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

Comparing AnyEvent/Changes (file contents):
Revision 1.230 by root, Sun Apr 19 12:09:46 2009 UTC vs.
Revision 1.568 by root, Sat Mar 23 14:41:21 2013 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
10TODO: invalid. and localhost. specialcasing inside AEDNS and not AESocket (rfc6761)
11TODO: maybe always try to downgrade in push_write.
12 - uts46data.pl couldn't be found due to wrong naming of the file
13 (reported by Fulko Hew).
14 - AnyEvent::Debug shell can now run coro shell commands, if
15 available.
16 - t/63* tests were wrongly in MANIFEST.
17 - mention IO::AIO as optional dependency.
18 - kernel.org's finger server went MIA, switch to freebsd.org.
19
207.04 Thu Nov 15 02:23:58 CET 2012
21 - debugging code left in.
22
237.03 Thu Nov 15 02:16:11 CET 2012
24 - AnyEvent::Socket::inet_aton did not work when dns resolution was
25 used to find the addresses (analyzed and patch by Maxime Soulé).
26 - fix a memory leak in the /etc/hosts lookup code when hosts don't
27 resolve and are not in hosts (analyzed by sten).
28
297.02 Tue Aug 14 04:27:58 CEST 2012
30 - AnyEvent::Util::run_cmd could block indefinitely (analyzed and test
31 program by Yorhel).
32 - verified that AnyEvent::Socket follows RFC5952.
33 - try to parse "ADDR#PORT" in addition to "ADDR PORT".
34
357.01 Sun May 13 01:03:17 CEST 2012
36 - fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but
37 not available, instead of throwing an exception.
38 - use File::Spec to get the tmpdir in t/*, to avoid needless
39 failures on (most, not mine :) windows boxes.
40 - new handle read types: tls_detect and tls_autostart.
41
427.0 Fri Apr 13 06:33:30 CEST 2012
43 - child watchers are broken in POE 1.352 (also many earlier
44 versions) and there seems to be no way to work around it, as POE
45 itself is inherently racy. Document this regression and add a delay
46 in t/68_poe_03_child.t for the time being.
47 - new module AnyEvent::IO, that is a frontend to either a pure-perl
48 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
49 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
50 - load /etc/hosts only when DNS has no answer.
51 - stat /etc/hosts on every access and reload it if it changed.
52 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
53 - fix a buggy croak in the dh parameter reading in AnyEvent::TLS.
54 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
55 a side effect, they now use true appending as opposed to libc
56 appending, and the file might not have been opened when the function
57 returns.
58 - the default logging level was not properly documented in a variety of
59 places, this has hopefully been rectified.
60 - updated uts46data.pl for unicode 6.1.0.
61 - made log messages generated by AnyEvent submodules not
62 include the package name anymore, as it will be logged
63 by default already.
64 - upgrade to the trick used by common-sense 3.5 to work
65 around extra warning torture/breakage under perl 5.15.x.
66 - log messages by anyevent are now ucfirst, are usually full
67 sentences and do no longer include the package name.
68 - the storable read type would sometimes throw an exception instead
69 of causing EBADMSG (analyzed by Maxime Soulé).
70 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
71 keysize (reported by Paul Howarth).
72 - documented Rocco fabricating statements about AnyEvent and me.
73
746.14 Tue Jan 31 20:00:24 CET 2012
75 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
76 probably an editing mistake (reported by Darin McBride).
77
786.13 Thu Jan 12 07:27:01 CET 2012
79 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
80 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
81 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
82 regex-matching inside the callback would change the parameter, despite
83 $1 being dynamically scoped per-block. this perl bug is now being
84 worked around (testcase by Cindy Wang).
85
866.12 Mon Dec 12 13:21:10 CET 2011
87 - $! was clobbered by subsequent calls in tcp_connect, due to
88 the postpone not saving/restoring it (analyzed by Richard Garnish).
89
906.11 Tue Nov 22 10:36:05 CET 2011
91 - Tk cannot create windows when tainted unless you set a title, so
92 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
93 - escape any nonprintable/nonascii characters when stringifying
94 backtraces.
95 - log the reason loading an interface module fails to load at level debug.
96
976.1 Tue Oct 4 19:44:30 CEST 2011
98 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
99 and above), and some messages inside AnyEvent have been elevated
100 to higher log levels to print by default.
101 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
102 - work around yet another signal race bug, newly introduced by POE. My
103 how that module sucks.
104 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
105 support ipv6-only operation even with defaults (also use ::1 instead
106 of 127.0.0.1).
107 - give better guidelines on how to chose a logging level.
108 - AnyEvent::Log can now cap (limit) log message priority by context.
109 - AE::log will now save and restore $! and $@, so callers don't have to.
110 - AE::Strict now checks that registered fds are still valid at regular
111 intervals.
112 - skip frequently called 'require' calls if the module appears loaded.
113 - increase signal latency test timer considerably for poor little osx,
114 which can't be bothered with basic posix functionality such as signal
115 delivery.
116 - mention some debugging aids in the tutorial.
117 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
118 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
119 of the context as title.
120 - AnyEvent::Strict would not restore read-write access to $_ when a callback
121 died - this is a user bug, but since AE::Strict exists to help dbeugging...
122 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
123 to work with it - the author is well aware of the issues but refuses
124 to fix them. This extreme step has been taken because the module actively
125 attacks the goals of AnyEvent and makes it harder for module authors to use
126 AnyEvent.
127
1286.02 Fri Aug 26 20:08:31 CEST 2011
129 - log_to_syslog now accepts facility strings.
130 - log_to_syslog would not allow facilities - "$facility|$level"
131 *literally* means that string. sheesh.
132 - the CPAN parser got confused and thought we provide Sys::Syslog, try
133 to work around.
134 - renamed AnyEvent::Impl::FLTK2 to FLTK.
135
1366.01 Fri Aug 26 07:04:11 CEST 2011
137 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
138 the TTL as fourth element - this affects ->resolve, ->request
139 and ->dns_unpack indirectly. Convenience functions such as
140 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
141 AnyEvent::Socket.
142 - INCOMPATIBLE CHANGE: backend authors now should not implement
143 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
144 methods.
145 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
146 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
147 - DEPRECIATION: please change
148 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
149 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
150 in your programs, for future compatibility.
151 - *blush* idle watcher emulation was borked.
152 - AnyEvent::Strict now checks the AE::xxx functions as well.
153 - querying too long domain names would barf deep inside AnyEvent::DNS - since
154 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
155 for too long domain names explicitly and passes an undef to the callback
156 instead (reported by James Bromberger).
157 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
158 - new functions: AnyEvent::postpone, AnyEvent::log.
159 - new module: AnyEvent::Log, for simple logging needs.
160 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
161 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
162 a debug shell.
163 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
164 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
165 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
166 /etc/hosts can be found.
167 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
168 configuration.
169 - env variables can now be specified using AE_xyz instead of
170 PERL_ANYEVENT_xyz, subject to some rules.
171 - lots of new AnyEvent::Debug functionality.
172 - AnyEvent::Strict now checks for modifications of $_ by the callback
173 (by setting it to readonly while executing the callback, if possible).
174 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
175 Invest extra overhead to work around this peculiar behaviour.
176 - split the pure perl backend into a loop module and an impl
177 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
178 - some read types (regex, netstring, json, storable) could
179 errornously keep an AE::Handle object alive on their own
180 (reported by Mohammad Toossi).
181 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
182 supports /etc/hosts overrides.
183 - add more workarounds around child watcher bugs in IO::Async
184 and (newly introduced) in POE.
185 - use glib child watchers and try to work around its limitations,
186 if possible (based on a patch by Kevin Ryde).
187 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
188 - many read types in AnyEvent::Handle would malfunction during
189 transitions to TLS because of a "delete $self->{rbuf}".
190 (reported by Gerald Galster).
191 - improve condvar blocking wait performance (probably only noticable
192 with faster backends), also make it easier for event loops that
193 do not support blocking waits.
194 - PERL_ANYEVENT_MODEL now also allows full module names, although
195 the usefulness of this feature is not yet clear.
196 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
197 - add t/80_ssltext.t.
198 - no longer use AUTOLOAD for the initial loading (for no very good reason).
199 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
200 - do not run event loop detection multiple times when programs call their cached
201 copies of AnyEvent::detect.
202 - parse_hostport now accepts absolute paths als unix domain sockets.
203 - suppress the idiotic warning inside POE by patching the POE::Kernel
204 run flags instead of asking the user to jump through contortions.
205 - add Task::Weaken as dependency (as recommends only at this time).
206 - replace PF_UNSPEC by 0, for good old minix.
207
2086.0 (not officially released, but was on CPAN for about 15 minutes :/)
209
2105.34 Thu May 12 09:59:41 CEST 2011
211 - AE::TLS didn't support both cert_file and key at the same time,
212 due to some copy&paste bug (reported and analyzed by vti).
213 - improve AE::Handle timeout documentation slightly.
214
2155.33 Mon Feb 7 21:51:41 CET 2011
216 - never appeared on CPAN, for unknown reasons.
217 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
218 object when it's already gone (analyzed by Ryan Bullock).
219 - update idna algorithm(s) to UTS#46 version 6.0.0.
220 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
221 uts46 says it's an error, and testsuite makes more sense.
222 - remove leading dots in idn_to_ascii - testsuite says to remove,
223 uts46 says it's an error, and testsuite makes more sense.
224 - fixed a bug in idn ignored character processing.
225 - passes the uts46 6.0.0 testsuite except for three cases which
226 apparently are bugs in the testsuite itself.
227 - passes the uts46 6.0.1 testsuite except for eleven cases which
228 apparently are bugs in the testsuite itself.
229
2305.31 Sun Jan 23 11:49:19 CET 2011
231 - work around a POE bug where POE would first poll for new events and
232 invoke handlers BEFORE handling the message queue.
233 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
234 - stop_read now does just that in TLS mode. see its
235 documentation for caveats.
236 - update to common::sense 3.4 warning set (some warnings in perl impose
237 an arbitrary 32kb limit on regex matches, which affects
238 AnyEvent::Handle).
239 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
240 - new experimental AE::Handle::resettls method.
241
2425.3 Fri Dec 31 05:48:13 CET 2010
243 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
244 issues with IO::Async have been solved in that version, including
245 the availabilty of a default loop!
246 - implement new wbuf_max parameter for AnyEvent::Handle.
247 - added Cocoa::EventLoop backend by Daisuke Murase.
248 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
249 - clarify handle on_prepare documentation.
250 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
251 to avoid leaking them to any child processes.
252 - try to force files into the arch-dependend lib directory, for the
253 half-"turly obsessive compulsive person"(s) that depend on it.
254
2555.29 Sun Dec 5 10:49:21 CET 2010
256 - convert EV backend to EV 4.00 API (so better upgrade EV too).
257 - AE::Handle: implement dynamic read_size adjustment, add
258 max_read_size parameter, reduce default read_size to 2048.
259 - add some FAQ entries for servers and/or on_eof vs. on_error.
260 - work around OS bugs (cygwin again) when lingering in AE::Handle.
261 - reduce memory usage slightly when connect was used in AE::Handle.
262 - wrok around more segfaults in Tk.
263 - document tls_ctx => undef as valid.
264 - detect WNOHANG value at compiletime.
265 - include some probably linux-only support for building AnyEvent
266 as part of the perl core.
267 - improve compatibility of t/08_idna.t to perl 5.8.
268 - make AnyEvent::Strict truly optional (in case somebody builds a
269 minimal perl...).
270
2715.28 Wed Oct 13 04:14:23 CEST 2010
272 - due to a glitch, AnyEvent's internal getprotobyname wasn't
273 actually used in tcp_server.
274 - implement AnyEvent::CondVar->new method, which was documented but
275 missing.
276 - support autoloading of read types in unshift_read.
277 - AnyEvent::Strict: do the test for invalid fh arguments differently,
278 resulting in more informative error messages.
279 - new AnyEvent::FAQ manpage.
280 - clarify that you can actually call push_read etc. in on_read.
281 - put keys and certs in /tmp, not into the current directory,
282 and document this in the SECURITY CONSIDERATIONS sections
283 of AnyEvent::TLS.
284 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
285 requested by Ron Isaacson.
286 - document $AnyEvent::DNS::RESOLVER variable.
287 - incorporated major typo patches by Abhijit Menon-Sen.
288 - AnyEvent::Handle now croaks when negative timeouts are passed.
289 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
290 undocumented, but who knows when they might come in handy.
291
2925.271 Tue Jun 8 12:05:46 CEST 2010
293 - backport to perl 5.8.x.
294
2955.27 Sun Jun 6 12:12:05 CEST 2010
296 - postpone differently in AnyEvent::Socket now, as
297 when not, canceling the connection attempt might fail
298 (found by Felix Antonius Wilhelm Ostmann).
299 - explicitly check for non-stream sockets in AE::Handle, too many
300 clueless people fell into the trap of this somehow working.
301 - simplified and reworked the "OTHER MODULES" section.
302 - better/more condvar examples.
303
3045.261 Wed Apr 28 16:13:36 CEST 2010
305 - AF_INET6 was not properly used from Socket6 during configuration
306 time (found by Andy Grundman).
307
3085.26 Mon Apr 12 04:49:35 CEST 2010
309 - don't generate spurious readyness notifications when select
310 returns EINTR in the pure perl backend (analysed and testcase
311 by Bas Denissen).
312 - give same examples in SYNOPSIS sections of AnyEvent and AE.
313 - provider faster implementations for AE::time/now/now_update for
314 EV and Perl backends.
315 - provide faster AE::cv and AE::time implementations for all
316 backends.
317 - no longer support lower/mixed-case signal names (this was
318 never documented nor universally supported).
319 - some more memory µ-optimisations, and somewhat less messy
320 code to cope with both AnyEvent and AE APIs.
321
3225.251 Sat Mar 13 00:58:21 CET 2010
323 - make test failed if it was run before make install on perl <=
324 5.8.8, and older versions of perl make our live very difficult,
325 so write constants.pl during Makefile.PL time.
326
3275.25 Sat Mar 13 00:23:14 CET 2010
328 - fix a race condition in AnyEvent::Handle that would cause
329 a "bio_read: unsupported method" error (in _another_
330 ssl connection) after user code threw an exception.
331 - added AnyEvent::Handle->destroyed method.
332 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
333 - the AnyEvent::Util::fh_nonblocking function worked only
334 by lucky accident on win32.
335 - smaller and faster AnyEvent::Util::fh_nonblocking.
336 - when the (required!) Time::HiRes module is missing, AnyEvent
337 did not fall back to built-in time properly.
338 - do not load Fcntl at runtime, saving memory and loading time.
339 - precompile a number of constants and use them instead of runtime
340 detection and eval.
341 - free detection code after detection and similar memory optimisations.
342 - Perl backend timer interval best effort drift has been improved
343 (same algorithm as EV).
344 - update unicode idna mapping table.
345
3465.24 Tue Jan 5 11:39:43 CET 2010
347 - cygwin never reports errors from failed connects. 1.5 just gives
348 you continous readyness and EAGAIN, 1.7 is even more broken and
349 just hangs. work around both issues in a cygwin-specific hack.
350 - improve idle watcher documentation slightly.
351
3525.23 Sun Dec 20 23:48:00 CET 2009
353 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
354 - implement punycode_encode/decode, idn_nameprep,
355 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
356 - provide $AE::VERSION.
357 - removed traces of "no strict 'refs'".
358
3595.22 Sat Dec 5 03:51:13 CET 2009
360 - downgrade-or-fail in AnyEvent::Handle::push_write, to
361 diagnose encoding failures earlier and more succinctly.
362 (this works around bugs in perl, throwing away encoding info
363 when passing scalar data to extensions).
364 - add more examples to AnyEvent::Socket manpage.
365 - upgrade internal warning set to the same as common::sense 2.03.
366 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
367 (John Beppu).
368 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
369 - add google public dns servers to fallback server set and make sure
370 we load-balance properly between the three sets. also add all
371 fallback dns servers, not just a random one, to each dns config.
372
3735.21 Thu Nov 19 02:48:47 CET 2009
374 - fix a problem where socket constants were called with parameters
375 (spotted by David Friedland).
376 - fork_call never use'd POSIX (reported by Daisuke Maki).
377 - improve perl 5.6 compatibility further (but it still won't work
378 unless you rip out everything but the core).
379 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
380 uses 10MB of RAM, but doesn't open a console window. *sigh*.
381
3825.202 Wed Oct 14 22:35:44 CEST 2009
383 - AnyEvent::DNS would unexpectedly clobber $_ under windows
384 (analysed by Matthias Waldorf).
385 - AnyEvent::Handle::run_cmd can now pass the PID of the
386 newly-created process, which is much less useful than it might
387 sound (based on patch by Yann Kerherve).
388
3895.201 Tue Sep 29 12:09:25 CEST 2009
390 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
391 (reported by Torsten Foertsch).
392 - common::sense 2.0 could cause tcp_server to throw an exception
393 (analysed by elmex).
394
3955.2 Mon Sep 14 07:04:49 CEST 2009
396 - INCOMPATIBLE CHANGE: do no longer support register_read_type
397 and register_write_type in AnyEvent::Handle, instead support
398 package names (the facility was mostly abused).
399 - implement "packagename-as-read/write type" support in
400 AnyEvent::Handle.
401 - AnyEvent::Handle: new options "keepalive" and "oobinline".
402 - oobinline set by default to avoid security issues.
403 - the pure-perl event loop backend wrongly detected times() fallback
404 support (spotted by Pavel Boldin).
405
4065.12 Tue Sep 1 20:26:50 CEST 2009
407 - be more lenient when parsing resolv.conf files, as some people
408 use hashmarks as comment indicator inside directives (reported
409 by Michael S. Fischer).
410 - use same set of warnings as common::sense 2.0.
411 - fix a potential 32 bit overflow issue due to perl having problems
412 with large hex constants in the Perl backend.
413 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
414 by default.
415 - fix some prototypes.
416 - (optionally) require Async::Interrupt 1.02.
417 - replace opendns fallbacks by verizon/level3 public dns servers.
418
4195.112 Fri Aug 21 13:59:04 CEST 2009
420 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
421 using starttls with an illegal/uncreatable tls context.
422 - port to broken win32 perls.
423
4245.111 Sun Aug 16 18:44:36 CEST 2009
425 - perl errornously requires sockaddr_un structures returned by the
426 kernel to have a specific length. work around this bug by padding
427 sockaddr_un structures (found to fail on OS X, as perl only uses
428 the correct code for GNU/Linux, but likely to fail on other
429 OSes as well).
430
4315.11 Wed Aug 12 17:49:37 CEST 2009
432 - fixed a bug in run_cmd with input from filename.
433 - create a json coder object when none is given in write direction
434 to reduce dependency on JSON.
435
4365.1 Tue Aug 11 03:17:32 CEST 2009
437 - tcp_connect now reports the correct error in $!, not an unrelated
438 one (was broken since 4.91).
439 - AE::Handle did not properly clear rtimeout/wtimeout watchers
440 when retrying a tcp_connect.
441 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
442 - fixed wrong documentation in AE::Socket.
443
4445.01 Mon Aug 10 03:16:32 CEST 2009
445 - last release broke EV child watchers :/.
446
4475.0 Sun Aug 9 17:32:09 CEST 2009
448 - officially introduce and document the AE API.
449 - lowering the timeout at runtime did not have immediate effect
450 in AnyEvent::Handle.
451 - AE::Handle now has separate and independent read and write timeout
452 settings.
453 - probe for Guard module when AnyEvent::Util is loaded, not at
454 runtime (Event doesn't like eval "use" from callbacks, and it avoids
455 exporting the probe function).
456 - Impl::IOAsync's io watchers did not use the filehandle emulation
457 code to work around it's watcher limits.
458 - work around IO::Async's problems with overloaded objects.
459 - try to disable wearnings inside IO::Async, it's just too buggy.
460 - do no longer use Event by default, only EV or Perl, Event is too buggy.
461 - hackishly provide fast AE:: interface when EV or Perl is the backend.
462 - inherit VERSION from AnyEvent::VERSION in most modules.
463
4644.91 Thu Aug 6 15:42:45 CEST 2009
465 - AE::Handle::starttls could get out of sync when the read buffer
466 already contains some TLS handshake.
467 - AE::Handle did not properly free the TLS session, causing bigger
468 memory leaks in Net::SSLeay than expected (reported by toaster).
469 - AE::Socket::tcp_connect will now artificially delay invoking
470 the callback to avoid returning after invoking the callback.
471 - convert many internal watcher uses to AE API.
472
4734.9 Sat Aug 1 11:07:01 CEST 2009
474 - Glib silently fails when registering a timeout with
475 a negative value, so avoid that.
476 - call condvar callback immediately when it is set
477 after the condition is already signalled.
478 - check rbuf_max condition only after trying to consume
479 data.
480 - normalise signal numbers to names when using the ae's
481 signal handling, but do not document this (yet).
482 - pure perl signal emulation did not properly set nonblocking
483 mode on the signal pipe on win32 (this is unlikely to have caused
484 any issues).
485 - new module: AnyEvent::Debug.
486 - AnyEvent::Strict now checks that a signal was specified
487 by name (not name or number).
488 - reduce memory footprint in typical cases by ~50kb by
489 only compiling signal code when necessary.
490 - add AnyEvent::Handle->rbuf_max.
491 - grab the AE namespace for future new API, implement stubs for
492 said future/alternative API.
493 - new function: AnyEvent::Socket::format_hostport.
494
4954.881 Tue Jul 28 12:51:53 CEST 2009
496 - work around a bug in local in pre-5.10 perls, causing
497 AnyEvent::Handle to recurse when it should not
498 (analysed by elmex).
499
5004.88 Tue Jul 28 04:04:37 CEST 2009
501 - re-bless the handle into a dummy package after calling
502 AnyEvent::Handle::destroy, so the user does not need to check
503 for errors after every push_write etc.
504 - do not attempt to run t/02_signals.t on obviously broken
505 platforms.
506
5074.87 Sun Jul 26 02:06:16 CEST 2009
508 - do not attempt to linger when there is no longer a valid fh.
509 - enforce tls mode to be either accept or connect, do not simply
510 segfault in Net::SSLeay.
511 - AnyEvent::Handle can now call tcp_connect itself (new parameters
512 connect, on_prepare, on_connect and on_connect_error). Updated
513 tutorial accordingly.
514 - add AnyEvent::Impl::Irssi backend.
515
5164.86 Mon Jul 20 23:52:29 CEST 2009
517 - since the verbose warning is not enough, explicitly document
518 that versions before 1.33 of Net::SSLeay are not secure.
519 - work around signal handling races in Event and (...) Event::Lib.
520 - try to align signal-race timer to full-second boundaries.
521 - work around Tk not liking negative timeouts.
522 - don't complain of different grades of environmental unfriendlyness
523 in IO::Async.
524
5254.85 Sat Jul 18 06:16:14 CEST 2009
526 - nail the signal race problem in perl once and for all
527 (see $AnyEvent::MAX_SIGNAL_LATENCY).
528 - take advantage of Async::Interrupt if it is available.
529 - load Time::HiRes and Guard modules on demand only.
530 - add optional/recommended modules section to AnyEvent
531 documentation.
532 - reduce memory usage considerably (and reduce startup penalty)
533 by not using "strict", "warnings" and "overload" modules.
534 - work around buggy windows/openbsd perls and provide EBADMSG
535 and EPROTO ourselves when missing.
536 - improve perl 5.6 compatibility of the core event loop.
537 - made Net::SSLeay version 1.33 a soft requirement.
538
5394.83 Fri Jul 17 16:56:26 CEST 2009
540 - implement AnyEvent::Socket::getprotobyname.
541 - AnyEvent::CondVar's will now detect recursive blocking
542 waits and will croak, as too many people fall into
543 this trap.
544 - AnyEvent::Handle will now call ->destroy on itself after
545 executing the on_error callback, instead of doing some
546 half-baked internal shutdown, for fatal errors.
547 - clarify on_eof behaviour w.r.t. the read queue and
548 on_read callbacks.
549 - ignore some possible spurious wake-ups in tcp_connect.
550
5514.82 Sat Jul 11 00:34:55 CEST 2009
552 - POE and Event backends didn't accept some callable objects as
553 callbacks.
554 - use Config module instead of POSIX module to detect signal names
555 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
556 module doesn't even have all POSIX signals :/.
557 - use more workarounds around the many refcnt/corruption bugs in
558 Event::Lib.
559 - work around a race condition in perl's select, causing t/03_child.t
560 to rarely fail.
561
5624.81 Thu Jul 9 10:30:30 CEST 2009
563 - AnyEvent::Handle didn't properly diagnose write errors
564 (it expected -1 from syswrite, how lame... :).
565 - support file descriptors in addition to file handles
566 in AnyEvent->io.
567 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
568 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
569 and PERL_ANYEVENT_CA_PATH.
570 - provide a sensible synopsis section for AnyEvent::TLS.
571 - add a "supported backends" section to the manpage.
572 - added simple io watcher test to testsuite, using a
573 portable_socketpair.
574 - tried to improve the stability of the Event::Lib backend,
575 YMMV.
576
5774.8 Mon Jul 6 23:45:16 CEST 2009
578 - AnyEvent::DNS did not properly follow CNAME records with
579 uppercase targets.
580 - AnyEvent::DNS would errornously return AAAA records
581 with v4 mapped addresses (a faulty record) as ipv4 addresses,
582 causing AnyEvent::Socket to throw an exception.
583 - added new module AnyEvent::TLS for easier SSL/TLS context
584 creation, with many options including hostname verification,
585 secure default configuration, lots of documentation and,
586 predefined diffie-hellman keys for perfect forward security
587 and much more. get it while it's still fresh!
588 - use AnyEvent::TLS in AnyEvent::Handle for context management.
589 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
590 so AnyEvent::Socket users have smaller memory footprint
591 in the common case.
592 - add AnyEvent::Handle->push_shutdown method.
593 - add an additional error message parameter to AnyEvent::Handle's
594 on_error callback (for TLS, $! is still available).
595 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
596 - make AnyEvent::Handle more robust against early conenction
597 failures (during new), and return C<undef> in that case
598 from the constructor.
599 - AnyEvent::Handle will now try to load only JSON::XS first,
600 then fall back to JSON.
601 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
602 and because it was documented that way.
603
6044.451 Fri Jul 3 00:28:58 CEST 2009
605 - do not clear rbuf when shutting down an AnyEvent::Handle
606 object - doing so breaks AnyEvent::HTTP.
607
6084.45 Mon Jun 29 22:59:26 CEST 2009
609 - a write error could cause AnyEvent::Handle to create
610 an I/O watcher with an undefined $fh.
611 - special-case mapped ipv4 addresses in both
612 AnyEvent::Socket::format_address and parse_address,
613 to treat them just like ipv4 addresses.
614 - updated and overhauled the AnyEvent::Intro doc.
615 - implement AnyEvent::Socket::format_ipv4/ipv6.
616 - slightly speed up portable_pipe/socketpair.
617 - expand condvar begin/end documentation.
618
6194.42 Fri Jun 26 08:32:18 CEST 2009
620 - preliminary and neccesarily incomplete support for IO::Async.
621 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
622 case it was set to IGNORE, to ensure we can catch child statuses
623 even when the calling env acts stupidly.
624 - updated benchmarks with IO::Async, which performs very well.
625
6264.412 Wed Jun 24 01:35:57 CEST 2009
627 - support an "untaint" attribute for AnyEvent::DNS
628 and set it on the default resolver.
629 - implement and document AnyEvent::Impl::Perl::loop.
630 - remove all anyevent-env variables from %ENV
631 when running in tainted mode.
632 - mention and extend the IO::Lambda benchmark.
633
6344.411 Sun Jun 7 18:48:02 CEST 2009
635 - do not try to use F_SETFD on windows.
636
6374.41 Thu May 14 06:40:11 CEST 2009
638 - work around issues in older perls (5.8.0?) when
639 a signal handler is deleted from the %SIG hash.
640 - use POSIX::_exit in child test, to avoid running
641 destructors.
642 - speed up CHLD handling by relying on SIGCHLD being
643 synchronously delivered, even when we roll our own
644 implementation.
645 - AnyEvent::DNS: add the "dname" resource record name
646 for cosmetic reasons.
647
6484.4 Sun Apr 26 20:12:33 CEST 2009
649 - implemented idle watchers, where applicable.
650 - AnyEvent->time died when Event backend was in use.
651 - fix a memleak in the Tk backend.
652 - sped up Tk timer handling.
653 - clip negative "after" values to 0 in AnyEvent::Impl::Event
654 to avoid spamming out warning messages.
655 - fix Qt timers without interval.
656 - avoid Qt zero-timeout special case, in old libqt's.
657
6584.352 Mon Apr 20 16:31:11 CEST 2009
3 - fix AnyEvent::Strict error messages for child watchers. 659 - fix AnyEvent::Strict error messages for child watchers.
4 - fix/update Makefile.PL "recommends" versions. 660 - fix/update Makefile.PL "recommends" versions.
5 - add AnyEvent->now_update. 661 - add AnyEvent->now_update.
6 662
74.351 Sat Apr 11 07:56:14 CEST 2009 6634.351 Sat Apr 11 07:56:14 CEST 2009
… …
694.3 Fri Oct 3 09:18:43 CEST 2008 7254.3 Fri Oct 3 09:18:43 CEST 2008
70 - AnyEvent will now install a no-op signal handler for SIGPIPE, 726 - AnyEvent will now install a no-op signal handler for SIGPIPE,
71 unless one has been installed already. 727 unless one has been installed already.
72 - warn about not loaidng AnyEvent::Impl::POE early enough 728 - warn about not loaidng AnyEvent::Impl::POE early enough
73 (patch by Adam Rosenstein). 729 (patch by Adam Rosenstein).
74 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 730 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
75 - fix a bug in where in SSL connect mode, the 731 - fix a bug in where in SSL connect mode, the
76 client would first wait for some data by the server and 732 client would first wait for some data by the server and
77 otherwise hang. (reported and analyzed in an absolutely 733 otherwise hang. (reported and analysed in an absolutely
78 exemplary manner by Adam Rosenstein). 734 exemplary manner by Adam Rosenstein).
79 - fix a bug in where SSL EOF would not be treated 735 - fix a bug in where SSL EOF would not be treated
80 as stream EOF, putting the connection into a hung state. 736 as stream EOF, putting the connection into a hung state.
81 (reported and analyzed in an absolutely exemplary manner by Adam 737 (reported and analysed in an absolutely exemplary manner by Adam
82 Rosenstein). 738 Rosenstein).
83 - fix a potential segfault when the TLS context 739 - fix a potential segfault when the TLS context
84 would go missing in a read callback (e.g. due to stoptls) - 740 would go missing in a read callback (e.g. due to stoptls) -
85 Net::SSLeay of course makes no type checking whatsoever. 741 Net::SSLeay of course makes no type checking whatsoever.
86 - AnyEvent::Handle will not stop the read watcher in TLS mode. 742 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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