1 |
Revision history for Perl extension AnyEvent. |
2 |
|
3 |
4.22 |
4 |
- new function AnyEvent::Socket::parse_hostport. |
5 |
|
6 |
4.21 Thu Jul 17 14:40:05 CEST 2008 |
7 |
- INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype |
8 |
which is incompatible to previous invocation syntax. |
9 |
- work around more windows perl bugs in fork_call. I cannot imagine |
10 |
why anybody would bother implementing fork in such an obviously |
11 |
unusable way. |
12 |
- avoid calling $do_retry if the request has already finished in |
13 |
AnyEvent::DNS, thus avoiding an exception (reported by Anatoly |
14 |
K. Sharifulin). |
15 |
|
16 |
4.2 Sat Jul 12 22:42:11 CEST 2008 |
17 |
- support an "interval" argument for repeating timers. |
18 |
- fix a bug in the pure perl event loop that caused timers to |
19 |
be delayed under some circumstances. |
20 |
- implement optional argument checking via PERL_ANYEVENT_STRICT. |
21 |
- not importing anything from Event causes it to malfunction |
22 |
and not use Time::HiRes, work around this bug. |
23 |
- more bug workarounds against the endlessly broken Tk module, |
24 |
seems to work now on 32 bit machines, but less so on 64 bit. |
25 |
- minor optimisations applied to most backends. |
26 |
|
27 |
4.161 Fri Jul 4 14:24:48 CEST 2008 |
28 |
- fixed the t/handle/02_write.t, it read after write, with the |
29 |
assumption that all data could be written before read |
30 |
(reported by Jim Razmus). |
31 |
|
32 |
4.160 Thu Jul 3 04:02:21 CEST 2008 |
33 |
- re-registering signal watchers after unregistering one did |
34 |
not work if a backend falls back on AnyEvent's default |
35 |
implementation (also affected child watchers). |
36 |
- new AnyEvent::Handle options: autocork and no_delay. |
37 |
|
38 |
4.152 Sun Jun 22 14:15:44 CEST 2008 |
39 |
- allow for 32-bit perls that implement shifts differently |
40 |
on different architectures in parse_ipv4 (reported and |
41 |
analysed by Keiichi DAIBA). |
42 |
|
43 |
4.151 Fri Jun 6 17:34:24 CEST 2008 |
44 |
- make sure specifying _only_ on_read and never pushing reads |
45 |
works. |
46 |
|
47 |
4.15 Fri Jun 6 13:00:46 CEST 2008 |
48 |
- the pure perl backend would keep some watchers alive when more than |
49 |
one watcher was registered for the same fd. |
50 |
- new "packstring" and "storable" read and write types |
51 |
for AnyEvent::Handle. |
52 |
- allow on_eof handler to be called after on_error with EPIPE returns. |
53 |
- do not immediately call on_read callback in handle constructor. |
54 |
|
55 |
4.14 Thu Jun 5 20:29:31 CEST 2008 |
56 |
- Fixed a bug in DNS SRV priority sorting. |
57 |
- AnyEvent::Util::guard now reports runtime errors while |
58 |
executing the guard block as warnings. |
59 |
- handle 0-byte-reads just before EOF correctly in AnyEvent::Handle. |
60 |
|
61 |
4.13 Thu Jun 5 00:47:59 CEST 2008 |
62 |
- AnyEvent::DNS only followed cname chains with length 2, |
63 |
contrary to documentation. bumped it up to 10, thanks to |
64 |
microsoft, the current broken-dns-config-king. |
65 |
- AnyEvent::DNS didn't check for socket return status |
66 |
but instead relied on perl not creating filehandles in that |
67 |
case - too bad it gives you a bogus file handle (reported |
68 |
and analysed by Vladimir Timofeev). |
69 |
- fix queue management logic in AnyEvent::Handle: |
70 |
when on_read was registered, the queue was empty |
71 |
and no progress could be made AnyEvent::Handle would |
72 |
enter an endless loop. |
73 |
- correctly start reading again when the handle became |
74 |
busy again after idling. |
75 |
- correctly treat tls shut-downs as EOF condition |
76 |
(for the time being, would be nice to have a callback for that). |
77 |
- correctly call eof callback under all conditions |
78 |
(hopefully). |
79 |
- the timeout callback did not expect that $self can go |
80 |
away any time. |
81 |
|
82 |
4.12 Tue Jun 3 10:58:04 CEST 2008 |
83 |
- include AnyEvent::Intro, a tutorial for anyevent, |
84 |
anyevent::socket and anyevent::handle. |
85 |
- allow more options in on_error. |
86 |
|
87 |
4.11 Fri May 30 23:42:25 CEST 2008 |
88 |
- INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide |
89 |
reverse_lookup and reverse_verify to replace it, support |
90 |
v4mapped and v4compat addresses. |
91 |
- provide more documentation for the resolver class. |
92 |
- really replace longest run of :0: by :: in format_address, |
93 |
also properly convert :: and ::1 again. |
94 |
- support NAPTR record name and decode it. |
95 |
- implement random weight sampling for SRV records, as per |
96 |
rfc 2782. |
97 |
- correctly abort on srv-record targets of ".". |
98 |
- added AnyEvent::DNS::wait_for_slot. |
99 |
- in the unlikely event of a virtual circuit connection |
100 |
being invalidated by a delayed udp reply, AnyEvent::DNS |
101 |
could die. this has been fixed. |
102 |
- plug a probable memleak in the DNS vc code. |
103 |
- use configured timeout also for VC request phase. |
104 |
- implement timeout and max_outstanding methods |
105 |
for resolver class. |
106 |
- update version numbers in all modules. |
107 |
|
108 |
4.1 Thu May 29 05:45:40 CEST 2008 |
109 |
- INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address |
110 |
which mirrors their purpose better (old names still available |
111 |
for a while). Also moved AnyEvent::DNS::addr to |
112 |
AnyEvent::Socket::resolve_sockaddr. |
113 |
- implement AnyEvent->time and AnyEvent->now. |
114 |
- fix IPv6 support in pack_sockaddr. |
115 |
- officially un-experimentalise all newly introduced code. |
116 |
- support unix domain sockets everywhere by specifying |
117 |
a host of "unix/" and the pathname as service. |
118 |
- implement an activity timeout in AnyEvent::Handle. |
119 |
- added a regex read type to AnyEvent::Handle. |
120 |
- added a json read/write type to AnyEvent::Handle. |
121 |
- always croak in on_error in AnyEvent::Handle. |
122 |
- document how to subclass AnyEvent::Handle. |
123 |
- implement AnyEvent::Util::fork_call. |
124 |
- add support for IPv6 nameservers and nameserver statements. |
125 |
- work around _yet_ _another_ windows perl bug in where empty |
126 |
select masks cause errors in the pure perl backend (this |
127 |
workaround cannot be implemented for the other event loops, |
128 |
but EV also works around this). |
129 |
- supply AnyEvent::Socket::address_type and make good use of it. |
130 |
- clarify and fix the Handle documentation w.r.t $self vs. $handle |
131 |
vs. callback arguments. |
132 |
- add some recommends to the META.yml. |
133 |
|
134 |
4.05 Mon May 26 19:44:06 CEST 2008 |
135 |
- some platforms ignore the rfc and prepend an extra sa_len member to |
136 |
the sockaddr structure, cater for those. also use sockaddr_family, |
137 |
didn't know it was there, but comes in handy for the workaround. |
138 |
- undo the import WIN32 hack. |
139 |
|
140 |
4.04 Mon May 26 08:03:31 CEST 2008 |
141 |
- try to work around yet another windows bug: failed connects are reported |
142 |
as if it were out-of-band data. windows users: you suck. Only |
143 |
EV, Glib, Event and the pure perl backend can handle this condition. |
144 |
- optimize write algorithm in AnyEvent::Handle. |
145 |
- properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly. |
146 |
- tcp_connect now properly iterates through all targets. |
147 |
- check wether IPv6 sockets can actually be created, otherwise, disable ipv6 |
148 |
support entirely. also hardcode some AF_INET6 constants for the |
149 |
sake of perl 5.8. |
150 |
- vastly improved nameserver/suffix detection algorithm for windows |
151 |
(still a hack...) |
152 |
- try to use a fallback dns server if no dns servers could be found. |
153 |
- splatter around some local $SIG{__DIE__}'s to work around bad |
154 |
die handlers. |
155 |
|
156 |
4.03 Sun May 25 05:05:57 CEST 2008 |
157 |
- work around perl bugs on windows, where perl returns undocumented |
158 |
error codes for sysread, syswrite etc. |
159 |
- fix AnyEvent::Handle::unshift_read to really unshift and not push. |
160 |
- AnyEvent::Handle could sometimes stumble over 0-byte writes. |
161 |
- fix unpack_sockaddr, now works for AF_INET6 sockaddr's. |
162 |
- updated documentation. |
163 |
- lots of workarounds for perl 5.8 quirks. |
164 |
- implement a more flexible type system for both reads and writes |
165 |
in AnyEvent::Handle. |
166 |
- support IPv6 in tcp_server. |
167 |
- pass local hostname and port to the prepare callback in tcp_server. |
168 |
- make types extendable in AnyEvent::Handle. |
169 |
- croak, not die, in AnyEvent::Handle; |
170 |
|
171 |
4.0 Sat May 24 19:58:08 CEST 2008 |
172 |
- added (experimental) AnyEvent::DNS module. |
173 |
- added (experimental) AnyEvent::Socket module. |
174 |
- added (experiemntal) support for TLS/SSL in AnyEvent::Handle. |
175 |
- make condvar->end call ->send as documented. |
176 |
- make condvar constructor support the documented cb => argument. |
177 |
- overload condvars so you can use them instead of code references. |
178 |
- support more of the POSIX forms instead of a pure dotted quad for |
179 |
verbatim ip addresses. |
180 |
- AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. |
181 |
- try very hard not to freeze on broken windows perls. |
182 |
- make on_eof optional in AnyEvent::Handle. |
183 |
- support service names in tcp_connect and tcp_server. |
184 |
- work around netbsd/openbsd bug causing unix sockets not to |
185 |
become writable until empty. |
186 |
- only make a dummy read for error detection when getpeername |
187 |
returns ENOTCONN. |
188 |
- defend AnyEvent::Util::inet_aton against Coro::LWP brutal |
189 |
override. |
190 |
- new AnyEvent::Guard::cancel. |
191 |
|
192 |
3.5 Sat May 17 23:17:14 CEST 2008 |
193 |
- gracefully deal with systems not defining CLOCK_MONOTONIC |
194 |
(as opposed to not supporting the clock itself, which worked in 3.41). |
195 |
- added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server |
196 |
helpers, to replace the ill-designed AnyEvent::Socket class. |
197 |
- added some examples (eg/connect, eg/handle). |
198 |
- added AnyEvent::Util::guard. |
199 |
- AnyEvent::Util now exports most of its functions. |
200 |
- fixed a bug in *_read_chunk, where the callback was called with $self as |
201 |
first and second argument. |
202 |
- simplified and fixed regex handling in AnyEvent::Handle::*_read_line. |
203 |
- fix manpage condvars examples to use new syntax. |
204 |
- considerably improve inet_aton with EV::ADNS backend to support |
205 |
cname chains (common dns configuration bug, e.g. with www.google.de) |
206 |
and aaaa records if no a record could be found. |
207 |
|
208 |
3.41 Sun May 11 19:53:13 CEST 2008 |
209 |
- work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. |
210 |
|
211 |
3.4 Sun May 11 00:29:25 CEST 2008 |
212 |
- complete re-implementation of condvars with a hopefully much mroe useful |
213 |
API (backward compatible functions exist but are undocumented). |
214 |
- AnyEvent::Handle fully rewritten with a hopefully more powerful API |
215 |
(still subject to change). |
216 |
- take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
217 |
- provide AnyEvent::post_detect and @AnyEvent::post_detect, which |
218 |
allows module authors to avoid forcing event loop detection. |
219 |
(used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example). |
220 |
- remove coro backends: Coro now provides generic support for AnyEvent |
221 |
via Coro::AnyEvent. |
222 |
|
223 |
3.3 Mon Apr 28 09:51:06 CEST 2008 |
224 |
- added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. |
225 |
- fix a bug in the pure perl backend that kept watchers alive |
226 |
when multiple watchers were registered for the same fh/poll combo. |
227 |
- add a benchmark section showing AnyEvent overhead and comparing |
228 |
the different event loops with each other. |
229 |
- prefer pure perl over tk when autoprobing, as it's about as fast, |
230 |
but doesn't crash with many watchers. |
231 |
- declare Qt support non-experimental. |
232 |
- clarify the confusing section about the file descriptor being |
233 |
kept alive. |
234 |
- document the race between loading of an event module and |
235 |
child processes exiting. |
236 |
- support POE as "backend" (with some caveats, POE is not generic |
237 |
enough, and darn slow). |
238 |
- support Wx and Prima through POE. |
239 |
- optimise perl backend to use 20% less memory and take advantage |
240 |
of typical timeout behaviour. It can now compete with |
241 |
select/poll-based C event loops in most cases (it is usually |
242 |
faster than Event and Glib :). |
243 |
- roughly cut EV memory use in half and increase its speed by 30%, |
244 |
by removing undocumented functionality. Did something similar |
245 |
to other event loops where possible. |
246 |
|
247 |
3.2 Thu Apr 24 10:10:40 CEST 2008 |
248 |
- do not die when anyevent watchers are destroyed while running callbacks |
249 |
in the pureperl backend (could only happen when two watchers are registered |
250 |
for the same fh/poll combo). |
251 |
- support autoloading for child watchers, was broken in all versions. |
252 |
- implement PERL_ANYEVENT_MODEL env variable. |
253 |
- (experimental) implement interface to the (very crashy/buggy) Event::Lib module. |
254 |
- (experimental) implement interface to the Qt module (cannot be autoprobed). |
255 |
- this release is IO:AnyEvent-proof. |
256 |
|
257 |
3.12 Tue Apr 22 07:11:46 CEST 2008 |
258 |
- reinstate AUTHOR section that got lost somehow. |
259 |
- do not hang in the testsuite with badly broken perls (activestate, |
260 |
strawberry...), but instead diagnose the problem and continue. |
261 |
- use INT instead of CHLD in an attempt to support broken windows perls |
262 |
better (this decreases the test quality, unfortunately). |
263 |
- do not send the signal to the process group |
264 |
(no problem for CHLD, bad for INT :). |
265 |
|
266 |
3.11 Sat Apr 19 06:57:31 CEST 2008 |
267 |
- major documentation rework. |
268 |
- document the fact that child watchers only watch for zombies. |
269 |
- fix the child watcher example. |
270 |
|
271 |
3.1 Wed Apr 16 17:09:01 CEST 2008 |
272 |
- work around recurring bugs in Tk by dup'ing filehandles, the |
273 |
only method with good success chances on Tk (the bugs apparently |
274 |
don't get fixed anytime soon). |
275 |
- lift the restriction of only one watcher per fh direction |
276 |
(as the Tk bug workaround also lifts it and only Tk imposed |
277 |
such strong limits). |
278 |
- changed probe order to prefer coro adaptors. |
279 |
- explain why recursion into the event loop is not supported |
280 |
unless the backend supports it (only Coro::EV does without |
281 |
any restrictions...). |
282 |
- add simple manpages for all backend modules. |
283 |
|
284 |
3.0 Mon Apr 7 21:30:23 CEST 2008 |
285 |
- Coro::Signal changed semantics, roll our own, also cleaning |
286 |
up the Coro implementation in general. |
287 |
- rename Coro backend to CoroEvent. |
288 |
- add some decision helping paragraph to the manpage that should |
289 |
help people to decide whether AnyEvent is the right thing for them. |
290 |
|
291 |
2.9 Mon Jan 28 13:31:54 CET 2008 |
292 |
- update for EV 3.0 API changes. |
293 |
|
294 |
2.8 Sun Nov 25 15:06:03 CET 2007 |
295 |
- waitpid can validly return 0. accept this fact of life |
296 |
instead of reporting it to any watchers. |
297 |
|
298 |
2.7 Fri Nov 23 11:41:14 CET 2007 |
299 |
- force use of AnyEvent::Impl::Perl in testsuite, there is |
300 |
too much breakage outside AnyEvent. |
301 |
- deliver signals synchronously in AnyEvent::Impl::Perl |
302 |
or any other event loop that relies on AnyEvents child |
303 |
watcher emulation. *Could* help with hanging testsuite |
304 |
(except when Event or EV are installed). |
305 |
|
306 |
2.6 Fri Nov 9 20:36:35 CET 2007 |
307 |
- fix bug in testsuite. |
308 |
- move EV adaptor modules to AnyEvent. |
309 |
- add Coro+EV adaptor module. |
310 |
|
311 |
2.55 Tue Nov 6 17:41:32 CET 2007 |
312 |
- add EV to the list of supported event models. |
313 |
- do not auto-reset pid watchers, pass pid and status to them. |
314 |
- allow a pid of zero to watch for all children in child watchers. |
315 |
|
316 |
2.54 Wed Jul 18 17:36:23 CEST 2007 |
317 |
- work around a perl bug that results in |
318 |
BEGIN not safe after errors--compilation aborted without |
319 |
any discernible reason or error message by once |
320 |
again not use'ing strict. |
321 |
|
322 |
2.53 Sun Jul 8 10:51:53 CEST 2007 |
323 |
- make Glib interface work again (spotted by elmex). |
324 |
|
325 |
2.52 Wed Mar 7 18:36:16 CET 2007 |
326 |
- child watchers stopped working permanently when |
327 |
all current child watchers were destroyed. |
328 |
|
329 |
2.51 Mon Dec 11 21:33:24 CET 2006 |
330 |
- work around bugs in perl where eval "require Module" returns |
331 |
true even if the module couldn't be loaded :/. |
332 |
|
333 |
2.5 Mon Dec 11 02:15:28 CET 2006 |
334 |
- avoid $AUTOLOAD because many perls corrupt it. |
335 |
- AnyEvent::detect forces autodetection. |
336 |
- implement signal watchers (experimental) |
337 |
(Tk does not support async signals (<= 804.027 at least), |
338 |
so they might get delayed indefinitely on Tk). |
339 |
- implement child watchers (experimental). |
340 |
- moved default condvar implementation into base module, simplifying |
341 |
most implementation modules. |
342 |
|
343 |
2.1 Fri Nov 24 15:50:48 CET 2006 |
344 |
- better docs. |
345 |
- simple testsuite. |
346 |
- added AnyEvent->one_event method for special purposes. |
347 |
|
348 |
2.0 Wed Nov 1 02:21:30 CET 2006 |
349 |
- INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both. |
350 |
- INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore. |
351 |
- EXPERIMENTAL: |
352 |
new pure-perl model added, AnyEvent now always finds a suitable |
353 |
event model. |
354 |
- improved documentation. |
355 |
- reduced cpu overhead. |
356 |
|
357 |
1.02 Fri Jan 13 14:15:40 CET 2006 |
358 |
- add COPYING to clarify license. |
359 |
|
360 |
1.01 Sun Jan 8 05:48:09 CET 2006 |
361 |
- correctly cancel Event timers, avoid memory leaks. |
362 |
|
363 |
1.0 Sun Jan 8 05:40:58 CET 2006 |
364 |
- allow third-party interfaces to be added at runtime. |
365 |
- fix version number. |
366 |
|
367 |
0.9 Fri Dec 30 02:25:41 CET 2005 |
368 |
- fixed errors in the documentation. |
369 |
- going back to using $VERSION as "load check", |
370 |
as suggested by Jörn Reder. Let's see why I stopped using |
371 |
this method. |
372 |
- add 'hup' to the list of events to receive for glib |
373 |
and generate 'rw' events for it, as some versions of |
374 |
glib seem to require this, while this breaks other |
375 |
(older) versions of glib. |
376 |
- implemented PERL_ANYEVENT_VERBOSE env variable. |
377 |
|
378 |
0.3 Sun Dec 4 10:43:33 CET 2005 |
379 |
- no changes to the code. |
380 |
- fixed documentation. |
381 |
- clarified some corner cases. |
382 |
- vastly improved documentation with a long example. |
383 |
|
384 |
0.2 Thu Dec 1 22:18:49 CET 2005 |
385 |
- check whether a valid method was called, to avoid endless loops. |
386 |
- move condvars into their own class (cleaner, safer). |
387 |
- don't die when some modules are not available. |
388 |
- Tk is too broken to allow for words. Will work once Tk works. |
389 |
|
390 |
0.1 Thu Dec 1 19:44:57 CET 2005 |
391 |
- spiced up and cleaned up for release. |
392 |
|
393 |
0.01 Tue Sep 21 18:12:09 CEST 2004 |
394 |
- original version; copied from Convert::Scalar. |
395 |
|