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

Comparing AnyEvent/Changes (file contents):
Revision 1.100 by root, Sun May 25 01:05:27 2008 UTC vs.
Revision 1.403 by root, Sun Jul 11 05:44:22 2010 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: document $AnyEvent::DNS::RESOLVER
45.271 Tue Jun 8 12:05:46 CEST 2010
5 - backport to perl 5.8.x.
6
75.27 Sun Jun 6 12:12:05 CEST 2010
8 - postpone differently in AnyEvent::Socket now, as
9 when not, canceling the connection attempt might fail
10 (found by Felix Antonius Wilhelm Ostmann).
11 - explicitly check for non-stream sockets in AE::Handle, too many
12 clueless people fell into the trap of this somehow working.
13 - simplified and reworked the "OTHER MODULES" section.
14 - better/more condvar examples.
15
165.261 Wed Apr 28 16:13:36 CEST 2010
17 - AF_INET6 was not properly used from Socket6 during configuration
18 time (found by Andy Grundman).
19
205.26 Mon Apr 12 04:49:35 CEST 2010
21 - don't generate spurious readyness notifications when select
22 returns EINTR in the pure perl backend (analysed and testcase
23 by Bas Denissen).
24 - give same examples in SYNOPSIS sections of AnyEvent and AE.
25 - provider faster implementations for AE::time/now/now_update for
26 EV and Perl backends.
27 - provide faster AE::cv and AE::time implementations for all
28 backends.
29 - no longer support lower/mixed-case signal names (this was
30 never documented nor universally supported).
31 - some more memory µ-optimisations, and somewhat less messy
32 code to cope with both AnyEvent and AE APIs.
33
345.251 Sat Mar 13 00:58:21 CET 2010
35 - make test failed if it was run before make install on perl <=
36 5.8.8, and older versions of perl make our live very difficult,
37 so write constants.pl during Makefile.PL time.
38
395.25 Sat Mar 13 00:23:14 CET 2010
40 - fix a race condition in AnyEvent::Handle that would cause
41 a "bio_read: unsupported method" error (in _another_
42 ssl connection) after user code threw an exception.
43 - added AnyEvent::Handle->destroyed method.
44 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
45 - the AnyEvent::Util::fh_nonblocking function worked only
46 by lucky accident on win32.
47 - smaller and faster AnyEvent::Util::fh_nonblocking.
48 - when the (required!) Time::HiRes module is missing, AnyEvent
49 did not fall back to built-in time properly.
50 - do not load Fcntl at runtime, saving memory and loading time.
51 - precompile a number of constants and use them instead of runtime
52 detection and eval.
53 - free detection code after detection and similar memory optimisations.
54 - Perl backend timer interval best effort drift has been improved
55 (same algorithm as EV).
56 - update unicode idna mapping table.
57
585.24 Tue Jan 5 11:39:43 CET 2010
59 - cygwin never reports errors from failed connects. 1.5 just gives
60 you continous readyness and EAGAIN, 1.7 is even more broken and
61 just hangs. work around both issues in a cygwin-specific hack.
62 - improve idle watcher documentation slightly.
63
645.23 Sun Dec 20 23:48:00 CET 2009
65 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
66 - implement punycode_encode/decode, idn_nameprep,
67 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
68 - provide $AE::VERSION.
69 - removed traces of "no strict 'refs'".
70
715.22 Sat Dec 5 03:51:13 CET 2009
72 - downgrade-or-fail in AnyEvent::Handle::push_write, to
73 diagnose encoding failures earlier and more succinctly.
74 (this works around bugs in perl, throwing away encoding info
75 when passing scalar data to extensions).
76 - add more examples to AnyEvent::Socket manpage.
77 - upgrade internal warning set to the same as common::sense 2.03.
78 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
79 (John Beppu).
80 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
81 - add google public dns servers to fallback server set and make sure
82 we load-balance properly between the three sets. also add all
83 fallback dns servers, not just a random one, to each dns config.
84
855.21 Thu Nov 19 02:48:47 CET 2009
86 - fix a problem where socket constants were called with parameters
87 (spotted by David Friedland).
88 - fork_call never use'd POSIX (reported by Daisuke Maki).
89 - improve perl 5.6 compatibility further (but it still won't work
90 unless you rip out everything but the core).
91 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
92 uses 10MB of RAM, but doesn't open a console window. *sigh*.
93
945.202 Wed Oct 14 22:35:44 CEST 2009
95 - AnyEvent::DNS would unexpectedly clobber $_ under windows
96 (analysed by Matthias Waldorf).
97 - AnyEvent::Handle::run_cmd can now pass the PID of the
98 newly-created process, which is much less useful than it might
99 sound (based on patch by Yann Kerherve).
100
1015.201 Tue Sep 29 12:09:25 CEST 2009
102 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
103 (reported by Torsten Foertsch).
104 - common::sense 2.0 could cause tcp_server to throw an exception
105 (analysed by elmex).
106
1075.2 Mon Sep 14 07:04:49 CEST 2009
108 - INCOMPATIBLE CHANGE: do no longer support register_read_type
109 and register_write_type in AnyEvent::Handle, instead support
110 package names (the facility was mostly abused).
111 - implement "packagename-as-read/write type" support in
112 AnyEvent::Handle.
113 - AnyEvent::Handle: new options "keepalive" and "oobinline".
114 - oobinline set by default to avoid security issues.
115 - the pure-perl event loop backend wrongly detected times() fallback
116 support (spotted by Pavel Boldin).
117
1185.12 Tue Sep 1 20:26:50 CEST 2009
119 - be more lenient when parsing resolv.conf files, as some people
120 use hashmarks as comment indicator inside directives (reported
121 by Michael S. Fischer).
122 - use same set of warnings as common::sense 2.0.
123 - fix a potential 32 bit overflow issue due to perl having problems
124 with large hex constants in the Perl backend.
125 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
126 by default.
127 - fix some prototypes.
128 - (optionally) require Async::Interrupt 1.02.
129 - replace opendns fallbacks by verizon/level3 public dns servers.
130
1315.112 Fri Aug 21 13:59:04 CEST 2009
132 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
133 using starttls with an illegal/uncreatable tls context.
134 - port to broken win32 perls.
135
1365.111 Sun Aug 16 18:44:36 CEST 2009
137 - perl errornously requires sockaddr_un structures returned by the
138 kernel to have a specific length. work around this bug by padding
139 sockaddr_un structures (found to fail on OS X, as perl only uses
140 the correct code for GNU/Linux, but likely to fail on other
141 OSes as well).
142
1435.11 Wed Aug 12 17:49:37 CEST 2009
144 - fixed a bug in run_cmd with input from filename.
145 - create a json coder object when none is given in write direction
146 to reduce dependency on JSON.
147
1485.1 Tue Aug 11 03:17:32 CEST 2009
149 - tcp_connect now reports the correct error in $!, not an unrelated
150 one (was broken since 4.91).
151 - AE::Handle did not properly clear rtimeout/wtimeout watchers
152 when retrying a tcp_connect.
153 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
154 - fixed wrong documentation in AE::Socket.
155
1565.01 Mon Aug 10 03:16:32 CEST 2009
157 - last release broke EV child watchers :/.
158
1595.0 Sun Aug 9 17:32:09 CEST 2009
160 - officially introduce and document the AE API.
161 - lowering the timeout at runtime did not have immediate effect
162 in AnyEvent::Handle.
163 - AE::Handle now has separate and independent read and write timeout
164 settings.
165 - probe for Guard module when AnyEvent::Util is loaded, not at
166 runtime (Event doesn't like eval "use" from callbacks, and it avoids
167 exporting the probe function).
168 - Impl::IOAsync's io watchers did not use the filehandle emulation
169 code to work around it's watcher limits.
170 - work around IO::Async's problems with overloaded objects.
171 - try to disable wearnings inside IO::Async, it's just too buggy.
172 - do no longer use Event by default, only EV or Perl, Event is too buggy.
173 - hackishly provide fast AE:: interface when EV or Perl is the backend.
174 - inherit VERSION from AnyEvent::VERSION in most modules.
175
1764.91 Thu Aug 6 15:42:45 CEST 2009
177 - AE::Handle::starttls could get out of sync when the read buffer
178 already contains some TLS handshake.
179 - AE::Handle did not properly free the TLS session, causing bigger
180 memory leaks in Net::SSLeay than expected (reported by toaster).
181 - AE::Socket::tcp_connect will now artificially delay invoking
182 the callback to avoid returning after invoking the callback.
183 - convert many internal watcher uses to AE API.
184
1854.9 Sat Aug 1 11:07:01 CEST 2009
186 - Glib silently fails when registering a timeout with
187 a negative value, so avoid that.
188 - call condvar callback immediately when it is set
189 after the condition is already signalled.
190 - check rbuf_max condition only after trying to consume
191 data.
192 - normalise signal numbers to names when using the ae's
193 signal handling, but do not document this (yet).
194 - pure perl signal emulation did not properly set nonblocking
195 mode on the signal pipe on win32 (this is unlikely to have caused
196 any issues).
197 - new module: AnyEvent::Debug.
198 - AnyEvent::Strict now checks that a signal was specified
199 by name (not name or number).
200 - reduce memory footprint in typical cases by ~50kb by
201 only compiling signal code when necessary.
202 - add AnyEvent::Handle->rbuf_max.
203 - grab the AE namespace for future new API, implement stubs for
204 said future/alternative API.
205 - new function: AnyEvent::Socket::format_hostport.
206
2074.881 Tue Jul 28 12:51:53 CEST 2009
208 - work around a bug in local in pre-5.10 perls, causing
209 AnyEvent::Handle to recurse when it should not
210 (analysed by elmex).
211
2124.88 Tue Jul 28 04:04:37 CEST 2009
213 - re-bless the handle into a dummy package after calling
214 AnyEvent::Handle::destroy, so the user does not need to check
215 for errors after every push_write etc.
216 - do not attempt to run t/02_signals.t on obviously broken
217 platforms.
218
2194.87 Sun Jul 26 02:06:16 CEST 2009
220 - do not attempt to linger when there is no longer a valid fh.
221 - enforce tls mode to be either accept or connect, do not simply
222 segfault in Net::SSLeay.
223 - AnyEvent::Handle can now call tcp_connect itself (new parameters
224 connect, on_prepare, on_connect and on_connect_error). Updated
225 tutorial accordingly.
226 - add AnyEvent::Impl::Irssi backend.
227
2284.86 Mon Jul 20 23:52:29 CEST 2009
229 - since the verbose warning is not enough, explicitly document
230 that versions before 1.33 of Net::SSLeay are not secure.
231 - work around signal handling races in Event and (...) Event::Lib.
232 - try to align signal-race timer to full-second boundaries.
233 - work around Tk not liking negative timeouts.
234 - don't complain of different grades of environmental unfriendlyness
235 in IO::Async.
236
2374.85 Sat Jul 18 06:16:14 CEST 2009
238 - nail the signal race problem in perl once and for all
239 (see $AnyEvent::MAX_SIGNAL_LATENCY).
240 - take advantage of Async::Interrupt if it is available.
241 - load Time::HiRes and Guard modules on demand only.
242 - add optional/recommended modules section to AnyEvent
243 documentation.
244 - reduce memory usage considerably (and reduce startup penalty)
245 by not using "strict", "warnings" and "overload" modules.
246 - work around buggy windows/openbsd perls and provide EBADMSG
247 and EPROTO ourselves when missing.
248 - improve perl 5.6 compatibility of the core event loop.
249 - made Net::SSLeay version 1.33 a soft requirement.
250
2514.83 Fri Jul 17 16:56:26 CEST 2009
252 - implement AnyEvent::Socket::getprotobyname.
253 - AnyEvent::CondVar's will now detect recursive blocking
254 waits and will croak, as too many people fall into
255 this trap.
256 - AnyEvent::Handle will now call ->destroy on itself after
257 executing the on_error callback, instead of doing some
258 half-baked internal shutdown, for fatal errors.
259 - clarify on_eof behaviour w.r.t. the read queue and
260 on_read callbacks.
261 - ignore some possible spurious wake-ups in tcp_connect.
262
2634.82 Sat Jul 11 00:34:55 CEST 2009
264 - POE and Event backends didn't accept some callable objects as
265 callbacks.
266 - use Config module instead of POSIX module to detect signal names
267 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
268 module doesn't even have all POSIX signals :/.
269 - use more workarounds around the many refcnt/corruption bugs in
270 Event::Lib.
271 - work around a race condition in perl's select, causing t/03_child.t
272 to rarely fail.
273
2744.81 Thu Jul 9 10:30:30 CEST 2009
275 - AnyEvent::Handle didn't properly diagnose write errors
276 (it expected -1 from syswrite, how lame... :).
277 - support file descriptors in addition to file handles
278 in AnyEvent->io.
279 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
280 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
281 and PERL_ANYEVENT_CA_PATH.
282 - provide a sensible synopsis section for AnyEvent::TLS.
283 - add a "supported backends" section to the manpage.
284 - added simple io watcher test to testsuite, using a
285 portable_socketpair.
286 - tried to improve the stability of the Event::Lib backend,
287 YMMV.
288
2894.8 Mon Jul 6 23:45:16 CEST 2009
290 - AnyEvent::DNS did not properly follow CNAME records with
291 uppercase targets.
292 - AnyEvent::DNS would errornously return AAAA records
293 with v4 mapped addresses (a faulty record) as ipv4 addresses,
294 causing AnyEvent::Socket to throw an exception.
295 - added new module AnyEvent::TLS for easier SSL/TLS context
296 creation, with many options including hostname verification,
297 secure default configuration, lots of documentation and,
298 predefined diffie-hellman keys for perfect forward security
299 and much more. get it while it's still fresh!
300 - use AnyEvent::TLS in AnyEvent::Handle for context management.
301 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
302 so AnyEvent::Socket users have smaller memory footprint
303 in the common case.
304 - add AnyEvent::Handle->push_shutdown method.
305 - add an additional error message parameter to AnyEvent::Handle's
306 on_error callback (for TLS, $! is still available).
307 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
308 - make AnyEvent::Handle more robust against early conenction
309 failures (during new), and return C<undef> in that case
310 from the constructor.
311 - AnyEvent::Handle will now try to load only JSON::XS first,
312 then fall back to JSON.
313 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
314 and because it was documented that way.
315
3164.451 Fri Jul 3 00:28:58 CEST 2009
317 - do not clear rbuf when shutting down an AnyEvent::Handle
318 object - doing so breaks AnyEvent::HTTP.
319
3204.45 Mon Jun 29 22:59:26 CEST 2009
321 - a write error could cause AnyEvent::Handle to create
322 an I/O watcher with an undefined $fh.
323 - special-case mapped ipv4 addresses in both
324 AnyEvent::Socket::format_address and parse_address,
325 to treat them just like ipv4 addresses.
326 - updated and overhauled the AnyEvent::Intro doc.
327 - implement AnyEvent::Socket::format_ipv4/ipv6.
328 - slightly speed up portable_pipe/socketpair.
329 - expand condvar begin/end documentation.
330
3314.42 Fri Jun 26 08:32:18 CEST 2009
332 - preliminary and neccesarily incomplete support for IO::Async.
333 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
334 case it was set to IGNORE, to ensure we can catch child statuses
335 even when the calling env acts stupidly.
336 - updated benchmarks with IO::Async, which performs very well.
337
3384.412 Wed Jun 24 01:35:57 CEST 2009
339 - support an "untaint" attribute for AnyEvent::DNS
340 and set it on the default resolver.
341 - implement and document AnyEvent::Impl::Perl::loop.
342 - remove all anyevent-env variables from %ENV
343 when running in tainted mode.
344 - mention and extend the IO::Lambda benchmark.
345
3464.411 Sun Jun 7 18:48:02 CEST 2009
347 - do not try to use F_SETFD on windows.
348
3494.41 Thu May 14 06:40:11 CEST 2009
350 - work around issues in older perls (5.8.0?) when
351 a signal handler is deleted from the %SIG hash.
352 - use POSIX::_exit in child test, to avoid running
353 destructors.
354 - speed up CHLD handling by relying on SIGCHLD being
355 synchronously delivered, even when we roll our own
356 implementation.
357 - AnyEvent::DNS: add the "dname" resource record name
358 for cosmetic reasons.
359
3604.4 Sun Apr 26 20:12:33 CEST 2009
361 - implemented idle watchers, where applicable.
362 - AnyEvent->time died when Event backend was in use.
363 - fix a memleak in the Tk backend.
364 - sped up Tk timer handling.
365 - clip negative "after" values to 0 in AnyEvent::Impl::Event
366 to avoid spamming out warning messages.
367 - fix Qt timers without interval.
368 - avoid Qt zero-timeout special case, in old libqt's.
369
3704.352 Mon Apr 20 16:31:11 CEST 2009
371 - fix AnyEvent::Strict error messages for child watchers.
372 - fix/update Makefile.PL "recommends" versions.
373 - add AnyEvent->now_update.
374
3754.351 Sat Apr 11 07:56:14 CEST 2009
376 - actually make the signal pipe work under win32.
377 - localise $! in signal handler to avoid changing $!, although
378 perl itself does not seem to save/restore errno either.
379 - set the cloexec flag on the signal pipe (normally set by perl too).
380
3814.35 Fri Mar 27 11:48:20 CET 2009
382 - event models relying on AnyEvent's signal watcher emulation did
383 invoke the callback asynchronously, contradicting documentation
384 and causing signals to get lost (this includes AnyEvent's own
385 event loop). AnyEvent now uses the standard pipe trick to make
386 callback execution synchronous to the event loop.
387 - AnyEvent::Handle didn't free TLS context data on DESTROY
388 (patch by Pavel Shaydo).
389 - work around the ever-incompatibly-changing API of MakeMaker.
390 - document that changing global variables without restoring them
391 is a bad idea in Perl (noted by Adam Rosenstein).
392 - AnyEvent::Strict now barfs if ->io is passed a file.
393
3944.34 Thu Feb 12 18:32:45 CET 2009
395 - separately buffer TLS read data, as otherwise the read queue could
396 deadlock as receiving data is not expected while draining the
397 read queue (which cna only happen with TLS).
398 - raise EBADMSG error on JSON decoding errors.
399 - fix some minor manpage bugs (reported by Maximilian Gaß).
400 - speed up select bitmask parsing quite a bit in the pure perl
401 backend.
402 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
403
4044.331 Tue Jan 6 21:07:25 CET 2009
405 - socketpair fails on many vista machines because vista has
406 completely broken accept/getpeername and getsockname functions,
407 so we provide our own socketpair emulation that kind of works
408 (AnyEvent::Util::portable_pipe).
409 - new function: AnyEvent::Util::portable_socketpair.
410 - take advantage of the Guard module if it exists.
411
4124.33 Fri Nov 21 02:35:40 CET 2008
413 - AnyEvent::Strict did errornously flag a fileno of 0 as
414 illegal.
415 - reduce memory usage and slightly speed up the pure perl backend
416 by only storing the file descriptor, not the file handle.
417 - add missing autocork method to AnyEvent::Handle (reported
418 by Adam Rosenstein).
419 - AnyEvent::DNS->resolve errornously documented an $rcode result
420 argument, but there isn't (spotted by Henrik Krohns).
421 - the naked truth about Tk - it's basically unmaintained.
422
4234.32 Mon Nov 3 22:46:32 CET 2008
424 - fix AnyEvent::Socket::resolve_sockaddr to properly support
425 unix sockets again and choose a proper default for
426 the socket type.
427 - fix call to resolve_sockaddr in tcp_connect.
428
4294.31 Thu Oct 30 04:41:48 CET 2008
430 - implemented AnyEvent::Handle->destroy method.
431 - hint about unexpected effects in TLS mode.
432 - speed up AnyEvent::Impl::Perl by using more arrays
433 instead of hashes without the slightest loss of readability :)
434 - work around a perl argument refcounting bug.
435 - with some perl patches applied, AnyEvent now seems leak-free.
436
4374.3 Fri Oct 3 09:18:43 CEST 2008
438 - AnyEvent will now install a no-op signal handler for SIGPIPE,
439 unless one has been installed already.
440 - warn about not loaidng AnyEvent::Impl::POE early enough
441 (patch by Adam Rosenstein).
442 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
443 - fix a bug in where in SSL connect mode, the
444 client would first wait for some data by the server and
445 otherwise hang. (reported and analysed in an absolutely
446 exemplary manner by Adam Rosenstein).
447 - fix a bug in where SSL EOF would not be treated
448 as stream EOF, putting the connection into a hung state.
449 (reported and analysed in an absolutely exemplary manner by Adam
450 Rosenstein).
451 - fix a potential segfault when the TLS context
452 would go missing in a read callback (e.g. due to stoptls) -
453 Net::SSLeay of course makes no type checking whatsoever.
454 - AnyEvent::Handle will not stop the read watcher in TLS mode.
455 - AnyEvent::Handle->stoptls will send a TLS close notify
456 instead of simply destroying the stream now.
457 - fix a bug where large blocks of data written in TLS mode
458 would not be sent unless triggered by receives.
459 - on_drain will now take the tls write buffer into account.
460 - SSL operations have been streamlined a bit and should be faster
461 now, more to come, though.
462 - added a FAQ section, will document nontrivial issues in there.
463
4644.234 Mon Sep 29 04:08:13 CEST 2008
465 - fix child watcher documentation: only child exits will be reported,
466 no trace events.
467 - mention SIGPIPE in AnyEvent::Handle.
468 - perl backend will now use POSIX::times as monotonic clock
469 when available and the monotonic clock option is not.
470 - run even without Time::HiRes in some cases now, but do not
471 provide sub-second accuracy (all *supported* perl versions have
472 Time::HiRes).
473
4744.233 Fri Aug 22 01:48:05 CEST 2008
475 - fix a bug in the testsuite, causing a hang.
476 - clarified lots of AnyEvent::Handle constructor arguments.
477
4784.232 Thu Aug 21 20:44:25 CEST 2008
479 - fix a bug in AnyEvent::Handle that could cause two fatal error
480 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
481 (testcase by Leon Brocard).
482 - clarified on_eof/on_error documentation a bit.
483
4844.231 Tue Jul 29 13:12:15 CEST 2008
485 - remove some debugging code left in AnyEvent::Util::fork_call
486 (and no, it's impossible to implement with the broken windows
487 perls without resource leaks or worse).
488
4894.23 Tue Jul 29 12:19:59 CEST 2008
490 - document the first parameter passed to condvar callbacks to be
491 the callback.
492 - add AnyEvent::Socket::{ntoa,aton} aliases.
493 - optimize the AE::Handle->push_read (line) for the default
494 eol marker.
495 - optimize push_read (packstring|storable) for small packets.
496 - invoke on_error callback when no on_eof callback is set.
497 - fix a bug in push_read (storable) of unknown impact.
498
4994.22 Sun Jul 20 16:34:13 CEST 2008
500 - new function AnyEvent::Socket::parse_hostport.
501 - as the bulkheads at microsoft can't even get getprotobyname reliably
502 working on their shitty fucking broken os we need to hardcode
503 some common protocol numbers in AnyEvent::Socket. How can
504 people even bother with such a pile of shit as windows.
505
5064.21 Thu Jul 17 14:40:05 CEST 2008
507 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
508 which is incompatible to previous invocation syntax.
509 - work around more windows perl bugs in fork_call. I cannot imagine
510 why anybody would bother implementing fork in such an obviously
511 unusable way.
512 - avoid calling $do_retry if the request has already finished in
513 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
514 K. Sharifulin).
515
5164.2 Sat Jul 12 22:42:11 CEST 2008
517 - support an "interval" argument for repeating timers.
518 - fix a bug in the pure perl event loop that caused timers to
519 be delayed under some circumstances.
520 - implement optional argument checking via PERL_ANYEVENT_STRICT.
521 - not importing anything from Event causes it to malfunction
522 and not use Time::HiRes, work around this bug.
523 - more bug workarounds against the endlessly broken Tk module,
524 seems to work now on 32 bit machines, but less so on 64 bit.
525 - minor optimisations applied to most backends.
526
5274.161 Fri Jul 4 14:24:48 CEST 2008
528 - fixed the t/handle/02_write.t, it read after write, with the
529 assumption that all data could be written before read
530 (reported by Jim Razmus).
531
5324.160 Thu Jul 3 04:02:21 CEST 2008
533 - re-registering signal watchers after unregistering one did
534 not work if a backend falls back on AnyEvent's default
535 implementation (also affected child watchers).
536 - new AnyEvent::Handle options: autocork and no_delay.
537
5384.152 Sun Jun 22 14:15:44 CEST 2008
539 - allow for 32-bit perls that implement shifts differently
540 on different architectures in parse_ipv4 (reported and
541 analysed by Keiichi DAIBA).
542
5434.151 Fri Jun 6 17:34:24 CEST 2008
544 - make sure specifying _only_ on_read and never pushing reads
545 works.
546
5474.15 Fri Jun 6 13:00:46 CEST 2008
548 - the pure perl backend would keep some watchers alive when more than
549 one watcher was registered for the same fd.
550 - new "packstring" and "storable" read and write types
551 for AnyEvent::Handle.
552 - allow on_eof handler to be called after on_error with EPIPE returns.
553 - do not immediately call on_read callback in handle constructor.
554
5554.14 Thu Jun 5 20:29:31 CEST 2008
556 - Fixed a bug in DNS SRV priority sorting.
557 - AnyEvent::Util::guard now reports runtime errors while
558 executing the guard block as warnings.
559 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
560
5614.13 Thu Jun 5 00:47:59 CEST 2008
562 - AnyEvent::DNS only followed cname chains with length 2,
563 contrary to documentation. bumped it up to 10, thanks to
564 microsoft, the current broken-dns-config-king.
565 - AnyEvent::DNS didn't check for socket return status
566 but instead relied on perl not creating filehandles in that
567 case - too bad it gives you a bogus file handle (reported
568 and analysed by Vladimir Timofeev).
569 - fix queue management logic in AnyEvent::Handle:
570 when on_read was registered, the queue was empty
571 and no progress could be made AnyEvent::Handle would
572 enter an endless loop.
573 - correctly start reading again when the handle became
574 busy again after idling.
575 - correctly treat tls shut-downs as EOF condition
576 (for the time being, would be nice to have a callback for that).
577 - correctly call eof callback under all conditions
578 (hopefully).
579 - the timeout callback did not expect that $self can go
580 away any time.
581
5824.12 Tue Jun 3 10:58:04 CEST 2008
583 - include AnyEvent::Intro, a tutorial for anyevent,
584 anyevent::socket and anyevent::handle.
585 - allow more options in on_error.
586
5874.11 Fri May 30 23:42:25 CEST 2008
588 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
589 reverse_lookup and reverse_verify to replace it, support
590 v4mapped and v4compat addresses.
591 - provide more documentation for the resolver class.
592 - really replace longest run of :0: by :: in format_address,
593 also properly convert :: and ::1 again.
594 - support NAPTR record name and decode it.
595 - implement random weight sampling for SRV records, as per
596 rfc 2782.
597 - correctly abort on srv-record targets of ".".
598 - added AnyEvent::DNS::wait_for_slot.
599 - in the unlikely event of a virtual circuit connection
600 being invalidated by a delayed udp reply, AnyEvent::DNS
601 could die. this has been fixed.
602 - plug a probable memleak in the DNS vc code.
603 - use configured timeout also for VC request phase.
604 - implement timeout and max_outstanding methods
605 for resolver class.
606 - update version numbers in all modules.
607
6084.1 Thu May 29 05:45:40 CEST 2008
609 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
610 which mirrors their purpose better (old names still available
611 for a while). Also moved AnyEvent::DNS::addr to
612 AnyEvent::Socket::resolve_sockaddr.
613 - implement AnyEvent->time and AnyEvent->now.
614 - fix IPv6 support in pack_sockaddr.
615 - officially un-experimentalise all newly introduced code.
616 - support unix domain sockets everywhere by specifying
617 a host of "unix/" and the pathname as service.
618 - implement an activity timeout in AnyEvent::Handle.
619 - added a regex read type to AnyEvent::Handle.
620 - added a json read/write type to AnyEvent::Handle.
621 - always croak in on_error in AnyEvent::Handle.
622 - document how to subclass AnyEvent::Handle.
623 - implement AnyEvent::Util::fork_call.
624 - add support for IPv6 nameservers and nameserver statements.
625 - work around _yet_ _another_ windows perl bug in where empty
626 select masks cause errors in the pure perl backend (this
627 workaround cannot be implemented for the other event loops,
628 but EV also works around this).
629 - supply AnyEvent::Socket::address_type and make good use of it.
630 - clarify and fix the Handle documentation w.r.t $self vs. $handle
631 vs. callback arguments.
632 - add some recommends to the META.yml.
633
6344.05 Mon May 26 19:44:06 CEST 2008
635 - some platforms ignore the rfc and prepend an extra sa_len member to
636 the sockaddr structure, cater for those. also use sockaddr_family,
637 didn't know it was there, but comes in handy for the workaround.
638 - undo the import WIN32 hack.
639
6404.04 Mon May 26 08:03:31 CEST 2008
641 - try to work around yet another windows bug: failed connects are reported
642 as if it were out-of-band data. windows users: you suck. Only
643 EV, Glib, Event and the pure perl backend can handle this condition.
644 - optimize write algorithm in AnyEvent::Handle.
645 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
646 - tcp_connect now properly iterates through all targets.
647 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
648 support entirely. also hardcode some AF_INET6 constants for the
649 sake of perl 5.8.
650 - vastly improved nameserver/suffix detection algorithm for windows
651 (still a hack...)
652 - try to use a fallback dns server if no dns servers could be found.
653 - splatter around some local $SIG{__DIE__}'s to work around bad
654 die handlers.
655
6564.03 Sun May 25 05:05:57 CEST 2008
3 - work around perl bugs on windows, where perl returns undocumented 657 - work around perl bugs on windows, where perl returns undocumented
4 error codes for sysread, syswrite etc. 658 error codes for sysread, syswrite etc.
5 - fix AnyEvent::Handle::unshift_read to really unshift and not push. 659 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
6 - AnyEvent::Handle could sometimes stumble over 0-byte writes. 660 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
661 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
7 - updated documentation. 662 - updated documentation.
663 - lots of workarounds for perl 5.8 quirks.
8 - implement a more flexible type system for both reads and writes 664 - implement a more flexible type system for both reads and writes
9 in AnyEvent::Handle. 665 in AnyEvent::Handle.
666 - support IPv6 in tcp_server.
667 - pass local hostname and port to the prepare callback in tcp_server.
10 - make types extendable in AnyEvent::Handle. 668 - make types extendable in AnyEvent::Handle.
11 - croak, not die, in AnyEvent::Handle; 669 - croak, not die, in AnyEvent::Handle;
12 670
134.0 Sat May 24 19:58:08 CEST 2008 6714.0 Sat May 24 19:58:08 CEST 2008
14 - added (experimental) AnyEvent::DNS module. 672 - added (experimental) AnyEvent::DNS module.
… …
22 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. 680 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
23 - try very hard not to freeze on broken windows perls. 681 - try very hard not to freeze on broken windows perls.
24 - make on_eof optional in AnyEvent::Handle. 682 - make on_eof optional in AnyEvent::Handle.
25 - support service names in tcp_connect and tcp_server. 683 - support service names in tcp_connect and tcp_server.
26 - work around netbsd/openbsd bug causing unix sockets not to 684 - work around netbsd/openbsd bug causing unix sockets not to
27 become writable untill empty. 685 become writable until empty.
28 - only make a dummy read for error detection when getpeername 686 - only make a dummy read for error detection when getpeername
29 returns ENOTCONN. 687 returns ENOTCONN.
30 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 688 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
31 override. 689 override.
32 - new AnyEvent::Guard::cancel. 690 - new AnyEvent::Guard::cancel.
33 691
343.5 Sat May 17 23:17:14 CEST 2008 6923.5 Sat May 17 23:17:14 CEST 2008
35 - gracefully deal with systems not defining CLOCK_MONOTONIC 693 - gracefully deal with systems not defining CLOCK_MONOTONIC
… …
49 707
503.41 Sun May 11 19:53:13 CEST 2008 7083.41 Sun May 11 19:53:13 CEST 2008
51 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 709 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
52 710
533.4 Sun May 11 00:29:25 CEST 2008 7113.4 Sun May 11 00:29:25 CEST 2008
54 - complete re-implementation of condvars with a hopefully much mroe useful 712 - complete re-implementation of condvars with a hopefully much more useful
55 API (backward compatible functions exist but are undocumented). 713 API (backward compatible functions exist but are undocumented).
56 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 714 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
57 (still subject to change). 715 (still subject to change).
58 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 716 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
59 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 717 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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