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

Comparing AnyEvent/Changes (file contents):
Revision 1.191 by root, Wed Oct 1 07:40:39 2008 UTC vs.
Revision 1.569 by root, Fri Apr 5 20:31:41 2013 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
34.3 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
10TODO: invalid. and localhost. specialcasing inside AEDNS and not AESocket (rfc6761)
11TODO: maybe always try to downgrade in push_write.
12.plan archives for this user: finger icculus?listarchives=1@icculus.org
13finger mikachu@
14 - uts46data.pl couldn't be found due to wrong naming of the file
15 (reported by Fulko Hew).
16 - AnyEvent::Debug shell can now run coro shell commands, if
17 available.
18 - t/63* tests were wrongly in MANIFEST.
19 - mention IO::AIO as optional dependency.
20 - kernel.org's finger server went MIA, switch to freebsd.org.
21
227.04 Thu Nov 15 02:23:58 CET 2012
23 - debugging code left in.
24
257.03 Thu Nov 15 02:16:11 CET 2012
26 - AnyEvent::Socket::inet_aton did not work when dns resolution was
27 used to find the addresses (analyzed and patch by Maxime Soulé).
28 - fix a memory leak in the /etc/hosts lookup code when hosts don't
29 resolve and are not in hosts (analyzed by sten).
30
317.02 Tue Aug 14 04:27:58 CEST 2012
32 - AnyEvent::Util::run_cmd could block indefinitely (analyzed and test
33 program by Yorhel).
34 - verified that AnyEvent::Socket follows RFC5952.
35 - try to parse "ADDR#PORT" in addition to "ADDR PORT".
36
377.01 Sun May 13 01:03:17 CEST 2012
38 - fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but
39 not available, instead of throwing an exception.
40 - use File::Spec to get the tmpdir in t/*, to avoid needless
41 failures on (most, not mine :) windows boxes.
42 - new handle read types: tls_detect and tls_autostart.
43
447.0 Fri Apr 13 06:33:30 CEST 2012
45 - child watchers are broken in POE 1.352 (also many earlier
46 versions) and there seems to be no way to work around it, as POE
47 itself is inherently racy. Document this regression and add a delay
48 in t/68_poe_03_child.t for the time being.
49 - new module AnyEvent::IO, that is a frontend to either a pure-perl
50 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
51 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
52 - load /etc/hosts only when DNS has no answer.
53 - stat /etc/hosts on every access and reload it if it changed.
54 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
55 - fix a buggy croak in the dh parameter reading in AnyEvent::TLS.
56 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
57 a side effect, they now use true appending as opposed to libc
58 appending, and the file might not have been opened when the function
59 returns.
60 - the default logging level was not properly documented in a variety of
61 places, this has hopefully been rectified.
62 - updated uts46data.pl for unicode 6.1.0.
63 - made log messages generated by AnyEvent submodules not
64 include the package name anymore, as it will be logged
65 by default already.
66 - upgrade to the trick used by common-sense 3.5 to work
67 around extra warning torture/breakage under perl 5.15.x.
68 - log messages by anyevent are now ucfirst, are usually full
69 sentences and do no longer include the package name.
70 - the storable read type would sometimes throw an exception instead
71 of causing EBADMSG (analyzed by Maxime Soulé).
72 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
73 keysize (reported by Paul Howarth).
74 - documented Rocco fabricating statements about AnyEvent and me.
75
766.14 Tue Jan 31 20:00:24 CET 2012
77 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
78 probably an editing mistake (reported by Darin McBride).
79
806.13 Thu Jan 12 07:27:01 CET 2012
81 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
82 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
83 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
84 regex-matching inside the callback would change the parameter, despite
85 $1 being dynamically scoped per-block. this perl bug is now being
86 worked around (testcase by Cindy Wang).
87
886.12 Mon Dec 12 13:21:10 CET 2011
89 - $! was clobbered by subsequent calls in tcp_connect, due to
90 the postpone not saving/restoring it (analyzed by Richard Garnish).
91
926.11 Tue Nov 22 10:36:05 CET 2011
93 - Tk cannot create windows when tainted unless you set a title, so
94 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
95 - escape any nonprintable/nonascii characters when stringifying
96 backtraces.
97 - log the reason loading an interface module fails to load at level debug.
98
996.1 Tue Oct 4 19:44:30 CEST 2011
100 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
101 and above), and some messages inside AnyEvent have been elevated
102 to higher log levels to print by default.
103 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
104 - work around yet another signal race bug, newly introduced by POE. My
105 how that module sucks.
106 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
107 support ipv6-only operation even with defaults (also use ::1 instead
108 of 127.0.0.1).
109 - give better guidelines on how to chose a logging level.
110 - AnyEvent::Log can now cap (limit) log message priority by context.
111 - AE::log will now save and restore $! and $@, so callers don't have to.
112 - AE::Strict now checks that registered fds are still valid at regular
113 intervals.
114 - skip frequently called 'require' calls if the module appears loaded.
115 - increase signal latency test timer considerably for poor little osx,
116 which can't be bothered with basic posix functionality such as signal
117 delivery.
118 - mention some debugging aids in the tutorial.
119 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
120 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
121 of the context as title.
122 - AnyEvent::Strict would not restore read-write access to $_ when a callback
123 died - this is a user bug, but since AE::Strict exists to help dbeugging...
124 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
125 to work with it - the author is well aware of the issues but refuses
126 to fix them. This extreme step has been taken because the module actively
127 attacks the goals of AnyEvent and makes it harder for module authors to use
128 AnyEvent.
129
1306.02 Fri Aug 26 20:08:31 CEST 2011
131 - log_to_syslog now accepts facility strings.
132 - log_to_syslog would not allow facilities - "$facility|$level"
133 *literally* means that string. sheesh.
134 - the CPAN parser got confused and thought we provide Sys::Syslog, try
135 to work around.
136 - renamed AnyEvent::Impl::FLTK2 to FLTK.
137
1386.01 Fri Aug 26 07:04:11 CEST 2011
139 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
140 the TTL as fourth element - this affects ->resolve, ->request
141 and ->dns_unpack indirectly. Convenience functions such as
142 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
143 AnyEvent::Socket.
144 - INCOMPATIBLE CHANGE: backend authors now should not implement
145 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
146 methods.
147 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
148 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
149 - DEPRECIATION: please change
150 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
151 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
152 in your programs, for future compatibility.
153 - *blush* idle watcher emulation was borked.
154 - AnyEvent::Strict now checks the AE::xxx functions as well.
155 - querying too long domain names would barf deep inside AnyEvent::DNS - since
156 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
157 for too long domain names explicitly and passes an undef to the callback
158 instead (reported by James Bromberger).
159 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
160 - new functions: AnyEvent::postpone, AnyEvent::log.
161 - new module: AnyEvent::Log, for simple logging needs.
162 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
163 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
164 a debug shell.
165 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
166 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
167 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
168 /etc/hosts can be found.
169 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
170 configuration.
171 - env variables can now be specified using AE_xyz instead of
172 PERL_ANYEVENT_xyz, subject to some rules.
173 - lots of new AnyEvent::Debug functionality.
174 - AnyEvent::Strict now checks for modifications of $_ by the callback
175 (by setting it to readonly while executing the callback, if possible).
176 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
177 Invest extra overhead to work around this peculiar behaviour.
178 - split the pure perl backend into a loop module and an impl
179 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
180 - some read types (regex, netstring, json, storable) could
181 errornously keep an AE::Handle object alive on their own
182 (reported by Mohammad Toossi).
183 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
184 supports /etc/hosts overrides.
185 - add more workarounds around child watcher bugs in IO::Async
186 and (newly introduced) in POE.
187 - use glib child watchers and try to work around its limitations,
188 if possible (based on a patch by Kevin Ryde).
189 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
190 - many read types in AnyEvent::Handle would malfunction during
191 transitions to TLS because of a "delete $self->{rbuf}".
192 (reported by Gerald Galster).
193 - improve condvar blocking wait performance (probably only noticable
194 with faster backends), also make it easier for event loops that
195 do not support blocking waits.
196 - PERL_ANYEVENT_MODEL now also allows full module names, although
197 the usefulness of this feature is not yet clear.
198 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
199 - add t/80_ssltext.t.
200 - no longer use AUTOLOAD for the initial loading (for no very good reason).
201 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
202 - do not run event loop detection multiple times when programs call their cached
203 copies of AnyEvent::detect.
204 - parse_hostport now accepts absolute paths als unix domain sockets.
205 - suppress the idiotic warning inside POE by patching the POE::Kernel
206 run flags instead of asking the user to jump through contortions.
207 - add Task::Weaken as dependency (as recommends only at this time).
208 - replace PF_UNSPEC by 0, for good old minix.
209
2106.0 (not officially released, but was on CPAN for about 15 minutes :/)
211
2125.34 Thu May 12 09:59:41 CEST 2011
213 - AE::TLS didn't support both cert_file and key at the same time,
214 due to some copy&paste bug (reported and analyzed by vti).
215 - improve AE::Handle timeout documentation slightly.
216
2175.33 Mon Feb 7 21:51:41 CET 2011
218 - never appeared on CPAN, for unknown reasons.
219 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
220 object when it's already gone (analyzed by Ryan Bullock).
221 - update idna algorithm(s) to UTS#46 version 6.0.0.
222 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
223 uts46 says it's an error, and testsuite makes more sense.
224 - remove leading dots in idn_to_ascii - testsuite says to remove,
225 uts46 says it's an error, and testsuite makes more sense.
226 - fixed a bug in idn ignored character processing.
227 - passes the uts46 6.0.0 testsuite except for three cases which
228 apparently are bugs in the testsuite itself.
229 - passes the uts46 6.0.1 testsuite except for eleven cases which
230 apparently are bugs in the testsuite itself.
231
2325.31 Sun Jan 23 11:49:19 CET 2011
233 - work around a POE bug where POE would first poll for new events and
234 invoke handlers BEFORE handling the message queue.
235 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
236 - stop_read now does just that in TLS mode. see its
237 documentation for caveats.
238 - update to common::sense 3.4 warning set (some warnings in perl impose
239 an arbitrary 32kb limit on regex matches, which affects
240 AnyEvent::Handle).
241 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
242 - new experimental AE::Handle::resettls method.
243
2445.3 Fri Dec 31 05:48:13 CET 2010
245 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
246 issues with IO::Async have been solved in that version, including
247 the availabilty of a default loop!
248 - implement new wbuf_max parameter for AnyEvent::Handle.
249 - added Cocoa::EventLoop backend by Daisuke Murase.
250 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
251 - clarify handle on_prepare documentation.
252 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
253 to avoid leaking them to any child processes.
254 - try to force files into the arch-dependend lib directory, for the
255 half-"turly obsessive compulsive person"(s) that depend on it.
256
2575.29 Sun Dec 5 10:49:21 CET 2010
258 - convert EV backend to EV 4.00 API (so better upgrade EV too).
259 - AE::Handle: implement dynamic read_size adjustment, add
260 max_read_size parameter, reduce default read_size to 2048.
261 - add some FAQ entries for servers and/or on_eof vs. on_error.
262 - work around OS bugs (cygwin again) when lingering in AE::Handle.
263 - reduce memory usage slightly when connect was used in AE::Handle.
264 - wrok around more segfaults in Tk.
265 - document tls_ctx => undef as valid.
266 - detect WNOHANG value at compiletime.
267 - include some probably linux-only support for building AnyEvent
268 as part of the perl core.
269 - improve compatibility of t/08_idna.t to perl 5.8.
270 - make AnyEvent::Strict truly optional (in case somebody builds a
271 minimal perl...).
272
2735.28 Wed Oct 13 04:14:23 CEST 2010
274 - due to a glitch, AnyEvent's internal getprotobyname wasn't
275 actually used in tcp_server.
276 - implement AnyEvent::CondVar->new method, which was documented but
277 missing.
278 - support autoloading of read types in unshift_read.
279 - AnyEvent::Strict: do the test for invalid fh arguments differently,
280 resulting in more informative error messages.
281 - new AnyEvent::FAQ manpage.
282 - clarify that you can actually call push_read etc. in on_read.
283 - put keys and certs in /tmp, not into the current directory,
284 and document this in the SECURITY CONSIDERATIONS sections
285 of AnyEvent::TLS.
286 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
287 requested by Ron Isaacson.
288 - document $AnyEvent::DNS::RESOLVER variable.
289 - incorporated major typo patches by Abhijit Menon-Sen.
290 - AnyEvent::Handle now croaks when negative timeouts are passed.
291 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
292 undocumented, but who knows when they might come in handy.
293
2945.271 Tue Jun 8 12:05:46 CEST 2010
295 - backport to perl 5.8.x.
296
2975.27 Sun Jun 6 12:12:05 CEST 2010
298 - postpone differently in AnyEvent::Socket now, as
299 when not, canceling the connection attempt might fail
300 (found by Felix Antonius Wilhelm Ostmann).
301 - explicitly check for non-stream sockets in AE::Handle, too many
302 clueless people fell into the trap of this somehow working.
303 - simplified and reworked the "OTHER MODULES" section.
304 - better/more condvar examples.
305
3065.261 Wed Apr 28 16:13:36 CEST 2010
307 - AF_INET6 was not properly used from Socket6 during configuration
308 time (found by Andy Grundman).
309
3105.26 Mon Apr 12 04:49:35 CEST 2010
311 - don't generate spurious readyness notifications when select
312 returns EINTR in the pure perl backend (analysed and testcase
313 by Bas Denissen).
314 - give same examples in SYNOPSIS sections of AnyEvent and AE.
315 - provider faster implementations for AE::time/now/now_update for
316 EV and Perl backends.
317 - provide faster AE::cv and AE::time implementations for all
318 backends.
319 - no longer support lower/mixed-case signal names (this was
320 never documented nor universally supported).
321 - some more memory µ-optimisations, and somewhat less messy
322 code to cope with both AnyEvent and AE APIs.
323
3245.251 Sat Mar 13 00:58:21 CET 2010
325 - make test failed if it was run before make install on perl <=
326 5.8.8, and older versions of perl make our live very difficult,
327 so write constants.pl during Makefile.PL time.
328
3295.25 Sat Mar 13 00:23:14 CET 2010
330 - fix a race condition in AnyEvent::Handle that would cause
331 a "bio_read: unsupported method" error (in _another_
332 ssl connection) after user code threw an exception.
333 - added AnyEvent::Handle->destroyed method.
334 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
335 - the AnyEvent::Util::fh_nonblocking function worked only
336 by lucky accident on win32.
337 - smaller and faster AnyEvent::Util::fh_nonblocking.
338 - when the (required!) Time::HiRes module is missing, AnyEvent
339 did not fall back to built-in time properly.
340 - do not load Fcntl at runtime, saving memory and loading time.
341 - precompile a number of constants and use them instead of runtime
342 detection and eval.
343 - free detection code after detection and similar memory optimisations.
344 - Perl backend timer interval best effort drift has been improved
345 (same algorithm as EV).
346 - update unicode idna mapping table.
347
3485.24 Tue Jan 5 11:39:43 CET 2010
349 - cygwin never reports errors from failed connects. 1.5 just gives
350 you continous readyness and EAGAIN, 1.7 is even more broken and
351 just hangs. work around both issues in a cygwin-specific hack.
352 - improve idle watcher documentation slightly.
353
3545.23 Sun Dec 20 23:48:00 CET 2009
355 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
356 - implement punycode_encode/decode, idn_nameprep,
357 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
358 - provide $AE::VERSION.
359 - removed traces of "no strict 'refs'".
360
3615.22 Sat Dec 5 03:51:13 CET 2009
362 - downgrade-or-fail in AnyEvent::Handle::push_write, to
363 diagnose encoding failures earlier and more succinctly.
364 (this works around bugs in perl, throwing away encoding info
365 when passing scalar data to extensions).
366 - add more examples to AnyEvent::Socket manpage.
367 - upgrade internal warning set to the same as common::sense 2.03.
368 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
369 (John Beppu).
370 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
371 - add google public dns servers to fallback server set and make sure
372 we load-balance properly between the three sets. also add all
373 fallback dns servers, not just a random one, to each dns config.
374
3755.21 Thu Nov 19 02:48:47 CET 2009
376 - fix a problem where socket constants were called with parameters
377 (spotted by David Friedland).
378 - fork_call never use'd POSIX (reported by Daisuke Maki).
379 - improve perl 5.6 compatibility further (but it still won't work
380 unless you rip out everything but the core).
381 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
382 uses 10MB of RAM, but doesn't open a console window. *sigh*.
383
3845.202 Wed Oct 14 22:35:44 CEST 2009
385 - AnyEvent::DNS would unexpectedly clobber $_ under windows
386 (analysed by Matthias Waldorf).
387 - AnyEvent::Handle::run_cmd can now pass the PID of the
388 newly-created process, which is much less useful than it might
389 sound (based on patch by Yann Kerherve).
390
3915.201 Tue Sep 29 12:09:25 CEST 2009
392 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
393 (reported by Torsten Foertsch).
394 - common::sense 2.0 could cause tcp_server to throw an exception
395 (analysed by elmex).
396
3975.2 Mon Sep 14 07:04:49 CEST 2009
398 - INCOMPATIBLE CHANGE: do no longer support register_read_type
399 and register_write_type in AnyEvent::Handle, instead support
400 package names (the facility was mostly abused).
401 - implement "packagename-as-read/write type" support in
402 AnyEvent::Handle.
403 - AnyEvent::Handle: new options "keepalive" and "oobinline".
404 - oobinline set by default to avoid security issues.
405 - the pure-perl event loop backend wrongly detected times() fallback
406 support (spotted by Pavel Boldin).
407
4085.12 Tue Sep 1 20:26:50 CEST 2009
409 - be more lenient when parsing resolv.conf files, as some people
410 use hashmarks as comment indicator inside directives (reported
411 by Michael S. Fischer).
412 - use same set of warnings as common::sense 2.0.
413 - fix a potential 32 bit overflow issue due to perl having problems
414 with large hex constants in the Perl backend.
415 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
416 by default.
417 - fix some prototypes.
418 - (optionally) require Async::Interrupt 1.02.
419 - replace opendns fallbacks by verizon/level3 public dns servers.
420
4215.112 Fri Aug 21 13:59:04 CEST 2009
422 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
423 using starttls with an illegal/uncreatable tls context.
424 - port to broken win32 perls.
425
4265.111 Sun Aug 16 18:44:36 CEST 2009
427 - perl errornously requires sockaddr_un structures returned by the
428 kernel to have a specific length. work around this bug by padding
429 sockaddr_un structures (found to fail on OS X, as perl only uses
430 the correct code for GNU/Linux, but likely to fail on other
431 OSes as well).
432
4335.11 Wed Aug 12 17:49:37 CEST 2009
434 - fixed a bug in run_cmd with input from filename.
435 - create a json coder object when none is given in write direction
436 to reduce dependency on JSON.
437
4385.1 Tue Aug 11 03:17:32 CEST 2009
439 - tcp_connect now reports the correct error in $!, not an unrelated
440 one (was broken since 4.91).
441 - AE::Handle did not properly clear rtimeout/wtimeout watchers
442 when retrying a tcp_connect.
443 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
444 - fixed wrong documentation in AE::Socket.
445
4465.01 Mon Aug 10 03:16:32 CEST 2009
447 - last release broke EV child watchers :/.
448
4495.0 Sun Aug 9 17:32:09 CEST 2009
450 - officially introduce and document the AE API.
451 - lowering the timeout at runtime did not have immediate effect
452 in AnyEvent::Handle.
453 - AE::Handle now has separate and independent read and write timeout
454 settings.
455 - probe for Guard module when AnyEvent::Util is loaded, not at
456 runtime (Event doesn't like eval "use" from callbacks, and it avoids
457 exporting the probe function).
458 - Impl::IOAsync's io watchers did not use the filehandle emulation
459 code to work around it's watcher limits.
460 - work around IO::Async's problems with overloaded objects.
461 - try to disable wearnings inside IO::Async, it's just too buggy.
462 - do no longer use Event by default, only EV or Perl, Event is too buggy.
463 - hackishly provide fast AE:: interface when EV or Perl is the backend.
464 - inherit VERSION from AnyEvent::VERSION in most modules.
465
4664.91 Thu Aug 6 15:42:45 CEST 2009
467 - AE::Handle::starttls could get out of sync when the read buffer
468 already contains some TLS handshake.
469 - AE::Handle did not properly free the TLS session, causing bigger
470 memory leaks in Net::SSLeay than expected (reported by toaster).
471 - AE::Socket::tcp_connect will now artificially delay invoking
472 the callback to avoid returning after invoking the callback.
473 - convert many internal watcher uses to AE API.
474
4754.9 Sat Aug 1 11:07:01 CEST 2009
476 - Glib silently fails when registering a timeout with
477 a negative value, so avoid that.
478 - call condvar callback immediately when it is set
479 after the condition is already signalled.
480 - check rbuf_max condition only after trying to consume
481 data.
482 - normalise signal numbers to names when using the ae's
483 signal handling, but do not document this (yet).
484 - pure perl signal emulation did not properly set nonblocking
485 mode on the signal pipe on win32 (this is unlikely to have caused
486 any issues).
487 - new module: AnyEvent::Debug.
488 - AnyEvent::Strict now checks that a signal was specified
489 by name (not name or number).
490 - reduce memory footprint in typical cases by ~50kb by
491 only compiling signal code when necessary.
492 - add AnyEvent::Handle->rbuf_max.
493 - grab the AE namespace for future new API, implement stubs for
494 said future/alternative API.
495 - new function: AnyEvent::Socket::format_hostport.
496
4974.881 Tue Jul 28 12:51:53 CEST 2009
498 - work around a bug in local in pre-5.10 perls, causing
499 AnyEvent::Handle to recurse when it should not
500 (analysed by elmex).
501
5024.88 Tue Jul 28 04:04:37 CEST 2009
503 - re-bless the handle into a dummy package after calling
504 AnyEvent::Handle::destroy, so the user does not need to check
505 for errors after every push_write etc.
506 - do not attempt to run t/02_signals.t on obviously broken
507 platforms.
508
5094.87 Sun Jul 26 02:06:16 CEST 2009
510 - do not attempt to linger when there is no longer a valid fh.
511 - enforce tls mode to be either accept or connect, do not simply
512 segfault in Net::SSLeay.
513 - AnyEvent::Handle can now call tcp_connect itself (new parameters
514 connect, on_prepare, on_connect and on_connect_error). Updated
515 tutorial accordingly.
516 - add AnyEvent::Impl::Irssi backend.
517
5184.86 Mon Jul 20 23:52:29 CEST 2009
519 - since the verbose warning is not enough, explicitly document
520 that versions before 1.33 of Net::SSLeay are not secure.
521 - work around signal handling races in Event and (...) Event::Lib.
522 - try to align signal-race timer to full-second boundaries.
523 - work around Tk not liking negative timeouts.
524 - don't complain of different grades of environmental unfriendlyness
525 in IO::Async.
526
5274.85 Sat Jul 18 06:16:14 CEST 2009
528 - nail the signal race problem in perl once and for all
529 (see $AnyEvent::MAX_SIGNAL_LATENCY).
530 - take advantage of Async::Interrupt if it is available.
531 - load Time::HiRes and Guard modules on demand only.
532 - add optional/recommended modules section to AnyEvent
533 documentation.
534 - reduce memory usage considerably (and reduce startup penalty)
535 by not using "strict", "warnings" and "overload" modules.
536 - work around buggy windows/openbsd perls and provide EBADMSG
537 and EPROTO ourselves when missing.
538 - improve perl 5.6 compatibility of the core event loop.
539 - made Net::SSLeay version 1.33 a soft requirement.
540
5414.83 Fri Jul 17 16:56:26 CEST 2009
542 - implement AnyEvent::Socket::getprotobyname.
543 - AnyEvent::CondVar's will now detect recursive blocking
544 waits and will croak, as too many people fall into
545 this trap.
546 - AnyEvent::Handle will now call ->destroy on itself after
547 executing the on_error callback, instead of doing some
548 half-baked internal shutdown, for fatal errors.
549 - clarify on_eof behaviour w.r.t. the read queue and
550 on_read callbacks.
551 - ignore some possible spurious wake-ups in tcp_connect.
552
5534.82 Sat Jul 11 00:34:55 CEST 2009
554 - POE and Event backends didn't accept some callable objects as
555 callbacks.
556 - use Config module instead of POSIX module to detect signal names
557 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
558 module doesn't even have all POSIX signals :/.
559 - use more workarounds around the many refcnt/corruption bugs in
560 Event::Lib.
561 - work around a race condition in perl's select, causing t/03_child.t
562 to rarely fail.
563
5644.81 Thu Jul 9 10:30:30 CEST 2009
565 - AnyEvent::Handle didn't properly diagnose write errors
566 (it expected -1 from syswrite, how lame... :).
567 - support file descriptors in addition to file handles
568 in AnyEvent->io.
569 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
570 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
571 and PERL_ANYEVENT_CA_PATH.
572 - provide a sensible synopsis section for AnyEvent::TLS.
573 - add a "supported backends" section to the manpage.
574 - added simple io watcher test to testsuite, using a
575 portable_socketpair.
576 - tried to improve the stability of the Event::Lib backend,
577 YMMV.
578
5794.8 Mon Jul 6 23:45:16 CEST 2009
580 - AnyEvent::DNS did not properly follow CNAME records with
581 uppercase targets.
582 - AnyEvent::DNS would errornously return AAAA records
583 with v4 mapped addresses (a faulty record) as ipv4 addresses,
584 causing AnyEvent::Socket to throw an exception.
585 - added new module AnyEvent::TLS for easier SSL/TLS context
586 creation, with many options including hostname verification,
587 secure default configuration, lots of documentation and,
588 predefined diffie-hellman keys for perfect forward security
589 and much more. get it while it's still fresh!
590 - use AnyEvent::TLS in AnyEvent::Handle for context management.
591 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
592 so AnyEvent::Socket users have smaller memory footprint
593 in the common case.
594 - add AnyEvent::Handle->push_shutdown method.
595 - add an additional error message parameter to AnyEvent::Handle's
596 on_error callback (for TLS, $! is still available).
597 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
598 - make AnyEvent::Handle more robust against early conenction
599 failures (during new), and return C<undef> in that case
600 from the constructor.
601 - AnyEvent::Handle will now try to load only JSON::XS first,
602 then fall back to JSON.
603 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
604 and because it was documented that way.
605
6064.451 Fri Jul 3 00:28:58 CEST 2009
607 - do not clear rbuf when shutting down an AnyEvent::Handle
608 object - doing so breaks AnyEvent::HTTP.
609
6104.45 Mon Jun 29 22:59:26 CEST 2009
611 - a write error could cause AnyEvent::Handle to create
612 an I/O watcher with an undefined $fh.
613 - special-case mapped ipv4 addresses in both
614 AnyEvent::Socket::format_address and parse_address,
615 to treat them just like ipv4 addresses.
616 - updated and overhauled the AnyEvent::Intro doc.
617 - implement AnyEvent::Socket::format_ipv4/ipv6.
618 - slightly speed up portable_pipe/socketpair.
619 - expand condvar begin/end documentation.
620
6214.42 Fri Jun 26 08:32:18 CEST 2009
622 - preliminary and neccesarily incomplete support for IO::Async.
623 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
624 case it was set to IGNORE, to ensure we can catch child statuses
625 even when the calling env acts stupidly.
626 - updated benchmarks with IO::Async, which performs very well.
627
6284.412 Wed Jun 24 01:35:57 CEST 2009
629 - support an "untaint" attribute for AnyEvent::DNS
630 and set it on the default resolver.
631 - implement and document AnyEvent::Impl::Perl::loop.
632 - remove all anyevent-env variables from %ENV
633 when running in tainted mode.
634 - mention and extend the IO::Lambda benchmark.
635
6364.411 Sun Jun 7 18:48:02 CEST 2009
637 - do not try to use F_SETFD on windows.
638
6394.41 Thu May 14 06:40:11 CEST 2009
640 - work around issues in older perls (5.8.0?) when
641 a signal handler is deleted from the %SIG hash.
642 - use POSIX::_exit in child test, to avoid running
643 destructors.
644 - speed up CHLD handling by relying on SIGCHLD being
645 synchronously delivered, even when we roll our own
646 implementation.
647 - AnyEvent::DNS: add the "dname" resource record name
648 for cosmetic reasons.
649
6504.4 Sun Apr 26 20:12:33 CEST 2009
651 - implemented idle watchers, where applicable.
652 - AnyEvent->time died when Event backend was in use.
653 - fix a memleak in the Tk backend.
654 - sped up Tk timer handling.
655 - clip negative "after" values to 0 in AnyEvent::Impl::Event
656 to avoid spamming out warning messages.
657 - fix Qt timers without interval.
658 - avoid Qt zero-timeout special case, in old libqt's.
659
6604.352 Mon Apr 20 16:31:11 CEST 2009
661 - fix AnyEvent::Strict error messages for child watchers.
662 - fix/update Makefile.PL "recommends" versions.
663 - add AnyEvent->now_update.
664
6654.351 Sat Apr 11 07:56:14 CEST 2009
666 - actually make the signal pipe work under win32.
667 - localise $! in signal handler to avoid changing $!, although
668 perl itself does not seem to save/restore errno either.
669 - set the cloexec flag on the signal pipe (normally set by perl too).
670
6714.35 Fri Mar 27 11:48:20 CET 2009
672 - event models relying on AnyEvent's signal watcher emulation did
673 invoke the callback asynchronously, contradicting documentation
674 and causing signals to get lost (this includes AnyEvent's own
675 event loop). AnyEvent now uses the standard pipe trick to make
676 callback execution synchronous to the event loop.
677 - AnyEvent::Handle didn't free TLS context data on DESTROY
678 (patch by Pavel Shaydo).
679 - work around the ever-incompatibly-changing API of MakeMaker.
680 - document that changing global variables without restoring them
681 is a bad idea in Perl (noted by Adam Rosenstein).
682 - AnyEvent::Strict now barfs if ->io is passed a file.
683
6844.34 Thu Feb 12 18:32:45 CET 2009
685 - separately buffer TLS read data, as otherwise the read queue could
686 deadlock as receiving data is not expected while draining the
687 read queue (which cna only happen with TLS).
688 - raise EBADMSG error on JSON decoding errors.
689 - fix some minor manpage bugs (reported by Maximilian Gaß).
690 - speed up select bitmask parsing quite a bit in the pure perl
691 backend.
692 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
693
6944.331 Tue Jan 6 21:07:25 CET 2009
695 - socketpair fails on many vista machines because vista has
696 completely broken accept/getpeername and getsockname functions,
697 so we provide our own socketpair emulation that kind of works
698 (AnyEvent::Util::portable_pipe).
699 - new function: AnyEvent::Util::portable_socketpair.
700 - take advantage of the Guard module if it exists.
701
7024.33 Fri Nov 21 02:35:40 CET 2008
703 - AnyEvent::Strict did errornously flag a fileno of 0 as
704 illegal.
705 - reduce memory usage and slightly speed up the pure perl backend
706 by only storing the file descriptor, not the file handle.
707 - add missing autocork method to AnyEvent::Handle (reported
708 by Adam Rosenstein).
709 - AnyEvent::DNS->resolve errornously documented an $rcode result
710 argument, but there isn't (spotted by Henrik Krohns).
711 - the naked truth about Tk - it's basically unmaintained.
712
7134.32 Mon Nov 3 22:46:32 CET 2008
714 - fix AnyEvent::Socket::resolve_sockaddr to properly support
715 unix sockets again and choose a proper default for
716 the socket type.
717 - fix call to resolve_sockaddr in tcp_connect.
718
7194.31 Thu Oct 30 04:41:48 CET 2008
720 - implemented AnyEvent::Handle->destroy method.
721 - hint about unexpected effects in TLS mode.
722 - speed up AnyEvent::Impl::Perl by using more arrays
723 instead of hashes without the slightest loss of readability :)
724 - work around a perl argument refcounting bug.
725 - with some perl patches applied, AnyEvent now seems leak-free.
726
7274.3 Fri Oct 3 09:18:43 CEST 2008
728 - AnyEvent will now install a no-op signal handler for SIGPIPE,
729 unless one has been installed already.
730 - warn about not loaidng AnyEvent::Impl::POE early enough
731 (patch by Adam Rosenstein).
732 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
4 - fix a bug in AnyEvent::Handle where in SSL connect mode, the 733 - fix a bug in where in SSL connect mode, the
5 client would first wait for some data by the server and 734 client would first wait for some data by the server and
6 otherwise hang. (Reported and analyzed in an absolutely 735 otherwise hang. (reported and analysed in an absolutely
7 exemplary manner by Adam Rosenstein). 736 exemplary manner by Adam Rosenstein).
8 - fix a bug in AnyEvent::Handle where SSL EOF would not be treated 737 - fix a bug in where SSL EOF would not be treated
9 as stream EOF, putting the connection into a hung state. 738 as stream EOF, putting the connection into a hung state.
10 (Reported and analyzed in an absolutely exemplary manner by Adam 739 (reported and analysed in an absolutely exemplary manner by Adam
11 Rosenstein). 740 Rosenstein).
12 - fix a potential segfault in AnyEvent::Handle when the TLS context 741 - fix a potential segfault when the TLS context
13 would go missing in a read callback (e.g. due to stoptls) - 742 would go missing in a read callback (e.g. due to stoptls) -
14 Net::SSLeay of course makes no type checking whatsoever. 743 Net::SSLeay of course makes no type checking whatsoever.
744 - AnyEvent::Handle will not stop the read watcher in TLS mode.
745 - AnyEvent::Handle->stoptls will send a TLS close notify
746 instead of simply destroying the stream now.
747 - fix a bug where large blocks of data written in TLS mode
748 would not be sent unless triggered by receives.
749 - on_drain will now take the tls write buffer into account.
750 - SSL operations have been streamlined a bit and should be faster
751 now, more to come, though.
752 - added a FAQ section, will document nontrivial issues in there.
15 753
164.234 Mon Sep 29 04:08:13 CEST 2008 7544.234 Mon Sep 29 04:08:13 CEST 2008
17 - fix child watcher documentation: only child exits will be reported, 755 - fix child watcher documentation: only child exits will be reported,
18 no trace events. 756 no trace events.
19 - mention SIGPIPE in AnyEvent::Handle. 757 - mention SIGPIPE in AnyEvent::Handle.

Diff Legend

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