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

Comparing AnyEvent/Changes (file contents):
Revision 1.96 by root, Sat May 24 22:27:11 2008 UTC vs.
Revision 1.379 by root, Wed Feb 10 13:33:43 2010 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
4 - the AnyEvent::Util::fh_nonblocking function worked only
5 by lucky accident on win32.
6 - smaller and faster AnyEvent::Util::fh_nonblocking.
7 - when the (required!) Time::HiRes module is missing, AnyEvent
8 did not fall back to built-in time properly.
9
105.24 Tue Jan 5 11:39:43 CET 2010
11 - cygwin never reports errors from failed connects. 1.5 just gives
12 you continous readyness and EAGAIN, 1.7 is even more broken and
13 just hangs. work around both issues in a cygwin-specific hack.
14 - improve idle watcher documentation slightly.
15
165.23 Sun Dec 20 23:48:00 CET 2009
17 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
18 - implement punycode_encode/decode, idn_nameprep,
19 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
20 - provide $AE::VERSION.
21 - removed traces of "no strict 'refs'".
22
235.22 Sat Dec 5 03:51:13 CET 2009
24 - downgrade-or-fail in AnyEvent::Handle::push_write, to
25 diagnose encoding failures earlier and more succinctly.
26 (this works around bugs in perl, throwing away encoding info
27 when passing scalar data to extensions).
28 - add more examples to AnyEvent::Socket manpage.
29 - upgrade internal warning set to the same as common::sense 2.03.
30 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
31 (John Beppu).
32 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
33 - add google public dns servers to fallback server set and make sure
34 we load-balance properly between the three sets. also add all
35 fallback dns servers, not just a random one, to each dns config.
36
375.21 Thu Nov 19 02:48:47 CET 2009
38 - fix a problem where socket constants were called with parameters
39 (spotted by David Friedland).
40 - fork_call never use'd POSIX (reported by Daisuke Maki).
41 - improve perl 5.6 compatibility further (but it still won't work
42 unless you rip out everything but the core).
43 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
44 uses 10MB of RAM, but doesn't open a console window. *sigh*.
45
465.202 Wed Oct 14 22:35:44 CEST 2009
47 - AnyEvent::DNS would unexpectedly clobber $_ under windows
48 (analysed by Matthias Waldorf).
49 - AnyEvent::Handle::run_cmd can now pass the PID of the
50 newly-created process, which is much less useful than it might
51 sound (based on patch by Yann Kerherve).
52
535.201 Tue Sep 29 12:09:25 CEST 2009
54 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
55 (reported by Torsten Foertsch).
56 - common::sense 2.0 could cause tcp_server to throw an exception
57 (analysed by elmex).
58
595.2 Mon Sep 14 07:04:49 CEST 2009
60 - INCOMPATIBLE CHANGE: do no longer support register_read_type
61 and register_write_type in AnyEvent::Handle, instead support
62 package names (the facility was mostly abused).
63 - implement "packagename-as-read/write type" support in
64 AnyEvent::Handle.
65 - AnyEvent::Handle: new options "keepalive" and "oobinline".
66 - oobinline set by default to avoid security issues.
67 - the pure-perl event loop backend wrongly detected times() fallback
68 support (spotted by Pavel Boldin).
69
705.12 Tue Sep 1 20:26:50 CEST 2009
71 - be more lenient when parsing resolv.conf files, as some people
72 use hashmarks as comment indicator inside directives (reported
73 by Michael S. Fischer).
74 - use same set of warnings as common::sense 2.0.
75 - fix a potential 32 bit overflow issue due to perl having problems
76 with large hex constants in the Perl backend.
77 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
78 by default.
79 - fix some prototypes.
80 - (optionally) require Async::Interrupt 1.02.
81 - replace opendns fallbacks by verizon/level3 public dns servers.
82
835.112 Fri Aug 21 13:59:04 CEST 2009
84 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
85 using starttls with an illegal/uncreatable tls context.
86 - port to broken win32 perls.
87
885.111 Sun Aug 16 18:44:36 CEST 2009
89 - perl errornously requires sockaddr_un structures returned by the
90 kernel to have a specific length. work around this bug by padding
91 sockaddr_un structures (found to fail on OS X, as perl only uses
92 the correct code for GNU/Linux, but likely to fail on other
93 OSes as well).
94
955.11 Wed Aug 12 17:49:37 CEST 2009
96 - fixed a bug in run_cmd with input from filename.
97 - create a json coder object when none is given in write direction
98 to reduce dependency on JSON.
99
1005.1 Tue Aug 11 03:17:32 CEST 2009
101 - tcp_connect now reports the correct error in $!, not an unrelated
102 one (was broken since 4.91).
103 - AE::Handle did not properly clear rtimeout/wtimeout watchers
104 when retrying a tcp_connect.
105 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
106 - fixed wrong documentation in AE::Socket.
107
1085.01 Mon Aug 10 03:16:32 CEST 2009
109 - last release broke EV child watchers :/.
110
1115.0 Sun Aug 9 17:32:09 CEST 2009
112 - officially introduce and document the AE API.
113 - lowering the timeout at runtime did not have immediate effect
114 in AnyEvent::Handle.
115 - AE::Handle now has separate and independent read and write timeout
116 settings.
117 - probe for Guard module when AnyEvent::Util is loaded, not at
118 runtime (Event doesn't like eval "use" from callbacks, and it avoids
119 exporting the probe function).
120 - Impl::IOAsync's io watchers did not use the filehandle emulation
121 code to work around it's watcher limits.
122 - work around IO::Async's problems with overloaded objects.
123 - try to disable wearnings inside IO::Async, it's just too buggy.
124 - do no longer use Event by default, only EV or Perl, Event is too buggy.
125 - hackishly provide fast AE:: interface when EV or Perl is the backend.
126 - inherit VERSION from AnyEvent::VERSION in most modules.
127
1284.91 Thu Aug 6 15:42:45 CEST 2009
129 - AE::Handle::starttls could get out of sync when the read buffer
130 already contains some TLS handshake.
131 - AE::Handle did not properly free the TLS session, causing bigger
132 memory leaks in Net::SSLeay than expected (reported by toaster).
133 - AE::Socket::tcp_connect will now artificially delay invoking
134 the callback to avoid returning after invoking the callback.
135 - convert many internal watcher uses to AE API.
136
1374.9 Sat Aug 1 11:07:01 CEST 2009
138 - Glib silently fails when registering a timeout with
139 a negative value, so avoid that.
140 - call condvar callback immediately when it is set
141 after the condition is already signalled.
142 - check rbuf_max condition only after trying to consume
143 data.
144 - normalise signal numbers to names when using the ae's
145 signal handling, but do not document this (yet).
146 - pure perl signal emulation did not properly set nonblocking
147 mode on the signal pipe on win32 (this is unlikely to have caused
148 any issues).
149 - new module: AnyEvent::Debug.
150 - AnyEvent::Strict now checks that a signal was specified
151 by name (not name or number).
152 - reduce memory footprint in typical cases by ~50kb by
153 only compiling signal code when necessary.
154 - add AnyEvent::Handle->rbuf_max.
155 - grab the AE namespace for future new API, implement stubs for
156 said future/alternative API.
157 - new function: AnyEvent::Socket::format_hostport.
158
1594.881 Tue Jul 28 12:51:53 CEST 2009
160 - work around a bug in local in pre-5.10 perls, causing
161 AnyEvent::Handle to recurse when it should not
162 (analysed by elmex).
163
1644.88 Tue Jul 28 04:04:37 CEST 2009
165 - re-bless the handle into a dummy package after calling
166 AnyEvent::Handle::destroy, so the user does not need to check
167 for errors after every push_write etc.
168 - do not attempt to run t/02_signals.t on obviously broken
169 platforms.
170
1714.87 Sun Jul 26 02:06:16 CEST 2009
172 - do not attempt to linger when there is no longer a valid fh.
173 - enforce tls mode to be either accept or connect, do not simply
174 segfault in Net::SSLeay.
175 - AnyEvent::Handle can now call tcp_connect itself (new parameters
176 connect, on_prepare, on_connect and on_connect_error). Updated
177 tutorial accordingly.
178 - add AnyEvent::Impl::Irssi backend.
179
1804.86 Mon Jul 20 23:52:29 CEST 2009
181 - since the verbose warning is not enough, explicitly document
182 that versions before 1.33 of Net::SSLeay are not secure.
183 - work around signal handling races in Event and (...) Event::Lib.
184 - try to align signal-race timer to full-second boundaries.
185 - work around Tk not liking negative timeouts.
186 - don't complain of different grades of environmental unfriendlyness
187 in IO::Async.
188
1894.85 Sat Jul 18 06:16:14 CEST 2009
190 - nail the signal race problem in perl once and for all
191 (see $AnyEvent::MAX_SIGNAL_LATENCY).
192 - take advantage of Async::Interrupt if it is available.
193 - load Time::HiRes and Guard modules on demand only.
194 - add optional/recommended modules section to AnyEvent
195 documentation.
196 - reduce memory usage considerably (and reduce startup penalty)
197 by not using "strict", "warnings" and "overload" modules.
198 - work around buggy windows/openbsd perls and provide EBADMSG
199 and EPROTO ourselves when missing.
200 - improve perl 5.6 compatibility of the core event loop.
201 - made Net::SSLeay version 1.33 a soft requirement.
202
2034.83 Fri Jul 17 16:56:26 CEST 2009
204 - implement AnyEvent::Socket::getprotobyname.
205 - AnyEvent::CondVar's will now detect recursive blocking
206 waits and will croak, as too many people fall into
207 this trap.
208 - AnyEvent::Handle will now call ->destroy on itself after
209 executing the on_error callback, instead of doing some
210 half-baked internal shutdown, for fatal errors.
211 - clarify on_eof behaviour w.r.t. the read queue and
212 on_read callbacks.
213 - ignore some possible spurious wake-ups in tcp_connect.
214
2154.82 Sat Jul 11 00:34:55 CEST 2009
216 - POE and Event backends didn't accept some callable objects as
217 callbacks.
218 - use Config module instead of POSIX module to detect signal names
219 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
220 module doesn't even have all POSIX signals :/.
221 - use more workarounds around the many refcnt/corruption bugs in
222 Event::Lib.
223 - work around a race condition in perl's select, causing t/03_child.t
224 to rarely fail.
225
2264.81 Thu Jul 9 10:30:30 CEST 2009
227 - AnyEvent::Handle didn't properly diagnose write errors
228 (it expected -1 from syswrite, how lame... :).
229 - support file descriptors in addition to file handles
230 in AnyEvent->io.
231 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
232 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
233 and PERL_ANYEVENT_CA_PATH.
234 - provide a sensible synopsis section for AnyEvent::TLS.
235 - add a "supported backends" section to the manpage.
236 - added simple io watcher test to testsuite, using a
237 portable_socketpair.
238 - tried to improve the stability of the Event::Lib backend,
239 YMMV.
240
2414.8 Mon Jul 6 23:45:16 CEST 2009
242 - AnyEvent::DNS did not properly follow CNAME records with
243 uppercase targets.
244 - AnyEvent::DNS would errornously return AAAA records
245 with v4 mapped addresses (a faulty record) as ipv4 addresses,
246 causing AnyEvent::Socket to throw an exception.
247 - added new module AnyEvent::TLS for easier SSL/TLS context
248 creation, with many options including hostname verification,
249 secure default configuration, lots of documentation and,
250 predefined diffie-hellman keys for perfect forward security
251 and much more. get it while it's still fresh!
252 - use AnyEvent::TLS in AnyEvent::Handle for context management.
253 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
254 so AnyEvent::Socket users have smaller memory footprint
255 in the common case.
256 - add AnyEvent::Handle->push_shutdown method.
257 - add an additional error message parameter to AnyEvent::Handle's
258 on_error callback (for TLS, $! is still available).
259 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
260 - make AnyEvent::Handle more robust against early conenction
261 failures (during new), and return C<undef> in that case
262 from the constructor.
263 - AnyEvent::Handle will now try to load only JSON::XS first,
264 then fall back to JSON.
265 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
266 and because it was documented that way.
267
2684.451 Fri Jul 3 00:28:58 CEST 2009
269 - do not clear rbuf when shutting down an AnyEvent::Handle
270 object - doing so breaks AnyEvent::HTTP.
271
2724.45 Mon Jun 29 22:59:26 CEST 2009
273 - a write error could cause AnyEvent::Handle to create
274 an I/O watcher with an undefined $fh.
275 - special-case mapped ipv4 addresses in both
276 AnyEvent::Socket::format_address and parse_address,
277 to treat them just like ipv4 addresses.
278 - updated and overhauled the AnyEvent::Intro doc.
279 - implement AnyEvent::Socket::format_ipv4/ipv6.
280 - slightly speed up portable_pipe/socketpair.
281 - expand condvar begin/end documentation.
282
2834.42 Fri Jun 26 08:32:18 CEST 2009
284 - preliminary and neccesarily incomplete support for IO::Async.
285 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
286 case it was set to IGNORE, to ensure we can catch child statuses
287 even when the calling env acts stupidly.
288 - updated benchmarks with IO::Async, which performs very well.
289
2904.412 Wed Jun 24 01:35:57 CEST 2009
291 - support an "untaint" attribute for AnyEvent::DNS
292 and set it on the default resolver.
293 - implement and document AnyEvent::Impl::Perl::loop.
294 - remove all anyevent-env variables from %ENV
295 when running in tainted mode.
296 - mention and extend the IO::Lambda benchmark.
297
2984.411 Sun Jun 7 18:48:02 CEST 2009
299 - do not try to use F_SETFD on windows.
300
3014.41 Thu May 14 06:40:11 CEST 2009
302 - work around issues in older perls (5.8.0?) when
303 a signal handler is deleted from the %SIG hash.
304 - use POSIX::_exit in child test, to avoid running
305 destructors.
306 - speed up CHLD handling by relying on SIGCHLD being
307 synchronously delivered, even when we roll our own
308 implementation.
309 - AnyEvent::DNS: add the "dname" resource record name
310 for cosmetic reasons.
311
3124.4 Sun Apr 26 20:12:33 CEST 2009
313 - implemented idle watchers, where applicable.
314 - AnyEvent->time died when Event backend was in use.
315 - fix a memleak in the Tk backend.
316 - sped up Tk timer handling.
317 - clip negative "after" values to 0 in AnyEvent::Impl::Event
318 to avoid spamming out warning messages.
319 - fix Qt timers without interval.
320 - avoid Qt zero-timeout special case, in old libqt's.
321
3224.352 Mon Apr 20 16:31:11 CEST 2009
323 - fix AnyEvent::Strict error messages for child watchers.
324 - fix/update Makefile.PL "recommends" versions.
325 - add AnyEvent->now_update.
326
3274.351 Sat Apr 11 07:56:14 CEST 2009
328 - actually make the signal pipe work under win32.
329 - localise $! in signal handler to avoid changing $!, although
330 perl itself does not seem to save/restore errno either.
331 - set the cloexec flag on the signal pipe (normally set by perl too).
332
3334.35 Fri Mar 27 11:48:20 CET 2009
334 - event models relying on AnyEvent's signal watcher emulation did
335 invoke the callback asynchronously, contradicting documentation
336 and causing signals to get lost (this includes AnyEvent's own
337 event loop). AnyEvent now uses the standard pipe trick to make
338 callback execution synchronous to the event loop.
339 - AnyEvent::Handle didn't free TLS context data on DESTROY
340 (patch by Pavel Shaydo).
341 - work around the ever-incompatibly-changing API of MakeMaker.
342 - document that changing global variables without restoring them
343 is a bad idea in Perl (noted by Adam Rosenstein).
344 - AnyEvent::Strict now barfs if ->io is passed a file.
345
3464.34 Thu Feb 12 18:32:45 CET 2009
347 - separately buffer TLS read data, as otherwise the read queue could
348 deadlock as receiving data is not expected while draining the
349 read queue (which cna only happen with TLS).
350 - raise EBADMSG error on JSON decoding errors.
351 - fix some minor manpage bugs (reported by Maximilian Gaß).
352 - speed up select bitmask parsing quite a bit in the pure perl
353 backend.
354 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
355
3564.331 Tue Jan 6 21:07:25 CET 2009
357 - socketpair fails on many vista machines because vista has
358 completely broken accept/getpeername and getsockname functions,
359 so we provide our own socketpair emulation that kind of works
360 (AnyEvent::Util::portable_pipe).
361 - new function: AnyEvent::Util::portable_socketpair.
362 - take advantage of the Guard module if it exists.
363
3644.33 Fri Nov 21 02:35:40 CET 2008
365 - AnyEvent::Strict did errornously flag a fileno of 0 as
366 illegal.
367 - reduce memory usage and slightly speed up the pure perl backend
368 by only storing the file descriptor, not the file handle.
369 - add missing autocork method to AnyEvent::Handle (reported
370 by Adam Rosenstein).
371 - AnyEvent::DNS->resolve errornously documented an $rcode result
372 argument, but there isn't (spotted by Henrik Krohns).
373 - the naked truth about Tk - it's basically unmaintained.
374
3754.32 Mon Nov 3 22:46:32 CET 2008
376 - fix AnyEvent::Socket::resolve_sockaddr to properly support
377 unix sockets again and choose a proper default for
378 the socket type.
379 - fix call to resolve_sockaddr in tcp_connect.
380
3814.31 Thu Oct 30 04:41:48 CET 2008
382 - implemented AnyEvent::Handle->destroy method.
383 - hint about unexpected effects in TLS mode.
384 - speed up AnyEvent::Impl::Perl by using more arrays
385 instead of hashes without the slightest loss of readability :)
386 - work around a perl argument refcounting bug.
387 - with some perl patches applied, AnyEvent now seems leak-free.
388
3894.3 Fri Oct 3 09:18:43 CEST 2008
390 - AnyEvent will now install a no-op signal handler for SIGPIPE,
391 unless one has been installed already.
392 - warn about not loaidng AnyEvent::Impl::POE early enough
393 (patch by Adam Rosenstein).
394 - fixed a great number of bugs and corner cases in AnyEvent::Handle:
395 - fix a bug in where in SSL connect mode, the
396 client would first wait for some data by the server and
397 otherwise hang. (reported and analysed in an absolutely
398 exemplary manner by Adam Rosenstein).
399 - fix a bug in where SSL EOF would not be treated
400 as stream EOF, putting the connection into a hung state.
401 (reported and analysed in an absolutely exemplary manner by Adam
402 Rosenstein).
403 - fix a potential segfault when the TLS context
404 would go missing in a read callback (e.g. due to stoptls) -
405 Net::SSLeay of course makes no type checking whatsoever.
406 - AnyEvent::Handle will not stop the read watcher in TLS mode.
407 - AnyEvent::Handle->stoptls will send a TLS close notify
408 instead of simply destroying the stream now.
409 - fix a bug where large blocks of data written in TLS mode
410 would not be sent unless triggered by receives.
411 - on_drain will now take the tls write buffer into account.
412 - SSL operations have been streamlined a bit and should be faster
413 now, more to come, though.
414 - added a FAQ section, will document nontrivial issues in there.
415
4164.234 Mon Sep 29 04:08:13 CEST 2008
417 - fix child watcher documentation: only child exits will be reported,
418 no trace events.
419 - mention SIGPIPE in AnyEvent::Handle.
420 - perl backend will now use POSIX::times as monotonic clock
421 when available and the monotonic clock option is not.
422 - run even without Time::HiRes in some cases now, but do not
423 provide sub-second accuracy (all *supported* perl versions have
424 Time::HiRes).
425
4264.233 Fri Aug 22 01:48:05 CEST 2008
427 - fix a bug in the testsuite, causing a hang.
428 - clarified lots of AnyEvent::Handle constructor arguments.
429
4304.232 Thu Aug 21 20:44:25 CEST 2008
431 - fix a bug in AnyEvent::Handle that could cause two fatal error
432 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
433 (testcase by Leon Brocard).
434 - clarified on_eof/on_error documentation a bit.
435
4364.231 Tue Jul 29 13:12:15 CEST 2008
437 - remove some debugging code left in AnyEvent::Util::fork_call
438 (and no, it's impossible to implement with the broken windows
439 perls without resource leaks or worse).
440
4414.23 Tue Jul 29 12:19:59 CEST 2008
442 - document the first parameter passed to condvar callbacks to be
443 the callback.
444 - add AnyEvent::Socket::{ntoa,aton} aliases.
445 - optimize the AE::Handle->push_read (line) for the default
446 eol marker.
447 - optimize push_read (packstring|storable) for small packets.
448 - invoke on_error callback when no on_eof callback is set.
449 - fix a bug in push_read (storable) of unknown impact.
450
4514.22 Sun Jul 20 16:34:13 CEST 2008
452 - new function AnyEvent::Socket::parse_hostport.
453 - as the bulkheads at microsoft can't even get getprotobyname reliably
454 working on their shitty fucking broken os we need to hardcode
455 some common protocol numbers in AnyEvent::Socket. How can
456 people even bother with such a pile of shit as windows.
457
4584.21 Thu Jul 17 14:40:05 CEST 2008
459 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
460 which is incompatible to previous invocation syntax.
461 - work around more windows perl bugs in fork_call. I cannot imagine
462 why anybody would bother implementing fork in such an obviously
463 unusable way.
464 - avoid calling $do_retry if the request has already finished in
465 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
466 K. Sharifulin).
467
4684.2 Sat Jul 12 22:42:11 CEST 2008
469 - support an "interval" argument for repeating timers.
470 - fix a bug in the pure perl event loop that caused timers to
471 be delayed under some circumstances.
472 - implement optional argument checking via PERL_ANYEVENT_STRICT.
473 - not importing anything from Event causes it to malfunction
474 and not use Time::HiRes, work around this bug.
475 - more bug workarounds against the endlessly broken Tk module,
476 seems to work now on 32 bit machines, but less so on 64 bit.
477 - minor optimisations applied to most backends.
478
4794.161 Fri Jul 4 14:24:48 CEST 2008
480 - fixed the t/handle/02_write.t, it read after write, with the
481 assumption that all data could be written before read
482 (reported by Jim Razmus).
483
4844.160 Thu Jul 3 04:02:21 CEST 2008
485 - re-registering signal watchers after unregistering one did
486 not work if a backend falls back on AnyEvent's default
487 implementation (also affected child watchers).
488 - new AnyEvent::Handle options: autocork and no_delay.
489
4904.152 Sun Jun 22 14:15:44 CEST 2008
491 - allow for 32-bit perls that implement shifts differently
492 on different architectures in parse_ipv4 (reported and
493 analysed by Keiichi DAIBA).
494
4954.151 Fri Jun 6 17:34:24 CEST 2008
496 - make sure specifying _only_ on_read and never pushing reads
497 works.
498
4994.15 Fri Jun 6 13:00:46 CEST 2008
500 - the pure perl backend would keep some watchers alive when more than
501 one watcher was registered for the same fd.
502 - new "packstring" and "storable" read and write types
503 for AnyEvent::Handle.
504 - allow on_eof handler to be called after on_error with EPIPE returns.
505 - do not immediately call on_read callback in handle constructor.
506
5074.14 Thu Jun 5 20:29:31 CEST 2008
508 - Fixed a bug in DNS SRV priority sorting.
509 - AnyEvent::Util::guard now reports runtime errors while
510 executing the guard block as warnings.
511 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
512
5134.13 Thu Jun 5 00:47:59 CEST 2008
514 - AnyEvent::DNS only followed cname chains with length 2,
515 contrary to documentation. bumped it up to 10, thanks to
516 microsoft, the current broken-dns-config-king.
517 - AnyEvent::DNS didn't check for socket return status
518 but instead relied on perl not creating filehandles in that
519 case - too bad it gives you a bogus file handle (reported
520 and analysed by Vladimir Timofeev).
521 - fix queue management logic in AnyEvent::Handle:
522 when on_read was registered, the queue was empty
523 and no progress could be made AnyEvent::Handle would
524 enter an endless loop.
525 - correctly start reading again when the handle became
526 busy again after idling.
527 - correctly treat tls shut-downs as EOF condition
528 (for the time being, would be nice to have a callback for that).
529 - correctly call eof callback under all conditions
530 (hopefully).
531 - the timeout callback did not expect that $self can go
532 away any time.
533
5344.12 Tue Jun 3 10:58:04 CEST 2008
535 - include AnyEvent::Intro, a tutorial for anyevent,
536 anyevent::socket and anyevent::handle.
537 - allow more options in on_error.
538
5394.11 Fri May 30 23:42:25 CEST 2008
540 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
541 reverse_lookup and reverse_verify to replace it, support
542 v4mapped and v4compat addresses.
543 - provide more documentation for the resolver class.
544 - really replace longest run of :0: by :: in format_address,
545 also properly convert :: and ::1 again.
546 - support NAPTR record name and decode it.
547 - implement random weight sampling for SRV records, as per
548 rfc 2782.
549 - correctly abort on srv-record targets of ".".
550 - added AnyEvent::DNS::wait_for_slot.
551 - in the unlikely event of a virtual circuit connection
552 being invalidated by a delayed udp reply, AnyEvent::DNS
553 could die. this has been fixed.
554 - plug a probable memleak in the DNS vc code.
555 - use configured timeout also for VC request phase.
556 - implement timeout and max_outstanding methods
557 for resolver class.
558 - update version numbers in all modules.
559
5604.1 Thu May 29 05:45:40 CEST 2008
561 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
562 which mirrors their purpose better (old names still available
563 for a while). Also moved AnyEvent::DNS::addr to
564 AnyEvent::Socket::resolve_sockaddr.
565 - implement AnyEvent->time and AnyEvent->now.
566 - fix IPv6 support in pack_sockaddr.
567 - officially un-experimentalise all newly introduced code.
568 - support unix domain sockets everywhere by specifying
569 a host of "unix/" and the pathname as service.
570 - implement an activity timeout in AnyEvent::Handle.
571 - added a regex read type to AnyEvent::Handle.
572 - added a json read/write type to AnyEvent::Handle.
573 - always croak in on_error in AnyEvent::Handle.
574 - document how to subclass AnyEvent::Handle.
575 - implement AnyEvent::Util::fork_call.
576 - add support for IPv6 nameservers and nameserver statements.
577 - work around _yet_ _another_ windows perl bug in where empty
578 select masks cause errors in the pure perl backend (this
579 workaround cannot be implemented for the other event loops,
580 but EV also works around this).
581 - supply AnyEvent::Socket::address_type and make good use of it.
582 - clarify and fix the Handle documentation w.r.t $self vs. $handle
583 vs. callback arguments.
584 - add some recommends to the META.yml.
585
5864.05 Mon May 26 19:44:06 CEST 2008
587 - some platforms ignore the rfc and prepend an extra sa_len member to
588 the sockaddr structure, cater for those. also use sockaddr_family,
589 didn't know it was there, but comes in handy for the workaround.
590 - undo the import WIN32 hack.
591
5924.04 Mon May 26 08:03:31 CEST 2008
593 - try to work around yet another windows bug: failed connects are reported
594 as if it were out-of-band data. windows users: you suck. Only
595 EV, Glib, Event and the pure perl backend can handle this condition.
596 - optimize write algorithm in AnyEvent::Handle.
597 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
598 - tcp_connect now properly iterates through all targets.
599 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
600 support entirely. also hardcode some AF_INET6 constants for the
601 sake of perl 5.8.
602 - vastly improved nameserver/suffix detection algorithm for windows
603 (still a hack...)
604 - try to use a fallback dns server if no dns servers could be found.
605 - splatter around some local $SIG{__DIE__}'s to work around bad
606 die handlers.
607
6084.03 Sun May 25 05:05:57 CEST 2008
609 - work around perl bugs on windows, where perl returns undocumented
610 error codes for sysread, syswrite etc.
3 - fix AnyEvent::Handle::unshift_read to really unshift and not push. 611 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
612 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
613 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
614 - updated documentation.
615 - lots of workarounds for perl 5.8 quirks.
616 - implement a more flexible type system for both reads and writes
617 in AnyEvent::Handle.
618 - support IPv6 in tcp_server.
619 - pass local hostname and port to the prepare callback in tcp_server.
620 - make types extendable in AnyEvent::Handle.
621 - croak, not die, in AnyEvent::Handle;
4 622
54.0 Sat May 24 19:58:08 CEST 2008 6234.0 Sat May 24 19:58:08 CEST 2008
6 - added (experimental) AnyEvent::DNS module. 624 - added (experimental) AnyEvent::DNS module.
7 - added (experimental) AnyEvent::Socket module. 625 - added (experimental) AnyEvent::Socket module.
8 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle. 626 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
… …
14 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. 632 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
15 - try very hard not to freeze on broken windows perls. 633 - try very hard not to freeze on broken windows perls.
16 - make on_eof optional in AnyEvent::Handle. 634 - make on_eof optional in AnyEvent::Handle.
17 - support service names in tcp_connect and tcp_server. 635 - support service names in tcp_connect and tcp_server.
18 - work around netbsd/openbsd bug causing unix sockets not to 636 - work around netbsd/openbsd bug causing unix sockets not to
19 become writable untill empty. 637 become writable until empty.
20 - only make a dummy read for error detection when getpeername 638 - only make a dummy read for error detection when getpeername
21 returns ENOTCONN. 639 returns ENOTCONN.
22 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 640 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
23 override. 641 override.
24 - new AnyEvent::Guard::cancel. 642 - new AnyEvent::Guard::cancel.
25 643
263.5 Sat May 17 23:17:14 CEST 2008 6443.5 Sat May 17 23:17:14 CEST 2008
27 - gracefully deal with systems not defining CLOCK_MONOTONIC 645 - gracefully deal with systems not defining CLOCK_MONOTONIC
… …
41 659
423.41 Sun May 11 19:53:13 CEST 2008 6603.41 Sun May 11 19:53:13 CEST 2008
43 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 661 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
44 662
453.4 Sun May 11 00:29:25 CEST 2008 6633.4 Sun May 11 00:29:25 CEST 2008
46 - complete re-implementation of condvars with a hopefully much mroe useful 664 - complete re-implementation of condvars with a hopefully much more useful
47 API (backward compatible functions exist but are undocumented). 665 API (backward compatible functions exist but are undocumented).
48 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 666 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
49 (still subject to change). 667 (still subject to change).
50 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 668 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
51 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 669 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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