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

Comparing AnyEvent/Changes (file contents):
Revision 1.150 by root, Thu Jun 5 16:28:48 2008 UTC vs.
Revision 1.424 by root, Thu Dec 23 11:59:35 2010 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
34.14 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: wbuf_max for anyevent
9TODO: fix with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if
105.29 Sun Dec 5 10:49:21 CET 2010
11 - convert EV backend to EV 4.00 API (so better upgrade EV too).
12 - AE::Handle: implement dynamic read_size adjustment, add
13 max_read_size parameter, reduce default read_size to 2048.
14 - add some FAQ entries for servers and/or on_eof vs. on_error.
15 - work around OS bugs (cygwin again) when lingering in AE::Handle.
16 - reduce memory usage slightly when connect was used in AE::Handle.
17 - wrok around more segfaults in Tk.
18 - document tls_ctx => undef as valid.
19 - detect WNOHANG value at compiletime.
20 - include some probably linux-only support for building AnyEvent
21 as part of the perl core.
22 - improve compatibility of t/08_idna.t to perl 5.8.
23 - make AnyEvent::Strict truly optional (in case somebody builds a
24 minimal perl...).
25
265.28 Wed Oct 13 04:14:23 CEST 2010
27 - due to a glitch, AnyEvent's internal getprotobyname wasn't
28 actually used in tcp_server.
29 - implement AnyEvent::CondVar->new method, which was documented but
30 missing.
31 - support autoloading of read types in unshift_read.
32 - AnyEvent::Strict: do the test for invalid fh arguments differently,
33 resulting in more informative error messages.
34 - new AnyEvent::FAQ manpage.
35 - clarify that you can actually call push_read etc. in on_read.
36 - put keys and certs in /tmp, not into the current directory,
37 and document this in the SECURITY CONSIDERATIONS sections
38 of AnyEvent::TLS.
39 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
40 requested by Ron Isaacson.
41 - document $AnyEvent::DNS::RESOLVER variable.
42 - incorporated major typo patches by Abhijit Menon-Sen.
43 - AnyEvent::Handle now croaks when negative timeouts are passed.
44 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
45 undocumented, but who knows when they might come in handy.
46
475.271 Tue Jun 8 12:05:46 CEST 2010
48 - backport to perl 5.8.x.
49
505.27 Sun Jun 6 12:12:05 CEST 2010
51 - postpone differently in AnyEvent::Socket now, as
52 when not, canceling the connection attempt might fail
53 (found by Felix Antonius Wilhelm Ostmann).
54 - explicitly check for non-stream sockets in AE::Handle, too many
55 clueless people fell into the trap of this somehow working.
56 - simplified and reworked the "OTHER MODULES" section.
57 - better/more condvar examples.
58
595.261 Wed Apr 28 16:13:36 CEST 2010
60 - AF_INET6 was not properly used from Socket6 during configuration
61 time (found by Andy Grundman).
62
635.26 Mon Apr 12 04:49:35 CEST 2010
64 - don't generate spurious readyness notifications when select
65 returns EINTR in the pure perl backend (analysed and testcase
66 by Bas Denissen).
67 - give same examples in SYNOPSIS sections of AnyEvent and AE.
68 - provider faster implementations for AE::time/now/now_update for
69 EV and Perl backends.
70 - provide faster AE::cv and AE::time implementations for all
71 backends.
72 - no longer support lower/mixed-case signal names (this was
73 never documented nor universally supported).
74 - some more memory µ-optimisations, and somewhat less messy
75 code to cope with both AnyEvent and AE APIs.
76
775.251 Sat Mar 13 00:58:21 CET 2010
78 - make test failed if it was run before make install on perl <=
79 5.8.8, and older versions of perl make our live very difficult,
80 so write constants.pl during Makefile.PL time.
81
825.25 Sat Mar 13 00:23:14 CET 2010
83 - fix a race condition in AnyEvent::Handle that would cause
84 a "bio_read: unsupported method" error (in _another_
85 ssl connection) after user code threw an exception.
86 - added AnyEvent::Handle->destroyed method.
87 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
88 - the AnyEvent::Util::fh_nonblocking function worked only
89 by lucky accident on win32.
90 - smaller and faster AnyEvent::Util::fh_nonblocking.
91 - when the (required!) Time::HiRes module is missing, AnyEvent
92 did not fall back to built-in time properly.
93 - do not load Fcntl at runtime, saving memory and loading time.
94 - precompile a number of constants and use them instead of runtime
95 detection and eval.
96 - free detection code after detection and similar memory optimisations.
97 - Perl backend timer interval best effort drift has been improved
98 (same algorithm as EV).
99 - update unicode idna mapping table.
100
1015.24 Tue Jan 5 11:39:43 CET 2010
102 - cygwin never reports errors from failed connects. 1.5 just gives
103 you continous readyness and EAGAIN, 1.7 is even more broken and
104 just hangs. work around both issues in a cygwin-specific hack.
105 - improve idle watcher documentation slightly.
106
1075.23 Sun Dec 20 23:48:00 CET 2009
108 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
109 - implement punycode_encode/decode, idn_nameprep,
110 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
111 - provide $AE::VERSION.
112 - removed traces of "no strict 'refs'".
113
1145.22 Sat Dec 5 03:51:13 CET 2009
115 - downgrade-or-fail in AnyEvent::Handle::push_write, to
116 diagnose encoding failures earlier and more succinctly.
117 (this works around bugs in perl, throwing away encoding info
118 when passing scalar data to extensions).
119 - add more examples to AnyEvent::Socket manpage.
120 - upgrade internal warning set to the same as common::sense 2.03.
121 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
122 (John Beppu).
123 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
124 - add google public dns servers to fallback server set and make sure
125 we load-balance properly between the three sets. also add all
126 fallback dns servers, not just a random one, to each dns config.
127
1285.21 Thu Nov 19 02:48:47 CET 2009
129 - fix a problem where socket constants were called with parameters
130 (spotted by David Friedland).
131 - fork_call never use'd POSIX (reported by Daisuke Maki).
132 - improve perl 5.6 compatibility further (but it still won't work
133 unless you rip out everything but the core).
134 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
135 uses 10MB of RAM, but doesn't open a console window. *sigh*.
136
1375.202 Wed Oct 14 22:35:44 CEST 2009
138 - AnyEvent::DNS would unexpectedly clobber $_ under windows
139 (analysed by Matthias Waldorf).
140 - AnyEvent::Handle::run_cmd can now pass the PID of the
141 newly-created process, which is much less useful than it might
142 sound (based on patch by Yann Kerherve).
143
1445.201 Tue Sep 29 12:09:25 CEST 2009
145 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
146 (reported by Torsten Foertsch).
147 - common::sense 2.0 could cause tcp_server to throw an exception
148 (analysed by elmex).
149
1505.2 Mon Sep 14 07:04:49 CEST 2009
151 - INCOMPATIBLE CHANGE: do no longer support register_read_type
152 and register_write_type in AnyEvent::Handle, instead support
153 package names (the facility was mostly abused).
154 - implement "packagename-as-read/write type" support in
155 AnyEvent::Handle.
156 - AnyEvent::Handle: new options "keepalive" and "oobinline".
157 - oobinline set by default to avoid security issues.
158 - the pure-perl event loop backend wrongly detected times() fallback
159 support (spotted by Pavel Boldin).
160
1615.12 Tue Sep 1 20:26:50 CEST 2009
162 - be more lenient when parsing resolv.conf files, as some people
163 use hashmarks as comment indicator inside directives (reported
164 by Michael S. Fischer).
165 - use same set of warnings as common::sense 2.0.
166 - fix a potential 32 bit overflow issue due to perl having problems
167 with large hex constants in the Perl backend.
168 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
169 by default.
170 - fix some prototypes.
171 - (optionally) require Async::Interrupt 1.02.
172 - replace opendns fallbacks by verizon/level3 public dns servers.
173
1745.112 Fri Aug 21 13:59:04 CEST 2009
175 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
176 using starttls with an illegal/uncreatable tls context.
177 - port to broken win32 perls.
178
1795.111 Sun Aug 16 18:44:36 CEST 2009
180 - perl errornously requires sockaddr_un structures returned by the
181 kernel to have a specific length. work around this bug by padding
182 sockaddr_un structures (found to fail on OS X, as perl only uses
183 the correct code for GNU/Linux, but likely to fail on other
184 OSes as well).
185
1865.11 Wed Aug 12 17:49:37 CEST 2009
187 - fixed a bug in run_cmd with input from filename.
188 - create a json coder object when none is given in write direction
189 to reduce dependency on JSON.
190
1915.1 Tue Aug 11 03:17:32 CEST 2009
192 - tcp_connect now reports the correct error in $!, not an unrelated
193 one (was broken since 4.91).
194 - AE::Handle did not properly clear rtimeout/wtimeout watchers
195 when retrying a tcp_connect.
196 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
197 - fixed wrong documentation in AE::Socket.
198
1995.01 Mon Aug 10 03:16:32 CEST 2009
200 - last release broke EV child watchers :/.
201
2025.0 Sun Aug 9 17:32:09 CEST 2009
203 - officially introduce and document the AE API.
204 - lowering the timeout at runtime did not have immediate effect
205 in AnyEvent::Handle.
206 - AE::Handle now has separate and independent read and write timeout
207 settings.
208 - probe for Guard module when AnyEvent::Util is loaded, not at
209 runtime (Event doesn't like eval "use" from callbacks, and it avoids
210 exporting the probe function).
211 - Impl::IOAsync's io watchers did not use the filehandle emulation
212 code to work around it's watcher limits.
213 - work around IO::Async's problems with overloaded objects.
214 - try to disable wearnings inside IO::Async, it's just too buggy.
215 - do no longer use Event by default, only EV or Perl, Event is too buggy.
216 - hackishly provide fast AE:: interface when EV or Perl is the backend.
217 - inherit VERSION from AnyEvent::VERSION in most modules.
218
2194.91 Thu Aug 6 15:42:45 CEST 2009
220 - AE::Handle::starttls could get out of sync when the read buffer
221 already contains some TLS handshake.
222 - AE::Handle did not properly free the TLS session, causing bigger
223 memory leaks in Net::SSLeay than expected (reported by toaster).
224 - AE::Socket::tcp_connect will now artificially delay invoking
225 the callback to avoid returning after invoking the callback.
226 - convert many internal watcher uses to AE API.
227
2284.9 Sat Aug 1 11:07:01 CEST 2009
229 - Glib silently fails when registering a timeout with
230 a negative value, so avoid that.
231 - call condvar callback immediately when it is set
232 after the condition is already signalled.
233 - check rbuf_max condition only after trying to consume
234 data.
235 - normalise signal numbers to names when using the ae's
236 signal handling, but do not document this (yet).
237 - pure perl signal emulation did not properly set nonblocking
238 mode on the signal pipe on win32 (this is unlikely to have caused
239 any issues).
240 - new module: AnyEvent::Debug.
241 - AnyEvent::Strict now checks that a signal was specified
242 by name (not name or number).
243 - reduce memory footprint in typical cases by ~50kb by
244 only compiling signal code when necessary.
245 - add AnyEvent::Handle->rbuf_max.
246 - grab the AE namespace for future new API, implement stubs for
247 said future/alternative API.
248 - new function: AnyEvent::Socket::format_hostport.
249
2504.881 Tue Jul 28 12:51:53 CEST 2009
251 - work around a bug in local in pre-5.10 perls, causing
252 AnyEvent::Handle to recurse when it should not
253 (analysed by elmex).
254
2554.88 Tue Jul 28 04:04:37 CEST 2009
256 - re-bless the handle into a dummy package after calling
257 AnyEvent::Handle::destroy, so the user does not need to check
258 for errors after every push_write etc.
259 - do not attempt to run t/02_signals.t on obviously broken
260 platforms.
261
2624.87 Sun Jul 26 02:06:16 CEST 2009
263 - do not attempt to linger when there is no longer a valid fh.
264 - enforce tls mode to be either accept or connect, do not simply
265 segfault in Net::SSLeay.
266 - AnyEvent::Handle can now call tcp_connect itself (new parameters
267 connect, on_prepare, on_connect and on_connect_error). Updated
268 tutorial accordingly.
269 - add AnyEvent::Impl::Irssi backend.
270
2714.86 Mon Jul 20 23:52:29 CEST 2009
272 - since the verbose warning is not enough, explicitly document
273 that versions before 1.33 of Net::SSLeay are not secure.
274 - work around signal handling races in Event and (...) Event::Lib.
275 - try to align signal-race timer to full-second boundaries.
276 - work around Tk not liking negative timeouts.
277 - don't complain of different grades of environmental unfriendlyness
278 in IO::Async.
279
2804.85 Sat Jul 18 06:16:14 CEST 2009
281 - nail the signal race problem in perl once and for all
282 (see $AnyEvent::MAX_SIGNAL_LATENCY).
283 - take advantage of Async::Interrupt if it is available.
284 - load Time::HiRes and Guard modules on demand only.
285 - add optional/recommended modules section to AnyEvent
286 documentation.
287 - reduce memory usage considerably (and reduce startup penalty)
288 by not using "strict", "warnings" and "overload" modules.
289 - work around buggy windows/openbsd perls and provide EBADMSG
290 and EPROTO ourselves when missing.
291 - improve perl 5.6 compatibility of the core event loop.
292 - made Net::SSLeay version 1.33 a soft requirement.
293
2944.83 Fri Jul 17 16:56:26 CEST 2009
295 - implement AnyEvent::Socket::getprotobyname.
296 - AnyEvent::CondVar's will now detect recursive blocking
297 waits and will croak, as too many people fall into
298 this trap.
299 - AnyEvent::Handle will now call ->destroy on itself after
300 executing the on_error callback, instead of doing some
301 half-baked internal shutdown, for fatal errors.
302 - clarify on_eof behaviour w.r.t. the read queue and
303 on_read callbacks.
304 - ignore some possible spurious wake-ups in tcp_connect.
305
3064.82 Sat Jul 11 00:34:55 CEST 2009
307 - POE and Event backends didn't accept some callable objects as
308 callbacks.
309 - use Config module instead of POSIX module to detect signal names
310 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
311 module doesn't even have all POSIX signals :/.
312 - use more workarounds around the many refcnt/corruption bugs in
313 Event::Lib.
314 - work around a race condition in perl's select, causing t/03_child.t
315 to rarely fail.
316
3174.81 Thu Jul 9 10:30:30 CEST 2009
318 - AnyEvent::Handle didn't properly diagnose write errors
319 (it expected -1 from syswrite, how lame... :).
320 - support file descriptors in addition to file handles
321 in AnyEvent->io.
322 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
323 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
324 and PERL_ANYEVENT_CA_PATH.
325 - provide a sensible synopsis section for AnyEvent::TLS.
326 - add a "supported backends" section to the manpage.
327 - added simple io watcher test to testsuite, using a
328 portable_socketpair.
329 - tried to improve the stability of the Event::Lib backend,
330 YMMV.
331
3324.8 Mon Jul 6 23:45:16 CEST 2009
333 - AnyEvent::DNS did not properly follow CNAME records with
334 uppercase targets.
335 - AnyEvent::DNS would errornously return AAAA records
336 with v4 mapped addresses (a faulty record) as ipv4 addresses,
337 causing AnyEvent::Socket to throw an exception.
338 - added new module AnyEvent::TLS for easier SSL/TLS context
339 creation, with many options including hostname verification,
340 secure default configuration, lots of documentation and,
341 predefined diffie-hellman keys for perfect forward security
342 and much more. get it while it's still fresh!
343 - use AnyEvent::TLS in AnyEvent::Handle for context management.
344 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
345 so AnyEvent::Socket users have smaller memory footprint
346 in the common case.
347 - add AnyEvent::Handle->push_shutdown method.
348 - add an additional error message parameter to AnyEvent::Handle's
349 on_error callback (for TLS, $! is still available).
350 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
351 - make AnyEvent::Handle more robust against early conenction
352 failures (during new), and return C<undef> in that case
353 from the constructor.
354 - AnyEvent::Handle will now try to load only JSON::XS first,
355 then fall back to JSON.
356 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
357 and because it was documented that way.
358
3594.451 Fri Jul 3 00:28:58 CEST 2009
360 - do not clear rbuf when shutting down an AnyEvent::Handle
361 object - doing so breaks AnyEvent::HTTP.
362
3634.45 Mon Jun 29 22:59:26 CEST 2009
364 - a write error could cause AnyEvent::Handle to create
365 an I/O watcher with an undefined $fh.
366 - special-case mapped ipv4 addresses in both
367 AnyEvent::Socket::format_address and parse_address,
368 to treat them just like ipv4 addresses.
369 - updated and overhauled the AnyEvent::Intro doc.
370 - implement AnyEvent::Socket::format_ipv4/ipv6.
371 - slightly speed up portable_pipe/socketpair.
372 - expand condvar begin/end documentation.
373
3744.42 Fri Jun 26 08:32:18 CEST 2009
375 - preliminary and neccesarily incomplete support for IO::Async.
376 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
377 case it was set to IGNORE, to ensure we can catch child statuses
378 even when the calling env acts stupidly.
379 - updated benchmarks with IO::Async, which performs very well.
380
3814.412 Wed Jun 24 01:35:57 CEST 2009
382 - support an "untaint" attribute for AnyEvent::DNS
383 and set it on the default resolver.
384 - implement and document AnyEvent::Impl::Perl::loop.
385 - remove all anyevent-env variables from %ENV
386 when running in tainted mode.
387 - mention and extend the IO::Lambda benchmark.
388
3894.411 Sun Jun 7 18:48:02 CEST 2009
390 - do not try to use F_SETFD on windows.
391
3924.41 Thu May 14 06:40:11 CEST 2009
393 - work around issues in older perls (5.8.0?) when
394 a signal handler is deleted from the %SIG hash.
395 - use POSIX::_exit in child test, to avoid running
396 destructors.
397 - speed up CHLD handling by relying on SIGCHLD being
398 synchronously delivered, even when we roll our own
399 implementation.
400 - AnyEvent::DNS: add the "dname" resource record name
401 for cosmetic reasons.
402
4034.4 Sun Apr 26 20:12:33 CEST 2009
404 - implemented idle watchers, where applicable.
405 - AnyEvent->time died when Event backend was in use.
406 - fix a memleak in the Tk backend.
407 - sped up Tk timer handling.
408 - clip negative "after" values to 0 in AnyEvent::Impl::Event
409 to avoid spamming out warning messages.
410 - fix Qt timers without interval.
411 - avoid Qt zero-timeout special case, in old libqt's.
412
4134.352 Mon Apr 20 16:31:11 CEST 2009
414 - fix AnyEvent::Strict error messages for child watchers.
415 - fix/update Makefile.PL "recommends" versions.
416 - add AnyEvent->now_update.
417
4184.351 Sat Apr 11 07:56:14 CEST 2009
419 - actually make the signal pipe work under win32.
420 - localise $! in signal handler to avoid changing $!, although
421 perl itself does not seem to save/restore errno either.
422 - set the cloexec flag on the signal pipe (normally set by perl too).
423
4244.35 Fri Mar 27 11:48:20 CET 2009
425 - event models relying on AnyEvent's signal watcher emulation did
426 invoke the callback asynchronously, contradicting documentation
427 and causing signals to get lost (this includes AnyEvent's own
428 event loop). AnyEvent now uses the standard pipe trick to make
429 callback execution synchronous to the event loop.
430 - AnyEvent::Handle didn't free TLS context data on DESTROY
431 (patch by Pavel Shaydo).
432 - work around the ever-incompatibly-changing API of MakeMaker.
433 - document that changing global variables without restoring them
434 is a bad idea in Perl (noted by Adam Rosenstein).
435 - AnyEvent::Strict now barfs if ->io is passed a file.
436
4374.34 Thu Feb 12 18:32:45 CET 2009
438 - separately buffer TLS read data, as otherwise the read queue could
439 deadlock as receiving data is not expected while draining the
440 read queue (which cna only happen with TLS).
441 - raise EBADMSG error on JSON decoding errors.
442 - fix some minor manpage bugs (reported by Maximilian Gaß).
443 - speed up select bitmask parsing quite a bit in the pure perl
444 backend.
445 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
446
4474.331 Tue Jan 6 21:07:25 CET 2009
448 - socketpair fails on many vista machines because vista has
449 completely broken accept/getpeername and getsockname functions,
450 so we provide our own socketpair emulation that kind of works
451 (AnyEvent::Util::portable_pipe).
452 - new function: AnyEvent::Util::portable_socketpair.
453 - take advantage of the Guard module if it exists.
454
4554.33 Fri Nov 21 02:35:40 CET 2008
456 - AnyEvent::Strict did errornously flag a fileno of 0 as
457 illegal.
458 - reduce memory usage and slightly speed up the pure perl backend
459 by only storing the file descriptor, not the file handle.
460 - add missing autocork method to AnyEvent::Handle (reported
461 by Adam Rosenstein).
462 - AnyEvent::DNS->resolve errornously documented an $rcode result
463 argument, but there isn't (spotted by Henrik Krohns).
464 - the naked truth about Tk - it's basically unmaintained.
465
4664.32 Mon Nov 3 22:46:32 CET 2008
467 - fix AnyEvent::Socket::resolve_sockaddr to properly support
468 unix sockets again and choose a proper default for
469 the socket type.
470 - fix call to resolve_sockaddr in tcp_connect.
471
4724.31 Thu Oct 30 04:41:48 CET 2008
473 - implemented AnyEvent::Handle->destroy method.
474 - hint about unexpected effects in TLS mode.
475 - speed up AnyEvent::Impl::Perl by using more arrays
476 instead of hashes without the slightest loss of readability :)
477 - work around a perl argument refcounting bug.
478 - with some perl patches applied, AnyEvent now seems leak-free.
479
4804.3 Fri Oct 3 09:18:43 CEST 2008
481 - AnyEvent will now install a no-op signal handler for SIGPIPE,
482 unless one has been installed already.
483 - warn about not loaidng AnyEvent::Impl::POE early enough
484 (patch by Adam Rosenstein).
485 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
486 - fix a bug in where in SSL connect mode, the
487 client would first wait for some data by the server and
488 otherwise hang. (reported and analysed in an absolutely
489 exemplary manner by Adam Rosenstein).
490 - fix a bug in where SSL EOF would not be treated
491 as stream EOF, putting the connection into a hung state.
492 (reported and analysed in an absolutely exemplary manner by Adam
493 Rosenstein).
494 - fix a potential segfault when the TLS context
495 would go missing in a read callback (e.g. due to stoptls) -
496 Net::SSLeay of course makes no type checking whatsoever.
497 - AnyEvent::Handle will not stop the read watcher in TLS mode.
498 - AnyEvent::Handle->stoptls will send a TLS close notify
499 instead of simply destroying the stream now.
500 - fix a bug where large blocks of data written in TLS mode
501 would not be sent unless triggered by receives.
502 - on_drain will now take the tls write buffer into account.
503 - SSL operations have been streamlined a bit and should be faster
504 now, more to come, though.
505 - added a FAQ section, will document nontrivial issues in there.
506
5074.234 Mon Sep 29 04:08:13 CEST 2008
508 - fix child watcher documentation: only child exits will be reported,
509 no trace events.
510 - mention SIGPIPE in AnyEvent::Handle.
511 - perl backend will now use POSIX::times as monotonic clock
512 when available and the monotonic clock option is not.
513 - run even without Time::HiRes in some cases now, but do not
514 provide sub-second accuracy (all *supported* perl versions have
515 Time::HiRes).
516
5174.233 Fri Aug 22 01:48:05 CEST 2008
518 - fix a bug in the testsuite, causing a hang.
519 - clarified lots of AnyEvent::Handle constructor arguments.
520
5214.232 Thu Aug 21 20:44:25 CEST 2008
522 - fix a bug in AnyEvent::Handle that could cause two fatal error
523 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
524 (testcase by Leon Brocard).
525 - clarified on_eof/on_error documentation a bit.
526
5274.231 Tue Jul 29 13:12:15 CEST 2008
528 - remove some debugging code left in AnyEvent::Util::fork_call
529 (and no, it's impossible to implement with the broken windows
530 perls without resource leaks or worse).
531
5324.23 Tue Jul 29 12:19:59 CEST 2008
533 - document the first parameter passed to condvar callbacks to be
534 the callback.
535 - add AnyEvent::Socket::{ntoa,aton} aliases.
536 - optimize the AE::Handle->push_read (line) for the default
537 eol marker.
538 - optimize push_read (packstring|storable) for small packets.
539 - invoke on_error callback when no on_eof callback is set.
540 - fix a bug in push_read (storable) of unknown impact.
541
5424.22 Sun Jul 20 16:34:13 CEST 2008
543 - new function AnyEvent::Socket::parse_hostport.
544 - as the bulkheads at microsoft can't even get getprotobyname reliably
545 working on their shitty fucking broken os we need to hardcode
546 some common protocol numbers in AnyEvent::Socket. How can
547 people even bother with such a pile of shit as windows.
548
5494.21 Thu Jul 17 14:40:05 CEST 2008
550 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
551 which is incompatible to previous invocation syntax.
552 - work around more windows perl bugs in fork_call. I cannot imagine
553 why anybody would bother implementing fork in such an obviously
554 unusable way.
555 - avoid calling $do_retry if the request has already finished in
556 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
557 K. Sharifulin).
558
5594.2 Sat Jul 12 22:42:11 CEST 2008
560 - support an "interval" argument for repeating timers.
561 - fix a bug in the pure perl event loop that caused timers to
562 be delayed under some circumstances.
563 - implement optional argument checking via PERL_ANYEVENT_STRICT.
564 - not importing anything from Event causes it to malfunction
565 and not use Time::HiRes, work around this bug.
566 - more bug workarounds against the endlessly broken Tk module,
567 seems to work now on 32 bit machines, but less so on 64 bit.
568 - minor optimisations applied to most backends.
569
5704.161 Fri Jul 4 14:24:48 CEST 2008
571 - fixed the t/handle/02_write.t, it read after write, with the
572 assumption that all data could be written before read
573 (reported by Jim Razmus).
574
5754.160 Thu Jul 3 04:02:21 CEST 2008
576 - re-registering signal watchers after unregistering one did
577 not work if a backend falls back on AnyEvent's default
578 implementation (also affected child watchers).
579 - new AnyEvent::Handle options: autocork and no_delay.
580
5814.152 Sun Jun 22 14:15:44 CEST 2008
582 - allow for 32-bit perls that implement shifts differently
583 on different architectures in parse_ipv4 (reported and
584 analysed by Keiichi DAIBA).
585
5864.151 Fri Jun 6 17:34:24 CEST 2008
587 - make sure specifying _only_ on_read and never pushing reads
588 works.
589
5904.15 Fri Jun 6 13:00:46 CEST 2008
591 - the pure perl backend would keep some watchers alive when more than
592 one watcher was registered for the same fd.
593 - new "packstring" and "storable" read and write types
594 for AnyEvent::Handle.
595 - allow on_eof handler to be called after on_error with EPIPE returns.
596 - do not immediately call on_read callback in handle constructor.
597
5984.14 Thu Jun 5 20:29:31 CEST 2008
4 - Fixed a bug in DNS SRV priority sorting. 599 - Fixed a bug in DNS SRV priority sorting.
5 - AnyEvent::Util::guard now reports runtime errors while 600 - AnyEvent::Util::guard now reports runtime errors while
6 executing the guard block as warnings. 601 executing the guard block as warnings.
602 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
7 603
84.13 Thu Jun 5 00:47:59 CEST 2008 6044.13 Thu Jun 5 00:47:59 CEST 2008
9 - AnyEvent::DNS only followed cname chains with length 2, 605 - AnyEvent::DNS only followed cname chains with length 2,
10 contrary to documentation. bumped it up to 10, thanks to 606 contrary to documentation. bumped it up to 10, thanks to
11 microsoft, the current broken-dns-config-king. 607 microsoft, the current broken-dns-config-king.
… …
130 - support service names in tcp_connect and tcp_server. 726 - support service names in tcp_connect and tcp_server.
131 - work around netbsd/openbsd bug causing unix sockets not to 727 - work around netbsd/openbsd bug causing unix sockets not to
132 become writable until empty. 728 become writable until empty.
133 - only make a dummy read for error detection when getpeername 729 - only make a dummy read for error detection when getpeername
134 returns ENOTCONN. 730 returns ENOTCONN.
135 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 731 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
136 override. 732 override.
137 - new AnyEvent::Guard::cancel. 733 - new AnyEvent::Guard::cancel.
138 734
1393.5 Sat May 17 23:17:14 CEST 2008 7353.5 Sat May 17 23:17:14 CEST 2008
140 - gracefully deal with systems not defining CLOCK_MONOTONIC 736 - gracefully deal with systems not defining CLOCK_MONOTONIC
… …
154 750
1553.41 Sun May 11 19:53:13 CEST 2008 7513.41 Sun May 11 19:53:13 CEST 2008
156 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 752 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
157 753
1583.4 Sun May 11 00:29:25 CEST 2008 7543.4 Sun May 11 00:29:25 CEST 2008
159 - complete re-implementation of condvars with a hopefully much mroe useful 755 - complete re-implementation of condvars with a hopefully much more useful
160 API (backward compatible functions exist but are undocumented). 756 API (backward compatible functions exist but are undocumented).
161 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 757 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
162 (still subject to change). 758 (still subject to change).
163 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 759 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
164 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 760 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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