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

Comparing AnyEvent/Changes (file contents):
Revision 1.143 by root, Mon Jun 2 11:03:40 2008 UTC vs.
Revision 1.450 by root, Sun Jul 24 13:10:43 2011 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3 - include AnyEvent::Intro as a beginner tutorial. 3TODO: document TCP_*** constants
4TODO: catch AE::xxx for anyevent strict and anyevent debug
5TODO: anyevent::trace? or debug tracing all watchers?
6TODO: anyevent strict checking for $_ changes?
7
8TODO: Kevin Ryde (3.8K) 1 inbox Glib child processes
9TODO: keeps alive, self-reference, storabel et al.
10 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
11 the TTL as fourth element - this affects ->resolve directly,
12 and ->request and ->dns_unpack indirectly. Convenience functions
13 such as AnyEvent::DNS::{a,txt,mx,srv} are not affected.
14 - some read types (regex, netstring, json, storable) could
15 errornously keep an AE::Handle object alive on their own
16 (reported by Mohammad Toossi).
17 - add support for (currently third-party) AnyEvent::Impl::FLTK.
18 - many read types in AnyEvent::Handle would malfunction during
19 transitions to TLS because of a "delete $self->{rbuf}".
20 (reported by Gerald Galster).
21
225.34 Thu May 12 09:59:41 CEST 2011
23 - AE::TLS didn't support both cert_file and key at the same time,
24 due to some copy&paste bug (reported and analyzed by vti).
25 - improve AE::Handle timeout documentation slightly.
26
275.33 Mon Feb 7 21:51:41 CET 2011
28 - never appeared on CPAN, for unknown reasons.
29 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
30 object when it's already gone (analyzed by Ryan Bullock).
31 - update idna algorithm(s) to UTS#46 version 6.0.0.
32 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
33 uts46 says it's an error, and testsuite makes more sense.
34 - remove leading dots in idn_to_ascii - testsuite says to remove,
35 uts46 says it's an error, and testsuite makes more sense.
36 - fixed a bug in idn ignored character processing.
37 - passes the uts46 6.0.0 testsuite except for three cases which
38 apparently are bugs in the testsuite itself.
39 - passes the uts46 6.0.1 testsuite except for eleven cases which
40 apparently are bugs in the testsuite itself.
41
425.31 Sun Jan 23 11:49:19 CET 2011
43 - work around a POE bug where POE would first poll for new events and
44 invoke handlers BEFORE handling the message queue.
45 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
46 - stop_read now does just that in TLS mode. see its
47 documentation for caveats.
48 - update to common::sense 3.4 warning set (some warnings in perl impose
49 an arbitrary 32kb limit on regex matches, which affects
50 AnyEvent::Handle).
51 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
52 - new experimental AE::Handle::resettls method.
53
545.3 Fri Dec 31 05:48:13 CET 2010
55 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
56 issues with IO::Async have been solved in that version, including
57 the availabilty of a default loop!
58 - implement new wbuf_max parameter for AnyEvent::Handle.
59 - added Cocoa::EventLoop backend by Daisuke Murase.
60 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
61 - clarify handle on_prepare documentation.
62 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
63 to avoid leaking them to any child processes.
64 - try to force files into the arch-dependend lib directory, for the
65 half-"turly obsessive compulsive person"(s) that depend on it.
66
675.29 Sun Dec 5 10:49:21 CET 2010
68 - convert EV backend to EV 4.00 API (so better upgrade EV too).
69 - AE::Handle: implement dynamic read_size adjustment, add
70 max_read_size parameter, reduce default read_size to 2048.
71 - add some FAQ entries for servers and/or on_eof vs. on_error.
72 - work around OS bugs (cygwin again) when lingering in AE::Handle.
73 - reduce memory usage slightly when connect was used in AE::Handle.
74 - wrok around more segfaults in Tk.
75 - document tls_ctx => undef as valid.
76 - detect WNOHANG value at compiletime.
77 - include some probably linux-only support for building AnyEvent
78 as part of the perl core.
79 - improve compatibility of t/08_idna.t to perl 5.8.
80 - make AnyEvent::Strict truly optional (in case somebody builds a
81 minimal perl...).
82
835.28 Wed Oct 13 04:14:23 CEST 2010
84 - due to a glitch, AnyEvent's internal getprotobyname wasn't
85 actually used in tcp_server.
86 - implement AnyEvent::CondVar->new method, which was documented but
87 missing.
88 - support autoloading of read types in unshift_read.
89 - AnyEvent::Strict: do the test for invalid fh arguments differently,
90 resulting in more informative error messages.
91 - new AnyEvent::FAQ manpage.
92 - clarify that you can actually call push_read etc. in on_read.
93 - put keys and certs in /tmp, not into the current directory,
94 and document this in the SECURITY CONSIDERATIONS sections
95 of AnyEvent::TLS.
96 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
97 requested by Ron Isaacson.
98 - document $AnyEvent::DNS::RESOLVER variable.
99 - incorporated major typo patches by Abhijit Menon-Sen.
100 - AnyEvent::Handle now croaks when negative timeouts are passed.
101 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
102 undocumented, but who knows when they might come in handy.
103
1045.271 Tue Jun 8 12:05:46 CEST 2010
105 - backport to perl 5.8.x.
106
1075.27 Sun Jun 6 12:12:05 CEST 2010
108 - postpone differently in AnyEvent::Socket now, as
109 when not, canceling the connection attempt might fail
110 (found by Felix Antonius Wilhelm Ostmann).
111 - explicitly check for non-stream sockets in AE::Handle, too many
112 clueless people fell into the trap of this somehow working.
113 - simplified and reworked the "OTHER MODULES" section.
114 - better/more condvar examples.
115
1165.261 Wed Apr 28 16:13:36 CEST 2010
117 - AF_INET6 was not properly used from Socket6 during configuration
118 time (found by Andy Grundman).
119
1205.26 Mon Apr 12 04:49:35 CEST 2010
121 - don't generate spurious readyness notifications when select
122 returns EINTR in the pure perl backend (analysed and testcase
123 by Bas Denissen).
124 - give same examples in SYNOPSIS sections of AnyEvent and AE.
125 - provider faster implementations for AE::time/now/now_update for
126 EV and Perl backends.
127 - provide faster AE::cv and AE::time implementations for all
128 backends.
129 - no longer support lower/mixed-case signal names (this was
130 never documented nor universally supported).
131 - some more memory µ-optimisations, and somewhat less messy
132 code to cope with both AnyEvent and AE APIs.
133
1345.251 Sat Mar 13 00:58:21 CET 2010
135 - make test failed if it was run before make install on perl <=
136 5.8.8, and older versions of perl make our live very difficult,
137 so write constants.pl during Makefile.PL time.
138
1395.25 Sat Mar 13 00:23:14 CET 2010
140 - fix a race condition in AnyEvent::Handle that would cause
141 a "bio_read: unsupported method" error (in _another_
142 ssl connection) after user code threw an exception.
143 - added AnyEvent::Handle->destroyed method.
144 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
145 - the AnyEvent::Util::fh_nonblocking function worked only
146 by lucky accident on win32.
147 - smaller and faster AnyEvent::Util::fh_nonblocking.
148 - when the (required!) Time::HiRes module is missing, AnyEvent
149 did not fall back to built-in time properly.
150 - do not load Fcntl at runtime, saving memory and loading time.
151 - precompile a number of constants and use them instead of runtime
152 detection and eval.
153 - free detection code after detection and similar memory optimisations.
154 - Perl backend timer interval best effort drift has been improved
155 (same algorithm as EV).
156 - update unicode idna mapping table.
157
1585.24 Tue Jan 5 11:39:43 CET 2010
159 - cygwin never reports errors from failed connects. 1.5 just gives
160 you continous readyness and EAGAIN, 1.7 is even more broken and
161 just hangs. work around both issues in a cygwin-specific hack.
162 - improve idle watcher documentation slightly.
163
1645.23 Sun Dec 20 23:48:00 CET 2009
165 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
166 - implement punycode_encode/decode, idn_nameprep,
167 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
168 - provide $AE::VERSION.
169 - removed traces of "no strict 'refs'".
170
1715.22 Sat Dec 5 03:51:13 CET 2009
172 - downgrade-or-fail in AnyEvent::Handle::push_write, to
173 diagnose encoding failures earlier and more succinctly.
174 (this works around bugs in perl, throwing away encoding info
175 when passing scalar data to extensions).
176 - add more examples to AnyEvent::Socket manpage.
177 - upgrade internal warning set to the same as common::sense 2.03.
178 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
179 (John Beppu).
180 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
181 - add google public dns servers to fallback server set and make sure
182 we load-balance properly between the three sets. also add all
183 fallback dns servers, not just a random one, to each dns config.
184
1855.21 Thu Nov 19 02:48:47 CET 2009
186 - fix a problem where socket constants were called with parameters
187 (spotted by David Friedland).
188 - fork_call never use'd POSIX (reported by Daisuke Maki).
189 - improve perl 5.6 compatibility further (but it still won't work
190 unless you rip out everything but the core).
191 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
192 uses 10MB of RAM, but doesn't open a console window. *sigh*.
193
1945.202 Wed Oct 14 22:35:44 CEST 2009
195 - AnyEvent::DNS would unexpectedly clobber $_ under windows
196 (analysed by Matthias Waldorf).
197 - AnyEvent::Handle::run_cmd can now pass the PID of the
198 newly-created process, which is much less useful than it might
199 sound (based on patch by Yann Kerherve).
200
2015.201 Tue Sep 29 12:09:25 CEST 2009
202 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
203 (reported by Torsten Foertsch).
204 - common::sense 2.0 could cause tcp_server to throw an exception
205 (analysed by elmex).
206
2075.2 Mon Sep 14 07:04:49 CEST 2009
208 - INCOMPATIBLE CHANGE: do no longer support register_read_type
209 and register_write_type in AnyEvent::Handle, instead support
210 package names (the facility was mostly abused).
211 - implement "packagename-as-read/write type" support in
212 AnyEvent::Handle.
213 - AnyEvent::Handle: new options "keepalive" and "oobinline".
214 - oobinline set by default to avoid security issues.
215 - the pure-perl event loop backend wrongly detected times() fallback
216 support (spotted by Pavel Boldin).
217
2185.12 Tue Sep 1 20:26:50 CEST 2009
219 - be more lenient when parsing resolv.conf files, as some people
220 use hashmarks as comment indicator inside directives (reported
221 by Michael S. Fischer).
222 - use same set of warnings as common::sense 2.0.
223 - fix a potential 32 bit overflow issue due to perl having problems
224 with large hex constants in the Perl backend.
225 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
226 by default.
227 - fix some prototypes.
228 - (optionally) require Async::Interrupt 1.02.
229 - replace opendns fallbacks by verizon/level3 public dns servers.
230
2315.112 Fri Aug 21 13:59:04 CEST 2009
232 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
233 using starttls with an illegal/uncreatable tls context.
234 - port to broken win32 perls.
235
2365.111 Sun Aug 16 18:44:36 CEST 2009
237 - perl errornously requires sockaddr_un structures returned by the
238 kernel to have a specific length. work around this bug by padding
239 sockaddr_un structures (found to fail on OS X, as perl only uses
240 the correct code for GNU/Linux, but likely to fail on other
241 OSes as well).
242
2435.11 Wed Aug 12 17:49:37 CEST 2009
244 - fixed a bug in run_cmd with input from filename.
245 - create a json coder object when none is given in write direction
246 to reduce dependency on JSON.
247
2485.1 Tue Aug 11 03:17:32 CEST 2009
249 - tcp_connect now reports the correct error in $!, not an unrelated
250 one (was broken since 4.91).
251 - AE::Handle did not properly clear rtimeout/wtimeout watchers
252 when retrying a tcp_connect.
253 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
254 - fixed wrong documentation in AE::Socket.
255
2565.01 Mon Aug 10 03:16:32 CEST 2009
257 - last release broke EV child watchers :/.
258
2595.0 Sun Aug 9 17:32:09 CEST 2009
260 - officially introduce and document the AE API.
261 - lowering the timeout at runtime did not have immediate effect
262 in AnyEvent::Handle.
263 - AE::Handle now has separate and independent read and write timeout
264 settings.
265 - probe for Guard module when AnyEvent::Util is loaded, not at
266 runtime (Event doesn't like eval "use" from callbacks, and it avoids
267 exporting the probe function).
268 - Impl::IOAsync's io watchers did not use the filehandle emulation
269 code to work around it's watcher limits.
270 - work around IO::Async's problems with overloaded objects.
271 - try to disable wearnings inside IO::Async, it's just too buggy.
272 - do no longer use Event by default, only EV or Perl, Event is too buggy.
273 - hackishly provide fast AE:: interface when EV or Perl is the backend.
274 - inherit VERSION from AnyEvent::VERSION in most modules.
275
2764.91 Thu Aug 6 15:42:45 CEST 2009
277 - AE::Handle::starttls could get out of sync when the read buffer
278 already contains some TLS handshake.
279 - AE::Handle did not properly free the TLS session, causing bigger
280 memory leaks in Net::SSLeay than expected (reported by toaster).
281 - AE::Socket::tcp_connect will now artificially delay invoking
282 the callback to avoid returning after invoking the callback.
283 - convert many internal watcher uses to AE API.
284
2854.9 Sat Aug 1 11:07:01 CEST 2009
286 - Glib silently fails when registering a timeout with
287 a negative value, so avoid that.
288 - call condvar callback immediately when it is set
289 after the condition is already signalled.
290 - check rbuf_max condition only after trying to consume
291 data.
292 - normalise signal numbers to names when using the ae's
293 signal handling, but do not document this (yet).
294 - pure perl signal emulation did not properly set nonblocking
295 mode on the signal pipe on win32 (this is unlikely to have caused
296 any issues).
297 - new module: AnyEvent::Debug.
298 - AnyEvent::Strict now checks that a signal was specified
299 by name (not name or number).
300 - reduce memory footprint in typical cases by ~50kb by
301 only compiling signal code when necessary.
302 - add AnyEvent::Handle->rbuf_max.
303 - grab the AE namespace for future new API, implement stubs for
304 said future/alternative API.
305 - new function: AnyEvent::Socket::format_hostport.
306
3074.881 Tue Jul 28 12:51:53 CEST 2009
308 - work around a bug in local in pre-5.10 perls, causing
309 AnyEvent::Handle to recurse when it should not
310 (analysed by elmex).
311
3124.88 Tue Jul 28 04:04:37 CEST 2009
313 - re-bless the handle into a dummy package after calling
314 AnyEvent::Handle::destroy, so the user does not need to check
315 for errors after every push_write etc.
316 - do not attempt to run t/02_signals.t on obviously broken
317 platforms.
318
3194.87 Sun Jul 26 02:06:16 CEST 2009
320 - do not attempt to linger when there is no longer a valid fh.
321 - enforce tls mode to be either accept or connect, do not simply
322 segfault in Net::SSLeay.
323 - AnyEvent::Handle can now call tcp_connect itself (new parameters
324 connect, on_prepare, on_connect and on_connect_error). Updated
325 tutorial accordingly.
326 - add AnyEvent::Impl::Irssi backend.
327
3284.86 Mon Jul 20 23:52:29 CEST 2009
329 - since the verbose warning is not enough, explicitly document
330 that versions before 1.33 of Net::SSLeay are not secure.
331 - work around signal handling races in Event and (...) Event::Lib.
332 - try to align signal-race timer to full-second boundaries.
333 - work around Tk not liking negative timeouts.
334 - don't complain of different grades of environmental unfriendlyness
335 in IO::Async.
336
3374.85 Sat Jul 18 06:16:14 CEST 2009
338 - nail the signal race problem in perl once and for all
339 (see $AnyEvent::MAX_SIGNAL_LATENCY).
340 - take advantage of Async::Interrupt if it is available.
341 - load Time::HiRes and Guard modules on demand only.
342 - add optional/recommended modules section to AnyEvent
343 documentation.
344 - reduce memory usage considerably (and reduce startup penalty)
345 by not using "strict", "warnings" and "overload" modules.
346 - work around buggy windows/openbsd perls and provide EBADMSG
347 and EPROTO ourselves when missing.
348 - improve perl 5.6 compatibility of the core event loop.
349 - made Net::SSLeay version 1.33 a soft requirement.
350
3514.83 Fri Jul 17 16:56:26 CEST 2009
352 - implement AnyEvent::Socket::getprotobyname.
353 - AnyEvent::CondVar's will now detect recursive blocking
354 waits and will croak, as too many people fall into
355 this trap.
356 - AnyEvent::Handle will now call ->destroy on itself after
357 executing the on_error callback, instead of doing some
358 half-baked internal shutdown, for fatal errors.
359 - clarify on_eof behaviour w.r.t. the read queue and
360 on_read callbacks.
361 - ignore some possible spurious wake-ups in tcp_connect.
362
3634.82 Sat Jul 11 00:34:55 CEST 2009
364 - POE and Event backends didn't accept some callable objects as
365 callbacks.
366 - use Config module instead of POSIX module to detect signal names
367 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
368 module doesn't even have all POSIX signals :/.
369 - use more workarounds around the many refcnt/corruption bugs in
370 Event::Lib.
371 - work around a race condition in perl's select, causing t/03_child.t
372 to rarely fail.
373
3744.81 Thu Jul 9 10:30:30 CEST 2009
375 - AnyEvent::Handle didn't properly diagnose write errors
376 (it expected -1 from syswrite, how lame... :).
377 - support file descriptors in addition to file handles
378 in AnyEvent->io.
379 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
380 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
381 and PERL_ANYEVENT_CA_PATH.
382 - provide a sensible synopsis section for AnyEvent::TLS.
383 - add a "supported backends" section to the manpage.
384 - added simple io watcher test to testsuite, using a
385 portable_socketpair.
386 - tried to improve the stability of the Event::Lib backend,
387 YMMV.
388
3894.8 Mon Jul 6 23:45:16 CEST 2009
390 - AnyEvent::DNS did not properly follow CNAME records with
391 uppercase targets.
392 - AnyEvent::DNS would errornously return AAAA records
393 with v4 mapped addresses (a faulty record) as ipv4 addresses,
394 causing AnyEvent::Socket to throw an exception.
395 - added new module AnyEvent::TLS for easier SSL/TLS context
396 creation, with many options including hostname verification,
397 secure default configuration, lots of documentation and,
398 predefined diffie-hellman keys for perfect forward security
399 and much more. get it while it's still fresh!
400 - use AnyEvent::TLS in AnyEvent::Handle for context management.
401 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
402 so AnyEvent::Socket users have smaller memory footprint
403 in the common case.
404 - add AnyEvent::Handle->push_shutdown method.
405 - add an additional error message parameter to AnyEvent::Handle's
406 on_error callback (for TLS, $! is still available).
407 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
408 - make AnyEvent::Handle more robust against early conenction
409 failures (during new), and return C<undef> in that case
410 from the constructor.
411 - AnyEvent::Handle will now try to load only JSON::XS first,
412 then fall back to JSON.
413 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
414 and because it was documented that way.
415
4164.451 Fri Jul 3 00:28:58 CEST 2009
417 - do not clear rbuf when shutting down an AnyEvent::Handle
418 object - doing so breaks AnyEvent::HTTP.
419
4204.45 Mon Jun 29 22:59:26 CEST 2009
421 - a write error could cause AnyEvent::Handle to create
422 an I/O watcher with an undefined $fh.
423 - special-case mapped ipv4 addresses in both
424 AnyEvent::Socket::format_address and parse_address,
425 to treat them just like ipv4 addresses.
426 - updated and overhauled the AnyEvent::Intro doc.
427 - implement AnyEvent::Socket::format_ipv4/ipv6.
428 - slightly speed up portable_pipe/socketpair.
429 - expand condvar begin/end documentation.
430
4314.42 Fri Jun 26 08:32:18 CEST 2009
432 - preliminary and neccesarily incomplete support for IO::Async.
433 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
434 case it was set to IGNORE, to ensure we can catch child statuses
435 even when the calling env acts stupidly.
436 - updated benchmarks with IO::Async, which performs very well.
437
4384.412 Wed Jun 24 01:35:57 CEST 2009
439 - support an "untaint" attribute for AnyEvent::DNS
440 and set it on the default resolver.
441 - implement and document AnyEvent::Impl::Perl::loop.
442 - remove all anyevent-env variables from %ENV
443 when running in tainted mode.
444 - mention and extend the IO::Lambda benchmark.
445
4464.411 Sun Jun 7 18:48:02 CEST 2009
447 - do not try to use F_SETFD on windows.
448
4494.41 Thu May 14 06:40:11 CEST 2009
450 - work around issues in older perls (5.8.0?) when
451 a signal handler is deleted from the %SIG hash.
452 - use POSIX::_exit in child test, to avoid running
453 destructors.
454 - speed up CHLD handling by relying on SIGCHLD being
455 synchronously delivered, even when we roll our own
456 implementation.
457 - AnyEvent::DNS: add the "dname" resource record name
458 for cosmetic reasons.
459
4604.4 Sun Apr 26 20:12:33 CEST 2009
461 - implemented idle watchers, where applicable.
462 - AnyEvent->time died when Event backend was in use.
463 - fix a memleak in the Tk backend.
464 - sped up Tk timer handling.
465 - clip negative "after" values to 0 in AnyEvent::Impl::Event
466 to avoid spamming out warning messages.
467 - fix Qt timers without interval.
468 - avoid Qt zero-timeout special case, in old libqt's.
469
4704.352 Mon Apr 20 16:31:11 CEST 2009
471 - fix AnyEvent::Strict error messages for child watchers.
472 - fix/update Makefile.PL "recommends" versions.
473 - add AnyEvent->now_update.
474
4754.351 Sat Apr 11 07:56:14 CEST 2009
476 - actually make the signal pipe work under win32.
477 - localise $! in signal handler to avoid changing $!, although
478 perl itself does not seem to save/restore errno either.
479 - set the cloexec flag on the signal pipe (normally set by perl too).
480
4814.35 Fri Mar 27 11:48:20 CET 2009
482 - event models relying on AnyEvent's signal watcher emulation did
483 invoke the callback asynchronously, contradicting documentation
484 and causing signals to get lost (this includes AnyEvent's own
485 event loop). AnyEvent now uses the standard pipe trick to make
486 callback execution synchronous to the event loop.
487 - AnyEvent::Handle didn't free TLS context data on DESTROY
488 (patch by Pavel Shaydo).
489 - work around the ever-incompatibly-changing API of MakeMaker.
490 - document that changing global variables without restoring them
491 is a bad idea in Perl (noted by Adam Rosenstein).
492 - AnyEvent::Strict now barfs if ->io is passed a file.
493
4944.34 Thu Feb 12 18:32:45 CET 2009
495 - separately buffer TLS read data, as otherwise the read queue could
496 deadlock as receiving data is not expected while draining the
497 read queue (which cna only happen with TLS).
498 - raise EBADMSG error on JSON decoding errors.
499 - fix some minor manpage bugs (reported by Maximilian Gaß).
500 - speed up select bitmask parsing quite a bit in the pure perl
501 backend.
502 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
503
5044.331 Tue Jan 6 21:07:25 CET 2009
505 - socketpair fails on many vista machines because vista has
506 completely broken accept/getpeername and getsockname functions,
507 so we provide our own socketpair emulation that kind of works
508 (AnyEvent::Util::portable_pipe).
509 - new function: AnyEvent::Util::portable_socketpair.
510 - take advantage of the Guard module if it exists.
511
5124.33 Fri Nov 21 02:35:40 CET 2008
513 - AnyEvent::Strict did errornously flag a fileno of 0 as
514 illegal.
515 - reduce memory usage and slightly speed up the pure perl backend
516 by only storing the file descriptor, not the file handle.
517 - add missing autocork method to AnyEvent::Handle (reported
518 by Adam Rosenstein).
519 - AnyEvent::DNS->resolve errornously documented an $rcode result
520 argument, but there isn't (spotted by Henrik Krohns).
521 - the naked truth about Tk - it's basically unmaintained.
522
5234.32 Mon Nov 3 22:46:32 CET 2008
524 - fix AnyEvent::Socket::resolve_sockaddr to properly support
525 unix sockets again and choose a proper default for
526 the socket type.
527 - fix call to resolve_sockaddr in tcp_connect.
528
5294.31 Thu Oct 30 04:41:48 CET 2008
530 - implemented AnyEvent::Handle->destroy method.
531 - hint about unexpected effects in TLS mode.
532 - speed up AnyEvent::Impl::Perl by using more arrays
533 instead of hashes without the slightest loss of readability :)
534 - work around a perl argument refcounting bug.
535 - with some perl patches applied, AnyEvent now seems leak-free.
536
5374.3 Fri Oct 3 09:18:43 CEST 2008
538 - AnyEvent will now install a no-op signal handler for SIGPIPE,
539 unless one has been installed already.
540 - warn about not loaidng AnyEvent::Impl::POE early enough
541 (patch by Adam Rosenstein).
542 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
543 - fix a bug in where in SSL connect mode, the
544 client would first wait for some data by the server and
545 otherwise hang. (reported and analysed in an absolutely
546 exemplary manner by Adam Rosenstein).
547 - fix a bug in where SSL EOF would not be treated
548 as stream EOF, putting the connection into a hung state.
549 (reported and analysed in an absolutely exemplary manner by Adam
550 Rosenstein).
551 - fix a potential segfault when the TLS context
552 would go missing in a read callback (e.g. due to stoptls) -
553 Net::SSLeay of course makes no type checking whatsoever.
554 - AnyEvent::Handle will not stop the read watcher in TLS mode.
555 - AnyEvent::Handle->stoptls will send a TLS close notify
556 instead of simply destroying the stream now.
557 - fix a bug where large blocks of data written in TLS mode
558 would not be sent unless triggered by receives.
559 - on_drain will now take the tls write buffer into account.
560 - SSL operations have been streamlined a bit and should be faster
561 now, more to come, though.
562 - added a FAQ section, will document nontrivial issues in there.
563
5644.234 Mon Sep 29 04:08:13 CEST 2008
565 - fix child watcher documentation: only child exits will be reported,
566 no trace events.
567 - mention SIGPIPE in AnyEvent::Handle.
568 - perl backend will now use POSIX::times as monotonic clock
569 when available and the monotonic clock option is not.
570 - run even without Time::HiRes in some cases now, but do not
571 provide sub-second accuracy (all *supported* perl versions have
572 Time::HiRes).
573
5744.233 Fri Aug 22 01:48:05 CEST 2008
575 - fix a bug in the testsuite, causing a hang.
576 - clarified lots of AnyEvent::Handle constructor arguments.
577
5784.232 Thu Aug 21 20:44:25 CEST 2008
579 - fix a bug in AnyEvent::Handle that could cause two fatal error
580 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
581 (testcase by Leon Brocard).
582 - clarified on_eof/on_error documentation a bit.
583
5844.231 Tue Jul 29 13:12:15 CEST 2008
585 - remove some debugging code left in AnyEvent::Util::fork_call
586 (and no, it's impossible to implement with the broken windows
587 perls without resource leaks or worse).
588
5894.23 Tue Jul 29 12:19:59 CEST 2008
590 - document the first parameter passed to condvar callbacks to be
591 the callback.
592 - add AnyEvent::Socket::{ntoa,aton} aliases.
593 - optimize the AE::Handle->push_read (line) for the default
594 eol marker.
595 - optimize push_read (packstring|storable) for small packets.
596 - invoke on_error callback when no on_eof callback is set.
597 - fix a bug in push_read (storable) of unknown impact.
598
5994.22 Sun Jul 20 16:34:13 CEST 2008
600 - new function AnyEvent::Socket::parse_hostport.
601 - as the bulkheads at microsoft can't even get getprotobyname reliably
602 working on their shitty fucking broken os we need to hardcode
603 some common protocol numbers in AnyEvent::Socket. How can
604 people even bother with such a pile of shit as windows.
605
6064.21 Thu Jul 17 14:40:05 CEST 2008
607 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
608 which is incompatible to previous invocation syntax.
609 - work around more windows perl bugs in fork_call. I cannot imagine
610 why anybody would bother implementing fork in such an obviously
611 unusable way.
612 - avoid calling $do_retry if the request has already finished in
613 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
614 K. Sharifulin).
615
6164.2 Sat Jul 12 22:42:11 CEST 2008
617 - support an "interval" argument for repeating timers.
618 - fix a bug in the pure perl event loop that caused timers to
619 be delayed under some circumstances.
620 - implement optional argument checking via PERL_ANYEVENT_STRICT.
621 - not importing anything from Event causes it to malfunction
622 and not use Time::HiRes, work around this bug.
623 - more bug workarounds against the endlessly broken Tk module,
624 seems to work now on 32 bit machines, but less so on 64 bit.
625 - minor optimisations applied to most backends.
626
6274.161 Fri Jul 4 14:24:48 CEST 2008
628 - fixed the t/handle/02_write.t, it read after write, with the
629 assumption that all data could be written before read
630 (reported by Jim Razmus).
631
6324.160 Thu Jul 3 04:02:21 CEST 2008
633 - re-registering signal watchers after unregistering one did
634 not work if a backend falls back on AnyEvent's default
635 implementation (also affected child watchers).
636 - new AnyEvent::Handle options: autocork and no_delay.
637
6384.152 Sun Jun 22 14:15:44 CEST 2008
639 - allow for 32-bit perls that implement shifts differently
640 on different architectures in parse_ipv4 (reported and
641 analysed by Keiichi DAIBA).
642
6434.151 Fri Jun 6 17:34:24 CEST 2008
644 - make sure specifying _only_ on_read and never pushing reads
645 works.
646
6474.15 Fri Jun 6 13:00:46 CEST 2008
648 - the pure perl backend would keep some watchers alive when more than
649 one watcher was registered for the same fd.
650 - new "packstring" and "storable" read and write types
651 for AnyEvent::Handle.
652 - allow on_eof handler to be called after on_error with EPIPE returns.
653 - do not immediately call on_read callback in handle constructor.
654
6554.14 Thu Jun 5 20:29:31 CEST 2008
656 - Fixed a bug in DNS SRV priority sorting.
657 - AnyEvent::Util::guard now reports runtime errors while
658 executing the guard block as warnings.
659 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
660
6614.13 Thu Jun 5 00:47:59 CEST 2008
662 - AnyEvent::DNS only followed cname chains with length 2,
663 contrary to documentation. bumped it up to 10, thanks to
664 microsoft, the current broken-dns-config-king.
665 - AnyEvent::DNS didn't check for socket return status
666 but instead relied on perl not creating filehandles in that
667 case - too bad it gives you a bogus file handle (reported
668 and analysed by Vladimir Timofeev).
669 - fix queue management logic in AnyEvent::Handle:
670 when on_read was registered, the queue was empty
671 and no progress could be made AnyEvent::Handle would
672 enter an endless loop.
673 - correctly start reading again when the handle became
674 busy again after idling.
675 - correctly treat tls shut-downs as EOF condition
676 (for the time being, would be nice to have a callback for that).
677 - correctly call eof callback under all conditions
678 (hopefully).
679 - the timeout callback did not expect that $self can go
680 away any time.
681
6824.12 Tue Jun 3 10:58:04 CEST 2008
683 - include AnyEvent::Intro, a tutorial for anyevent,
684 anyevent::socket and anyevent::handle.
4 - allow more options in on_error. 685 - allow more options in on_error.
5 686
64.11 Fri May 30 23:42:25 CEST 2008 6874.11 Fri May 30 23:42:25 CEST 2008
7 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide 688 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
8 reverse_lookup and reverse_verify to replace it, support 689 reverse_lookup and reverse_verify to replace it, support
… …
102 - support service names in tcp_connect and tcp_server. 783 - support service names in tcp_connect and tcp_server.
103 - work around netbsd/openbsd bug causing unix sockets not to 784 - work around netbsd/openbsd bug causing unix sockets not to
104 become writable until empty. 785 become writable until empty.
105 - only make a dummy read for error detection when getpeername 786 - only make a dummy read for error detection when getpeername
106 returns ENOTCONN. 787 returns ENOTCONN.
107 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 788 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
108 override. 789 override.
109 - new AnyEvent::Guard::cancel. 790 - new AnyEvent::Guard::cancel.
110 791
1113.5 Sat May 17 23:17:14 CEST 2008 7923.5 Sat May 17 23:17:14 CEST 2008
112 - gracefully deal with systems not defining CLOCK_MONOTONIC 793 - gracefully deal with systems not defining CLOCK_MONOTONIC
… …
126 807
1273.41 Sun May 11 19:53:13 CEST 2008 8083.41 Sun May 11 19:53:13 CEST 2008
128 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 809 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
129 810
1303.4 Sun May 11 00:29:25 CEST 2008 8113.4 Sun May 11 00:29:25 CEST 2008
131 - complete re-implementation of condvars with a hopefully much mroe useful 812 - complete re-implementation of condvars with a hopefully much more useful
132 API (backward compatible functions exist but are undocumented). 813 API (backward compatible functions exist but are undocumented).
133 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 814 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
134 (still subject to change). 815 (still subject to change).
135 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 816 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
136 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 817 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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