1 |
Revision history for Perl extension AnyEvent. |
2 |
|
3 |
TODO: packagename for ae types |
4 |
|
5 |
- INCOMPATIBLE CHANGE: do no longer support register_read_type |
6 |
and register_write_type in AnyEvent::Handle, instead support |
7 |
package names (the facility was mostly abused). |
8 |
- implement "packagename-as-read/write type" support in AnyEvent::Handle. |
9 |
- AnyEvent::Handle: new options "keepalive" and "oobinline". |
10 |
- oobinline set by default to avoid security issues. |
11 |
|
12 |
5.12 Tue Sep 1 20:26:50 CEST 2009 |
13 |
- be more lenient when parsing resolv.conf files, as some people |
14 |
use hashmarks as comment indicator inside directives (reported |
15 |
by Michael S. Fischer). |
16 |
- use same set of warnings as common::sense 2.0. |
17 |
- fix a potential 32 bit overflow issue due to perl having problems |
18 |
with large hex constants in the Perl backend. |
19 |
- do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS |
20 |
by default. |
21 |
- fix some prototypes. |
22 |
- (optionally) require Async::Interrupt 1.02. |
23 |
- replace opendns fallbacks by verizon/level3 public dns servers. |
24 |
|
25 |
5.112 Fri Aug 21 13:59:04 CEST 2009 |
26 |
- AnyEvent::Handle could cause hard crashes in Net::SSLeay when |
27 |
using starttls with an illegal/uncreatable tls context. |
28 |
- port to broken win32 perls. |
29 |
|
30 |
5.111 Sun Aug 16 18:44:36 CEST 2009 |
31 |
- perl errornously requires sockaddr_un structures returned by the |
32 |
kernel to have a specific length. work around this bug by padding |
33 |
sockaddr_un structures (found to fail on OS X, as perl only uses |
34 |
the correct code for GNU/Linux, but likely to fail on other |
35 |
OSes as well). |
36 |
|
37 |
5.11 Wed Aug 12 17:49:37 CEST 2009 |
38 |
- fixed a bug in run_cmd with input from filename. |
39 |
- create a json coder object when none is given in write direction |
40 |
to reduce dependency on JSON. |
41 |
|
42 |
5.1 Tue Aug 11 03:17:32 CEST 2009 |
43 |
- tcp_connect now reports the correct error in $!, not an unrelated |
44 |
one (was broken since 4.91). |
45 |
- AE::Handle did not properly clear rtimeout/wtimeout watchers |
46 |
when retrying a tcp_connect. |
47 |
- new functions: AnyEvent::Util::run_cmd and close_all_fds_except. |
48 |
- fixed wrong documentation in AE::Socket. |
49 |
|
50 |
5.01 Mon Aug 10 03:16:32 CEST 2009 |
51 |
- last release broke EV child watchers :/. |
52 |
|
53 |
5.0 Sun Aug 9 17:32:09 CEST 2009 |
54 |
- officially introduce and document the AE API. |
55 |
- lowering the timeout at runtime did not have immediate effect |
56 |
in AnyEvent::Handle. |
57 |
- AE::Handle now has separate and independent read and write timeout |
58 |
settings. |
59 |
- probe for Guard module when AnyEvent::Util is loaded, not at |
60 |
runtime (Event doesn't like eval "use" from callbacks, and it avoids |
61 |
exporting the probe function). |
62 |
- Impl::IOAsync's io watchers did not use the filehandle emulation |
63 |
code to work around it's watcher limits. |
64 |
- work around IO::Async's problems with overloaded objects. |
65 |
- try to disable wearnings inside IO::Async, it's just too buggy. |
66 |
- do no longer use Event by default, only EV or Perl, Event is too buggy. |
67 |
- hackishly provide fast AE:: interface when EV or Perl is the backend. |
68 |
- inherit VERSION from AnyEvent::VERSION in most modules. |
69 |
|
70 |
4.91 Thu Aug 6 15:42:45 CEST 2009 |
71 |
- AE::Handle::starttls could get out of sync when the read buffer |
72 |
already contains some TLS handshake. |
73 |
- AE::Handle did not properly free the TLS session, causing bigger |
74 |
memory leaks in Net::SSLeay than expected (reported by toaster). |
75 |
- AE::Socket::tcp_connect will now artificially delay invoking |
76 |
the callback to avoid returning after invoking the callback. |
77 |
- convert many internal watcher uses to AE API. |
78 |
|
79 |
4.9 Sat Aug 1 11:07:01 CEST 2009 |
80 |
- Glib silently fails when registering a timeout with |
81 |
a negative value, so avoid that. |
82 |
- call condvar callback immediately when it is set |
83 |
after the condition is already signalled. |
84 |
- check rbuf_max condition only after trying to consume |
85 |
data. |
86 |
- normalise signal numbers to names when using the ae's |
87 |
signal handling, but do not document this (yet). |
88 |
- pure perl signal emulation did not properly set nonblocking |
89 |
mode on the signal pipe on win32 (this is unlikely to have caused |
90 |
any issues). |
91 |
- new module: AnyEvent::Debug. |
92 |
- AnyEvent::Strict now checks that a signal was specified |
93 |
by name (not name or number). |
94 |
- reduce memory footprint in typical cases by ~50kb by |
95 |
only compiling signal code when necessary. |
96 |
- add AnyEvent::Handle->rbuf_max. |
97 |
- grab the AE namespace for future new API, implement stubs for |
98 |
said future/alternative API. |
99 |
- new function: AnyEvent::Socket::format_hostport. |
100 |
|
101 |
4.881 Tue Jul 28 12:51:53 CEST 2009 |
102 |
- work around a bug in local in pre-5.10 perls, causing |
103 |
AnyEvent::Handle to recurse when it should not |
104 |
(analyzed by elmex). |
105 |
|
106 |
4.88 Tue Jul 28 04:04:37 CEST 2009 |
107 |
- re-bless the handle into a dummy package after calling |
108 |
AnyEvent::Handle::destroy, so the user does not need to check |
109 |
for errors after every push_write etc. |
110 |
- do not attempt to run t/02_signals.t on obviously broken |
111 |
platforms. |
112 |
|
113 |
4.87 Sun Jul 26 02:06:16 CEST 2009 |
114 |
- do not attempt to linger when there is no longer a valid fh. |
115 |
- enforce tls mode to be either accept or connect, do not simply |
116 |
segfault in Net::SSLeay. |
117 |
- AnyEvent::Handle can now call tcp_connect itself (new parameters |
118 |
connect, on_prepare, on_connect and on_connect_error). Updated |
119 |
tutorial accordingly. |
120 |
- add AnyEvent::Impl::Irssi backend. |
121 |
|
122 |
4.86 Mon Jul 20 23:52:29 CEST 2009 |
123 |
- since the verbose warning is not enough, explicitly document |
124 |
that versions before 1.33 of Net::SSLeay are not secure. |
125 |
- work around signal handling races in Event and (...) Event::Lib. |
126 |
- try to align signal-race timer to full-second boundaries. |
127 |
- work around Tk not liking negative timeouts. |
128 |
- don't complain of different grades of environmental unfriendlyness |
129 |
in IO::Async. |
130 |
|
131 |
4.85 Sat Jul 18 06:16:14 CEST 2009 |
132 |
- nail the signal race problem in perl once and for all |
133 |
(see $AnyEvent::MAX_SIGNAL_LATENCY). |
134 |
- take advantage of Async::Interrupt if it is available. |
135 |
- load Time::HiRes and Guard modules on demand only. |
136 |
- add optional/recommended modules section to AnyEvent |
137 |
documentation. |
138 |
- reduce memory usage considerably (and reduce startup penalty) |
139 |
by not using "strict", "warnings" and "overload" modules. |
140 |
- work around buggy windows/openbsd perls and provide EBADMSG |
141 |
and EPROTO ourselves when missing. |
142 |
- improve perl 5.6 compatibility of the core event loop. |
143 |
- made Net::SSLeay version 1.33 a soft requirement. |
144 |
|
145 |
4.83 Fri Jul 17 16:56:26 CEST 2009 |
146 |
- implement AnyEvent::Socket::getprotobyname. |
147 |
- AnyEvent::CondVar's will now detect recursive blocking |
148 |
waits and will croak, as too many people fall into |
149 |
this trap. |
150 |
- AnyEvent::Handle will now call ->destroy on itself after |
151 |
executing the on_error callback, instead of doing some |
152 |
half-baked internal shutdown, for fatal errors. |
153 |
- clarify on_eof behaviour w.r.t. the read queue and |
154 |
on_read callbacks. |
155 |
- ignore some possible spurious wake-ups in tcp_connect. |
156 |
|
157 |
4.82 Sat Jul 11 00:34:55 CEST 2009 |
158 |
- POE and Event backends didn't accept some callable objects as |
159 |
callbacks. |
160 |
- use Config module instead of POSIX module to detect signal names |
161 |
in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX |
162 |
module doesn't even have all POSIX signals :/. |
163 |
- use more workarounds around the many refcnt/corruption bugs in |
164 |
Event::Lib. |
165 |
- work around a race condition in perl's select, causing t/03_child.t |
166 |
to rarely fail. |
167 |
|
168 |
4.81 Thu Jul 9 10:30:30 CEST 2009 |
169 |
- AnyEvent::Handle didn't properly diagnose write errors |
170 |
(it expected -1 from syswrite, how lame... :). |
171 |
- support file descriptors in addition to file handles |
172 |
in AnyEvent->io. |
173 |
- new env variables: PERL_ANYEVENT_RESOLV_CONF, |
174 |
PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE |
175 |
and PERL_ANYEVENT_CA_PATH. |
176 |
- provide a sensible synopsis section for AnyEvent::TLS. |
177 |
- add a "supported backends" section to the manpage. |
178 |
- added simple io watcher test to testsuite, using a |
179 |
portable_socketpair. |
180 |
- tried to improve the stability of the Event::Lib backend, |
181 |
YMMV. |
182 |
|
183 |
4.8 Mon Jul 6 23:45:16 CEST 2009 |
184 |
- AnyEvent::DNS did not properly follow CNAME records with |
185 |
uppercase targets. |
186 |
- AnyEvent::DNS would errornously return AAAA records |
187 |
with v4 mapped addresses (a faulty record) as ipv4 addresses, |
188 |
causing AnyEvent::Socket to throw an exception. |
189 |
- added new module AnyEvent::TLS for easier SSL/TLS context |
190 |
creation, with many options including hostname verification, |
191 |
secure default configuration, lots of documentation and, |
192 |
predefined diffie-hellman keys for perfect forward security |
193 |
and much more. get it while it's still fresh! |
194 |
- use AnyEvent::TLS in AnyEvent::Handle for context management. |
195 |
- load AnyEvent::Handle only on demand in AnyEvent::DNS, |
196 |
so AnyEvent::Socket users have smaller memory footprint |
197 |
in the common case. |
198 |
- add AnyEvent::Handle->push_shutdown method. |
199 |
- add an additional error message parameter to AnyEvent::Handle's |
200 |
on_error callback (for TLS, $! is still available). |
201 |
- add AnyEvent::Handle on_starttls/on_stoptls callbacks. |
202 |
- make AnyEvent::Handle more robust against early conenction |
203 |
failures (during new), and return C<undef> in that case |
204 |
from the constructor. |
205 |
- AnyEvent::Handle will now try to load only JSON::XS first, |
206 |
then fall back to JSON. |
207 |
- format_ipv4/format_ipv6 are now exported by default, for symmetry, |
208 |
and because it was documented that way. |
209 |
|
210 |
4.451 Fri Jul 3 00:28:58 CEST 2009 |
211 |
- do not clear rbuf when shutting down an AnyEvent::Handle |
212 |
object - doing so breaks AnyEvent::HTTP. |
213 |
|
214 |
4.45 Mon Jun 29 22:59:26 CEST 2009 |
215 |
- a write error could cause AnyEvent::Handle to create |
216 |
an I/O watcher with an undefined $fh. |
217 |
- special-case mapped ipv4 addresses in both |
218 |
AnyEvent::Socket::format_address and parse_address, |
219 |
to treat them just like ipv4 addresses. |
220 |
- updated and overhauled the AnyEvent::Intro doc. |
221 |
- implement AnyEvent::Socket::format_ipv4/ipv6. |
222 |
- slightly speed up portable_pipe/socketpair. |
223 |
- expand condvar begin/end documentation. |
224 |
|
225 |
4.42 Fri Jun 26 08:32:18 CEST 2009 |
226 |
- preliminary and neccesarily incomplete support for IO::Async. |
227 |
- reset SIGCHLD to DEFAULT when AnyEvent is loaded, in |
228 |
case it was set to IGNORE, to ensure we can catch child statuses |
229 |
even when the calling env acts stupidly. |
230 |
- updated benchmarks with IO::Async, which performs very well. |
231 |
|
232 |
4.412 Wed Jun 24 01:35:57 CEST 2009 |
233 |
- support an "untaint" attribute for AnyEvent::DNS |
234 |
and set it on the default resolver. |
235 |
- implement and document AnyEvent::Impl::Perl::loop. |
236 |
- remove all anyevent-env variables from %ENV |
237 |
when running in tainted mode. |
238 |
- mention and extend the IO::Lambda benchmark. |
239 |
|
240 |
4.411 Sun Jun 7 18:48:02 CEST 2009 |
241 |
- do not try to use F_SETFD on windows. |
242 |
|
243 |
4.41 Thu May 14 06:40:11 CEST 2009 |
244 |
- work around issues in older perls (5.8.0?) when |
245 |
a signal handler is deleted from the %SIG hash. |
246 |
- use POSIX::_exit in child test, to avoid running |
247 |
destructors. |
248 |
- speed up CHLD handling by relying on SIGCHLD being |
249 |
synchronously delivered, even when we roll our own |
250 |
implementation. |
251 |
- AnyEvent::DNS: add the "dname" resource record name |
252 |
for cosmetic reasons. |
253 |
|
254 |
4.4 Sun Apr 26 20:12:33 CEST 2009 |
255 |
- implemented idle watchers, where applicable. |
256 |
- AnyEvent->time died when Event backend was in use. |
257 |
- fix a memleak in the Tk backend. |
258 |
- sped up Tk timer handling. |
259 |
- clip negative "after" values to 0 in AnyEvent::Impl::Event |
260 |
to avoid spamming out warning messages. |
261 |
- fix Qt timers without interval. |
262 |
- avoid Qt zero-timeout special case, in old libqt's. |
263 |
|
264 |
4.352 Mon Apr 20 16:31:11 CEST 2009 |
265 |
- fix AnyEvent::Strict error messages for child watchers. |
266 |
- fix/update Makefile.PL "recommends" versions. |
267 |
- add AnyEvent->now_update. |
268 |
|
269 |
4.351 Sat Apr 11 07:56:14 CEST 2009 |
270 |
- actually make the signal pipe work under win32. |
271 |
- localise $! in signal handler to avoid changing $!, although |
272 |
perl itself does not seem to save/restore errno either. |
273 |
- set the cloexec flag on the signal pipe (normally set by perl too). |
274 |
|
275 |
4.35 Fri Mar 27 11:48:20 CET 2009 |
276 |
- event models relying on AnyEvent's signal watcher emulation did |
277 |
invoke the callback asynchronously, contradicting documentation |
278 |
and causing signals to get lost (this includes AnyEvent's own |
279 |
event loop). AnyEvent now uses the standard pipe trick to make |
280 |
callback execution synchronous to the event loop. |
281 |
- AnyEvent::Handle didn't free TLS context data on DESTROY |
282 |
(patch by Pavel Shaydo). |
283 |
- work around the ever-incompatibly-changing API of MakeMaker. |
284 |
- document that changing global variables without restoring them |
285 |
is a bad idea in Perl (noted by Adam Rosenstein). |
286 |
- AnyEvent::Strict now barfs if ->io is passed a file. |
287 |
|
288 |
4.34 Thu Feb 12 18:32:45 CET 2009 |
289 |
- separately buffer TLS read data, as otherwise the read queue could |
290 |
deadlock as receiving data is not expected while draining the |
291 |
read queue (which cna only happen with TLS). |
292 |
- raise EBADMSG error on JSON decoding errors. |
293 |
- fix some minor manpage bugs (reported by Maximilian Gaß). |
294 |
- speed up select bitmask parsing quite a bit in the pure perl |
295 |
backend. |
296 |
- use CORE::select instead of just select in AnyEvent::Impl::Perl. |
297 |
|
298 |
4.331 Tue Jan 6 21:07:25 CET 2009 |
299 |
- socketpair fails on many vista machines because vista has |
300 |
completely broken accept/getpeername and getsockname functions, |
301 |
so we provide our own socketpair emulation that kind of works |
302 |
(AnyEvent::Util::portable_pipe). |
303 |
- new function: AnyEvent::Util::portable_socketpair. |
304 |
- take advantage of the Guard module if it exists. |
305 |
|
306 |
4.33 Fri Nov 21 02:35:40 CET 2008 |
307 |
- AnyEvent::Strict did errornously flag a fileno of 0 as |
308 |
illegal. |
309 |
- reduce memory usage and slightly speed up the pure perl backend |
310 |
by only storing the file descriptor, not the file handle. |
311 |
- add missing autocork method to AnyEvent::Handle (reported |
312 |
by Adam Rosenstein). |
313 |
- AnyEvent::DNS->resolve errornously documented an $rcode result |
314 |
argument, but there isn't (spotted by Henrik Krohns). |
315 |
- the naked truth about Tk - it's basically unmaintained. |
316 |
|
317 |
4.32 Mon Nov 3 22:46:32 CET 2008 |
318 |
- fix AnyEvent::Socket::resolve_sockaddr to properly support |
319 |
unix sockets again and choose a proper default for |
320 |
the socket type. |
321 |
- fix call to resolve_sockaddr in tcp_connect. |
322 |
|
323 |
4.31 Thu Oct 30 04:41:48 CET 2008 |
324 |
- implemented AnyEvent::Handle->destroy method. |
325 |
- hint about unexpected effects in TLS mode. |
326 |
- speed up AnyEvent::Impl::Perl by using more arrays |
327 |
instead of hashes without the slightest loss of readability :) |
328 |
- work around a perl argument refcounting bug. |
329 |
- with some perl patches applied, AnyEvent now seems leak-free. |
330 |
|
331 |
4.3 Fri Oct 3 09:18:43 CEST 2008 |
332 |
- AnyEvent will now install a no-op signal handler for SIGPIPE, |
333 |
unless one has been installed already. |
334 |
- warn about not loaidng AnyEvent::Impl::POE early enough |
335 |
(patch by Adam Rosenstein). |
336 |
- fixed a great number of bugs and corner cases in AnyEvent::Handle: |
337 |
- fix a bug in where in SSL connect mode, the |
338 |
client would first wait for some data by the server and |
339 |
otherwise hang. (reported and analyzed in an absolutely |
340 |
exemplary manner by Adam Rosenstein). |
341 |
- fix a bug in where SSL EOF would not be treated |
342 |
as stream EOF, putting the connection into a hung state. |
343 |
(reported and analyzed in an absolutely exemplary manner by Adam |
344 |
Rosenstein). |
345 |
- fix a potential segfault when the TLS context |
346 |
would go missing in a read callback (e.g. due to stoptls) - |
347 |
Net::SSLeay of course makes no type checking whatsoever. |
348 |
- AnyEvent::Handle will not stop the read watcher in TLS mode. |
349 |
- AnyEvent::Handle->stoptls will send a TLS close notify |
350 |
instead of simply destroying the stream now. |
351 |
- fix a bug where large blocks of data written in TLS mode |
352 |
would not be sent unless triggered by receives. |
353 |
- on_drain will now take the tls write buffer into account. |
354 |
- SSL operations have been streamlined a bit and should be faster |
355 |
now, more to come, though. |
356 |
- added a FAQ section, will document nontrivial issues in there. |
357 |
|
358 |
4.234 Mon Sep 29 04:08:13 CEST 2008 |
359 |
- fix child watcher documentation: only child exits will be reported, |
360 |
no trace events. |
361 |
- mention SIGPIPE in AnyEvent::Handle. |
362 |
- perl backend will now use POSIX::times as monotonic clock |
363 |
when available and the monotonic clock option is not. |
364 |
- run even without Time::HiRes in some cases now, but do not |
365 |
provide sub-second accuracy (all *supported* perl versions have |
366 |
Time::HiRes). |
367 |
|
368 |
4.233 Fri Aug 22 01:48:05 CEST 2008 |
369 |
- fix a bug in the testsuite, causing a hang. |
370 |
- clarified lots of AnyEvent::Handle constructor arguments. |
371 |
|
372 |
4.232 Thu Aug 21 20:44:25 CEST 2008 |
373 |
- fix a bug in AnyEvent::Handle that could cause two fatal error |
374 |
callbacks on EOF, which caused AnyEvent::HTTP to signal success twice |
375 |
(testcase by Leon Brocard). |
376 |
- clarified on_eof/on_error documentation a bit. |
377 |
|
378 |
4.231 Tue Jul 29 13:12:15 CEST 2008 |
379 |
- remove some debugging code left in AnyEvent::Util::fork_call |
380 |
(and no, it's impossible to implement with the broken windows |
381 |
perls without resource leaks or worse). |
382 |
|
383 |
4.23 Tue Jul 29 12:19:59 CEST 2008 |
384 |
- document the first parameter passed to condvar callbacks to be |
385 |
the callback. |
386 |
- add AnyEvent::Socket::{ntoa,aton} aliases. |
387 |
- optimize the AE::Handle->push_read (line) for the default |
388 |
eol marker. |
389 |
- optimize push_read (packstring|storable) for small packets. |
390 |
- invoke on_error callback when no on_eof callback is set. |
391 |
- fix a bug in push_read (storable) of unknown impact. |
392 |
|
393 |
4.22 Sun Jul 20 16:34:13 CEST 2008 |
394 |
- new function AnyEvent::Socket::parse_hostport. |
395 |
- as the bulkheads at microsoft can't even get getprotobyname reliably |
396 |
working on their shitty fucking broken os we need to hardcode |
397 |
some common protocol numbers in AnyEvent::Socket. How can |
398 |
people even bother with such a pile of shit as windows. |
399 |
|
400 |
4.21 Thu Jul 17 14:40:05 CEST 2008 |
401 |
- INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype |
402 |
which is incompatible to previous invocation syntax. |
403 |
- work around more windows perl bugs in fork_call. I cannot imagine |
404 |
why anybody would bother implementing fork in such an obviously |
405 |
unusable way. |
406 |
- avoid calling $do_retry if the request has already finished in |
407 |
AnyEvent::DNS, thus avoiding an exception (reported by Anatoly |
408 |
K. Sharifulin). |
409 |
|
410 |
4.2 Sat Jul 12 22:42:11 CEST 2008 |
411 |
- support an "interval" argument for repeating timers. |
412 |
- fix a bug in the pure perl event loop that caused timers to |
413 |
be delayed under some circumstances. |
414 |
- implement optional argument checking via PERL_ANYEVENT_STRICT. |
415 |
- not importing anything from Event causes it to malfunction |
416 |
and not use Time::HiRes, work around this bug. |
417 |
- more bug workarounds against the endlessly broken Tk module, |
418 |
seems to work now on 32 bit machines, but less so on 64 bit. |
419 |
- minor optimisations applied to most backends. |
420 |
|
421 |
4.161 Fri Jul 4 14:24:48 CEST 2008 |
422 |
- fixed the t/handle/02_write.t, it read after write, with the |
423 |
assumption that all data could be written before read |
424 |
(reported by Jim Razmus). |
425 |
|
426 |
4.160 Thu Jul 3 04:02:21 CEST 2008 |
427 |
- re-registering signal watchers after unregistering one did |
428 |
not work if a backend falls back on AnyEvent's default |
429 |
implementation (also affected child watchers). |
430 |
- new AnyEvent::Handle options: autocork and no_delay. |
431 |
|
432 |
4.152 Sun Jun 22 14:15:44 CEST 2008 |
433 |
- allow for 32-bit perls that implement shifts differently |
434 |
on different architectures in parse_ipv4 (reported and |
435 |
analysed by Keiichi DAIBA). |
436 |
|
437 |
4.151 Fri Jun 6 17:34:24 CEST 2008 |
438 |
- make sure specifying _only_ on_read and never pushing reads |
439 |
works. |
440 |
|
441 |
4.15 Fri Jun 6 13:00:46 CEST 2008 |
442 |
- the pure perl backend would keep some watchers alive when more than |
443 |
one watcher was registered for the same fd. |
444 |
- new "packstring" and "storable" read and write types |
445 |
for AnyEvent::Handle. |
446 |
- allow on_eof handler to be called after on_error with EPIPE returns. |
447 |
- do not immediately call on_read callback in handle constructor. |
448 |
|
449 |
4.14 Thu Jun 5 20:29:31 CEST 2008 |
450 |
- Fixed a bug in DNS SRV priority sorting. |
451 |
- AnyEvent::Util::guard now reports runtime errors while |
452 |
executing the guard block as warnings. |
453 |
- handle 0-byte-reads just before EOF correctly in AnyEvent::Handle. |
454 |
|
455 |
4.13 Thu Jun 5 00:47:59 CEST 2008 |
456 |
- AnyEvent::DNS only followed cname chains with length 2, |
457 |
contrary to documentation. bumped it up to 10, thanks to |
458 |
microsoft, the current broken-dns-config-king. |
459 |
- AnyEvent::DNS didn't check for socket return status |
460 |
but instead relied on perl not creating filehandles in that |
461 |
case - too bad it gives you a bogus file handle (reported |
462 |
and analysed by Vladimir Timofeev). |
463 |
- fix queue management logic in AnyEvent::Handle: |
464 |
when on_read was registered, the queue was empty |
465 |
and no progress could be made AnyEvent::Handle would |
466 |
enter an endless loop. |
467 |
- correctly start reading again when the handle became |
468 |
busy again after idling. |
469 |
- correctly treat tls shut-downs as EOF condition |
470 |
(for the time being, would be nice to have a callback for that). |
471 |
- correctly call eof callback under all conditions |
472 |
(hopefully). |
473 |
- the timeout callback did not expect that $self can go |
474 |
away any time. |
475 |
|
476 |
4.12 Tue Jun 3 10:58:04 CEST 2008 |
477 |
- include AnyEvent::Intro, a tutorial for anyevent, |
478 |
anyevent::socket and anyevent::handle. |
479 |
- allow more options in on_error. |
480 |
|
481 |
4.11 Fri May 30 23:42:25 CEST 2008 |
482 |
- INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide |
483 |
reverse_lookup and reverse_verify to replace it, support |
484 |
v4mapped and v4compat addresses. |
485 |
- provide more documentation for the resolver class. |
486 |
- really replace longest run of :0: by :: in format_address, |
487 |
also properly convert :: and ::1 again. |
488 |
- support NAPTR record name and decode it. |
489 |
- implement random weight sampling for SRV records, as per |
490 |
rfc 2782. |
491 |
- correctly abort on srv-record targets of ".". |
492 |
- added AnyEvent::DNS::wait_for_slot. |
493 |
- in the unlikely event of a virtual circuit connection |
494 |
being invalidated by a delayed udp reply, AnyEvent::DNS |
495 |
could die. this has been fixed. |
496 |
- plug a probable memleak in the DNS vc code. |
497 |
- use configured timeout also for VC request phase. |
498 |
- implement timeout and max_outstanding methods |
499 |
for resolver class. |
500 |
- update version numbers in all modules. |
501 |
|
502 |
4.1 Thu May 29 05:45:40 CEST 2008 |
503 |
- INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address |
504 |
which mirrors their purpose better (old names still available |
505 |
for a while). Also moved AnyEvent::DNS::addr to |
506 |
AnyEvent::Socket::resolve_sockaddr. |
507 |
- implement AnyEvent->time and AnyEvent->now. |
508 |
- fix IPv6 support in pack_sockaddr. |
509 |
- officially un-experimentalise all newly introduced code. |
510 |
- support unix domain sockets everywhere by specifying |
511 |
a host of "unix/" and the pathname as service. |
512 |
- implement an activity timeout in AnyEvent::Handle. |
513 |
- added a regex read type to AnyEvent::Handle. |
514 |
- added a json read/write type to AnyEvent::Handle. |
515 |
- always croak in on_error in AnyEvent::Handle. |
516 |
- document how to subclass AnyEvent::Handle. |
517 |
- implement AnyEvent::Util::fork_call. |
518 |
- add support for IPv6 nameservers and nameserver statements. |
519 |
- work around _yet_ _another_ windows perl bug in where empty |
520 |
select masks cause errors in the pure perl backend (this |
521 |
workaround cannot be implemented for the other event loops, |
522 |
but EV also works around this). |
523 |
- supply AnyEvent::Socket::address_type and make good use of it. |
524 |
- clarify and fix the Handle documentation w.r.t $self vs. $handle |
525 |
vs. callback arguments. |
526 |
- add some recommends to the META.yml. |
527 |
|
528 |
4.05 Mon May 26 19:44:06 CEST 2008 |
529 |
- some platforms ignore the rfc and prepend an extra sa_len member to |
530 |
the sockaddr structure, cater for those. also use sockaddr_family, |
531 |
didn't know it was there, but comes in handy for the workaround. |
532 |
- undo the import WIN32 hack. |
533 |
|
534 |
4.04 Mon May 26 08:03:31 CEST 2008 |
535 |
- try to work around yet another windows bug: failed connects are reported |
536 |
as if it were out-of-band data. windows users: you suck. Only |
537 |
EV, Glib, Event and the pure perl backend can handle this condition. |
538 |
- optimize write algorithm in AnyEvent::Handle. |
539 |
- properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly. |
540 |
- tcp_connect now properly iterates through all targets. |
541 |
- check wether IPv6 sockets can actually be created, otherwise, disable ipv6 |
542 |
support entirely. also hardcode some AF_INET6 constants for the |
543 |
sake of perl 5.8. |
544 |
- vastly improved nameserver/suffix detection algorithm for windows |
545 |
(still a hack...) |
546 |
- try to use a fallback dns server if no dns servers could be found. |
547 |
- splatter around some local $SIG{__DIE__}'s to work around bad |
548 |
die handlers. |
549 |
|
550 |
4.03 Sun May 25 05:05:57 CEST 2008 |
551 |
- work around perl bugs on windows, where perl returns undocumented |
552 |
error codes for sysread, syswrite etc. |
553 |
- fix AnyEvent::Handle::unshift_read to really unshift and not push. |
554 |
- AnyEvent::Handle could sometimes stumble over 0-byte writes. |
555 |
- fix unpack_sockaddr, now works for AF_INET6 sockaddr's. |
556 |
- updated documentation. |
557 |
- lots of workarounds for perl 5.8 quirks. |
558 |
- implement a more flexible type system for both reads and writes |
559 |
in AnyEvent::Handle. |
560 |
- support IPv6 in tcp_server. |
561 |
- pass local hostname and port to the prepare callback in tcp_server. |
562 |
- make types extendable in AnyEvent::Handle. |
563 |
- croak, not die, in AnyEvent::Handle; |
564 |
|
565 |
4.0 Sat May 24 19:58:08 CEST 2008 |
566 |
- added (experimental) AnyEvent::DNS module. |
567 |
- added (experimental) AnyEvent::Socket module. |
568 |
- added (experiemntal) support for TLS/SSL in AnyEvent::Handle. |
569 |
- make condvar->end call ->send as documented. |
570 |
- make condvar constructor support the documented cb => argument. |
571 |
- overload condvars so you can use them instead of code references. |
572 |
- support more of the POSIX forms instead of a pure dotted quad for |
573 |
verbatim ip addresses. |
574 |
- AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. |
575 |
- try very hard not to freeze on broken windows perls. |
576 |
- make on_eof optional in AnyEvent::Handle. |
577 |
- support service names in tcp_connect and tcp_server. |
578 |
- work around netbsd/openbsd bug causing unix sockets not to |
579 |
become writable until empty. |
580 |
- only make a dummy read for error detection when getpeername |
581 |
returns ENOTCONN. |
582 |
- defend AnyEvent::Util::inet_aton against Coro::LWP's brutal |
583 |
override. |
584 |
- new AnyEvent::Guard::cancel. |
585 |
|
586 |
3.5 Sat May 17 23:17:14 CEST 2008 |
587 |
- gracefully deal with systems not defining CLOCK_MONOTONIC |
588 |
(as opposed to not supporting the clock itself, which worked in 3.41). |
589 |
- added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server |
590 |
helpers, to replace the ill-designed AnyEvent::Socket class. |
591 |
- added some examples (eg/connect, eg/handle). |
592 |
- added AnyEvent::Util::guard. |
593 |
- AnyEvent::Util now exports most of its functions. |
594 |
- fixed a bug in *_read_chunk, where the callback was called with $self as |
595 |
first and second argument. |
596 |
- simplified and fixed regex handling in AnyEvent::Handle::*_read_line. |
597 |
- fix manpage condvars examples to use new syntax. |
598 |
- considerably improve inet_aton with EV::ADNS backend to support |
599 |
cname chains (common dns configuration bug, e.g. with www.google.de) |
600 |
and aaaa records if no a record could be found. |
601 |
|
602 |
3.41 Sun May 11 19:53:13 CEST 2008 |
603 |
- work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. |
604 |
|
605 |
3.4 Sun May 11 00:29:25 CEST 2008 |
606 |
- complete re-implementation of condvars with a hopefully much more useful |
607 |
API (backward compatible functions exist but are undocumented). |
608 |
- AnyEvent::Handle fully rewritten with a hopefully more powerful API |
609 |
(still subject to change). |
610 |
- take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
611 |
- provide AnyEvent::post_detect and @AnyEvent::post_detect, which |
612 |
allows module authors to avoid forcing event loop detection. |
613 |
(used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example). |
614 |
- remove coro backends: Coro now provides generic support for AnyEvent |
615 |
via Coro::AnyEvent. |
616 |
|
617 |
3.3 Mon Apr 28 09:51:06 CEST 2008 |
618 |
- added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. |
619 |
- fix a bug in the pure perl backend that kept watchers alive |
620 |
when multiple watchers were registered for the same fh/poll combo. |
621 |
- add a benchmark section showing AnyEvent overhead and comparing |
622 |
the different event loops with each other. |
623 |
- prefer pure perl over tk when autoprobing, as it's about as fast, |
624 |
but doesn't crash with many watchers. |
625 |
- declare Qt support non-experimental. |
626 |
- clarify the confusing section about the file descriptor being |
627 |
kept alive. |
628 |
- document the race between loading of an event module and |
629 |
child processes exiting. |
630 |
- support POE as "backend" (with some caveats, POE is not generic |
631 |
enough, and darn slow). |
632 |
- support Wx and Prima through POE. |
633 |
- optimise perl backend to use 20% less memory and take advantage |
634 |
of typical timeout behaviour. It can now compete with |
635 |
select/poll-based C event loops in most cases (it is usually |
636 |
faster than Event and Glib :). |
637 |
- roughly cut EV memory use in half and increase its speed by 30%, |
638 |
by removing undocumented functionality. Did something similar |
639 |
to other event loops where possible. |
640 |
|
641 |
3.2 Thu Apr 24 10:10:40 CEST 2008 |
642 |
- do not die when anyevent watchers are destroyed while running callbacks |
643 |
in the pureperl backend (could only happen when two watchers are registered |
644 |
for the same fh/poll combo). |
645 |
- support autoloading for child watchers, was broken in all versions. |
646 |
- implement PERL_ANYEVENT_MODEL env variable. |
647 |
- (experimental) implement interface to the (very crashy/buggy) Event::Lib module. |
648 |
- (experimental) implement interface to the Qt module (cannot be autoprobed). |
649 |
- this release is IO:AnyEvent-proof. |
650 |
|
651 |
3.12 Tue Apr 22 07:11:46 CEST 2008 |
652 |
- reinstate AUTHOR section that got lost somehow. |
653 |
- do not hang in the testsuite with badly broken perls (activestate, |
654 |
strawberry...), but instead diagnose the problem and continue. |
655 |
- use INT instead of CHLD in an attempt to support broken windows perls |
656 |
better (this decreases the test quality, unfortunately). |
657 |
- do not send the signal to the process group |
658 |
(no problem for CHLD, bad for INT :). |
659 |
|
660 |
3.11 Sat Apr 19 06:57:31 CEST 2008 |
661 |
- major documentation rework. |
662 |
- document the fact that child watchers only watch for zombies. |
663 |
- fix the child watcher example. |
664 |
|
665 |
3.1 Wed Apr 16 17:09:01 CEST 2008 |
666 |
- work around recurring bugs in Tk by dup'ing filehandles, the |
667 |
only method with good success chances on Tk (the bugs apparently |
668 |
don't get fixed anytime soon). |
669 |
- lift the restriction of only one watcher per fh direction |
670 |
(as the Tk bug workaround also lifts it and only Tk imposed |
671 |
such strong limits). |
672 |
- changed probe order to prefer coro adaptors. |
673 |
- explain why recursion into the event loop is not supported |
674 |
unless the backend supports it (only Coro::EV does without |
675 |
any restrictions...). |
676 |
- add simple manpages for all backend modules. |
677 |
|
678 |
3.0 Mon Apr 7 21:30:23 CEST 2008 |
679 |
- Coro::Signal changed semantics, roll our own, also cleaning |
680 |
up the Coro implementation in general. |
681 |
- rename Coro backend to CoroEvent. |
682 |
- add some decision helping paragraph to the manpage that should |
683 |
help people to decide whether AnyEvent is the right thing for them. |
684 |
|
685 |
2.9 Mon Jan 28 13:31:54 CET 2008 |
686 |
- update for EV 3.0 API changes. |
687 |
|
688 |
2.8 Sun Nov 25 15:06:03 CET 2007 |
689 |
- waitpid can validly return 0. accept this fact of life |
690 |
instead of reporting it to any watchers. |
691 |
|
692 |
2.7 Fri Nov 23 11:41:14 CET 2007 |
693 |
- force use of AnyEvent::Impl::Perl in testsuite, there is |
694 |
too much breakage outside AnyEvent. |
695 |
- deliver signals synchronously in AnyEvent::Impl::Perl |
696 |
or any other event loop that relies on AnyEvents child |
697 |
watcher emulation. *Could* help with hanging testsuite |
698 |
(except when Event or EV are installed). |
699 |
|
700 |
2.6 Fri Nov 9 20:36:35 CET 2007 |
701 |
- fix bug in testsuite. |
702 |
- move EV adaptor modules to AnyEvent. |
703 |
- add Coro+EV adaptor module. |
704 |
|
705 |
2.55 Tue Nov 6 17:41:32 CET 2007 |
706 |
- add EV to the list of supported event models. |
707 |
- do not auto-reset pid watchers, pass pid and status to them. |
708 |
- allow a pid of zero to watch for all children in child watchers. |
709 |
|
710 |
2.54 Wed Jul 18 17:36:23 CEST 2007 |
711 |
- work around a perl bug that results in |
712 |
BEGIN not safe after errors--compilation aborted without |
713 |
any discernible reason or error message by once |
714 |
again not use'ing strict. |
715 |
|
716 |
2.53 Sun Jul 8 10:51:53 CEST 2007 |
717 |
- make Glib interface work again (spotted by elmex). |
718 |
|
719 |
2.52 Wed Mar 7 18:36:16 CET 2007 |
720 |
- child watchers stopped working permanently when |
721 |
all current child watchers were destroyed. |
722 |
|
723 |
2.51 Mon Dec 11 21:33:24 CET 2006 |
724 |
- work around bugs in perl where eval "require Module" returns |
725 |
true even if the module couldn't be loaded :/. |
726 |
|
727 |
2.5 Mon Dec 11 02:15:28 CET 2006 |
728 |
- avoid $AUTOLOAD because many perls corrupt it. |
729 |
- AnyEvent::detect forces autodetection. |
730 |
- implement signal watchers (experimental) |
731 |
(Tk does not support async signals (<= 804.027 at least), |
732 |
so they might get delayed indefinitely on Tk). |
733 |
- implement child watchers (experimental). |
734 |
- moved default condvar implementation into base module, simplifying |
735 |
most implementation modules. |
736 |
|
737 |
2.1 Fri Nov 24 15:50:48 CET 2006 |
738 |
- better docs. |
739 |
- simple testsuite. |
740 |
- added AnyEvent->one_event method for special purposes. |
741 |
|
742 |
2.0 Wed Nov 1 02:21:30 CET 2006 |
743 |
- INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both. |
744 |
- INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore. |
745 |
- EXPERIMENTAL: |
746 |
new pure-perl model added, AnyEvent now always finds a suitable |
747 |
event model. |
748 |
- improved documentation. |
749 |
- reduced cpu overhead. |
750 |
|
751 |
1.02 Fri Jan 13 14:15:40 CET 2006 |
752 |
- add COPYING to clarify license. |
753 |
|
754 |
1.01 Sun Jan 8 05:48:09 CET 2006 |
755 |
- correctly cancel Event timers, avoid memory leaks. |
756 |
|
757 |
1.0 Sun Jan 8 05:40:58 CET 2006 |
758 |
- allow third-party interfaces to be added at runtime. |
759 |
- fix version number. |
760 |
|
761 |
0.9 Fri Dec 30 02:25:41 CET 2005 |
762 |
- fixed errors in the documentation. |
763 |
- going back to using $VERSION as "load check", |
764 |
as suggested by Jörn Reder. Let's see why I stopped using |
765 |
this method. |
766 |
- add 'hup' to the list of events to receive for glib |
767 |
and generate 'rw' events for it, as some versions of |
768 |
glib seem to require this, while this breaks other |
769 |
(older) versions of glib. |
770 |
- implemented PERL_ANYEVENT_VERBOSE env variable. |
771 |
|
772 |
0.3 Sun Dec 4 10:43:33 CET 2005 |
773 |
- no changes to the code. |
774 |
- fixed documentation. |
775 |
- clarified some corner cases. |
776 |
- vastly improved documentation with a long example. |
777 |
|
778 |
0.2 Thu Dec 1 22:18:49 CET 2005 |
779 |
- check whether a valid method was called, to avoid endless loops. |
780 |
- move condvars into their own class (cleaner, safer). |
781 |
- don't die when some modules are not available. |
782 |
- Tk is too broken to allow for words. Will work once Tk works. |
783 |
|
784 |
0.1 Thu Dec 1 19:44:57 CET 2005 |
785 |
- spiced up and cleaned up for release. |
786 |
|
787 |
0.01 Tue Sep 21 18:12:09 CEST 2004 |
788 |
- original version; copied from Convert::Scalar. |
789 |
|