1 |
Revision history for Perl extension AnyEvent. |
2 |
|
3 |
ZERO_RETURN - stoptls? |
4 |
- added new module AnyEvent::TLS for easier SSL/TLS context |
5 |
creation, with many options including hostname verification, |
6 |
secure default configuration, lots of documentation and, |
7 |
predefined diffie-hellman keys for perfect forward security |
8 |
and much more. get it while it's still fresh! |
9 |
- use AnyEvent::TLS in AnyEvent::Handle for context management. |
10 |
- load AnyEvent::Handle only on demand in AnyEvent::DNS, |
11 |
so AnyEvent::Socket users have smaller memory footprint |
12 |
in the common case. |
13 |
- add AnyEvent::Handle->push_shutdown method. |
14 |
- add an additional error message parameter to AnyEvent::Handle's |
15 |
on_error callback (for TLS, $! is still available). |
16 |
- make AnyEvent::Handle more robust against early conenction |
17 |
failures (during new), and return C<undef> in that case |
18 |
from the constructor. |
19 |
- AnyEvent::Handle will now try to load only JSON::XS first, |
20 |
then fall back to JSON. |
21 |
|
22 |
4.451 Fri Jul 3 00:28:58 CEST 2009 |
23 |
- do not clear rbuf when shutting down an AnyEvent::Handle |
24 |
object - doing so breaks AnyEvent::HTTP. |
25 |
|
26 |
4.45 Mon Jun 29 22:59:26 CEST 2009 |
27 |
- a write error could cause AnyEvent::Handle to create |
28 |
an I/O watcher with an undefined $fh. |
29 |
- special-case mapped ipv4 addresses in both |
30 |
AnyEvent::Socket::format_address and parse_address, |
31 |
to treat them just like ipv4 addresses. |
32 |
- updated and overhauled the AnyEvent::Intro doc. |
33 |
- implement AnyEvent::Socket::format_ipv4/ipv6. |
34 |
- slightly speed up portable_pipe/socketpair. |
35 |
- expand condvar begin/end documentation. |
36 |
|
37 |
4.42 Fri Jun 26 08:32:18 CEST 2009 |
38 |
- preliminary and neccesarily incomplete support for IO::Async. |
39 |
- reset SIGCHLD to DEFAULT when AnyEvent is loaded, in |
40 |
case it was set to IGNORE, to ensure we can catch child statuses |
41 |
even when the calling env acts stupidly. |
42 |
- updated benchmarks with IO::Async, which performs very well. |
43 |
|
44 |
4.412 Wed Jun 24 01:35:57 CEST 2009 |
45 |
- support an "untaint" attribute for AnyEvent::DNS |
46 |
and set it on the default resolver. |
47 |
- implement and document AnyEvent::Impl::Perl::loop. |
48 |
- remove all anyevent-env variables from %ENV |
49 |
when running in tainted mode. |
50 |
- mention and extend the IO::Lambda benchmark. |
51 |
|
52 |
4.411 Sun Jun 7 18:48:02 CEST 2009 |
53 |
- do not try to use F_SETFD on windows. |
54 |
|
55 |
4.41 Thu May 14 06:40:11 CEST 2009 |
56 |
- work around issues in older perls (5.8.0?) when |
57 |
a signal handler is deleted from the %SIG hash. |
58 |
- use POSIX::_exit in child test, to avoid running |
59 |
destructors. |
60 |
- speed up CHLD handling by relying on SIGCHLD being |
61 |
synchronously delivered, even when we roll our own |
62 |
implementation. |
63 |
- AnyEvent::DNS: add the "dname" resource record name |
64 |
for cosmetic reasons. |
65 |
|
66 |
4.4 Sun Apr 26 20:12:33 CEST 2009 |
67 |
- implemented idle watchers, where applicable. |
68 |
- AnyEvent->time died when Event backend was in use. |
69 |
- fix a memleak in the Tk backend. |
70 |
- sped up Tk timer handling. |
71 |
- clip negative "after" values to 0 in AnyEvent::Impl::Event |
72 |
to avoid spamming out warning messages. |
73 |
- fix Qt timers without interval. |
74 |
- avoid Qt zero-timeout special case, in old libqt's. |
75 |
|
76 |
4.352 Mon Apr 20 16:31:11 CEST 2009 |
77 |
- fix AnyEvent::Strict error messages for child watchers. |
78 |
- fix/update Makefile.PL "recommends" versions. |
79 |
- add AnyEvent->now_update. |
80 |
|
81 |
4.351 Sat Apr 11 07:56:14 CEST 2009 |
82 |
- actually make the signal pipe work under win32. |
83 |
- localise $! in signal handler to avoid changing $!, although |
84 |
perl itself does not seem to save/restore errno either. |
85 |
- set the cloexec flag on the signal pipe (normally set by perl too). |
86 |
|
87 |
4.35 Fri Mar 27 11:48:20 CET 2009 |
88 |
- event models relying on AnyEvent's signal watcher emulation did |
89 |
invoke the callback asynchronously, contradicting documentation |
90 |
and causing signals to get lost (this includes AnyEvent's own |
91 |
event loop). AnyEvent now uses the standard pipe trick to make |
92 |
callback execution synchronous to the event loop. |
93 |
- AnyEvent::Handle didn't free TLS context data on DESTROY |
94 |
(patch by Pavel Shaydo). |
95 |
- work around the ever-incompatibly-changing API of MakeMaker. |
96 |
- document that changing global variables without restoring them |
97 |
is a bad idea in Perl (noted by Adam Rosenstein). |
98 |
- AnyEvent::Strict now barfs if ->io is passed a file. |
99 |
|
100 |
4.34 Thu Feb 12 18:32:45 CET 2009 |
101 |
- separately buffer TLS read data, as otherwise the read queue could |
102 |
deadlock as receiving data is not expected while draining the |
103 |
read queue (which cna only happen with TLS). |
104 |
- raise EBADMSG error on JSON decoding errors. |
105 |
- fix some minor manpage bugs (reported by Maximilian Gaß). |
106 |
- speed up select bitmask parsing quite a bit in the pure perl |
107 |
backend. |
108 |
- use CORE::select instead of just select in AnyEvent::Impl::Perl. |
109 |
|
110 |
4.331 Tue Jan 6 21:07:25 CET 2009 |
111 |
- socketpair fails on many vista machines because vista has |
112 |
completely broken accept/getpeername and getsockname functions, |
113 |
so we provide our own socketpair emulation that kind of works |
114 |
(AnyEvent::Util::portable_pipe). |
115 |
- new function: AnyEvent::Util::portable_socketpair. |
116 |
- take advantage of the Guard module if it exists. |
117 |
|
118 |
4.33 Fri Nov 21 02:35:40 CET 2008 |
119 |
- AnyEvent::Strict did errornously flag a fileno of 0 as |
120 |
illegal. |
121 |
- reduce memory usage and slightly speed up the pure perl backend |
122 |
by only storing the file descriptor, not the file handle. |
123 |
- add missing autocork method to AnyEvent::Handle (reported |
124 |
by Adam Rosenstein). |
125 |
- AnyEvent::DNS->resolve errornously documented an $rcode result |
126 |
argument, but there isn't (spotted by Henrik Krohns). |
127 |
- the naked truth about Tk - it's basically unmaintained. |
128 |
|
129 |
4.32 Mon Nov 3 22:46:32 CET 2008 |
130 |
- fix AnyEvent::Socket::resolve_sockaddr to properly support |
131 |
unix sockets again and choose a proper default for |
132 |
the socket type. |
133 |
- fix call to resolve_sockaddr in tcp_connect. |
134 |
|
135 |
4.31 Thu Oct 30 04:41:48 CET 2008 |
136 |
- implemented AnyEvent::Handle->destroy method. |
137 |
- hint about unexpected effects in TLS mode. |
138 |
- speed up AnyEvent::Impl::Perl by using more arrays |
139 |
instead of hashes without the slightest loss of readability :) |
140 |
- work around a perl argument refcounting bug. |
141 |
- with some perl patches applied, AnyEvent now seems leak-free. |
142 |
|
143 |
4.3 Fri Oct 3 09:18:43 CEST 2008 |
144 |
- AnyEvent will now install a no-op signal handler for SIGPIPE, |
145 |
unless one has been installed already. |
146 |
- warn about not loaidng AnyEvent::Impl::POE early enough |
147 |
(patch by Adam Rosenstein). |
148 |
- fixed a great number of bugs and corner cases in AnyEvent::Handle: |
149 |
- fix a bug in where in SSL connect mode, the |
150 |
client would first wait for some data by the server and |
151 |
otherwise hang. (reported and analyzed in an absolutely |
152 |
exemplary manner by Adam Rosenstein). |
153 |
- fix a bug in where SSL EOF would not be treated |
154 |
as stream EOF, putting the connection into a hung state. |
155 |
(reported and analyzed in an absolutely exemplary manner by Adam |
156 |
Rosenstein). |
157 |
- fix a potential segfault when the TLS context |
158 |
would go missing in a read callback (e.g. due to stoptls) - |
159 |
Net::SSLeay of course makes no type checking whatsoever. |
160 |
- AnyEvent::Handle will not stop the read watcher in TLS mode. |
161 |
- AnyEvent::Handle->stoptls will send a TLS close notify |
162 |
instead of simply destroying the stream now. |
163 |
- fix a bug where large blocks of data written in TLS mode |
164 |
would not be sent unless triggered by receives. |
165 |
- on_drain will now take the tls write buffer into account. |
166 |
- SSL operations have been streamlined a bit and should be faster |
167 |
now, more to come, though. |
168 |
- added a FAQ section, will document nontrivial issues in there. |
169 |
|
170 |
4.234 Mon Sep 29 04:08:13 CEST 2008 |
171 |
- fix child watcher documentation: only child exits will be reported, |
172 |
no trace events. |
173 |
- mention SIGPIPE in AnyEvent::Handle. |
174 |
- perl backend will now use POSIX::times as monotonic clock |
175 |
when available and the monotonic clock option is not. |
176 |
- run even without Time::HiRes in some cases now, but do not |
177 |
provide sub-second accuracy (all *supported* perl versions have |
178 |
Time::HiRes). |
179 |
|
180 |
4.233 Fri Aug 22 01:48:05 CEST 2008 |
181 |
- fix a bug in the testsuite, causing a hang. |
182 |
- clarified lots of AnyEvent::Handle constructor arguments. |
183 |
|
184 |
4.232 Thu Aug 21 20:44:25 CEST 2008 |
185 |
- fix a bug in AnyEvent::Handle that could cause two fatal error |
186 |
callbacks on EOF, which caused AnyEvent::HTTP to signal success twice |
187 |
(testcase by Leon Brocard). |
188 |
- clarified on_eof/on_error documentation a bit. |
189 |
|
190 |
4.231 Tue Jul 29 13:12:15 CEST 2008 |
191 |
- remove some debugging code left in AnyEvent::Util::fork_call |
192 |
(and no, it's impossible to implement with the broken windows |
193 |
perls without resource leaks or worse). |
194 |
|
195 |
4.23 Tue Jul 29 12:19:59 CEST 2008 |
196 |
- document the first parameter passed to condvar callbacks to be |
197 |
the callback. |
198 |
- add AnyEvent::Socket::{ntoa,aton} aliases. |
199 |
- optimize the AE::Handle->push_read (line) for the default |
200 |
eol marker. |
201 |
- optimize push_read (packstring|storable) for small packets. |
202 |
- invoke on_error callback when no on_eof callback is set. |
203 |
- fix a bug in push_read (storable) of unknown impact. |
204 |
|
205 |
4.22 Sun Jul 20 16:34:13 CEST 2008 |
206 |
- new function AnyEvent::Socket::parse_hostport. |
207 |
- as the bulkheads at microsoft can't even get getprotobyname reliably |
208 |
working on their shitty fucking broken os we need to hardcode |
209 |
some common protocol numbers in AnyEvent::Socket. How can |
210 |
people even bother with such a pile of shit as windows. |
211 |
|
212 |
4.21 Thu Jul 17 14:40:05 CEST 2008 |
213 |
- INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype |
214 |
which is incompatible to previous invocation syntax. |
215 |
- work around more windows perl bugs in fork_call. I cannot imagine |
216 |
why anybody would bother implementing fork in such an obviously |
217 |
unusable way. |
218 |
- avoid calling $do_retry if the request has already finished in |
219 |
AnyEvent::DNS, thus avoiding an exception (reported by Anatoly |
220 |
K. Sharifulin). |
221 |
|
222 |
4.2 Sat Jul 12 22:42:11 CEST 2008 |
223 |
- support an "interval" argument for repeating timers. |
224 |
- fix a bug in the pure perl event loop that caused timers to |
225 |
be delayed under some circumstances. |
226 |
- implement optional argument checking via PERL_ANYEVENT_STRICT. |
227 |
- not importing anything from Event causes it to malfunction |
228 |
and not use Time::HiRes, work around this bug. |
229 |
- more bug workarounds against the endlessly broken Tk module, |
230 |
seems to work now on 32 bit machines, but less so on 64 bit. |
231 |
- minor optimisations applied to most backends. |
232 |
|
233 |
4.161 Fri Jul 4 14:24:48 CEST 2008 |
234 |
- fixed the t/handle/02_write.t, it read after write, with the |
235 |
assumption that all data could be written before read |
236 |
(reported by Jim Razmus). |
237 |
|
238 |
4.160 Thu Jul 3 04:02:21 CEST 2008 |
239 |
- re-registering signal watchers after unregistering one did |
240 |
not work if a backend falls back on AnyEvent's default |
241 |
implementation (also affected child watchers). |
242 |
- new AnyEvent::Handle options: autocork and no_delay. |
243 |
|
244 |
4.152 Sun Jun 22 14:15:44 CEST 2008 |
245 |
- allow for 32-bit perls that implement shifts differently |
246 |
on different architectures in parse_ipv4 (reported and |
247 |
analysed by Keiichi DAIBA). |
248 |
|
249 |
4.151 Fri Jun 6 17:34:24 CEST 2008 |
250 |
- make sure specifying _only_ on_read and never pushing reads |
251 |
works. |
252 |
|
253 |
4.15 Fri Jun 6 13:00:46 CEST 2008 |
254 |
- the pure perl backend would keep some watchers alive when more than |
255 |
one watcher was registered for the same fd. |
256 |
- new "packstring" and "storable" read and write types |
257 |
for AnyEvent::Handle. |
258 |
- allow on_eof handler to be called after on_error with EPIPE returns. |
259 |
- do not immediately call on_read callback in handle constructor. |
260 |
|
261 |
4.14 Thu Jun 5 20:29:31 CEST 2008 |
262 |
- Fixed a bug in DNS SRV priority sorting. |
263 |
- AnyEvent::Util::guard now reports runtime errors while |
264 |
executing the guard block as warnings. |
265 |
- handle 0-byte-reads just before EOF correctly in AnyEvent::Handle. |
266 |
|
267 |
4.13 Thu Jun 5 00:47:59 CEST 2008 |
268 |
- AnyEvent::DNS only followed cname chains with length 2, |
269 |
contrary to documentation. bumped it up to 10, thanks to |
270 |
microsoft, the current broken-dns-config-king. |
271 |
- AnyEvent::DNS didn't check for socket return status |
272 |
but instead relied on perl not creating filehandles in that |
273 |
case - too bad it gives you a bogus file handle (reported |
274 |
and analysed by Vladimir Timofeev). |
275 |
- fix queue management logic in AnyEvent::Handle: |
276 |
when on_read was registered, the queue was empty |
277 |
and no progress could be made AnyEvent::Handle would |
278 |
enter an endless loop. |
279 |
- correctly start reading again when the handle became |
280 |
busy again after idling. |
281 |
- correctly treat tls shut-downs as EOF condition |
282 |
(for the time being, would be nice to have a callback for that). |
283 |
- correctly call eof callback under all conditions |
284 |
(hopefully). |
285 |
- the timeout callback did not expect that $self can go |
286 |
away any time. |
287 |
|
288 |
4.12 Tue Jun 3 10:58:04 CEST 2008 |
289 |
- include AnyEvent::Intro, a tutorial for anyevent, |
290 |
anyevent::socket and anyevent::handle. |
291 |
- allow more options in on_error. |
292 |
|
293 |
4.11 Fri May 30 23:42:25 CEST 2008 |
294 |
- INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide |
295 |
reverse_lookup and reverse_verify to replace it, support |
296 |
v4mapped and v4compat addresses. |
297 |
- provide more documentation for the resolver class. |
298 |
- really replace longest run of :0: by :: in format_address, |
299 |
also properly convert :: and ::1 again. |
300 |
- support NAPTR record name and decode it. |
301 |
- implement random weight sampling for SRV records, as per |
302 |
rfc 2782. |
303 |
- correctly abort on srv-record targets of ".". |
304 |
- added AnyEvent::DNS::wait_for_slot. |
305 |
- in the unlikely event of a virtual circuit connection |
306 |
being invalidated by a delayed udp reply, AnyEvent::DNS |
307 |
could die. this has been fixed. |
308 |
- plug a probable memleak in the DNS vc code. |
309 |
- use configured timeout also for VC request phase. |
310 |
- implement timeout and max_outstanding methods |
311 |
for resolver class. |
312 |
- update version numbers in all modules. |
313 |
|
314 |
4.1 Thu May 29 05:45:40 CEST 2008 |
315 |
- INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address |
316 |
which mirrors their purpose better (old names still available |
317 |
for a while). Also moved AnyEvent::DNS::addr to |
318 |
AnyEvent::Socket::resolve_sockaddr. |
319 |
- implement AnyEvent->time and AnyEvent->now. |
320 |
- fix IPv6 support in pack_sockaddr. |
321 |
- officially un-experimentalise all newly introduced code. |
322 |
- support unix domain sockets everywhere by specifying |
323 |
a host of "unix/" and the pathname as service. |
324 |
- implement an activity timeout in AnyEvent::Handle. |
325 |
- added a regex read type to AnyEvent::Handle. |
326 |
- added a json read/write type to AnyEvent::Handle. |
327 |
- always croak in on_error in AnyEvent::Handle. |
328 |
- document how to subclass AnyEvent::Handle. |
329 |
- implement AnyEvent::Util::fork_call. |
330 |
- add support for IPv6 nameservers and nameserver statements. |
331 |
- work around _yet_ _another_ windows perl bug in where empty |
332 |
select masks cause errors in the pure perl backend (this |
333 |
workaround cannot be implemented for the other event loops, |
334 |
but EV also works around this). |
335 |
- supply AnyEvent::Socket::address_type and make good use of it. |
336 |
- clarify and fix the Handle documentation w.r.t $self vs. $handle |
337 |
vs. callback arguments. |
338 |
- add some recommends to the META.yml. |
339 |
|
340 |
4.05 Mon May 26 19:44:06 CEST 2008 |
341 |
- some platforms ignore the rfc and prepend an extra sa_len member to |
342 |
the sockaddr structure, cater for those. also use sockaddr_family, |
343 |
didn't know it was there, but comes in handy for the workaround. |
344 |
- undo the import WIN32 hack. |
345 |
|
346 |
4.04 Mon May 26 08:03:31 CEST 2008 |
347 |
- try to work around yet another windows bug: failed connects are reported |
348 |
as if it were out-of-band data. windows users: you suck. Only |
349 |
EV, Glib, Event and the pure perl backend can handle this condition. |
350 |
- optimize write algorithm in AnyEvent::Handle. |
351 |
- properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly. |
352 |
- tcp_connect now properly iterates through all targets. |
353 |
- check wether IPv6 sockets can actually be created, otherwise, disable ipv6 |
354 |
support entirely. also hardcode some AF_INET6 constants for the |
355 |
sake of perl 5.8. |
356 |
- vastly improved nameserver/suffix detection algorithm for windows |
357 |
(still a hack...) |
358 |
- try to use a fallback dns server if no dns servers could be found. |
359 |
- splatter around some local $SIG{__DIE__}'s to work around bad |
360 |
die handlers. |
361 |
|
362 |
4.03 Sun May 25 05:05:57 CEST 2008 |
363 |
- work around perl bugs on windows, where perl returns undocumented |
364 |
error codes for sysread, syswrite etc. |
365 |
- fix AnyEvent::Handle::unshift_read to really unshift and not push. |
366 |
- AnyEvent::Handle could sometimes stumble over 0-byte writes. |
367 |
- fix unpack_sockaddr, now works for AF_INET6 sockaddr's. |
368 |
- updated documentation. |
369 |
- lots of workarounds for perl 5.8 quirks. |
370 |
- implement a more flexible type system for both reads and writes |
371 |
in AnyEvent::Handle. |
372 |
- support IPv6 in tcp_server. |
373 |
- pass local hostname and port to the prepare callback in tcp_server. |
374 |
- make types extendable in AnyEvent::Handle. |
375 |
- croak, not die, in AnyEvent::Handle; |
376 |
|
377 |
4.0 Sat May 24 19:58:08 CEST 2008 |
378 |
- added (experimental) AnyEvent::DNS module. |
379 |
- added (experimental) AnyEvent::Socket module. |
380 |
- added (experiemntal) support for TLS/SSL in AnyEvent::Handle. |
381 |
- make condvar->end call ->send as documented. |
382 |
- make condvar constructor support the documented cb => argument. |
383 |
- overload condvars so you can use them instead of code references. |
384 |
- support more of the POSIX forms instead of a pure dotted quad for |
385 |
verbatim ip addresses. |
386 |
- AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. |
387 |
- try very hard not to freeze on broken windows perls. |
388 |
- make on_eof optional in AnyEvent::Handle. |
389 |
- support service names in tcp_connect and tcp_server. |
390 |
- work around netbsd/openbsd bug causing unix sockets not to |
391 |
become writable until empty. |
392 |
- only make a dummy read for error detection when getpeername |
393 |
returns ENOTCONN. |
394 |
- defend AnyEvent::Util::inet_aton against Coro::LWP's brutal |
395 |
override. |
396 |
- new AnyEvent::Guard::cancel. |
397 |
|
398 |
3.5 Sat May 17 23:17:14 CEST 2008 |
399 |
- gracefully deal with systems not defining CLOCK_MONOTONIC |
400 |
(as opposed to not supporting the clock itself, which worked in 3.41). |
401 |
- added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server |
402 |
helpers, to replace the ill-designed AnyEvent::Socket class. |
403 |
- added some examples (eg/connect, eg/handle). |
404 |
- added AnyEvent::Util::guard. |
405 |
- AnyEvent::Util now exports most of its functions. |
406 |
- fixed a bug in *_read_chunk, where the callback was called with $self as |
407 |
first and second argument. |
408 |
- simplified and fixed regex handling in AnyEvent::Handle::*_read_line. |
409 |
- fix manpage condvars examples to use new syntax. |
410 |
- considerably improve inet_aton with EV::ADNS backend to support |
411 |
cname chains (common dns configuration bug, e.g. with www.google.de) |
412 |
and aaaa records if no a record could be found. |
413 |
|
414 |
3.41 Sun May 11 19:53:13 CEST 2008 |
415 |
- work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. |
416 |
|
417 |
3.4 Sun May 11 00:29:25 CEST 2008 |
418 |
- complete re-implementation of condvars with a hopefully much more useful |
419 |
API (backward compatible functions exist but are undocumented). |
420 |
- AnyEvent::Handle fully rewritten with a hopefully more powerful API |
421 |
(still subject to change). |
422 |
- take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
423 |
- provide AnyEvent::post_detect and @AnyEvent::post_detect, which |
424 |
allows module authors to avoid forcing event loop detection. |
425 |
(used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example). |
426 |
- remove coro backends: Coro now provides generic support for AnyEvent |
427 |
via Coro::AnyEvent. |
428 |
|
429 |
3.3 Mon Apr 28 09:51:06 CEST 2008 |
430 |
- added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. |
431 |
- fix a bug in the pure perl backend that kept watchers alive |
432 |
when multiple watchers were registered for the same fh/poll combo. |
433 |
- add a benchmark section showing AnyEvent overhead and comparing |
434 |
the different event loops with each other. |
435 |
- prefer pure perl over tk when autoprobing, as it's about as fast, |
436 |
but doesn't crash with many watchers. |
437 |
- declare Qt support non-experimental. |
438 |
- clarify the confusing section about the file descriptor being |
439 |
kept alive. |
440 |
- document the race between loading of an event module and |
441 |
child processes exiting. |
442 |
- support POE as "backend" (with some caveats, POE is not generic |
443 |
enough, and darn slow). |
444 |
- support Wx and Prima through POE. |
445 |
- optimise perl backend to use 20% less memory and take advantage |
446 |
of typical timeout behaviour. It can now compete with |
447 |
select/poll-based C event loops in most cases (it is usually |
448 |
faster than Event and Glib :). |
449 |
- roughly cut EV memory use in half and increase its speed by 30%, |
450 |
by removing undocumented functionality. Did something similar |
451 |
to other event loops where possible. |
452 |
|
453 |
3.2 Thu Apr 24 10:10:40 CEST 2008 |
454 |
- do not die when anyevent watchers are destroyed while running callbacks |
455 |
in the pureperl backend (could only happen when two watchers are registered |
456 |
for the same fh/poll combo). |
457 |
- support autoloading for child watchers, was broken in all versions. |
458 |
- implement PERL_ANYEVENT_MODEL env variable. |
459 |
- (experimental) implement interface to the (very crashy/buggy) Event::Lib module. |
460 |
- (experimental) implement interface to the Qt module (cannot be autoprobed). |
461 |
- this release is IO:AnyEvent-proof. |
462 |
|
463 |
3.12 Tue Apr 22 07:11:46 CEST 2008 |
464 |
- reinstate AUTHOR section that got lost somehow. |
465 |
- do not hang in the testsuite with badly broken perls (activestate, |
466 |
strawberry...), but instead diagnose the problem and continue. |
467 |
- use INT instead of CHLD in an attempt to support broken windows perls |
468 |
better (this decreases the test quality, unfortunately). |
469 |
- do not send the signal to the process group |
470 |
(no problem for CHLD, bad for INT :). |
471 |
|
472 |
3.11 Sat Apr 19 06:57:31 CEST 2008 |
473 |
- major documentation rework. |
474 |
- document the fact that child watchers only watch for zombies. |
475 |
- fix the child watcher example. |
476 |
|
477 |
3.1 Wed Apr 16 17:09:01 CEST 2008 |
478 |
- work around recurring bugs in Tk by dup'ing filehandles, the |
479 |
only method with good success chances on Tk (the bugs apparently |
480 |
don't get fixed anytime soon). |
481 |
- lift the restriction of only one watcher per fh direction |
482 |
(as the Tk bug workaround also lifts it and only Tk imposed |
483 |
such strong limits). |
484 |
- changed probe order to prefer coro adaptors. |
485 |
- explain why recursion into the event loop is not supported |
486 |
unless the backend supports it (only Coro::EV does without |
487 |
any restrictions...). |
488 |
- add simple manpages for all backend modules. |
489 |
|
490 |
3.0 Mon Apr 7 21:30:23 CEST 2008 |
491 |
- Coro::Signal changed semantics, roll our own, also cleaning |
492 |
up the Coro implementation in general. |
493 |
- rename Coro backend to CoroEvent. |
494 |
- add some decision helping paragraph to the manpage that should |
495 |
help people to decide whether AnyEvent is the right thing for them. |
496 |
|
497 |
2.9 Mon Jan 28 13:31:54 CET 2008 |
498 |
- update for EV 3.0 API changes. |
499 |
|
500 |
2.8 Sun Nov 25 15:06:03 CET 2007 |
501 |
- waitpid can validly return 0. accept this fact of life |
502 |
instead of reporting it to any watchers. |
503 |
|
504 |
2.7 Fri Nov 23 11:41:14 CET 2007 |
505 |
- force use of AnyEvent::Impl::Perl in testsuite, there is |
506 |
too much breakage outside AnyEvent. |
507 |
- deliver signals synchronously in AnyEvent::Impl::Perl |
508 |
or any other event loop that relies on AnyEvents child |
509 |
watcher emulation. *Could* help with hanging testsuite |
510 |
(except when Event or EV are installed). |
511 |
|
512 |
2.6 Fri Nov 9 20:36:35 CET 2007 |
513 |
- fix bug in testsuite. |
514 |
- move EV adaptor modules to AnyEvent. |
515 |
- add Coro+EV adaptor module. |
516 |
|
517 |
2.55 Tue Nov 6 17:41:32 CET 2007 |
518 |
- add EV to the list of supported event models. |
519 |
- do not auto-reset pid watchers, pass pid and status to them. |
520 |
- allow a pid of zero to watch for all children in child watchers. |
521 |
|
522 |
2.54 Wed Jul 18 17:36:23 CEST 2007 |
523 |
- work around a perl bug that results in |
524 |
BEGIN not safe after errors--compilation aborted without |
525 |
any discernible reason or error message by once |
526 |
again not use'ing strict. |
527 |
|
528 |
2.53 Sun Jul 8 10:51:53 CEST 2007 |
529 |
- make Glib interface work again (spotted by elmex). |
530 |
|
531 |
2.52 Wed Mar 7 18:36:16 CET 2007 |
532 |
- child watchers stopped working permanently when |
533 |
all current child watchers were destroyed. |
534 |
|
535 |
2.51 Mon Dec 11 21:33:24 CET 2006 |
536 |
- work around bugs in perl where eval "require Module" returns |
537 |
true even if the module couldn't be loaded :/. |
538 |
|
539 |
2.5 Mon Dec 11 02:15:28 CET 2006 |
540 |
- avoid $AUTOLOAD because many perls corrupt it. |
541 |
- AnyEvent::detect forces autodetection. |
542 |
- implement signal watchers (experimental) |
543 |
(Tk does not support async signals (<= 804.027 at least), |
544 |
so they might get delayed indefinitely on Tk). |
545 |
- implement child watchers (experimental). |
546 |
- moved default condvar implementation into base module, simplifying |
547 |
most implementation modules. |
548 |
|
549 |
2.1 Fri Nov 24 15:50:48 CET 2006 |
550 |
- better docs. |
551 |
- simple testsuite. |
552 |
- added AnyEvent->one_event method for special purposes. |
553 |
|
554 |
2.0 Wed Nov 1 02:21:30 CET 2006 |
555 |
- INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both. |
556 |
- INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore. |
557 |
- EXPERIMENTAL: |
558 |
new pure-perl model added, AnyEvent now always finds a suitable |
559 |
event model. |
560 |
- improved documentation. |
561 |
- reduced cpu overhead. |
562 |
|
563 |
1.02 Fri Jan 13 14:15:40 CET 2006 |
564 |
- add COPYING to clarify license. |
565 |
|
566 |
1.01 Sun Jan 8 05:48:09 CET 2006 |
567 |
- correctly cancel Event timers, avoid memory leaks. |
568 |
|
569 |
1.0 Sun Jan 8 05:40:58 CET 2006 |
570 |
- allow third-party interfaces to be added at runtime. |
571 |
- fix version number. |
572 |
|
573 |
0.9 Fri Dec 30 02:25:41 CET 2005 |
574 |
- fixed errors in the documentation. |
575 |
- going back to using $VERSION as "load check", |
576 |
as suggested by Jörn Reder. Let's see why I stopped using |
577 |
this method. |
578 |
- add 'hup' to the list of events to receive for glib |
579 |
and generate 'rw' events for it, as some versions of |
580 |
glib seem to require this, while this breaks other |
581 |
(older) versions of glib. |
582 |
- implemented PERL_ANYEVENT_VERBOSE env variable. |
583 |
|
584 |
0.3 Sun Dec 4 10:43:33 CET 2005 |
585 |
- no changes to the code. |
586 |
- fixed documentation. |
587 |
- clarified some corner cases. |
588 |
- vastly improved documentation with a long example. |
589 |
|
590 |
0.2 Thu Dec 1 22:18:49 CET 2005 |
591 |
- check whether a valid method was called, to avoid endless loops. |
592 |
- move condvars into their own class (cleaner, safer). |
593 |
- don't die when some modules are not available. |
594 |
- Tk is too broken to allow for words. Will work once Tk works. |
595 |
|
596 |
0.1 Thu Dec 1 19:44:57 CET 2005 |
597 |
- spiced up and cleaned up for release. |
598 |
|
599 |
0.01 Tue Sep 21 18:12:09 CEST 2004 |
600 |
- original version; copied from Convert::Scalar. |
601 |
|