it is available.
523 - load Time::HiRes and Guard modules on demand only.
524 - add optional/recommended modules section to AnyEvent
525 documentation.
526 - reduce memory usage considerably (and reduce startup penalty)
527 by not using "strict", "warnings" and "overload" modules.
528 - work around buggy windows/openbsd perls and provide EBADMSG
529 and EPROTO ourselves when missing.
530 - improve perl 5.6 compatibility of the core event loop.
531 - made Net::SSLeay version 1.33 a soft requirement.
532
5334.83 Fri Jul 17 16:56:26 CEST 2009
534 - implement AnyEvent::Socket::getprotobyname.
535 - AnyEvent::CondVar's will now detect recursive blocking
536 waits and will croak, as too many people fall into
537 this trap.
538 - AnyEvent::Handle will now call ->destroy on itself after
539 executing the on_error callback, instead of doing some
540 half-baked internal shutdown, for fatal errors.
541 - clarify on_eof behaviour w.r.t. the read queue and
542 on_read callbacks.
543 - ignore some possible spurious wake-ups in tcp_connect.
544
5454.82 Sat Jul 11 00:34:55 CEST 2009
546 - POE and Event backends didn't accept some callable objects as
547 callbacks.
548 - use Config module instead of POSIX module to detect signal names
549 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
550 module doesn't even have all POSIX signals :/.
551 - use more workarounds around the many refcnt/corruption bugs in
552 Event::Lib.
553 - work around a race condition in perl's select, causing t/03_child.t
554 to rarely fail.
555
5564.81 Thu Jul 9 10:30:30 CEST 2009
557 - AnyEvent::Handle didn't properly diagnose write errors
558 (it expected -1 from syswrite, how lame... :).
559 - support file descriptors in addition to file handles
560 in AnyEvent->io.
561 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
562 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
563 and PERL_ANYEVENT_CA_PATH.
564 - provide a sensible synopsis section for AnyEvent::TLS.
565 - add a "supported backends" section to the manpage.
566 - added simple io watcher test to testsuite, using a
567 portable_socketpair.
568 - tried to improve the stability of the Event::Lib backend,
569 YMMV.
570
5714.8 Mon Jul 6 23:45:16 CEST 2009
572 - AnyEvent::DNS did not properly follow CNAME records with
573 uppercase targets.
574 - AnyEvent::DNS would errornously return AAAA records
575 with v4 mapped addresses (a faulty record) as ipv4 addresses,
576 causing AnyEvent::Socket to throw an exception.
577 - added new module AnyEvent::TLS for easier SSL/TLS context
578 creation, with many options including hostname verification,
579 secure default configuration, lots of documentation and,
580 predefined diffie-hellman keys for perfect forward security
581 and much more. get it while it's still fresh!
582 - use AnyEvent::TLS in AnyEvent::Handle for context management.
583 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
584 so AnyEvent::Socket users have smaller memory footprint
585 in the common case.
586 - add AnyEvent::Handle->push_shutdown method.
587 - add an additional error message parameter to AnyEvent::Handle's
588 on_error callback (for TLS, $! is still available).
589 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
590 - make AnyEvent::Handle more robust against early conenction
591 failures (during new), and return C<undef> in that case
592 from the constructor.
593 - AnyEvent::Handle will now try to load only JSON::XS first,
594 then fall back to JSON.
595 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
596 and because it was documented that way.
597
5984.451 Fri Jul 3 00:28:58 CEST 2009
599 - do not clear rbuf when shutting down an AnyEvent::Handle
600 object - doing so breaks AnyEvent::HTTP.
601
6024.45 Mon Jun 29 22:59:26 CEST 2009
603 - a write error could cause AnyEvent::Handle to create
604 an I/O watcher with an undefined $fh.
605 - special-case mapped ipv4 addresses in both
606 AnyEvent::Socket::format_address and parse_address,
607 to treat them just like ipv4 addresses.
608 - updated and overhauled the AnyEvent::Intro doc.
609 - implement AnyEvent::Socket::format_ipv4/ipv6.
610 - slightly speed up portable_pipe/socketpair.
611 - expand condvar begin/end documentation.
612
6134.42 Fri Jun 26 08:32:18 CEST 2009
614 - preliminary and neccesarily incomplete support for IO::Async.
615 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
616 case it was set to IGNORE, to ensure we can catch child statuses
617 even when the calling env acts stupidly.
618 - updated benchmarks with IO::Async, which performs very well.
619
6204.412 Wed Jun 24 01:35:57 CEST 2009
621 - support an "untaint" attribute for AnyEvent::DNS
622 and set it on the default resolver.
623 - implement and document AnyEvent::Impl::Perl::loop.
624 - remove all anyevent-env variables from %ENV
625 when running in tainted mode.
626 - mention and extend the IO::Lambda benchmark.
627
6284.411 Sun Jun 7 18:48:02 CEST 2009
629 - do not try to use F_SETFD on windows.
630
6314.41 Thu May 14 06:40:11 CEST 2009
632 - work around issues in older perls (5.8.0?) when
633 a signal handler is deleted from the %SIG hash.
634 - use POSIX::_exit in child test, to avoid running
635 destructors.
636 - speed up CHLD handling by relying on SIGCHLD being
637 synchronously delivered, even when we roll our own
638 implementation.
639 - AnyEvent::DNS: add the "dname" resource record name
640 for cosmetic reasons.
641
6424.4 Sun Apr 26 20:12:33 CEST 2009
643 - implemented idle watchers, where applicable.
644 - AnyEvent->time died when Event backend was in use.
645 - fix a memleak in the Tk backend.
646 - sped up Tk timer handling.
647 - clip negative "after" values to 0 in AnyEvent::Impl::Event
648 to avoid spamming out warning messages.
649 - fix Qt timers without interval.
650 - avoid Qt zero-timeout special case, in old libqt's.
651
6524.352 Mon Apr 20 16:31:11 CEST 2009
653 - fix AnyEvent::Strict error messages for child watchers.
654 - fix/update Makefile.PL "recommends" versions.
655 - add AnyEvent->now_update.
656
6574.351 Sat Apr 11 07:56:14 CEST 2009
658 - actually make the signal pipe work under win32.
659 - localise $! in signal handler to avoid changing $!, although
660 perl itself does not seem to save/restore errno either.
661 - set the cloexec flag on the signal pipe (normally set by perl too).
662
6634.35 Fri Mar 27 11:48:20 CET 2009
664 - event models relying on AnyEvent's signal watcher emulation did
665 invoke the callback asynchronously, contradicting documentation
666 and causing signals to get lost (this includes AnyEvent's own
667 event loop). AnyEvent now uses the standard pipe trick to make
668 callback execution synchronous to the event loop.
669 - AnyEvent::Handle didn't free TLS context data on DESTROY
670 (patch by Pavel Shaydo).
671 - work around the ever-incompatibly-changing API of MakeMaker.
672 - document that changing global variables without restoring them
673 is a bad idea in Perl (noted by Adam Rosenstein).
674 - AnyEvent::Strict now barfs if ->io is passed a file.
675
6764.34 Thu Feb 12 18:32:45 CET 2009
677 - separately buffer TLS read data, as otherwise the read queue could
678 deadlock as receiving data is not expected while draining the
679 read queue (which cna only happen with TLS).
680 - raise EBADMSG error on JSON decoding errors.
681 - fix some minor manpage bugs (reported by Maximilian Gaß).
682 - speed up select bitmask parsing quite a bit in the pure perl
683 backend.
684 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
685
6864.331 Tue Jan 6 21:07:25 CET 2009
687 - socketpair fails on many vista machines because vista has
688 completely broken accept/getpeername and getsockname functions,
689 so we provide our own socketpair emulation that kind of works
690 (AnyEvent::Util::portable_pipe).
691 - new function: AnyEvent::Util::portable_socketpair.
692 - take advantage of the Guard module if it exists.
693
6944.33 Fri Nov 21 02:35:40 CET 2008
695 - AnyEvent::Strict did errornously flag a fileno of 0 as
696 illegal.
697 - reduce memory usage and slightly speed up the pure perl backend
698 by only storing the file descriptor, not the file handle.
699 - add missing autocork method to AnyEvent::Handle (reported
700 by Adam Rosenstein).
701 - AnyEvent::DNS->resolve errornously documented an $rcode result
702 argument, but there isn't (spotted by Henrik Krohns).
703 - the naked truth about Tk - it's basically unmaintained.
704
7054.32 Mon Nov 3 22:46:32 CET 2008
706 - fix AnyEvent::Socket::resolve_sockaddr to properly support
707 unix sockets again and choose a proper default for
708 the socket type.
709 - fix call to resolve_sockaddr in tcp_connect.
710
7114.31 Thu Oct 30 04:41:48 CET 2008
712 - implemented AnyEvent::Handle->destroy method.
713 - hint about unexpected effects in TLS mode.
714 - speed up AnyEvent::Impl::Perl by using more arrays
715 instead of hashes without the slightest loss of readability :)
716 - work around a perl argument refcounting bug.
717 - with some perl patches applied, AnyEvent now seems leak-free.
718
7194.3 Fri Oct 3 09:18:43 CEST 2008
720 - AnyEvent will now install a no-op signal handler for SIGPIPE,
721 unless one has been installed already.
722 - warn about not loaidng AnyEvent::Impl::POE early enough
723 (patch by Adam Rosenstein).
724 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
725 - fix a bug in where in SSL connect mode, the
726 client would first wait for some data by the server and
727 otherwise hang. (reported and analysed in an absolutely
728 exemplary manner by Adam Rosenstein).
729 - fix a bug in where SSL EOF would not be treated
730 as stream EOF, putting the connection into a hung state.
731 (reported and analysed in an absolutely exemplary manner by Adam
732 Rosenstein).
733 - fix a potential segfault when the TLS context
734 would go missing in a read callback (e.g. due to stoptls) -
735 Net::SSLeay of course makes no type checking whatsoever.
736 - AnyEvent::Handle will not stop the read watcher in TLS mode.
737 - AnyEvent::Handle->stoptls will send a TLS close notify
738 instead of simply destroying the stream now.
739 - fix a bug where large blocks of data written in TLS mode
740 would not be sent unless triggered by receives.
741 - on_drain will now take the tls write buffer into account.
742 - SSL operations have been streamlined a bit and should be faster
743 now, more to come, though.
744 - added a FAQ section, will document nontrivial issues in there.
745
7464.234 Mon Sep 29 04:08:13 CEST 2008
747 - fix child watcher documentation: only child exits will be reported,
748 no trace events.
749 - mention SIGPIPE in AnyEvent::Handle.
750 - perl backend will now use POSIX::times as monotonic clock
751 when available and the monotonic clock option is not.
752 - run even without Time::HiRes in some cases now, but do not
753 provide sub-second accuracy (all *supported* perl versions have
754 Time::HiRes).
755
7564.233 Fri Aug 22 01:48:05 CEST 2008
757 - fix a bug in the testsuite, causing a hang.
758 - clarified lots of AnyEvent::Handle constructor arguments.
759
7604.232 Thu Aug 21 20:44:25 CEST 2008
761 - fix a bug in AnyEvent::Handle that could cause two fatal error
762 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
763 (testcase by Leon Brocard).
764 - clarified on_eof/on_error documentation a bit.
765
7664.231 Tue Jul 29 13:12:15 CEST 2008
767 - remove some debugging code left in AnyEvent::Util::fork_call
768 (and no, it's impossible to implement with the broken windows
769 perls without resource leaks or worse).
770
7714.23 Tue Jul 29 12:19:59 CEST 2008
772 - document the first parameter passed to condvar callbacks to be
773 the callback.
774 - add AnyEvent::Socket::{ntoa,aton} aliases.
775 - optimize the AE::Handle->push_read (line) for the default
776 eol marker.
777 - optimize push_read (packstring|storable) for small packets.
778 - invoke on_error callback when no on_eof callback is set.
779 - fix a bug in push_read (storable) of unknown impact.
780
7814.22 Sun Jul 20 16:34:13 CEST 2008
782 - new function AnyEvent::Socket::parse_hostport.
783 - as the bulkheads at microsoft can't even get getprotobyname reliably
784 working on their shitty fucking broken os we need to hardcode
785 some common protocol numbers in AnyEvent::Socket. How can
786 people even bother with such a pile of shit as windows.
787
7884.21 Thu Jul 17 14:40:05 CEST 2008
789 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
790 which is incompatible to previous invocation syntax.
791 - work around more windows perl bugs in fork_call. I cannot imagine
792 why anybody would bother implementing fork in such an obviously
793 unusable way.
794 - avoid calling $do_retry if the request has already finished in
795 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
796 K. Sharifulin).
797
7984.2 Sat Jul 12 22:42:11 CEST 2008
799 - support an "interval" argument for repeating timers.
800 - fix a bug in the pure perl event loop that caused timers to
801 be delayed under some circumstances.
802 - implement optional argument checking via PERL_ANYEVENT_STRICT.
803 - not importing anything from Event causes it to malfunction
804 and not use Time::HiRes, work around this bug.
805 - more bug workarounds against the endlessly broken Tk module,
806 seems to work now on 32 bit machines, but less so on 64 bit.
807 - minor optimisations applied to most backends.
808
8094.161 Fri Jul 4 14:24:48 CEST 2008
810 - fixed the t/handle/02_write.t, it read after write, with the
811 assumption that all data could be written before read
812 (reported by Jim Razmus).
813
8144.160 Thu Jul 3 04:02:21 CEST 2008
815 - re-registering signal watchers after unregistering one did
816 not work if a backend falls back on AnyEvent's default
817 implementation (also affected child watchers).
818 - new AnyEvent::Handle options: autocork and no_delay.
819
8204.152 Sun Jun 22 14:15:44 CEST 2008
821 - allow for 32-bit perls that implement shifts differently
822 on different architectures in parse_ipv4 (reported and
823 analysed by Keiichi DAIBA).
824
8254.151 Fri Jun 6 17:34:24 CEST 2008
826 - make sure specifying _only_ on_read and never pushing reads
827 works.
828
8294.15 Fri Jun 6 13:00:46 CEST 2008
830 - the pure perl backend would keep some watchers alive when more than
831 one watcher was registered for the same fd.
832 - new "packstring" and "storable" read and write types
833 for AnyEvent::Handle.
834 - allow on_eof handler to be called after on_error with EPIPE returns.
835 - do not immediately call on_read callback in handle constructor.
836
8374.14 Thu Jun 5 20:29:31 CEST 2008
838 - Fixed a bug in DNS SRV priority sorting.
839 - AnyEvent::Util::guard now reports runtime errors while
840 executing the guard block as warnings.
841 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
842
8434.13 Thu Jun 5 00:47:59 CEST 2008
844 - AnyEvent::DNS only followed cname chains with length 2,
845 contrary to documentation. bumped it up to 10, thanks to
846 microsoft, the current broken-dns-config-king.
847 - AnyEvent::DNS didn't check for socket return status
848 but instead relied on perl not creating filehandles in that
849 case - too bad it gives you a bogus file handle (reported
850 and analysed by Vladimir Timofeev).
851 - fix queue management logic in AnyEvent::Handle:
852 when on_read was registered, the queue was empty
853 and no progress could be made AnyEvent::Handle would
854 enter an endless loop.
855 - correctly start reading again when the handle became
856 busy again after idling.
857 - correctly treat tls shut-downs as EOF condition
858 (for the time being, would be nice to have a callback for that).
859 - correctly call eof callback under all conditions
860 (hopefully).
861 - the timeout callback did not expect that $self can go
862 away any time.
863
8644.12 Tue Jun 3 10:58:04 CEST 2008
865 - include AnyEvent::Intro, a tutorial for anyevent,
866 anyevent::socket and anyevent::handle.
867 - allow more options in on_error.
868
8694.11 Fri May 30 23:42:25 CEST 2008
3 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide 870 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
4 reverse_lookup and reverse_verify to replace it, support 871 reverse_lookup and reverse_verify to replace it, support
5 v4mapped and v4compat addresses. 872 v4mapped and v4compat addresses.
6 - provide more documentation for the resolver class. 873 - provide more documentation for the resolver class.
7 - really replace longest run of :0: by :: in format_address, 874 - really replace longest run of :0: by :: in format_address,
8 also properly convert :: and ::1 again. 875 also properly convert :: and ::1 again.
9 - support NAPTR record name and decode it. 876 - support NAPTR record name and decode it.
10 - implement random weight sampling for SRV records, as per 877 - implement random weight sampling for SRV records, as per
11 rfc 2782. 878 rfc 2782.
879 - correctly abort on srv-record targets of ".".
880 - added AnyEvent::DNS::wait_for_slot.
881 - in the unlikely event of a virtual circuit connection
882 being invalidated by a delayed udp reply, AnyEvent::DNS
883 could die. this has been fixed.
884 - plug a probable memleak in the DNS vc code.
885 - use configured timeout also for VC request phase.
886 - implement timeout and max_outstanding methods
887 for resolver class.
888 - update version numbers in all modules.
12 889
134.1 Thu May 29 05:45:40 CEST 2008 8904.1 Thu May 29 05:45:40 CEST 2008
14 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address 891 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
15 which mirrors their purpose better (old names still available 892 which mirrors their purpose better (old names still available
16 for a while). Also moved AnyEvent::DNS::addr to 893 for a while). Also moved AnyEvent::DNS::addr to
… …
88 - support service names in tcp_connect and tcp_server. 965 - support service names in tcp_connect and tcp_server.
89 - work around netbsd/openbsd bug causing unix sockets not to 966 - work around netbsd/openbsd bug causing unix sockets not to
90 become writable until empty. 967 become writable until empty.
91 - only make a dummy read for error detection when getpeername 968 - only make a dummy read for error detection when getpeername
92 returns ENOTCONN. 969 returns ENOTCONN.
93 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 970 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
94 override. 971 override.
95 - new AnyEvent::Guard::cancel. 972 - new AnyEvent::Guard::cancel.
96 973
973.5 Sat May 17 23:17:14 CEST 2008 9743.5 Sat May 17 23:17:14 CEST 2008
98 - gracefully deal with systems not defining CLOCK_MONOTONIC 975 - gracefully deal with systems not defining CLOCK_MONOTONIC
… …
112 989
1133.41 Sun May 11 19:53:13 CEST 2008 9903.41 Sun May 11 19:53:13 CEST 2008
114 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 991 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
115 992
1163.4 Sun May 11 00:29:25 CEST 2008 9933.4 Sun May 11 00:29:25 CEST 2008
117 - complete re-implementation of condvars with a hopefully much mroe useful 994 - complete re-implementation of condvars with a hopefully much more useful
118 API (backward compatible functions exist but are undocumented). 995 API (backward compatible functions exist but are undocumented).
119 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 996 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
120 (still subject to change). 997 (still subject to change).
121 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 998 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
122 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 999 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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