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

Comparing AnyEvent/Changes (file contents):
Revision 1.147 by root, Wed Jun 4 22:47:27 2008 UTC vs.
Revision 1.551 by root, Wed Apr 18 09:44:09 2012 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: document TCP_*** constants
4TODO: add debug/trace logging to some modules
5TODO: use the eval-with-cb-call trick also to speed up JSON?
6TODO: common pitfalls? invoke-before-return
7TODO: AnyEvent::Handle with AnyEvent::IO?
8
97.01
10 - fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but
11 not available, instead of throwing an exception.
12 - use File::Spec to get the tmpdir in t/*, to avoid needless
13 failures on (most, not mine :) windows boxes.
14
157.0 Fri Apr 13 06:33:30 CEST 2012
16 - child watchers are broken in POE 1.352 (also many earlier
17 versions) and there seems to be no way to work around it, as POE
18 itself is inherently racy. Document this regression and add a delay
19 in t/68_poe_03_child.t for the time being.
20 - new module AnyEvent::IO, that is a frontend to either a pure-perl
21 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
22 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
23 - load /etc/hosts only when DNS has no answer.
24 - stat /etc/hosts on every access and reload it if it changed.
25 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
26 - fix a buggy croak in the dh parameter reading in AnyEvent::TLS.
27 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
28 a side effect, they now use true appending as opposed to libc
29 appending, and the file might not have been opened when the function
30 returns.
31 - the default logging level was not properly documented in a variety of
32 places, this has hopefully been rectified.
33 - updated uts46data.pl for unicode 6.1.0.
34 - made log messages generated by AnyEvent submodules not
35 include the package name anymore, as it will be logged
36 by default already.
37 - upgrade to the trick used by common-sense 3.5 to work
38 around extra warning torture/breakage under perl 5.15.x.
39 - log messages by anyevent are now ucfirst, are usually full
40 sentences and do no longer include the package name.
41 - the storable read type would sometimes throw an exception instead
42 of causing EBADMSG (analyzed by Maxime Soulé).
43 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
44 keysize (reported by Paul Howarth).
45 - documented Rocco fabricating statements about AnyEvent and me.
46
476.14 Tue Jan 31 20:00:24 CET 2012
48 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
49 probably an editing mistake (reported by Darin McBride).
50
516.13 Thu Jan 12 07:27:01 CET 2012
52 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
53 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
54 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
55 regex-matching inside the callback would change the parameter, despite
56 $1 being dynamically scoped per-block. this perl bug is now being
57 worked around (testcase by Cindy Wang).
58
596.12 Mon Dec 12 13:21:10 CET 2011
60 - $! was clobbered by subsequent calls in tcp_connect, due to
61 the postpone not saving/restoring it (analyzed by Richard Garnish).
62
636.11 Tue Nov 22 10:36:05 CET 2011
64 - Tk cannot create windows when tainted unless you set a title, so
65 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
66 - escape any nonprintable/nonascii characters when stringifying
67 backtraces.
68 - log the reason loading an interface module fails to load at level debug.
69
706.1 Tue Oct 4 19:44:30 CEST 2011
71 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
72 and above), and some messages inside AnyEvent have been elevated
73 to higher log levels to print by default.
74 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
75 - work around yet another signal race bug, newly introduced by POE. My
76 how that module sucks.
77 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
78 support ipv6-only operation even with defaults (also use ::1 instead
79 of 127.0.0.1).
80 - give better guidelines on how to chose a logging level.
81 - AnyEvent::Log can now cap (limit) log message priority by context.
82 - AE::log will now save and restore $! and $@, so callers don't have to.
83 - AE::Strict now checks that registered fds are still valid at regular
84 intervals.
85 - skip frequently called 'require' calls if the module appears loaded.
86 - increase signal latency test timer considerably for poor little osx,
87 which can't be bothered with basic posix functionality such as signal
88 delivery.
89 - mention some debugging aids in the tutorial.
90 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
91 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
92 of the context as title.
93 - AnyEvent::Strict would not restore read-write access to $_ when a callback
94 died - this is a user bug, but since AE::Strict exists to help dbeugging...
95 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
96 to work with it - the author is well aware of the issues but refuses
97 to fix them. This extreme step has been taken because the module actively
98 attacks the goals of AnyEvent and makes it harder for module authors to use
99 AnyEvent.
100
1016.02 Fri Aug 26 20:08:31 CEST 2011
102 - log_to_syslog now accepts facility strings.
103 - log_to_syslog would not allow facilities - "$facility|$level"
104 *literally* means that string. sheesh.
105 - the CPAN parser got confused and thought we provide Sys::Syslog, try
106 to work around.
107 - renamed AnyEvent::Impl::FLTK2 to FLTK.
108
1096.01 Fri Aug 26 07:04:11 CEST 2011
110 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
111 the TTL as fourth element - this affects ->resolve, ->request
112 and ->dns_unpack indirectly. Convenience functions such as
113 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
114 AnyEvent::Socket.
115 - INCOMPATIBLE CHANGE: backend authors now should not implement
116 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
117 methods.
118 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
119 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
120 - DEPRECIATION: please change
121 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
122 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
123 in your programs, for future compatibility.
124 - *blush* idle watcher emulation was borked.
125 - AnyEvent::Strict now checks the AE::xxx functions as well.
126 - querying too long domain names would barf deep inside AnyEvent::DNS - since
127 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
128 for too long domain names explicitly and passes an undef to the callback
129 instead (reported by James Bromberger).
130 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
131 - new functions: AnyEvent::postpone, AnyEvent::log.
132 - new module: AnyEvent::Log, for simple logging needs.
133 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
134 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
135 a debug shell.
136 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
137 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
138 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
139 /etc/hosts can be found.
140 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
141 configuration.
142 - env variables can now be specified using AE_xyz instead of
143 PERL_ANYEVENT_xyz, subject to some rules.
144 - lots of new AnyEvent::Debug functionality.
145 - AnyEvent::Strict now checks for modifications of $_ by the callback
146 (by setting it to readonly while executing the callback, if possible).
147 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
148 Invest extra overhead to work around this peculiar behaviour.
149 - split the pure perl backend into a loop module and an impl
150 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
151 - some read types (regex, netstring, json, storable) could
152 errornously keep an AE::Handle object alive on their own
153 (reported by Mohammad Toossi).
154 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
155 supports /etc/hosts overrides.
156 - add more workarounds around child watcher bugs in IO::Async
157 and (newly introduced) in POE.
158 - use glib child watchers and try to work around its limitations,
159 if possible (based on a patch by Kevin Ryde).
160 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
161 - many read types in AnyEvent::Handle would malfunction during
162 transitions to TLS because of a "delete $self->{rbuf}".
163 (reported by Gerald Galster).
164 - improve condvar blocking wait performance (probably only noticable
165 with faster backends), also make it easier for event loops that
166 do not support blocking waits.
167 - PERL_ANYEVENT_MODEL now also allows full module names, although
168 the usefulness of this feature is not yet clear.
169 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
170 - add t/80_ssltext.t.
171 - no longer use AUTOLOAD for the initial loading (for no very good reason).
172 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
173 - do not run event loop detection multiple times when programs call their cached
174 copies of AnyEvent::detect.
175 - parse_hostport now accepts absolute paths als unix domain sockets.
176 - suppress the idiotic warning inside POE by patching the POE::Kernel
177 run flags instead of asking the user to jump through contortions.
178 - add Task::Weaken as dependency (as recommends only at this time).
179 - replace PF_UNSPEC by 0, for good old minix.
180
1816.0 (not officially released, but was on CPAN for about 15 minutes :/)
182
1835.34 Thu May 12 09:59:41 CEST 2011
184 - AE::TLS didn't support both cert_file and key at the same time,
185 due to some copy&paste bug (reported and analyzed by vti).
186 - improve AE::Handle timeout documentation slightly.
187
1885.33 Mon Feb 7 21:51:41 CET 2011
189 - never appeared on CPAN, for unknown reasons.
190 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
191 object when it's already gone (analyzed by Ryan Bullock).
192 - update idna algorithm(s) to UTS#46 version 6.0.0.
193 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
194 uts46 says it's an error, and testsuite makes more sense.
195 - remove leading dots in idn_to_ascii - testsuite says to remove,
196 uts46 says it's an error, and testsuite makes more sense.
197 - fixed a bug in idn ignored character processing.
198 - passes the uts46 6.0.0 testsuite except for three cases which
199 apparently are bugs in the testsuite itself.
200 - passes the uts46 6.0.1 testsuite except for eleven cases which
201 apparently are bugs in the testsuite itself.
202
2035.31 Sun Jan 23 11:49:19 CET 2011
204 - work around a POE bug where POE would first poll for new events and
205 invoke handlers BEFORE handling the message queue.
206 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
207 - stop_read now does just that in TLS mode. see its
208 documentation for caveats.
209 - update to common::sense 3.4 warning set (some warnings in perl impose
210 an arbitrary 32kb limit on regex matches, which affects
211 AnyEvent::Handle).
212 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
213 - new experimental AE::Handle::resettls method.
214
2155.3 Fri Dec 31 05:48:13 CET 2010
216 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
217 issues with IO::Async have been solved in that version, including
218 the availabilty of a default loop!
219 - implement new wbuf_max parameter for AnyEvent::Handle.
220 - added Cocoa::EventLoop backend by Daisuke Murase.
221 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
222 - clarify handle on_prepare documentation.
223 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
224 to avoid leaking them to any child processes.
225 - try to force files into the arch-dependend lib directory, for the
226 half-"turly obsessive compulsive person"(s) that depend on it.
227
2285.29 Sun Dec 5 10:49:21 CET 2010
229 - convert EV backend to EV 4.00 API (so better upgrade EV too).
230 - AE::Handle: implement dynamic read_size adjustment, add
231 max_read_size parameter, reduce default read_size to 2048.
232 - add some FAQ entries for servers and/or on_eof vs. on_error.
233 - work around OS bugs (cygwin again) when lingering in AE::Handle.
234 - reduce memory usage slightly when connect was used in AE::Handle.
235 - wrok around more segfaults in Tk.
236 - document tls_ctx => undef as valid.
237 - detect WNOHANG value at compiletime.
238 - include some probably linux-only support for building AnyEvent
239 as part of the perl core.
240 - improve compatibility of t/08_idna.t to perl 5.8.
241 - make AnyEvent::Strict truly optional (in case somebody builds a
242 minimal perl...).
243
2445.28 Wed Oct 13 04:14:23 CEST 2010
245 - due to a glitch, AnyEvent's internal getprotobyname wasn't
246 actually used in tcp_server.
247 - implement AnyEvent::CondVar->new method, which was documented but
248 missing.
249 - support autoloading of read types in unshift_read.
250 - AnyEvent::Strict: do the test for invalid fh arguments differently,
251 resulting in more informative error messages.
252 - new AnyEvent::FAQ manpage.
253 - clarify that you can actually call push_read etc. in on_read.
254 - put keys and certs in /tmp, not into the current directory,
255 and document this in the SECURITY CONSIDERATIONS sections
256 of AnyEvent::TLS.
257 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
258 requested by Ron Isaacson.
259 - document $AnyEvent::DNS::RESOLVER variable.
260 - incorporated major typo patches by Abhijit Menon-Sen.
261 - AnyEvent::Handle now croaks when negative timeouts are passed.
262 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
263 undocumented, but who knows when they might come in handy.
264
2655.271 Tue Jun 8 12:05:46 CEST 2010
266 - backport to perl 5.8.x.
267
2685.27 Sun Jun 6 12:12:05 CEST 2010
269 - postpone differently in AnyEvent::Socket now, as
270 when not, canceling the connection attempt might fail
271 (found by Felix Antonius Wilhelm Ostmann).
272 - explicitly check for non-stream sockets in AE::Handle, too many
273 clueless people fell into the trap of this somehow working.
274 - simplified and reworked the "OTHER MODULES" section.
275 - better/more condvar examples.
276
2775.261 Wed Apr 28 16:13:36 CEST 2010
278 - AF_INET6 was not properly used from Socket6 during configuration
279 time (found by Andy Grundman).
280
2815.26 Mon Apr 12 04:49:35 CEST 2010
282 - don't generate spurious readyness notifications when select
283 returns EINTR in the pure perl backend (analysed and testcase
284 by Bas Denissen).
285 - give same examples in SYNOPSIS sections of AnyEvent and AE.
286 - provider faster implementations for AE::time/now/now_update for
287 EV and Perl backends.
288 - provide faster AE::cv and AE::time implementations for all
289 backends.
290 - no longer support lower/mixed-case signal names (this was
291 never documented nor universally supported).
292 - some more memory µ-optimisations, and somewhat less messy
293 code to cope with both AnyEvent and AE APIs.
294
2955.251 Sat Mar 13 00:58:21 CET 2010
296 - make test failed if it was run before make install on perl <=
297 5.8.8, and older versions of perl make our live very difficult,
298 so write constants.pl during Makefile.PL time.
299
3005.25 Sat Mar 13 00:23:14 CET 2010
301 - fix a race condition in AnyEvent::Handle that would cause
302 a "bio_read: unsupported method" error (in _another_
303 ssl connection) after user code threw an exception.
304 - added AnyEvent::Handle->destroyed method.
305 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
306 - the AnyEvent::Util::fh_nonblocking function worked only
307 by lucky accident on win32.
308 - smaller and faster AnyEvent::Util::fh_nonblocking.
309 - when the (required!) Time::HiRes module is missing, AnyEvent
310 did not fall back to built-in time properly.
311 - do not load Fcntl at runtime, saving memory and loading time.
312 - precompile a number of constants and use them instead of runtime
313 detection and eval.
314 - free detection code after detection and similar memory optimisations.
315 - Perl backend timer interval best effort drift has been improved
316 (same algorithm as EV).
317 - update unicode idna mapping table.
318
3195.24 Tue Jan 5 11:39:43 CET 2010
320 - cygwin never reports errors from failed connects. 1.5 just gives
321 you continous readyness and EAGAIN, 1.7 is even more broken and
322 just hangs. work around both issues in a cygwin-specific hack.
323 - improve idle watcher documentation slightly.
324
3255.23 Sun Dec 20 23:48:00 CET 2009
326 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
327 - implement punycode_encode/decode, idn_nameprep,
328 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
329 - provide $AE::VERSION.
330 - removed traces of "no strict 'refs'".
331
3325.22 Sat Dec 5 03:51:13 CET 2009
333 - downgrade-or-fail in AnyEvent::Handle::push_write, to
334 diagnose encoding failures earlier and more succinctly.
335 (this works around bugs in perl, throwing away encoding info
336 when passing scalar data to extensions).
337 - add more examples to AnyEvent::Socket manpage.
338 - upgrade internal warning set to the same as common::sense 2.03.
339 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
340 (John Beppu).
341 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
342 - add google public dns servers to fallback server set and make sure
343 we load-balance properly between the three sets. also add all
344 fallback dns servers, not just a random one, to each dns config.
345
3465.21 Thu Nov 19 02:48:47 CET 2009
347 - fix a problem where socket constants were called with parameters
348 (spotted by David Friedland).
349 - fork_call never use'd POSIX (reported by Daisuke Maki).
350 - improve perl 5.6 compatibility further (but it still won't work
351 unless you rip out everything but the core).
352 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
353 uses 10MB of RAM, but doesn't open a console window. *sigh*.
354
3555.202 Wed Oct 14 22:35:44 CEST 2009
356 - AnyEvent::DNS would unexpectedly clobber $_ under windows
357 (analysed by Matthias Waldorf).
358 - AnyEvent::Handle::run_cmd can now pass the PID of the
359 newly-created process, which is much less useful than it might
360 sound (based on patch by Yann Kerherve).
361
3625.201 Tue Sep 29 12:09:25 CEST 2009
363 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
364 (reported by Torsten Foertsch).
365 - common::sense 2.0 could cause tcp_server to throw an exception
366 (analysed by elmex).
367
3685.2 Mon Sep 14 07:04:49 CEST 2009
369 - INCOMPATIBLE CHANGE: do no longer support register_read_type
370 and register_write_type in AnyEvent::Handle, instead support
371 package names (the facility was mostly abused).
372 - implement "packagename-as-read/write type" support in
373 AnyEvent::Handle.
374 - AnyEvent::Handle: new options "keepalive" and "oobinline".
375 - oobinline set by default to avoid security issues.
376 - the pure-perl event loop backend wrongly detected times() fallback
377 support (spotted by Pavel Boldin).
378
3795.12 Tue Sep 1 20:26:50 CEST 2009
380 - be more lenient when parsing resolv.conf files, as some people
381 use hashmarks as comment indicator inside directives (reported
382 by Michael S. Fischer).
383 - use same set of warnings as common::sense 2.0.
384 - fix a potential 32 bit overflow issue due to perl having problems
385 with large hex constants in the Perl backend.
386 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
387 by default.
388 - fix some prototypes.
389 - (optionally) require Async::Interrupt 1.02.
390 - replace opendns fallbacks by verizon/level3 public dns servers.
391
3925.112 Fri Aug 21 13:59:04 CEST 2009
393 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
394 using starttls with an illegal/uncreatable tls context.
395 - port to broken win32 perls.
396
3975.111 Sun Aug 16 18:44:36 CEST 2009
398 - perl errornously requires sockaddr_un structures returned by the
399 kernel to have a specific length. work around this bug by padding
400 sockaddr_un structures (found to fail on OS X, as perl only uses
401 the correct code for GNU/Linux, but likely to fail on other
402 OSes as well).
403
4045.11 Wed Aug 12 17:49:37 CEST 2009
405 - fixed a bug in run_cmd with input from filename.
406 - create a json coder object when none is given in write direction
407 to reduce dependency on JSON.
408
4095.1 Tue Aug 11 03:17:32 CEST 2009
410 - tcp_connect now reports the correct error in $!, not an unrelated
411 one (was broken since 4.91).
412 - AE::Handle did not properly clear rtimeout/wtimeout watchers
413 when retrying a tcp_connect.
414 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
415 - fixed wrong documentation in AE::Socket.
416
4175.01 Mon Aug 10 03:16:32 CEST 2009
418 - last release broke EV child watchers :/.
419
4205.0 Sun Aug 9 17:32:09 CEST 2009
421 - officially introduce and document the AE API.
422 - lowering the timeout at runtime did not have immediate effect
423 in AnyEvent::Handle.
424 - AE::Handle now has separate and independent read and write timeout
425 settings.
426 - probe for Guard module when AnyEvent::Util is loaded, not at
427 runtime (Event doesn't like eval "use" from callbacks, and it avoids
428 exporting the probe function).
429 - Impl::IOAsync's io watchers did not use the filehandle emulation
430 code to work around it's watcher limits.
431 - work around IO::Async's problems with overloaded objects.
432 - try to disable wearnings inside IO::Async, it's just too buggy.
433 - do no longer use Event by default, only EV or Perl, Event is too buggy.
434 - hackishly provide fast AE:: interface when EV or Perl is the backend.
435 - inherit VERSION from AnyEvent::VERSION in most modules.
436
4374.91 Thu Aug 6 15:42:45 CEST 2009
438 - AE::Handle::starttls could get out of sync when the read buffer
439 already contains some TLS handshake.
440 - AE::Handle did not properly free the TLS session, causing bigger
441 memory leaks in Net::SSLeay than expected (reported by toaster).
442 - AE::Socket::tcp_connect will now artificially delay invoking
443 the callback to avoid returning after invoking the callback.
444 - convert many internal watcher uses to AE API.
445
4464.9 Sat Aug 1 11:07:01 CEST 2009
447 - Glib silently fails when registering a timeout with
448 a negative value, so avoid that.
449 - call condvar callback immediately when it is set
450 after the condition is already signalled.
451 - check rbuf_max condition only after trying to consume
452 data.
453 - normalise signal numbers to names when using the ae's
454 signal handling, but do not document this (yet).
455 - pure perl signal emulation did not properly set nonblocking
456 mode on the signal pipe on win32 (this is unlikely to have caused
457 any issues).
458 - new module: AnyEvent::Debug.
459 - AnyEvent::Strict now checks that a signal was specified
460 by name (not name or number).
461 - reduce memory footprint in typical cases by ~50kb by
462 only compiling signal code when necessary.
463 - add AnyEvent::Handle->rbuf_max.
464 - grab the AE namespace for future new API, implement stubs for
465 said future/alternative API.
466 - new function: AnyEvent::Socket::format_hostport.
467
4684.881 Tue Jul 28 12:51:53 CEST 2009
469 - work around a bug in local in pre-5.10 perls, causing
470 AnyEvent::Handle to recurse when it should not
471 (analysed by elmex).
472
4734.88 Tue Jul 28 04:04:37 CEST 2009
474 - re-bless the handle into a dummy package after calling
475 AnyEvent::Handle::destroy, so the user does not need to check
476 for errors after every push_write etc.
477 - do not attempt to run t/02_signals.t on obviously broken
478 platforms.
479
4804.87 Sun Jul 26 02:06:16 CEST 2009
481 - do not attempt to linger when there is no longer a valid fh.
482 - enforce tls mode to be either accept or connect, do not simply
483 segfault in Net::SSLeay.
484 - AnyEvent::Handle can now call tcp_connect itself (new parameters
485 connect, on_prepare, on_connect and on_connect_error). Updated
486 tutorial accordingly.
487 - add AnyEvent::Impl::Irssi backend.
488
4894.86 Mon Jul 20 23:52:29 CEST 2009
490 - since the verbose warning is not enough, explicitly document
491 that versions before 1.33 of Net::SSLeay are not secure.
492 - work around signal handling races in Event and (...) Event::Lib.
493 - try to align signal-race timer to full-second boundaries.
494 - work around Tk not liking negative timeouts.
495 - don't complain of different grades of environmental unfriendlyness
496 in IO::Async.
497
4984.85 Sat Jul 18 06:16:14 CEST 2009
499 - nail the signal race problem in perl once and for all
500 (see $AnyEvent::MAX_SIGNAL_LATENCY).
501 - take advantage of Async::Interrupt if it is available.
502 - load Time::HiRes and Guard modules on demand only.
503 - add optional/recommended modules section to AnyEvent
504 documentation.
505 - reduce memory usage considerably (and reduce startup penalty)
506 by not using "strict", "warnings" and "overload" modules.
507 - work around buggy windows/openbsd perls and provide EBADMSG
508 and EPROTO ourselves when missing.
509 - improve perl 5.6 compatibility of the core event loop.
510 - made Net::SSLeay version 1.33 a soft requirement.
511
5124.83 Fri Jul 17 16:56:26 CEST 2009
513 - implement AnyEvent::Socket::getprotobyname.
514 - AnyEvent::CondVar's will now detect recursive blocking
515 waits and will croak, as too many people fall into
516 this trap.
517 - AnyEvent::Handle will now call ->destroy on itself after
518 executing the on_error callback, instead of doing some
519 half-baked internal shutdown, for fatal errors.
520 - clarify on_eof behaviour w.r.t. the read queue and
521 on_read callbacks.
522 - ignore some possible spurious wake-ups in tcp_connect.
523
5244.82 Sat Jul 11 00:34:55 CEST 2009
525 - POE and Event backends didn't accept some callable objects as
526 callbacks.
527 - use Config module instead of POSIX module to detect signal names
528 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
529 module doesn't even have all POSIX signals :/.
530 - use more workarounds around the many refcnt/corruption bugs in
531 Event::Lib.
532 - work around a race condition in perl's select, causing t/03_child.t
533 to rarely fail.
534
5354.81 Thu Jul 9 10:30:30 CEST 2009
536 - AnyEvent::Handle didn't properly diagnose write errors
537 (it expected -1 from syswrite, how lame... :).
538 - support file descriptors in addition to file handles
539 in AnyEvent->io.
540 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
541 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
542 and PERL_ANYEVENT_CA_PATH.
543 - provide a sensible synopsis section for AnyEvent::TLS.
544 - add a "supported backends" section to the manpage.
545 - added simple io watcher test to testsuite, using a
546 portable_socketpair.
547 - tried to improve the stability of the Event::Lib backend,
548 YMMV.
549
5504.8 Mon Jul 6 23:45:16 CEST 2009
551 - AnyEvent::DNS did not properly follow CNAME records with
552 uppercase targets.
553 - AnyEvent::DNS would errornously return AAAA records
554 with v4 mapped addresses (a faulty record) as ipv4 addresses,
555 causing AnyEvent::Socket to throw an exception.
556 - added new module AnyEvent::TLS for easier SSL/TLS context
557 creation, with many options including hostname verification,
558 secure default configuration, lots of documentation and,
559 predefined diffie-hellman keys for perfect forward security
560 and much more. get it while it's still fresh!
561 - use AnyEvent::TLS in AnyEvent::Handle for context management.
562 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
563 so AnyEvent::Socket users have smaller memory footprint
564 in the common case.
565 - add AnyEvent::Handle->push_shutdown method.
566 - add an additional error message parameter to AnyEvent::Handle's
567 on_error callback (for TLS, $! is still available).
568 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
569 - make AnyEvent::Handle more robust against early conenction
570 failures (during new), and return C<undef> in that case
571 from the constructor.
572 - AnyEvent::Handle will now try to load only JSON::XS first,
573 then fall back to JSON.
574 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
575 and because it was documented that way.
576
5774.451 Fri Jul 3 00:28:58 CEST 2009
578 - do not clear rbuf when shutting down an AnyEvent::Handle
579 object - doing so breaks AnyEvent::HTTP.
580
5814.45 Mon Jun 29 22:59:26 CEST 2009
582 - a write error could cause AnyEvent::Handle to create
583 an I/O watcher with an undefined $fh.
584 - special-case mapped ipv4 addresses in both
585 AnyEvent::Socket::format_address and parse_address,
586 to treat them just like ipv4 addresses.
587 - updated and overhauled the AnyEvent::Intro doc.
588 - implement AnyEvent::Socket::format_ipv4/ipv6.
589 - slightly speed up portable_pipe/socketpair.
590 - expand condvar begin/end documentation.
591
5924.42 Fri Jun 26 08:32:18 CEST 2009
593 - preliminary and neccesarily incomplete support for IO::Async.
594 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
595 case it was set to IGNORE, to ensure we can catch child statuses
596 even when the calling env acts stupidly.
597 - updated benchmarks with IO::Async, which performs very well.
598
5994.412 Wed Jun 24 01:35:57 CEST 2009
600 - support an "untaint" attribute for AnyEvent::DNS
601 and set it on the default resolver.
602 - implement and document AnyEvent::Impl::Perl::loop.
603 - remove all anyevent-env variables from %ENV
604 when running in tainted mode.
605 - mention and extend the IO::Lambda benchmark.
606
6074.411 Sun Jun 7 18:48:02 CEST 2009
608 - do not try to use F_SETFD on windows.
609
6104.41 Thu May 14 06:40:11 CEST 2009
611 - work around issues in older perls (5.8.0?) when
612 a signal handler is deleted from the %SIG hash.
613 - use POSIX::_exit in child test, to avoid running
614 destructors.
615 - speed up CHLD handling by relying on SIGCHLD being
616 synchronously delivered, even when we roll our own
617 implementation.
618 - AnyEvent::DNS: add the "dname" resource record name
619 for cosmetic reasons.
620
6214.4 Sun Apr 26 20:12:33 CEST 2009
622 - implemented idle watchers, where applicable.
623 - AnyEvent->time died when Event backend was in use.
624 - fix a memleak in the Tk backend.
625 - sped up Tk timer handling.
626 - clip negative "after" values to 0 in AnyEvent::Impl::Event
627 to avoid spamming out warning messages.
628 - fix Qt timers without interval.
629 - avoid Qt zero-timeout special case, in old libqt's.
630
6314.352 Mon Apr 20 16:31:11 CEST 2009
632 - fix AnyEvent::Strict error messages for child watchers.
633 - fix/update Makefile.PL "recommends" versions.
634 - add AnyEvent->now_update.
635
6364.351 Sat Apr 11 07:56:14 CEST 2009
637 - actually make the signal pipe work under win32.
638 - localise $! in signal handler to avoid changing $!, although
639 perl itself does not seem to save/restore errno either.
640 - set the cloexec flag on the signal pipe (normally set by perl too).
641
6424.35 Fri Mar 27 11:48:20 CET 2009
643 - event models relying on AnyEvent's signal watcher emulation did
644 invoke the callback asynchronously, contradicting documentation
645 and causing signals to get lost (this includes AnyEvent's own
646 event loop). AnyEvent now uses the standard pipe trick to make
647 callback execution synchronous to the event loop.
648 - AnyEvent::Handle didn't free TLS context data on DESTROY
649 (patch by Pavel Shaydo).
650 - work around the ever-incompatibly-changing API of MakeMaker.
651 - document that changing global variables without restoring them
652 is a bad idea in Perl (noted by Adam Rosenstein).
653 - AnyEvent::Strict now barfs if ->io is passed a file.
654
6554.34 Thu Feb 12 18:32:45 CET 2009
656 - separately buffer TLS read data, as otherwise the read queue could
657 deadlock as receiving data is not expected while draining the
658 read queue (which cna only happen with TLS).
659 - raise EBADMSG error on JSON decoding errors.
660 - fix some minor manpage bugs (reported by Maximilian Gaß).
661 - speed up select bitmask parsing quite a bit in the pure perl
662 backend.
663 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
664
6654.331 Tue Jan 6 21:07:25 CET 2009
666 - socketpair fails on many vista machines because vista has
667 completely broken accept/getpeername and getsockname functions,
668 so we provide our own socketpair emulation that kind of works
669 (AnyEvent::Util::portable_pipe).
670 - new function: AnyEvent::Util::portable_socketpair.
671 - take advantage of the Guard module if it exists.
672
6734.33 Fri Nov 21 02:35:40 CET 2008
674 - AnyEvent::Strict did errornously flag a fileno of 0 as
675 illegal.
676 - reduce memory usage and slightly speed up the pure perl backend
677 by only storing the file descriptor, not the file handle.
678 - add missing autocork method to AnyEvent::Handle (reported
679 by Adam Rosenstein).
680 - AnyEvent::DNS->resolve errornously documented an $rcode result
681 argument, but there isn't (spotted by Henrik Krohns).
682 - the naked truth about Tk - it's basically unmaintained.
683
6844.32 Mon Nov 3 22:46:32 CET 2008
685 - fix AnyEvent::Socket::resolve_sockaddr to properly support
686 unix sockets again and choose a proper default for
687 the socket type.
688 - fix call to resolve_sockaddr in tcp_connect.
689
6904.31 Thu Oct 30 04:41:48 CET 2008
691 - implemented AnyEvent::Handle->destroy method.
692 - hint about unexpected effects in TLS mode.
693 - speed up AnyEvent::Impl::Perl by using more arrays
694 instead of hashes without the slightest loss of readability :)
695 - work around a perl argument refcounting bug.
696 - with some perl patches applied, AnyEvent now seems leak-free.
697
6984.3 Fri Oct 3 09:18:43 CEST 2008
699 - AnyEvent will now install a no-op signal handler for SIGPIPE,
700 unless one has been installed already.
701 - warn about not loaidng AnyEvent::Impl::POE early enough
702 (patch by Adam Rosenstein).
703 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
704 - fix a bug in where in SSL connect mode, the
705 client would first wait for some data by the server and
706 otherwise hang. (reported and analysed in an absolutely
707 exemplary manner by Adam Rosenstein).
708 - fix a bug in where SSL EOF would not be treated
709 as stream EOF, putting the connection into a hung state.
710 (reported and analysed in an absolutely exemplary manner by Adam
711 Rosenstein).
712 - fix a potential segfault when the TLS context
713 would go missing in a read callback (e.g. due to stoptls) -
714 Net::SSLeay of course makes no type checking whatsoever.
715 - AnyEvent::Handle will not stop the read watcher in TLS mode.
716 - AnyEvent::Handle->stoptls will send a TLS close notify
717 instead of simply destroying the stream now.
718 - fix a bug where large blocks of data written in TLS mode
719 would not be sent unless triggered by receives.
720 - on_drain will now take the tls write buffer into account.
721 - SSL operations have been streamlined a bit and should be faster
722 now, more to come, though.
723 - added a FAQ section, will document nontrivial issues in there.
724
7254.234 Mon Sep 29 04:08:13 CEST 2008
726 - fix child watcher documentation: only child exits will be reported,
727 no trace events.
728 - mention SIGPIPE in AnyEvent::Handle.
729 - perl backend will now use POSIX::times as monotonic clock
730 when available and the monotonic clock option is not.
731 - run even without Time::HiRes in some cases now, but do not
732 provide sub-second accuracy (all *supported* perl versions have
733 Time::HiRes).
734
7354.233 Fri Aug 22 01:48:05 CEST 2008
736 - fix a bug in the testsuite, causing a hang.
737 - clarified lots of AnyEvent::Handle constructor arguments.
738
7394.232 Thu Aug 21 20:44:25 CEST 2008
740 - fix a bug in AnyEvent::Handle that could cause two fatal error
741 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
742 (testcase by Leon Brocard).
743 - clarified on_eof/on_error documentation a bit.
744
7454.231 Tue Jul 29 13:12:15 CEST 2008
746 - remove some debugging code left in AnyEvent::Util::fork_call
747 (and no, it's impossible to implement with the broken windows
748 perls without resource leaks or worse).
749
7504.23 Tue Jul 29 12:19:59 CEST 2008
751 - document the first parameter passed to condvar callbacks to be
752 the callback.
753 - add AnyEvent::Socket::{ntoa,aton} aliases.
754 - optimize the AE::Handle->push_read (line) for the default
755 eol marker.
756 - optimize push_read (packstring|storable) for small packets.
757 - invoke on_error callback when no on_eof callback is set.
758 - fix a bug in push_read (storable) of unknown impact.
759
7604.22 Sun Jul 20 16:34:13 CEST 2008
761 - new function AnyEvent::Socket::parse_hostport.
762 - as the bulkheads at microsoft can't even get getprotobyname reliably
763 working on their shitty fucking broken os we need to hardcode
764 some common protocol numbers in AnyEvent::Socket. How can
765 people even bother with such a pile of shit as windows.
766
7674.21 Thu Jul 17 14:40:05 CEST 2008
768 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
769 which is incompatible to previous invocation syntax.
770 - work around more windows perl bugs in fork_call. I cannot imagine
771 why anybody would bother implementing fork in such an obviously
772 unusable way.
773 - avoid calling $do_retry if the request has already finished in
774 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
775 K. Sharifulin).
776
7774.2 Sat Jul 12 22:42:11 CEST 2008
778 - support an "interval" argument for repeating timers.
779 - fix a bug in the pure perl event loop that caused timers to
780 be delayed under some circumstances.
781 - implement optional argument checking via PERL_ANYEVENT_STRICT.
782 - not importing anything from Event causes it to malfunction
783 and not use Time::HiRes, work around this bug.
784 - more bug workarounds against the endlessly broken Tk module,
785 seems to work now on 32 bit machines, but less so on 64 bit.
786 - minor optimisations applied to most backends.
787
7884.161 Fri Jul 4 14:24:48 CEST 2008
789 - fixed the t/handle/02_write.t, it read after write, with the
790 assumption that all data could be written before read
791 (reported by Jim Razmus).
792
7934.160 Thu Jul 3 04:02:21 CEST 2008
794 - re-registering signal watchers after unregistering one did
795 not work if a backend falls back on AnyEvent's default
796 implementation (also affected child watchers).
797 - new AnyEvent::Handle options: autocork and no_delay.
798
7994.152 Sun Jun 22 14:15:44 CEST 2008
800 - allow for 32-bit perls that implement shifts differently
801 on different architectures in parse_ipv4 (reported and
802 analysed by Keiichi DAIBA).
803
8044.151 Fri Jun 6 17:34:24 CEST 2008
805 - make sure specifying _only_ on_read and never pushing reads
806 works.
807
8084.15 Fri Jun 6 13:00:46 CEST 2008
809 - the pure perl backend would keep some watchers alive when more than
810 one watcher was registered for the same fd.
811 - new "packstring" and "storable" read and write types
812 for AnyEvent::Handle.
813 - allow on_eof handler to be called after on_error with EPIPE returns.
814 - do not immediately call on_read callback in handle constructor.
815
8164.14 Thu Jun 5 20:29:31 CEST 2008
817 - Fixed a bug in DNS SRV priority sorting.
818 - AnyEvent::Util::guard now reports runtime errors while
819 executing the guard block as warnings.
820 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
821
8224.13 Thu Jun 5 00:47:59 CEST 2008
3 - AnyEvent::DNS only followed cname chains with length 2, 823 - AnyEvent::DNS only followed cname chains with length 2,
4 contrary to documentation. bumped it up to 10, thanks to 824 contrary to documentation. bumped it up to 10, thanks to
5 microsoft, the current broken-dns-config-king. 825 microsoft, the current broken-dns-config-king.
6 - AnyEvent::DNS didn't check for socket return status 826 - AnyEvent::DNS didn't check for socket return status
7 but instead relied on perl not creating filehandles in that 827 but instead relied on perl not creating filehandles in that
… …
124 - support service names in tcp_connect and tcp_server. 944 - support service names in tcp_connect and tcp_server.
125 - work around netbsd/openbsd bug causing unix sockets not to 945 - work around netbsd/openbsd bug causing unix sockets not to
126 become writable until empty. 946 become writable until empty.
127 - only make a dummy read for error detection when getpeername 947 - only make a dummy read for error detection when getpeername
128 returns ENOTCONN. 948 returns ENOTCONN.
129 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 949 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
130 override. 950 override.
131 - new AnyEvent::Guard::cancel. 951 - new AnyEvent::Guard::cancel.
132 952
1333.5 Sat May 17 23:17:14 CEST 2008 9533.5 Sat May 17 23:17:14 CEST 2008
134 - gracefully deal with systems not defining CLOCK_MONOTONIC 954 - gracefully deal with systems not defining CLOCK_MONOTONIC
… …
148 968
1493.41 Sun May 11 19:53:13 CEST 2008 9693.41 Sun May 11 19:53:13 CEST 2008
150 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 970 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
151 971
1523.4 Sun May 11 00:29:25 CEST 2008 9723.4 Sun May 11 00:29:25 CEST 2008
153 - complete re-implementation of condvars with a hopefully much mroe useful 973 - complete re-implementation of condvars with a hopefully much more useful
154 API (backward compatible functions exist but are undocumented). 974 API (backward compatible functions exist but are undocumented).
155 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 975 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
156 (still subject to change). 976 (still subject to change).
157 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 977 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
158 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 978 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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