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

Comparing AnyEvent/Changes (file contents):
Revision 1.94 by root, Sat May 24 17:48:38 2008 UTC vs.
Revision 1.497 by root, Fri Aug 26 16:18:01 2011 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
34.0 3TODO: document TCP_*** constants
4
5 - log_to_syslog now accepts facility strings.
6 - log_to_syslog would not allow facilities - "$facility|$level"
7 *literally* means that string. sheesh.
8 - the CPAN parser got confused and thought we provide Sys::Syslog, try
9 to work around.
10
116.01 Fri Aug 26 07:04:11 CEST 2011
12 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
13 the TTL as fourth element - this affects ->resolve, ->request
14 and ->dns_unpack indirectly. Convenience functions such as
15 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
16 AnyEvent::Socket.
17 - INCOMPATIBLE CHANGE: backend authors now should not implement
18 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
19 methods.
20 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
21 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
22 - DEPRECIATION: please change
23 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
24 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
25 in your programs, for future compatibility.
26 - *blush* idle watcher emulation was borked.
27 - AnyEvent::Strict now checks the AE::xxx functions as well.
28 - querying too long domain names would barf deep inside AnyEvent::DNS - since
29 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
30 for too long domain names explicitly and passes an undef to the callback
31 instead (reported by James Bromberger).
32 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
33 - new functions: AnyEvent::postpone, AnyEvent::log.
34 - new module: AnyEvent::Log, for simple logging needs.
35 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
36 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
37 a debug shell.
38 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
39 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
40 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
41 F</etc/hosts> can be found.
42 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
43 configuration.
44 - env variables can now be specified using AE_xyz instead of
45 PERL_ANYEVENT_xyz, subject to some rules.
46 - lots of new AnyEvent::Debug functionality.
47 - AnyEvent::Strict now checks for modifications $_ by the callback
48 (by setting it to readonly while executing the callback, if possible).
49 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
50 Invest extra overhead to work around this peculiar behaviour.
51 - split the pure perl backend into a loop module and an impl
52 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
53 - some read types (regex, netstring, json, storable) could
54 errornously keep an AE::Handle object alive on their own
55 (reported by Mohammad Toossi).
56 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
57 supports /etc/hosts overrides.
58 - add more workarounds around child watcher bugs in IO::Async
59 and (newly introduced) in POE.
60 - use glib child watchers and try to work around its limitations,
61 if possible (based on a patch by Kevin Ryde).
62 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
63 - many read types in AnyEvent::Handle would malfunction during
64 transitions to TLS because of a "delete $self->{rbuf}".
65 (reported by Gerald Galster).
66 - improve condvar blocking wait performance (probably only noticable
67 with faster backends), also make it easier for event loops that
68 do not support blocking waits.
69 - PERL_ANYEVENT_MODEL now also allows full module names, although
70 the usefulness of this feature is not yet clear.
71 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
72 - add t/80_ssltext.t.
73 - no longer use AUTOLOAD for the initial loading (for no very good reason).
74 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
75 - do not run event loop detection multiple times when programs call their cached
76 copies of AnyEvent::detect.
77 - parse_hostport now accepts absolute paths als unix domain sockets.
78 - suppress the idiotic warning inside POE by patching the POE::Kernel
79 run flags instead of asking the user to jump through contortions.
80 - add Task::Weaken as dependency (as recommends only at this time).
81 - replace PF_UNSPEC by 0, for good old minix.
82
836.0 (not officially released, but was on CPAN for about 15 minutes :/)
84
855.34 Thu May 12 09:59:41 CEST 2011
86 - AE::TLS didn't support both cert_file and key at the same time,
87 due to some copy&paste bug (reported and analyzed by vti).
88 - improve AE::Handle timeout documentation slightly.
89
905.33 Mon Feb 7 21:51:41 CET 2011
91 - never appeared on CPAN, for unknown reasons.
92 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
93 object when it's already gone (analyzed by Ryan Bullock).
94 - update idna algorithm(s) to UTS#46 version 6.0.0.
95 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
96 uts46 says it's an error, and testsuite makes more sense.
97 - remove leading dots in idn_to_ascii - testsuite says to remove,
98 uts46 says it's an error, and testsuite makes more sense.
99 - fixed a bug in idn ignored character processing.
100 - passes the uts46 6.0.0 testsuite except for three cases which
101 apparently are bugs in the testsuite itself.
102 - passes the uts46 6.0.1 testsuite except for eleven cases which
103 apparently are bugs in the testsuite itself.
104
1055.31 Sun Jan 23 11:49:19 CET 2011
106 - work around a POE bug where POE would first poll for new events and
107 invoke handlers BEFORE handling the message queue.
108 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
109 - stop_read now does just that in TLS mode. see its
110 documentation for caveats.
111 - update to common::sense 3.4 warning set (some warnings in perl impose
112 an arbitrary 32kb limit on regex matches, which affects
113 AnyEvent::Handle).
114 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
115 - new experimental AE::Handle::resettls method.
116
1175.3 Fri Dec 31 05:48:13 CET 2010
118 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
119 issues with IO::Async have been solved in that version, including
120 the availabilty of a default loop!
121 - implement new wbuf_max parameter for AnyEvent::Handle.
122 - added Cocoa::EventLoop backend by Daisuke Murase.
123 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
124 - clarify handle on_prepare documentation.
125 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
126 to avoid leaking them to any child processes.
127 - try to force files into the arch-dependend lib directory, for the
128 half-"turly obsessive compulsive person"(s) that depend on it.
129
1305.29 Sun Dec 5 10:49:21 CET 2010
131 - convert EV backend to EV 4.00 API (so better upgrade EV too).
132 - AE::Handle: implement dynamic read_size adjustment, add
133 max_read_size parameter, reduce default read_size to 2048.
134 - add some FAQ entries for servers and/or on_eof vs. on_error.
135 - work around OS bugs (cygwin again) when lingering in AE::Handle.
136 - reduce memory usage slightly when connect was used in AE::Handle.
137 - wrok around more segfaults in Tk.
138 - document tls_ctx => undef as valid.
139 - detect WNOHANG value at compiletime.
140 - include some probably linux-only support for building AnyEvent
141 as part of the perl core.
142 - improve compatibility of t/08_idna.t to perl 5.8.
143 - make AnyEvent::Strict truly optional (in case somebody builds a
144 minimal perl...).
145
1465.28 Wed Oct 13 04:14:23 CEST 2010
147 - due to a glitch, AnyEvent's internal getprotobyname wasn't
148 actually used in tcp_server.
149 - implement AnyEvent::CondVar->new method, which was documented but
150 missing.
151 - support autoloading of read types in unshift_read.
152 - AnyEvent::Strict: do the test for invalid fh arguments differently,
153 resulting in more informative error messages.
154 - new AnyEvent::FAQ manpage.
155 - clarify that you can actually call push_read etc. in on_read.
156 - put keys and certs in /tmp, not into the current directory,
157 and document this in the SECURITY CONSIDERATIONS sections
158 of AnyEvent::TLS.
159 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
160 requested by Ron Isaacson.
161 - document $AnyEvent::DNS::RESOLVER variable.
162 - incorporated major typo patches by Abhijit Menon-Sen.
163 - AnyEvent::Handle now croaks when negative timeouts are passed.
164 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
165 undocumented, but who knows when they might come in handy.
166
1675.271 Tue Jun 8 12:05:46 CEST 2010
168 - backport to perl 5.8.x.
169
1705.27 Sun Jun 6 12:12:05 CEST 2010
171 - postpone differently in AnyEvent::Socket now, as
172 when not, canceling the connection attempt might fail
173 (found by Felix Antonius Wilhelm Ostmann).
174 - explicitly check for non-stream sockets in AE::Handle, too many
175 clueless people fell into the trap of this somehow working.
176 - simplified and reworked the "OTHER MODULES" section.
177 - better/more condvar examples.
178
1795.261 Wed Apr 28 16:13:36 CEST 2010
180 - AF_INET6 was not properly used from Socket6 during configuration
181 time (found by Andy Grundman).
182
1835.26 Mon Apr 12 04:49:35 CEST 2010
184 - don't generate spurious readyness notifications when select
185 returns EINTR in the pure perl backend (analysed and testcase
186 by Bas Denissen).
187 - give same examples in SYNOPSIS sections of AnyEvent and AE.
188 - provider faster implementations for AE::time/now/now_update for
189 EV and Perl backends.
190 - provide faster AE::cv and AE::time implementations for all
191 backends.
192 - no longer support lower/mixed-case signal names (this was
193 never documented nor universally supported).
194 - some more memory µ-optimisations, and somewhat less messy
195 code to cope with both AnyEvent and AE APIs.
196
1975.251 Sat Mar 13 00:58:21 CET 2010
198 - make test failed if it was run before make install on perl <=
199 5.8.8, and older versions of perl make our live very difficult,
200 so write constants.pl during Makefile.PL time.
201
2025.25 Sat Mar 13 00:23:14 CET 2010
203 - fix a race condition in AnyEvent::Handle that would cause
204 a "bio_read: unsupported method" error (in _another_
205 ssl connection) after user code threw an exception.
206 - added AnyEvent::Handle->destroyed method.
207 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
208 - the AnyEvent::Util::fh_nonblocking function worked only
209 by lucky accident on win32.
210 - smaller and faster AnyEvent::Util::fh_nonblocking.
211 - when the (required!) Time::HiRes module is missing, AnyEvent
212 did not fall back to built-in time properly.
213 - do not load Fcntl at runtime, saving memory and loading time.
214 - precompile a number of constants and use them instead of runtime
215 detection and eval.
216 - free detection code after detection and similar memory optimisations.
217 - Perl backend timer interval best effort drift has been improved
218 (same algorithm as EV).
219 - update unicode idna mapping table.
220
2215.24 Tue Jan 5 11:39:43 CET 2010
222 - cygwin never reports errors from failed connects. 1.5 just gives
223 you continous readyness and EAGAIN, 1.7 is even more broken and
224 just hangs. work around both issues in a cygwin-specific hack.
225 - improve idle watcher documentation slightly.
226
2275.23 Sun Dec 20 23:48:00 CET 2009
228 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
229 - implement punycode_encode/decode, idn_nameprep,
230 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
231 - provide $AE::VERSION.
232 - removed traces of "no strict 'refs'".
233
2345.22 Sat Dec 5 03:51:13 CET 2009
235 - downgrade-or-fail in AnyEvent::Handle::push_write, to
236 diagnose encoding failures earlier and more succinctly.
237 (this works around bugs in perl, throwing away encoding info
238 when passing scalar data to extensions).
239 - add more examples to AnyEvent::Socket manpage.
240 - upgrade internal warning set to the same as common::sense 2.03.
241 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
242 (John Beppu).
243 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
244 - add google public dns servers to fallback server set and make sure
245 we load-balance properly between the three sets. also add all
246 fallback dns servers, not just a random one, to each dns config.
247
2485.21 Thu Nov 19 02:48:47 CET 2009
249 - fix a problem where socket constants were called with parameters
250 (spotted by David Friedland).
251 - fork_call never use'd POSIX (reported by Daisuke Maki).
252 - improve perl 5.6 compatibility further (but it still won't work
253 unless you rip out everything but the core).
254 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
255 uses 10MB of RAM, but doesn't open a console window. *sigh*.
256
2575.202 Wed Oct 14 22:35:44 CEST 2009
258 - AnyEvent::DNS would unexpectedly clobber $_ under windows
259 (analysed by Matthias Waldorf).
260 - AnyEvent::Handle::run_cmd can now pass the PID of the
261 newly-created process, which is much less useful than it might
262 sound (based on patch by Yann Kerherve).
263
2645.201 Tue Sep 29 12:09:25 CEST 2009
265 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
266 (reported by Torsten Foertsch).
267 - common::sense 2.0 could cause tcp_server to throw an exception
268 (analysed by elmex).
269
2705.2 Mon Sep 14 07:04:49 CEST 2009
271 - INCOMPATIBLE CHANGE: do no longer support register_read_type
272 and register_write_type in AnyEvent::Handle, instead support
273 package names (the facility was mostly abused).
274 - implement "packagename-as-read/write type" support in
275 AnyEvent::Handle.
276 - AnyEvent::Handle: new options "keepalive" and "oobinline".
277 - oobinline set by default to avoid security issues.
278 - the pure-perl event loop backend wrongly detected times() fallback
279 support (spotted by Pavel Boldin).
280
2815.12 Tue Sep 1 20:26:50 CEST 2009
282 - be more lenient when parsing resolv.conf files, as some people
283 use hashmarks as comment indicator inside directives (reported
284 by Michael S. Fischer).
285 - use same set of warnings as common::sense 2.0.
286 - fix a potential 32 bit overflow issue due to perl having problems
287 with large hex constants in the Perl backend.
288 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
289 by default.
290 - fix some prototypes.
291 - (optionally) require Async::Interrupt 1.02.
292 - replace opendns fallbacks by verizon/level3 public dns servers.
293
2945.112 Fri Aug 21 13:59:04 CEST 2009
295 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
296 using starttls with an illegal/uncreatable tls context.
297 - port to broken win32 perls.
298
2995.111 Sun Aug 16 18:44:36 CEST 2009
300 - perl errornously requires sockaddr_un structures returned by the
301 kernel to have a specific length. work around this bug by padding
302 sockaddr_un structures (found to fail on OS X, as perl only uses
303 the correct code for GNU/Linux, but likely to fail on other
304 OSes as well).
305
3065.11 Wed Aug 12 17:49:37 CEST 2009
307 - fixed a bug in run_cmd with input from filename.
308 - create a json coder object when none is given in write direction
309 to reduce dependency on JSON.
310
3115.1 Tue Aug 11 03:17:32 CEST 2009
312 - tcp_connect now reports the correct error in $!, not an unrelated
313 one (was broken since 4.91).
314 - AE::Handle did not properly clear rtimeout/wtimeout watchers
315 when retrying a tcp_connect.
316 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
317 - fixed wrong documentation in AE::Socket.
318
3195.01 Mon Aug 10 03:16:32 CEST 2009
320 - last release broke EV child watchers :/.
321
3225.0 Sun Aug 9 17:32:09 CEST 2009
323 - officially introduce and document the AE API.
324 - lowering the timeout at runtime did not have immediate effect
325 in AnyEvent::Handle.
326 - AE::Handle now has separate and independent read and write timeout
327 settings.
328 - probe for Guard module when AnyEvent::Util is loaded, not at
329 runtime (Event doesn't like eval "use" from callbacks, and it avoids
330 exporting the probe function).
331 - Impl::IOAsync's io watchers did not use the filehandle emulation
332 code to work around it's watcher limits.
333 - work around IO::Async's problems with overloaded objects.
334 - try to disable wearnings inside IO::Async, it's just too buggy.
335 - do no longer use Event by default, only EV or Perl, Event is too buggy.
336 - hackishly provide fast AE:: interface when EV or Perl is the backend.
337 - inherit VERSION from AnyEvent::VERSION in most modules.
338
3394.91 Thu Aug 6 15:42:45 CEST 2009
340 - AE::Handle::starttls could get out of sync when the read buffer
341 already contains some TLS handshake.
342 - AE::Handle did not properly free the TLS session, causing bigger
343 memory leaks in Net::SSLeay than expected (reported by toaster).
344 - AE::Socket::tcp_connect will now artificially delay invoking
345 the callback to avoid returning after invoking the callback.
346 - convert many internal watcher uses to AE API.
347
3484.9 Sat Aug 1 11:07:01 CEST 2009
349 - Glib silently fails when registering a timeout with
350 a negative value, so avoid that.
351 - call condvar callback immediately when it is set
352 after the condition is already signalled.
353 - check rbuf_max condition only after trying to consume
354 data.
355 - normalise signal numbers to names when using the ae's
356 signal handling, but do not document this (yet).
357 - pure perl signal emulation did not properly set nonblocking
358 mode on the signal pipe on win32 (this is unlikely to have caused
359 any issues).
360 - new module: AnyEvent::Debug.
361 - AnyEvent::Strict now checks that a signal was specified
362 by name (not name or number).
363 - reduce memory footprint in typical cases by ~50kb by
364 only compiling signal code when necessary.
365 - add AnyEvent::Handle->rbuf_max.
366 - grab the AE namespace for future new API, implement stubs for
367 said future/alternative API.
368 - new function: AnyEvent::Socket::format_hostport.
369
3704.881 Tue Jul 28 12:51:53 CEST 2009
371 - work around a bug in local in pre-5.10 perls, causing
372 AnyEvent::Handle to recurse when it should not
373 (analysed by elmex).
374
3754.88 Tue Jul 28 04:04:37 CEST 2009
376 - re-bless the handle into a dummy package after calling
377 AnyEvent::Handle::destroy, so the user does not need to check
378 for errors after every push_write etc.
379 - do not attempt to run t/02_signals.t on obviously broken
380 platforms.
381
3824.87 Sun Jul 26 02:06:16 CEST 2009
383 - do not attempt to linger when there is no longer a valid fh.
384 - enforce tls mode to be either accept or connect, do not simply
385 segfault in Net::SSLeay.
386 - AnyEvent::Handle can now call tcp_connect itself (new parameters
387 connect, on_prepare, on_connect and on_connect_error). Updated
388 tutorial accordingly.
389 - add AnyEvent::Impl::Irssi backend.
390
3914.86 Mon Jul 20 23:52:29 CEST 2009
392 - since the verbose warning is not enough, explicitly document
393 that versions before 1.33 of Net::SSLeay are not secure.
394 - work around signal handling races in Event and (...) Event::Lib.
395 - try to align signal-race timer to full-second boundaries.
396 - work around Tk not liking negative timeouts.
397 - don't complain of different grades of environmental unfriendlyness
398 in IO::Async.
399
4004.85 Sat Jul 18 06:16:14 CEST 2009
401 - nail the signal race problem in perl once and for all
402 (see $AnyEvent::MAX_SIGNAL_LATENCY).
403 - take advantage of Async::Interrupt if it is available.
404 - load Time::HiRes and Guard modules on demand only.
405 - add optional/recommended modules section to AnyEvent
406 documentation.
407 - reduce memory usage considerably (and reduce startup penalty)
408 by not using "strict", "warnings" and "overload" modules.
409 - work around buggy windows/openbsd perls and provide EBADMSG
410 and EPROTO ourselves when missing.
411 - improve perl 5.6 compatibility of the core event loop.
412 - made Net::SSLeay version 1.33 a soft requirement.
413
4144.83 Fri Jul 17 16:56:26 CEST 2009
415 - implement AnyEvent::Socket::getprotobyname.
416 - AnyEvent::CondVar's will now detect recursive blocking
417 waits and will croak, as too many people fall into
418 this trap.
419 - AnyEvent::Handle will now call ->destroy on itself after
420 executing the on_error callback, instead of doing some
421 half-baked internal shutdown, for fatal errors.
422 - clarify on_eof behaviour w.r.t. the read queue and
423 on_read callbacks.
424 - ignore some possible spurious wake-ups in tcp_connect.
425
4264.82 Sat Jul 11 00:34:55 CEST 2009
427 - POE and Event backends didn't accept some callable objects as
428 callbacks.
429 - use Config module instead of POSIX module to detect signal names
430 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
431 module doesn't even have all POSIX signals :/.
432 - use more workarounds around the many refcnt/corruption bugs in
433 Event::Lib.
434 - work around a race condition in perl's select, causing t/03_child.t
435 to rarely fail.
436
4374.81 Thu Jul 9 10:30:30 CEST 2009
438 - AnyEvent::Handle didn't properly diagnose write errors
439 (it expected -1 from syswrite, how lame... :).
440 - support file descriptors in addition to file handles
441 in AnyEvent->io.
442 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
443 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
444 and PERL_ANYEVENT_CA_PATH.
445 - provide a sensible synopsis section for AnyEvent::TLS.
446 - add a "supported backends" section to the manpage.
447 - added simple io watcher test to testsuite, using a
448 portable_socketpair.
449 - tried to improve the stability of the Event::Lib backend,
450 YMMV.
451
4524.8 Mon Jul 6 23:45:16 CEST 2009
453 - AnyEvent::DNS did not properly follow CNAME records with
454 uppercase targets.
455 - AnyEvent::DNS would errornously return AAAA records
456 with v4 mapped addresses (a faulty record) as ipv4 addresses,
457 causing AnyEvent::Socket to throw an exception.
458 - added new module AnyEvent::TLS for easier SSL/TLS context
459 creation, with many options including hostname verification,
460 secure default configuration, lots of documentation and,
461 predefined diffie-hellman keys for perfect forward security
462 and much more. get it while it's still fresh!
463 - use AnyEvent::TLS in AnyEvent::Handle for context management.
464 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
465 so AnyEvent::Socket users have smaller memory footprint
466 in the common case.
467 - add AnyEvent::Handle->push_shutdown method.
468 - add an additional error message parameter to AnyEvent::Handle's
469 on_error callback (for TLS, $! is still available).
470 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
471 - make AnyEvent::Handle more robust against early conenction
472 failures (during new), and return C<undef> in that case
473 from the constructor.
474 - AnyEvent::Handle will now try to load only JSON::XS first,
475 then fall back to JSON.
476 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
477 and because it was documented that way.
478
4794.451 Fri Jul 3 00:28:58 CEST 2009
480 - do not clear rbuf when shutting down an AnyEvent::Handle
481 object - doing so breaks AnyEvent::HTTP.
482
4834.45 Mon Jun 29 22:59:26 CEST 2009
484 - a write error could cause AnyEvent::Handle to create
485 an I/O watcher with an undefined $fh.
486 - special-case mapped ipv4 addresses in both
487 AnyEvent::Socket::format_address and parse_address,
488 to treat them just like ipv4 addresses.
489 - updated and overhauled the AnyEvent::Intro doc.
490 - implement AnyEvent::Socket::format_ipv4/ipv6.
491 - slightly speed up portable_pipe/socketpair.
492 - expand condvar begin/end documentation.
493
4944.42 Fri Jun 26 08:32:18 CEST 2009
495 - preliminary and neccesarily incomplete support for IO::Async.
496 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
497 case it was set to IGNORE, to ensure we can catch child statuses
498 even when the calling env acts stupidly.
499 - updated benchmarks with IO::Async, which performs very well.
500
5014.412 Wed Jun 24 01:35:57 CEST 2009
502 - support an "untaint" attribute for AnyEvent::DNS
503 and set it on the default resolver.
504 - implement and document AnyEvent::Impl::Perl::loop.
505 - remove all anyevent-env variables from %ENV
506 when running in tainted mode.
507 - mention and extend the IO::Lambda benchmark.
508
5094.411 Sun Jun 7 18:48:02 CEST 2009
510 - do not try to use F_SETFD on windows.
511
5124.41 Thu May 14 06:40:11 CEST 2009
513 - work around issues in older perls (5.8.0?) when
514 a signal handler is deleted from the %SIG hash.
515 - use POSIX::_exit in child test, to avoid running
516 destructors.
517 - speed up CHLD handling by relying on SIGCHLD being
518 synchronously delivered, even when we roll our own
519 implementation.
520 - AnyEvent::DNS: add the "dname" resource record name
521 for cosmetic reasons.
522
5234.4 Sun Apr 26 20:12:33 CEST 2009
524 - implemented idle watchers, where applicable.
525 - AnyEvent->time died when Event backend was in use.
526 - fix a memleak in the Tk backend.
527 - sped up Tk timer handling.
528 - clip negative "after" values to 0 in AnyEvent::Impl::Event
529 to avoid spamming out warning messages.
530 - fix Qt timers without interval.
531 - avoid Qt zero-timeout special case, in old libqt's.
532
5334.352 Mon Apr 20 16:31:11 CEST 2009
534 - fix AnyEvent::Strict error messages for child watchers.
535 - fix/update Makefile.PL "recommends" versions.
536 - add AnyEvent->now_update.
537
5384.351 Sat Apr 11 07:56:14 CEST 2009
539 - actually make the signal pipe work under win32.
540 - localise $! in signal handler to avoid changing $!, although
541 perl itself does not seem to save/restore errno either.
542 - set the cloexec flag on the signal pipe (normally set by perl too).
543
5444.35 Fri Mar 27 11:48:20 CET 2009
545 - event models relying on AnyEvent's signal watcher emulation did
546 invoke the callback asynchronously, contradicting documentation
547 and causing signals to get lost (this includes AnyEvent's own
548 event loop). AnyEvent now uses the standard pipe trick to make
549 callback execution synchronous to the event loop.
550 - AnyEvent::Handle didn't free TLS context data on DESTROY
551 (patch by Pavel Shaydo).
552 - work around the ever-incompatibly-changing API of MakeMaker.
553 - document that changing global variables without restoring them
554 is a bad idea in Perl (noted by Adam Rosenstein).
555 - AnyEvent::Strict now barfs if ->io is passed a file.
556
5574.34 Thu Feb 12 18:32:45 CET 2009
558 - separately buffer TLS read data, as otherwise the read queue could
559 deadlock as receiving data is not expected while draining the
560 read queue (which cna only happen with TLS).
561 - raise EBADMSG error on JSON decoding errors.
562 - fix some minor manpage bugs (reported by Maximilian Gaß).
563 - speed up select bitmask parsing quite a bit in the pure perl
564 backend.
565 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
566
5674.331 Tue Jan 6 21:07:25 CET 2009
568 - socketpair fails on many vista machines because vista has
569 completely broken accept/getpeername and getsockname functions,
570 so we provide our own socketpair emulation that kind of works
571 (AnyEvent::Util::portable_pipe).
572 - new function: AnyEvent::Util::portable_socketpair.
573 - take advantage of the Guard module if it exists.
574
5754.33 Fri Nov 21 02:35:40 CET 2008
576 - AnyEvent::Strict did errornously flag a fileno of 0 as
577 illegal.
578 - reduce memory usage and slightly speed up the pure perl backend
579 by only storing the file descriptor, not the file handle.
580 - add missing autocork method to AnyEvent::Handle (reported
581 by Adam Rosenstein).
582 - AnyEvent::DNS->resolve errornously documented an $rcode result
583 argument, but there isn't (spotted by Henrik Krohns).
584 - the naked truth about Tk - it's basically unmaintained.
585
5864.32 Mon Nov 3 22:46:32 CET 2008
587 - fix AnyEvent::Socket::resolve_sockaddr to properly support
588 unix sockets again and choose a proper default for
589 the socket type.
590 - fix call to resolve_sockaddr in tcp_connect.
591
5924.31 Thu Oct 30 04:41:48 CET 2008
593 - implemented AnyEvent::Handle->destroy method.
594 - hint about unexpected effects in TLS mode.
595 - speed up AnyEvent::Impl::Perl by using more arrays
596 instead of hashes without the slightest loss of readability :)
597 - work around a perl argument refcounting bug.
598 - with some perl patches applied, AnyEvent now seems leak-free.
599
6004.3 Fri Oct 3 09:18:43 CEST 2008
601 - AnyEvent will now install a no-op signal handler for SIGPIPE,
602 unless one has been installed already.
603 - warn about not loaidng AnyEvent::Impl::POE early enough
604 (patch by Adam Rosenstein).
605 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
606 - fix a bug in where in SSL connect mode, the
607 client would first wait for some data by the server and
608 otherwise hang. (reported and analysed in an absolutely
609 exemplary manner by Adam Rosenstein).
610 - fix a bug in where SSL EOF would not be treated
611 as stream EOF, putting the connection into a hung state.
612 (reported and analysed in an absolutely exemplary manner by Adam
613 Rosenstein).
614 - fix a potential segfault when the TLS context
615 would go missing in a read callback (e.g. due to stoptls) -
616 Net::SSLeay of course makes no type checking whatsoever.
617 - AnyEvent::Handle will not stop the read watcher in TLS mode.
618 - AnyEvent::Handle->stoptls will send a TLS close notify
619 instead of simply destroying the stream now.
620 - fix a bug where large blocks of data written in TLS mode
621 would not be sent unless triggered by receives.
622 - on_drain will now take the tls write buffer into account.
623 - SSL operations have been streamlined a bit and should be faster
624 now, more to come, though.
625 - added a FAQ section, will document nontrivial issues in there.
626
6274.234 Mon Sep 29 04:08:13 CEST 2008
628 - fix child watcher documentation: only child exits will be reported,
629 no trace events.
630 - mention SIGPIPE in AnyEvent::Handle.
631 - perl backend will now use POSIX::times as monotonic clock
632 when available and the monotonic clock option is not.
633 - run even without Time::HiRes in some cases now, but do not
634 provide sub-second accuracy (all *supported* perl versions have
635 Time::HiRes).
636
6374.233 Fri Aug 22 01:48:05 CEST 2008
638 - fix a bug in the testsuite, causing a hang.
639 - clarified lots of AnyEvent::Handle constructor arguments.
640
6414.232 Thu Aug 21 20:44:25 CEST 2008
642 - fix a bug in AnyEvent::Handle that could cause two fatal error
643 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
644 (testcase by Leon Brocard).
645 - clarified on_eof/on_error documentation a bit.
646
6474.231 Tue Jul 29 13:12:15 CEST 2008
648 - remove some debugging code left in AnyEvent::Util::fork_call
649 (and no, it's impossible to implement with the broken windows
650 perls without resource leaks or worse).
651
6524.23 Tue Jul 29 12:19:59 CEST 2008
653 - document the first parameter passed to condvar callbacks to be
654 the callback.
655 - add AnyEvent::Socket::{ntoa,aton} aliases.
656 - optimize the AE::Handle->push_read (line) for the default
657 eol marker.
658 - optimize push_read (packstring|storable) for small packets.
659 - invoke on_error callback when no on_eof callback is set.
660 - fix a bug in push_read (storable) of unknown impact.
661
6624.22 Sun Jul 20 16:34:13 CEST 2008
663 - new function AnyEvent::Socket::parse_hostport.
664 - as the bulkheads at microsoft can't even get getprotobyname reliably
665 working on their shitty fucking broken os we need to hardcode
666 some common protocol numbers in AnyEvent::Socket. How can
667 people even bother with such a pile of shit as windows.
668
6694.21 Thu Jul 17 14:40:05 CEST 2008
670 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
671 which is incompatible to previous invocation syntax.
672 - work around more windows perl bugs in fork_call. I cannot imagine
673 why anybody would bother implementing fork in such an obviously
674 unusable way.
675 - avoid calling $do_retry if the request has already finished in
676 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
677 K. Sharifulin).
678
6794.2 Sat Jul 12 22:42:11 CEST 2008
680 - support an "interval" argument for repeating timers.
681 - fix a bug in the pure perl event loop that caused timers to
682 be delayed under some circumstances.
683 - implement optional argument checking via PERL_ANYEVENT_STRICT.
684 - not importing anything from Event causes it to malfunction
685 and not use Time::HiRes, work around this bug.
686 - more bug workarounds against the endlessly broken Tk module,
687 seems to work now on 32 bit machines, but less so on 64 bit.
688 - minor optimisations applied to most backends.
689
6904.161 Fri Jul 4 14:24:48 CEST 2008
691 - fixed the t/handle/02_write.t, it read after write, with the
692 assumption that all data could be written before read
693 (reported by Jim Razmus).
694
6954.160 Thu Jul 3 04:02:21 CEST 2008
696 - re-registering signal watchers after unregistering one did
697 not work if a backend falls back on AnyEvent's default
698 implementation (also affected child watchers).
699 - new AnyEvent::Handle options: autocork and no_delay.
700
7014.152 Sun Jun 22 14:15:44 CEST 2008
702 - allow for 32-bit perls that implement shifts differently
703 on different architectures in parse_ipv4 (reported and
704 analysed by Keiichi DAIBA).
705
7064.151 Fri Jun 6 17:34:24 CEST 2008
707 - make sure specifying _only_ on_read and never pushing reads
708 works.
709
7104.15 Fri Jun 6 13:00:46 CEST 2008
711 - the pure perl backend would keep some watchers alive when more than
712 one watcher was registered for the same fd.
713 - new "packstring" and "storable" read and write types
714 for AnyEvent::Handle.
715 - allow on_eof handler to be called after on_error with EPIPE returns.
716 - do not immediately call on_read callback in handle constructor.
717
7184.14 Thu Jun 5 20:29:31 CEST 2008
719 - Fixed a bug in DNS SRV priority sorting.
720 - AnyEvent::Util::guard now reports runtime errors while
721 executing the guard block as warnings.
722 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
723
7244.13 Thu Jun 5 00:47:59 CEST 2008
725 - AnyEvent::DNS only followed cname chains with length 2,
726 contrary to documentation. bumped it up to 10, thanks to
727 microsoft, the current broken-dns-config-king.
728 - AnyEvent::DNS didn't check for socket return status
729 but instead relied on perl not creating filehandles in that
730 case - too bad it gives you a bogus file handle (reported
731 and analysed by Vladimir Timofeev).
732 - fix queue management logic in AnyEvent::Handle:
733 when on_read was registered, the queue was empty
734 and no progress could be made AnyEvent::Handle would
735 enter an endless loop.
736 - correctly start reading again when the handle became
737 busy again after idling.
738 - correctly treat tls shut-downs as EOF condition
739 (for the time being, would be nice to have a callback for that).
740 - correctly call eof callback under all conditions
741 (hopefully).
742 - the timeout callback did not expect that $self can go
743 away any time.
744
7454.12 Tue Jun 3 10:58:04 CEST 2008
746 - include AnyEvent::Intro, a tutorial for anyevent,
747 anyevent::socket and anyevent::handle.
748 - allow more options in on_error.
749
7504.11 Fri May 30 23:42:25 CEST 2008
751 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
752 reverse_lookup and reverse_verify to replace it, support
753 v4mapped and v4compat addresses.
754 - provide more documentation for the resolver class.
755 - really replace longest run of :0: by :: in format_address,
756 also properly convert :: and ::1 again.
757 - support NAPTR record name and decode it.
758 - implement random weight sampling for SRV records, as per
759 rfc 2782.
760 - correctly abort on srv-record targets of ".".
761 - added AnyEvent::DNS::wait_for_slot.
762 - in the unlikely event of a virtual circuit connection
763 being invalidated by a delayed udp reply, AnyEvent::DNS
764 could die. this has been fixed.
765 - plug a probable memleak in the DNS vc code.
766 - use configured timeout also for VC request phase.
767 - implement timeout and max_outstanding methods
768 for resolver class.
769 - update version numbers in all modules.
770
7714.1 Thu May 29 05:45:40 CEST 2008
772 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
773 which mirrors their purpose better (old names still available
774 for a while). Also moved AnyEvent::DNS::addr to
775 AnyEvent::Socket::resolve_sockaddr.
776 - implement AnyEvent->time and AnyEvent->now.
777 - fix IPv6 support in pack_sockaddr.
778 - officially un-experimentalise all newly introduced code.
779 - support unix domain sockets everywhere by specifying
780 a host of "unix/" and the pathname as service.
781 - implement an activity timeout in AnyEvent::Handle.
782 - added a regex read type to AnyEvent::Handle.
783 - added a json read/write type to AnyEvent::Handle.
784 - always croak in on_error in AnyEvent::Handle.
785 - document how to subclass AnyEvent::Handle.
786 - implement AnyEvent::Util::fork_call.
787 - add support for IPv6 nameservers and nameserver statements.
788 - work around _yet_ _another_ windows perl bug in where empty
789 select masks cause errors in the pure perl backend (this
790 workaround cannot be implemented for the other event loops,
791 but EV also works around this).
792 - supply AnyEvent::Socket::address_type and make good use of it.
793 - clarify and fix the Handle documentation w.r.t $self vs. $handle
794 vs. callback arguments.
795 - add some recommends to the META.yml.
796
7974.05 Mon May 26 19:44:06 CEST 2008
798 - some platforms ignore the rfc and prepend an extra sa_len member to
799 the sockaddr structure, cater for those. also use sockaddr_family,
800 didn't know it was there, but comes in handy for the workaround.
801 - undo the import WIN32 hack.
802
8034.04 Mon May 26 08:03:31 CEST 2008
804 - try to work around yet another windows bug: failed connects are reported
805 as if it were out-of-band data. windows users: you suck. Only
806 EV, Glib, Event and the pure perl backend can handle this condition.
807 - optimize write algorithm in AnyEvent::Handle.
808 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
809 - tcp_connect now properly iterates through all targets.
810 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
811 support entirely. also hardcode some AF_INET6 constants for the
812 sake of perl 5.8.
813 - vastly improved nameserver/suffix detection algorithm for windows
814 (still a hack...)
815 - try to use a fallback dns server if no dns servers could be found.
816 - splatter around some local $SIG{__DIE__}'s to work around bad
817 die handlers.
818
8194.03 Sun May 25 05:05:57 CEST 2008
820 - work around perl bugs on windows, where perl returns undocumented
821 error codes for sysread, syswrite etc.
822 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
823 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
824 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
825 - updated documentation.
826 - lots of workarounds for perl 5.8 quirks.
827 - implement a more flexible type system for both reads and writes
828 in AnyEvent::Handle.
829 - support IPv6 in tcp_server.
830 - pass local hostname and port to the prepare callback in tcp_server.
831 - make types extendable in AnyEvent::Handle.
832 - croak, not die, in AnyEvent::Handle;
833
8344.0 Sat May 24 19:58:08 CEST 2008
4 - added (experimental) AnyEvent::DNS module. 835 - added (experimental) AnyEvent::DNS module.
5 - added (experimental) AnyEvent::Socket module. 836 - added (experimental) AnyEvent::Socket module.
6 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle. 837 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
7 - make condvar->end call ->send as documented. 838 - make condvar->end call ->send as documented.
8 - make condvar constructor support the documented cb => argument. 839 - make condvar constructor support the documented cb => argument.
… …
12 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. 843 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
13 - try very hard not to freeze on broken windows perls. 844 - try very hard not to freeze on broken windows perls.
14 - make on_eof optional in AnyEvent::Handle. 845 - make on_eof optional in AnyEvent::Handle.
15 - support service names in tcp_connect and tcp_server. 846 - support service names in tcp_connect and tcp_server.
16 - work around netbsd/openbsd bug causing unix sockets not to 847 - work around netbsd/openbsd bug causing unix sockets not to
17 become writable untill empty. 848 become writable until empty.
18 - only make a dummy read for error detection when getpeername 849 - only make a dummy read for error detection when getpeername
19 returns ENOTCONN. 850 returns ENOTCONN.
20 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 851 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
21 override. 852 override.
22 - new AnyEvent::Guard::cancel. 853 - new AnyEvent::Guard::cancel.
23 854
243.5 Sat May 17 23:17:14 CEST 2008 8553.5 Sat May 17 23:17:14 CEST 2008
25 - gracefully deal with systems not defining CLOCK_MONOTONIC 856 - gracefully deal with systems not defining CLOCK_MONOTONIC
… …
39 870
403.41 Sun May 11 19:53:13 CEST 2008 8713.41 Sun May 11 19:53:13 CEST 2008
41 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 872 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
42 873
433.4 Sun May 11 00:29:25 CEST 2008 8743.4 Sun May 11 00:29:25 CEST 2008
44 - complete re-implementation of condvars with a hopefully much mroe useful 875 - complete re-implementation of condvars with a hopefully much more useful
45 API (backward compatible functions exist but are undocumented). 876 API (backward compatible functions exist but are undocumented).
46 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 877 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
47 (still subject to change). 878 (still subject to change).
48 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 879 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
49 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 880 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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