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

Comparing AnyEvent/Changes (file contents):
Revision 1.246 by root, Wed Jun 24 10:03:42 2009 UTC vs.
Revision 1.586 by root, Tue Dec 17 18:36:22 2013 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: document ->loop, and maybe one_event. 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?
9TODO: invalid. and localhost. specialcasing inside AEDNS and not AESocket (rfc6761)
10TODO: maybe implement env variable to give hosts precedence
11
127.07 Tue Dec 17 17:45:02 CET 2013
13 - the documentation for custom tls verify schemes was wrong. make it agree
14 with the code (reported by Maxime Soulé).
15 - added cbor read and write types to AnyEvent::Handle (using CBOR::XS).
16 - work around an API change in openssl that could cause wrong tls connection
17 aborts, likely on windows only (analyzed by sten).
18 - calling AnyEvent->now_update with AnyEvent::Impl::Perl caused an
19 endless loop (reported by Dietrich Rebmann).
20 - add tlsv1_1 and tlsv1_2 protocols to AnyEvent::TLS
21 (patch by Maxime Soulé).
22 - document AnyEvent::Impl::IOAsync::set_loop and
23 $AnyEvent::Impl::IOAsync::LOOP. Though only documented now, this
24 functionality has _always_ been available.
25 - force a toplevel domain name in t/81_hosts.t (analyzed by
26 David Jack Wange Olrik).
27 - document that AnyEvent::Log uses AnyEvent::IO.
28 - warn about AnyEvent::Filesys::Notify performance.
29 - praise the joys of AnyEvent::Fork::*.
30 - time for an =encoding directive.
31 - do no longer use JSON to create a default json coder, use
32 JSON::XS or JSON::PP directly.
33
347.05 Wed Aug 21 10:38:08 CEST 2013
35 - uts46data.pl couldn't be found due to wrong naming of the file
36 (reported by Fulko Hew).
37 - handle lone \015's properly in AE::Handle's default line read
38 (reported by various people).
39 - untaint ip addresses found in /etc/hosts (patch by José Micó).
40 - the memleak fix in 7.03 caused resolving via /etc/hosts to always
41 fail on first use (reported and testcase by Andrew Whatson).
42 - expose AnyEvent::Log::format_time, and allow users to redefine it.
43 - expose AnyEvent::Log::default_format, and allow redefinition.
44 - expose AnyEvent::Log::fatal_exit, to allow redefinition.
45 - AnyEvent::Debug shell can now run coro shell commands, if available.
46 - t/63* tests were wrongly in MANIFEST.
47 - kernel.org's finger server went MIA, switch to freebsd.org and
48 icculus.org.
49 - clarify that IO::AIO and AnyEvent::AIO are needed for AnyEvent::IO
50 to function asynchronously (reported by Daniel Carosone).
51 - hard-disable $^W in most tests, it generates too much garbage output.
52 - use a (hopefully) more future-proof method to emulate common::sense.
53 - upgrade to UTS-46:6.2.0.
54 - switch to INSTLIB from INSTLIBDIR, as INSTLIBDIR was wrongly
55 documented. should not affect anything.
56
577.04 Thu Nov 15 02:23:58 CET 2012
58 - debugging code left in.
59
607.03 Thu Nov 15 02:16:11 CET 2012
61 - AnyEvent::Socket::inet_aton did not work when dns resolution was
62 used to find the addresses (analyzed and patch by Maxime Soulé).
63 - fix a memory leak in the /etc/hosts lookup code when hosts don't
64 resolve and are not in hosts (analyzed by sten).
65
667.02 Tue Aug 14 04:27:58 CEST 2012
67 - AnyEvent::Util::run_cmd could block indefinitely (analyzed and test
68 program by Yorhel).
69 - verified that AnyEvent::Socket follows RFC5952.
70 - try to parse "ADDR#PORT" in addition to "ADDR PORT".
71
727.01 Sun May 13 01:03:17 CEST 2012
73 - fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but
74 not available, instead of throwing an exception.
75 - use File::Spec to get the tmpdir in t/*, to avoid needless
76 failures on (most, not mine :) windows boxes.
77 - new handle read types: tls_detect and tls_autostart.
78
797.0 Fri Apr 13 06:33:30 CEST 2012
80 - child watchers are broken in POE 1.352 (also many earlier
81 versions) and there seems to be no way to work around it, as POE
82 itself is inherently racy. Document this regression and add a delay
83 in t/68_poe_03_child.t for the time being.
84 - new module AnyEvent::IO, that is a frontend to either a pure-perl
85 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
86 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
87 - load /etc/hosts only when DNS has no answer.
88 - stat /etc/hosts on every access and reload it if it changed.
89 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
90 - fix a buggy croak in the dh parameter reading in AnyEvent::TLS.
91 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
92 a side effect, they now use true appending as opposed to libc
93 appending, and the file might not have been opened when the function
94 returns.
95 - the default logging level was not properly documented in a variety of
96 places, this has hopefully been rectified.
97 - updated uts46data.pl for unicode 6.1.0.
98 - made log messages generated by AnyEvent submodules not
99 include the package name anymore, as it will be logged
100 by default already.
101 - upgrade to the trick used by common-sense 3.5 to work
102 around extra warning torture/breakage under perl 5.15.x.
103 - log messages by anyevent are now ucfirst, are usually full
104 sentences and do no longer include the package name.
105 - the storable read type would sometimes throw an exception instead
106 of causing EBADMSG (analyzed by Maxime Soulé).
107 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
108 keysize (reported by Paul Howarth).
109 - documented Rocco fabricating statements about AnyEvent and me.
110
1116.14 Tue Jan 31 20:00:24 CET 2012
112 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
113 probably an editing mistake (reported by Darin McBride).
114
1156.13 Thu Jan 12 07:27:01 CET 2012
116 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
117 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
118 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
119 regex-matching inside the callback would change the parameter, despite
120 $1 being dynamically scoped per-block. this perl bug is now being
121 worked around (testcase by Cindy Wang).
122
1236.12 Mon Dec 12 13:21:10 CET 2011
124 - $! was clobbered by subsequent calls in tcp_connect, due to
125 the postpone not saving/restoring it (analyzed by Richard Garnish).
126
1276.11 Tue Nov 22 10:36:05 CET 2011
128 - Tk cannot create windows when tainted unless you set a title, so
129 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
130 - escape any nonprintable/nonascii characters when stringifying
131 backtraces.
132 - log the reason loading an interface module fails to load at level debug.
133
1346.1 Tue Oct 4 19:44:30 CEST 2011
135 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
136 and above), and some messages inside AnyEvent have been elevated
137 to higher log levels to print by default.
138 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
139 - work around yet another signal race bug, newly introduced by POE. My
140 how that module sucks.
141 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
142 support ipv6-only operation even with defaults (also use ::1 instead
143 of 127.0.0.1).
144 - give better guidelines on how to chose a logging level.
145 - AnyEvent::Log can now cap (limit) log message priority by context.
146 - AE::log will now save and restore $! and $@, so callers don't have to.
147 - AE::Strict now checks that registered fds are still valid at regular
148 intervals.
149 - skip frequently called 'require' calls if the module appears loaded.
150 - increase signal latency test timer considerably for poor little osx,
151 which can't be bothered with basic posix functionality such as signal
152 delivery.
153 - mention some debugging aids in the tutorial.
154 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
155 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
156 of the context as title.
157 - AnyEvent::Strict would not restore read-write access to $_ when a callback
158 died - this is a user bug, but since AE::Strict exists to help dbeugging...
159 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
160 to work with it - the author is well aware of the issues but refuses
161 to fix them. This extreme step has been taken because the module actively
162 attacks the goals of AnyEvent and makes it harder for module authors to use
163 AnyEvent.
164
1656.02 Fri Aug 26 20:08:31 CEST 2011
166 - log_to_syslog now accepts facility strings.
167 - log_to_syslog would not allow facilities - "$facility|$level"
168 *literally* means that string. sheesh.
169 - the CPAN parser got confused and thought we provide Sys::Syslog, try
170 to work around.
171 - renamed AnyEvent::Impl::FLTK2 to FLTK.
172
1736.01 Fri Aug 26 07:04:11 CEST 2011
174 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
175 the TTL as fourth element - this affects ->resolve, ->request
176 and ->dns_unpack indirectly. Convenience functions such as
177 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
178 AnyEvent::Socket.
179 - INCOMPATIBLE CHANGE: backend authors now should not implement
180 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
181 methods.
182 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
183 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
184 - DEPRECIATION: please change
185 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
186 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
187 in your programs, for future compatibility.
188 - *blush* idle watcher emulation was borked.
189 - AnyEvent::Strict now checks the AE::xxx functions as well.
190 - querying too long domain names would barf deep inside AnyEvent::DNS - since
191 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
192 for too long domain names explicitly and passes an undef to the callback
193 instead (reported by James Bromberger).
194 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
195 - new functions: AnyEvent::postpone, AnyEvent::log.
196 - new module: AnyEvent::Log, for simple logging needs.
197 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
198 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
199 a debug shell.
200 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
201 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
202 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
203 /etc/hosts can be found.
204 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
205 configuration.
206 - env variables can now be specified using AE_xyz instead of
207 PERL_ANYEVENT_xyz, subject to some rules.
208 - lots of new AnyEvent::Debug functionality.
209 - AnyEvent::Strict now checks for modifications of $_ by the callback
210 (by setting it to readonly while executing the callback, if possible).
211 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
212 Invest extra overhead to work around this peculiar behaviour.
213 - split the pure perl backend into a loop module and an impl
214 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
215 - some read types (regex, netstring, json, storable) could
216 errornously keep an AE::Handle object alive on their own
217 (reported by Mohammad Toossi).
218 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
219 supports /etc/hosts overrides.
220 - add more workarounds around child watcher bugs in IO::Async
221 and (newly introduced) in POE.
222 - use glib child watchers and try to work around its limitations,
223 if possible (based on a patch by Kevin Ryde).
224 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
225 - many read types in AnyEvent::Handle would malfunction during
226 transitions to TLS because of a "delete $self->{rbuf}".
227 (reported by Gerald Galster).
228 - improve condvar blocking wait performance (probably only noticable
229 with faster backends), also make it easier for event loops that
230 do not support blocking waits.
231 - PERL_ANYEVENT_MODEL now also allows full module names, although
232 the usefulness of this feature is not yet clear.
233 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
234 - add t/80_ssltext.t.
235 - no longer use AUTOLOAD for the initial loading (for no very good reason).
236 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
237 - do not run event loop detection multiple times when programs call their cached
238 copies of AnyEvent::detect.
239 - parse_hostport now accepts absolute paths als unix domain sockets.
240 - suppress the idiotic warning inside POE by patching the POE::Kernel
241 run flags instead of asking the user to jump through contortions.
242 - add Task::Weaken as dependency (as recommends only at this time).
243 - replace PF_UNSPEC by 0, for good old minix.
244
2456.0 (not officially released, but was on CPAN for about 15 minutes :/)
246
2475.34 Thu May 12 09:59:41 CEST 2011
248 - AE::TLS didn't support both cert_file and key at the same time,
249 due to some copy&paste bug (reported and analyzed by vti).
250 - improve AE::Handle timeout documentation slightly.
251
2525.33 Mon Feb 7 21:51:41 CET 2011
253 - never appeared on CPAN, for unknown reasons.
254 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
255 object when it's already gone (analyzed by Ryan Bullock).
256 - update idna algorithm(s) to UTS#46 version 6.0.0.
257 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
258 uts46 says it's an error, and testsuite makes more sense.
259 - remove leading dots in idn_to_ascii - testsuite says to remove,
260 uts46 says it's an error, and testsuite makes more sense.
261 - fixed a bug in idn ignored character processing.
262 - passes the uts46 6.0.0 testsuite except for three cases which
263 apparently are bugs in the testsuite itself.
264 - passes the uts46 6.0.1 testsuite except for eleven cases which
265 apparently are bugs in the testsuite itself.
266
2675.31 Sun Jan 23 11:49:19 CET 2011
268 - work around a POE bug where POE would first poll for new events and
269 invoke handlers BEFORE handling the message queue.
270 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
271 - stop_read now does just that in TLS mode. see its
272 documentation for caveats.
273 - update to common::sense 3.4 warning set (some warnings in perl impose
274 an arbitrary 32kb limit on regex matches, which affects
275 AnyEvent::Handle).
276 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
277 - new experimental AE::Handle::resettls method.
278
2795.3 Fri Dec 31 05:48:13 CET 2010
280 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
281 issues with IO::Async have been solved in that version, including
282 the availabilty of a default loop!
283 - implement new wbuf_max parameter for AnyEvent::Handle.
284 - added Cocoa::EventLoop backend by Daisuke Murase.
285 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
286 - clarify handle on_prepare documentation.
287 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
288 to avoid leaking them to any child processes.
289 - try to force files into the arch-dependend lib directory, for the
290 half-"turly obsessive compulsive person"(s) that depend on it.
291
2925.29 Sun Dec 5 10:49:21 CET 2010
293 - convert EV backend to EV 4.00 API (so better upgrade EV too).
294 - AE::Handle: implement dynamic read_size adjustment, add
295 max_read_size parameter, reduce default read_size to 2048.
296 - add some FAQ entries for servers and/or on_eof vs. on_error.
297 - work around OS bugs (cygwin again) when lingering in AE::Handle.
298 - reduce memory usage slightly when connect was used in AE::Handle.
299 - wrok around more segfaults in Tk.
300 - document tls_ctx => undef as valid.
301 - detect WNOHANG value at compiletime.
302 - include some probably linux-only support for building AnyEvent
303 as part of the perl core.
304 - improve compatibility of t/08_idna.t to perl 5.8.
305 - make AnyEvent::Strict truly optional (in case somebody builds a
306 minimal perl...).
307
3085.28 Wed Oct 13 04:14:23 CEST 2010
309 - due to a glitch, AnyEvent's internal getprotobyname wasn't
310 actually used in tcp_server.
311 - implement AnyEvent::CondVar->new method, which was documented but
312 missing.
313 - support autoloading of read types in unshift_read.
314 - AnyEvent::Strict: do the test for invalid fh arguments differently,
315 resulting in more informative error messages.
316 - new AnyEvent::FAQ manpage.
317 - clarify that you can actually call push_read etc. in on_read.
318 - put keys and certs in /tmp, not into the current directory,
319 and document this in the SECURITY CONSIDERATIONS sections
320 of AnyEvent::TLS.
321 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
322 requested by Ron Isaacson.
323 - document $AnyEvent::DNS::RESOLVER variable.
324 - incorporated major typo patches by Abhijit Menon-Sen.
325 - AnyEvent::Handle now croaks when negative timeouts are passed.
326 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
327 undocumented, but who knows when they might come in handy.
328
3295.271 Tue Jun 8 12:05:46 CEST 2010
330 - backport to perl 5.8.x.
331
3325.27 Sun Jun 6 12:12:05 CEST 2010
333 - postpone differently in AnyEvent::Socket now, as
334 when not, canceling the connection attempt might fail
335 (found by Felix Antonius Wilhelm Ostmann).
336 - explicitly check for non-stream sockets in AE::Handle, too many
337 clueless people fell into the trap of this somehow working.
338 - simplified and reworked the "OTHER MODULES" section.
339 - better/more condvar examples.
340
3415.261 Wed Apr 28 16:13:36 CEST 2010
342 - AF_INET6 was not properly used from Socket6 during configuration
343 time (found by Andy Grundman).
344
3455.26 Mon Apr 12 04:49:35 CEST 2010
346 - don't generate spurious readyness notifications when select
347 returns EINTR in the pure perl backend (analysed and testcase
348 by Bas Denissen).
349 - give same examples in SYNOPSIS sections of AnyEvent and AE.
350 - provider faster implementations for AE::time/now/now_update for
351 EV and Perl backends.
352 - provide faster AE::cv and AE::time implementations for all
353 backends.
354 - no longer support lower/mixed-case signal names (this was
355 never documented nor universally supported).
356 - some more memory µ-optimisations, and somewhat less messy
357 code to cope with both AnyEvent and AE APIs.
358
3595.251 Sat Mar 13 00:58:21 CET 2010
360 - make test failed if it was run before make install on perl <=
361 5.8.8, and older versions of perl make our live very difficult,
362 so write constants.pl during Makefile.PL time.
363
3645.25 Sat Mar 13 00:23:14 CET 2010
365 - fix a race condition in AnyEvent::Handle that would cause
366 a "bio_read: unsupported method" error (in _another_
367 ssl connection) after user code threw an exception.
368 - added AnyEvent::Handle->destroyed method.
369 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
370 - the AnyEvent::Util::fh_nonblocking function worked only
371 by lucky accident on win32.
372 - smaller and faster AnyEvent::Util::fh_nonblocking.
373 - when the (required!) Time::HiRes module is missing, AnyEvent
374 did not fall back to built-in time properly.
375 - do not load Fcntl at runtime, saving memory and loading time.
376 - precompile a number of constants and use them instead of runtime
377 detection and eval.
378 - free detection code after detection and similar memory optimisations.
379 - Perl backend timer interval best effort drift has been improved
380 (same algorithm as EV).
381 - update unicode idna mapping table.
382
3835.24 Tue Jan 5 11:39:43 CET 2010
384 - cygwin never reports errors from failed connects. 1.5 just gives
385 you continous readyness and EAGAIN, 1.7 is even more broken and
386 just hangs. work around both issues in a cygwin-specific hack.
387 - improve idle watcher documentation slightly.
388
3895.23 Sun Dec 20 23:48:00 CET 2009
390 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
391 - implement punycode_encode/decode, idn_nameprep,
392 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
393 - provide $AE::VERSION.
394 - removed traces of "no strict 'refs'".
395
3965.22 Sat Dec 5 03:51:13 CET 2009
397 - downgrade-or-fail in AnyEvent::Handle::push_write, to
398 diagnose encoding failures earlier and more succinctly.
399 (this works around bugs in perl, throwing away encoding info
400 when passing scalar data to extensions).
401 - add more examples to AnyEvent::Socket manpage.
402 - upgrade internal warning set to the same as common::sense 2.03.
403 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
404 (John Beppu).
405 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
406 - add google public dns servers to fallback server set and make sure
407 we load-balance properly between the three sets. also add all
408 fallback dns servers, not just a random one, to each dns config.
409
4105.21 Thu Nov 19 02:48:47 CET 2009
411 - fix a problem where socket constants were called with parameters
412 (spotted by David Friedland).
413 - fork_call never use'd POSIX (reported by Daisuke Maki).
414 - improve perl 5.6 compatibility further (but it still won't work
415 unless you rip out everything but the core).
416 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
417 uses 10MB of RAM, but doesn't open a console window. *sigh*.
418
4195.202 Wed Oct 14 22:35:44 CEST 2009
420 - AnyEvent::DNS would unexpectedly clobber $_ under windows
421 (analysed by Matthias Waldorf).
422 - AnyEvent::Handle::run_cmd can now pass the PID of the
423 newly-created process, which is much less useful than it might
424 sound (based on patch by Yann Kerherve).
425
4265.201 Tue Sep 29 12:09:25 CEST 2009
427 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
428 (reported by Torsten Foertsch).
429 - common::sense 2.0 could cause tcp_server to throw an exception
430 (analysed by elmex).
431
4325.2 Mon Sep 14 07:04:49 CEST 2009
433 - INCOMPATIBLE CHANGE: do no longer support register_read_type
434 and register_write_type in AnyEvent::Handle, instead support
435 package names (the facility was mostly abused).
436 - implement "packagename-as-read/write type" support in
437 AnyEvent::Handle.
438 - AnyEvent::Handle: new options "keepalive" and "oobinline".
439 - oobinline set by default to avoid security issues.
440 - the pure-perl event loop backend wrongly detected times() fallback
441 support (spotted by Pavel Boldin).
442
4435.12 Tue Sep 1 20:26:50 CEST 2009
444 - be more lenient when parsing resolv.conf files, as some people
445 use hashmarks as comment indicator inside directives (reported
446 by Michael S. Fischer).
447 - use same set of warnings as common::sense 2.0.
448 - fix a potential 32 bit overflow issue due to perl having problems
449 with large hex constants in the Perl backend.
450 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
451 by default.
452 - fix some prototypes.
453 - (optionally) require Async::Interrupt 1.02.
454 - replace opendns fallbacks by verizon/level3 public dns servers.
455
4565.112 Fri Aug 21 13:59:04 CEST 2009
457 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
458 using starttls with an illegal/uncreatable tls context.
459 - port to broken win32 perls.
460
4615.111 Sun Aug 16 18:44:36 CEST 2009
462 - perl errornously requires sockaddr_un structures returned by the
463 kernel to have a specific length. work around this bug by padding
464 sockaddr_un structures (found to fail on OS X, as perl only uses
465 the correct code for GNU/Linux, but likely to fail on other
466 OSes as well).
467
4685.11 Wed Aug 12 17:49:37 CEST 2009
469 - fixed a bug in run_cmd with input from filename.
470 - create a json coder object when none is given in write direction
471 to reduce dependency on JSON.
472
4735.1 Tue Aug 11 03:17:32 CEST 2009
474 - tcp_connect now reports the correct error in $!, not an unrelated
475 one (was broken since 4.91).
476 - AE::Handle did not properly clear rtimeout/wtimeout watchers
477 when retrying a tcp_connect.
478 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
479 - fixed wrong documentation in AE::Socket.
480
4815.01 Mon Aug 10 03:16:32 CEST 2009
482 - last release broke EV child watchers :/.
483
4845.0 Sun Aug 9 17:32:09 CEST 2009
485 - officially introduce and document the AE API.
486 - lowering the timeout at runtime did not have immediate effect
487 in AnyEvent::Handle.
488 - AE::Handle now has separate and independent read and write timeout
489 settings.
490 - probe for Guard module when AnyEvent::Util is loaded, not at
491 runtime (Event doesn't like eval "use" from callbacks, and it avoids
492 exporting the probe function).
493 - Impl::IOAsync's io watchers did not use the filehandle emulation
494 code to work around it's watcher limits.
495 - work around IO::Async's problems with overloaded objects.
496 - try to disable wearnings inside IO::Async, it's just too buggy.
497 - do no longer use Event by default, only EV or Perl, Event is too buggy.
498 - hackishly provide fast AE:: interface when EV or Perl is the backend.
499 - inherit VERSION from AnyEvent::VERSION in most modules.
500
5014.91 Thu Aug 6 15:42:45 CEST 2009
502 - AE::Handle::starttls could get out of sync when the read buffer
503 already contains some TLS handshake.
504 - AE::Handle did not properly free the TLS session, causing bigger
505 memory leaks in Net::SSLeay than expected (reported by toaster).
506 - AE::Socket::tcp_connect will now artificially delay invoking
507 the callback to avoid returning after invoking the callback.
508 - convert many internal watcher uses to AE API.
509
5104.9 Sat Aug 1 11:07:01 CEST 2009
511 - Glib silently fails when registering a timeout with
512 a negative value, so avoid that.
513 - call condvar callback immediately when it is set
514 after the condition is already signalled.
515 - check rbuf_max condition only after trying to consume
516 data.
517 - normalise signal numbers to names when using the ae's
518 signal handling, but do not document this (yet).
519 - pure perl signal emulation did not properly set nonblocking
520 mode on the signal pipe on win32 (this is unlikely to have caused
521 any issues).
522 - new module: AnyEvent::Debug.
523 - AnyEvent::Strict now checks that a signal was specified
524 by name (not name or number).
525 - reduce memory footprint in typical cases by ~50kb by
526 only compiling signal code when necessary.
527 - add AnyEvent::Handle->rbuf_max.
528 - grab the AE namespace for future new API, implement stubs for
529 said future/alternative API.
530 - new function: AnyEvent::Socket::format_hostport.
531
5324.881 Tue Jul 28 12:51:53 CEST 2009
533 - work around a bug in local in pre-5.10 perls, causing
534 AnyEvent::Handle to recurse when it should not
535 (analysed by elmex).
536
5374.88 Tue Jul 28 04:04:37 CEST 2009
538 - re-bless the handle into a dummy package after calling
539 AnyEvent::Handle::destroy, so the user does not need to check
540 for errors after every push_write etc.
541 - do not attempt to run t/02_signals.t on obviously broken
542 platforms.
543
5444.87 Sun Jul 26 02:06:16 CEST 2009
545 - do not attempt to linger when there is no longer a valid fh.
546 - enforce tls mode to be either accept or connect, do not simply
547 segfault in Net::SSLeay.
548 - AnyEvent::Handle can now call tcp_connect itself (new parameters
549 connect, on_prepare, on_connect and on_connect_error). Updated
550 tutorial accordingly.
551 - add AnyEvent::Impl::Irssi backend.
552
5534.86 Mon Jul 20 23:52:29 CEST 2009
554 - since the verbose warning is not enough, explicitly document
555 that versions before 1.33 of Net::SSLeay are not secure.
556 - work around signal handling races in Event and (...) Event::Lib.
557 - try to align signal-race timer to full-second boundaries.
558 - work around Tk not liking negative timeouts.
559 - don't complain of different grades of environmental unfriendlyness
560 in IO::Async.
561
5624.85 Sat Jul 18 06:16:14 CEST 2009
563 - nail the signal race problem in perl once and for all
564 (see $AnyEvent::MAX_SIGNAL_LATENCY).
565 - take advantage of Async::Interrupt if it is available.
566 - load Time::HiRes and Guard modules on demand only.
567 - add optional/recommended modules section to AnyEvent
568 documentation.
569 - reduce memory usage considerably (and reduce startup penalty)
570 by not using "strict", "warnings" and "overload" modules.
571 - work around buggy windows/openbsd perls and provide EBADMSG
572 and EPROTO ourselves when missing.
573 - improve perl 5.6 compatibility of the core event loop.
574 - made Net::SSLeay version 1.33 a soft requirement.
575
5764.83 Fri Jul 17 16:56:26 CEST 2009
577 - implement AnyEvent::Socket::getprotobyname.
578 - AnyEvent::CondVar's will now detect recursive blocking
579 waits and will croak, as too many people fall into
580 this trap.
581 - AnyEvent::Handle will now call ->destroy on itself after
582 executing the on_error callback, instead of doing some
583 half-baked internal shutdown, for fatal errors.
584 - clarify on_eof behaviour w.r.t. the read queue and
585 on_read callbacks.
586 - ignore some possible spurious wake-ups in tcp_connect.
587
5884.82 Sat Jul 11 00:34:55 CEST 2009
589 - POE and Event backends didn't accept some callable objects as
590 callbacks.
591 - use Config module instead of POSIX module to detect signal names
592 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
593 module doesn't even have all POSIX signals :/.
594 - use more workarounds around the many refcnt/corruption bugs in
595 Event::Lib.
596 - work around a race condition in perl's select, causing t/03_child.t
597 to rarely fail.
598
5994.81 Thu Jul 9 10:30:30 CEST 2009
600 - AnyEvent::Handle didn't properly diagnose write errors
601 (it expected -1 from syswrite, how lame... :).
602 - support file descriptors in addition to file handles
603 in AnyEvent->io.
604 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
605 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
606 and PERL_ANYEVENT_CA_PATH.
607 - provide a sensible synopsis section for AnyEvent::TLS.
608 - add a "supported backends" section to the manpage.
609 - added simple io watcher test to testsuite, using a
610 portable_socketpair.
611 - tried to improve the stability of the Event::Lib backend,
612 YMMV.
613
6144.8 Mon Jul 6 23:45:16 CEST 2009
615 - AnyEvent::DNS did not properly follow CNAME records with
616 uppercase targets.
617 - AnyEvent::DNS would errornously return AAAA records
618 with v4 mapped addresses (a faulty record) as ipv4 addresses,
619 causing AnyEvent::Socket to throw an exception.
620 - added new module AnyEvent::TLS for easier SSL/TLS context
621 creation, with many options including hostname verification,
622 secure default configuration, lots of documentation and,
623 predefined diffie-hellman keys for perfect forward security
624 and much more. get it while it's still fresh!
625 - use AnyEvent::TLS in AnyEvent::Handle for context management.
626 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
627 so AnyEvent::Socket users have smaller memory footprint
628 in the common case.
629 - add AnyEvent::Handle->push_shutdown method.
630 - add an additional error message parameter to AnyEvent::Handle's
631 on_error callback (for TLS, $! is still available).
632 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
633 - make AnyEvent::Handle more robust against early conenction
634 failures (during new), and return C<undef> in that case
635 from the constructor.
636 - AnyEvent::Handle will now try to load only JSON::XS first,
637 then fall back to JSON.
638 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
639 and because it was documented that way.
640
6414.451 Fri Jul 3 00:28:58 CEST 2009
642 - do not clear rbuf when shutting down an AnyEvent::Handle
643 object - doing so breaks AnyEvent::HTTP.
644
6454.45 Mon Jun 29 22:59:26 CEST 2009
646 - a write error could cause AnyEvent::Handle to create
647 an I/O watcher with an undefined $fh.
648 - special-case mapped ipv4 addresses in both
649 AnyEvent::Socket::format_address and parse_address,
650 to treat them just like ipv4 addresses.
651 - updated and overhauled the AnyEvent::Intro doc.
652 - implement AnyEvent::Socket::format_ipv4/ipv6.
653 - slightly speed up portable_pipe/socketpair.
654 - expand condvar begin/end documentation.
655
6564.42 Fri Jun 26 08:32:18 CEST 2009
657 - preliminary and neccesarily incomplete support for IO::Async.
658 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
659 case it was set to IGNORE, to ensure we can catch child statuses
660 even when the calling env acts stupidly.
661 - updated benchmarks with IO::Async, which performs very well.
662
44.412 Wed Jun 24 01:35:57 CEST 2009 6634.412 Wed Jun 24 01:35:57 CEST 2009
5 - support an "untaint" attribute for AnyEvent::DNS 664 - support an "untaint" attribute for AnyEvent::DNS
6 and set it on the default resolver. 665 and set it on the default resolver.
7 - implement and document AnyEvent::Impl::Perl::loop. 666 - implement and document AnyEvent::Impl::Perl::loop.
8 - remove all anyevent-env variables from %ENV 667 - remove all anyevent-env variables from %ENV
… …
1034.3 Fri Oct 3 09:18:43 CEST 2008 7624.3 Fri Oct 3 09:18:43 CEST 2008
104 - AnyEvent will now install a no-op signal handler for SIGPIPE, 763 - AnyEvent will now install a no-op signal handler for SIGPIPE,
105 unless one has been installed already. 764 unless one has been installed already.
106 - warn about not loaidng AnyEvent::Impl::POE early enough 765 - warn about not loaidng AnyEvent::Impl::POE early enough
107 (patch by Adam Rosenstein). 766 (patch by Adam Rosenstein).
108 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 767 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
109 - fix a bug in where in SSL connect mode, the 768 - fix a bug in where in SSL connect mode, the
110 client would first wait for some data by the server and 769 client would first wait for some data by the server and
111 otherwise hang. (reported and analyzed in an absolutely 770 otherwise hang. (reported and analysed in an absolutely
112 exemplary manner by Adam Rosenstein). 771 exemplary manner by Adam Rosenstein).
113 - fix a bug in where SSL EOF would not be treated 772 - fix a bug in where SSL EOF would not be treated
114 as stream EOF, putting the connection into a hung state. 773 as stream EOF, putting the connection into a hung state.
115 (reported and analyzed in an absolutely exemplary manner by Adam 774 (reported and analysed in an absolutely exemplary manner by Adam
116 Rosenstein). 775 Rosenstein).
117 - fix a potential segfault when the TLS context 776 - fix a potential segfault when the TLS context
118 would go missing in a read callback (e.g. due to stoptls) - 777 would go missing in a read callback (e.g. due to stoptls) -
119 Net::SSLeay of course makes no type checking whatsoever. 778 Net::SSLeay of course makes no type checking whatsoever.
120 - AnyEvent::Handle will not stop the read watcher in TLS mode. 779 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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