1 |
Revision history for Perl extension AnyEvent. |
2 |
|
3 |
TODO: docs signal for condvar |
4 |
|
5 |
- postpone differently in AnyEvent::Socket now, as |
6 |
when not, canceling the connection attempt might fail |
7 |
(found by Felix Antonius Wilhelm Ostmann). |
8 |
- explicitly check for non-stream sockets in AE::Handle, too many |
9 |
clueless people fell into the trap of this somehow working. |
10 |
- simplified and reworked the "OTHER MODULES" section. |
11 |
|
12 |
5.261 Wed Apr 28 16:13:36 CEST 2010 |
13 |
- AF_INET6 was not properly used from Socket6 during configuration |
14 |
time (found by Andy Grundman). |
15 |
|
16 |
5.26 Mon Apr 12 04:49:35 CEST 2010 |
17 |
- don't generate spurious readyness notifications when select |
18 |
returns EINTR in the pure perl backend (analysed and testcase |
19 |
by Bas Denissen). |
20 |
- give same examples in SYNOPSIS sections of AnyEvent and AE. |
21 |
- provider faster implementations for AE::time/now/now_update for |
22 |
EV and Perl backends. |
23 |
- provide faster AE::cv and AE::time implementations for all |
24 |
backends. |
25 |
- no longer support lower/mixed-case signal names (this was |
26 |
never documented nor universally supported). |
27 |
- some more memory µ-optimisations, and somewhat less messy |
28 |
code to cope with both AnyEvent and AE APIs. |
29 |
|
30 |
5.251 Sat Mar 13 00:58:21 CET 2010 |
31 |
- make test failed if it was run before make install on perl <= |
32 |
5.8.8, and older versions of perl make our live very difficult, |
33 |
so write constants.pl during Makefile.PL time. |
34 |
|
35 |
5.25 Sat Mar 13 00:23:14 CET 2010 |
36 |
- fix a race condition in AnyEvent::Handle that would cause |
37 |
a "bio_read: unsupported method" error (in _another_ |
38 |
ssl connection) after user code threw an exception. |
39 |
- added AnyEvent::Handle->destroyed method. |
40 |
- speed up AnyEvent::Socket::format_address and ::format_ipv6. |
41 |
- the AnyEvent::Util::fh_nonblocking function worked only |
42 |
by lucky accident on win32. |
43 |
- smaller and faster AnyEvent::Util::fh_nonblocking. |
44 |
- when the (required!) Time::HiRes module is missing, AnyEvent |
45 |
did not fall back to built-in time properly. |
46 |
- do not load Fcntl at runtime, saving memory and loading time. |
47 |
- precompile a number of constants and use them instead of runtime |
48 |
detection and eval. |
49 |
- free detection code after detection and similar memory optimisations. |
50 |
- Perl backend timer interval best effort drift has been improved |
51 |
(same algorithm as EV). |
52 |
- update unicode idna mapping table. |
53 |
|
54 |
5.24 Tue Jan 5 11:39:43 CET 2010 |
55 |
- cygwin never reports errors from failed connects. 1.5 just gives |
56 |
you continous readyness and EAGAIN, 1.7 is even more broken and |
57 |
just hangs. work around both issues in a cygwin-specific hack. |
58 |
- improve idle watcher documentation slightly. |
59 |
|
60 |
5.23 Sun Dec 20 23:48:00 CET 2009 |
61 |
- support IDNs in resolve_sockaddr, and therefore in tcp_connect. |
62 |
- implement punycode_encode/decode, idn_nameprep, |
63 |
idn_to_ascii and idn_to_unicode operations in AnyEvent::Util. |
64 |
- provide $AE::VERSION. |
65 |
- removed traces of "no strict 'refs'". |
66 |
|
67 |
5.22 Sat Dec 5 03:51:13 CET 2009 |
68 |
- downgrade-or-fail in AnyEvent::Handle::push_write, to |
69 |
diagnose encoding failures earlier and more succinctly. |
70 |
(this works around bugs in perl, throwing away encoding info |
71 |
when passing scalar data to extensions). |
72 |
- add more examples to AnyEvent::Socket manpage. |
73 |
- upgrade internal warning set to the same as common::sense 2.03. |
74 |
- use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS |
75 |
(John Beppu). |
76 |
- AnyEvent::Socket::inet_aton now properly supports ipv6, as documented. |
77 |
- add google public dns servers to fallback server set and make sure |
78 |
we load-balance properly between the three sets. also add all |
79 |
fallback dns servers, not just a random one, to each dns config. |
80 |
|
81 |
5.21 Thu Nov 19 02:48:47 CET 2009 |
82 |
- fix a problem where socket constants were called with parameters |
83 |
(spotted by David Friedland). |
84 |
- fork_call never use'd POSIX (reported by Daisuke Maki). |
85 |
- improve perl 5.6 compatibility further (but it still won't work |
86 |
unless you rip out everything but the core). |
87 |
- prefer Net::DNS::Resolver over ipconfig, if installed, on win32. |
88 |
uses 10MB of RAM, but doesn't open a console window. *sigh*. |
89 |
|
90 |
5.202 Wed Oct 14 22:35:44 CEST 2009 |
91 |
- AnyEvent::DNS would unexpectedly clobber $_ under windows |
92 |
(analysed by Matthias Waldorf). |
93 |
- AnyEvent::Handle::run_cmd can now pass the PID of the |
94 |
newly-created process, which is much less useful than it might |
95 |
sound (based on patch by Yann Kerherve). |
96 |
|
97 |
5.201 Tue Sep 29 12:09:25 CEST 2009 |
98 |
- AnyEvent:Handle::on_starttls/on_stoptls methods were broken |
99 |
(reported by Torsten Foertsch). |
100 |
- common::sense 2.0 could cause tcp_server to throw an exception |
101 |
(analysed by elmex). |
102 |
|
103 |
5.2 Mon Sep 14 07:04:49 CEST 2009 |
104 |
- INCOMPATIBLE CHANGE: do no longer support register_read_type |
105 |
and register_write_type in AnyEvent::Handle, instead support |
106 |
package names (the facility was mostly abused). |
107 |
- implement "packagename-as-read/write type" support in |
108 |
AnyEvent::Handle. |
109 |
- AnyEvent::Handle: new options "keepalive" and "oobinline". |
110 |
- oobinline set by default to avoid security issues. |
111 |
- the pure-perl event loop backend wrongly detected times() fallback |
112 |
support (spotted by Pavel Boldin). |
113 |
|
114 |
5.12 Tue Sep 1 20:26:50 CEST 2009 |
115 |
- be more lenient when parsing resolv.conf files, as some people |
116 |
use hashmarks as comment indicator inside directives (reported |
117 |
by Michael S. Fischer). |
118 |
- use same set of warnings as common::sense 2.0. |
119 |
- fix a potential 32 bit overflow issue due to perl having problems |
120 |
with large hex constants in the Perl backend. |
121 |
- do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS |
122 |
by default. |
123 |
- fix some prototypes. |
124 |
- (optionally) require Async::Interrupt 1.02. |
125 |
- replace opendns fallbacks by verizon/level3 public dns servers. |
126 |
|
127 |
5.112 Fri Aug 21 13:59:04 CEST 2009 |
128 |
- AnyEvent::Handle could cause hard crashes in Net::SSLeay when |
129 |
using starttls with an illegal/uncreatable tls context. |
130 |
- port to broken win32 perls. |
131 |
|
132 |
5.111 Sun Aug 16 18:44:36 CEST 2009 |
133 |
- perl errornously requires sockaddr_un structures returned by the |
134 |
kernel to have a specific length. work around this bug by padding |
135 |
sockaddr_un structures (found to fail on OS X, as perl only uses |
136 |
the correct code for GNU/Linux, but likely to fail on other |
137 |
OSes as well). |
138 |
|
139 |
5.11 Wed Aug 12 17:49:37 CEST 2009 |
140 |
- fixed a bug in run_cmd with input from filename. |
141 |
- create a json coder object when none is given in write direction |
142 |
to reduce dependency on JSON. |
143 |
|
144 |
5.1 Tue Aug 11 03:17:32 CEST 2009 |
145 |
- tcp_connect now reports the correct error in $!, not an unrelated |
146 |
one (was broken since 4.91). |
147 |
- AE::Handle did not properly clear rtimeout/wtimeout watchers |
148 |
when retrying a tcp_connect. |
149 |
- new functions: AnyEvent::Util::run_cmd and close_all_fds_except. |
150 |
- fixed wrong documentation in AE::Socket. |
151 |
|
152 |
5.01 Mon Aug 10 03:16:32 CEST 2009 |
153 |
- last release broke EV child watchers :/. |
154 |
|
155 |
5.0 Sun Aug 9 17:32:09 CEST 2009 |
156 |
- officially introduce and document the AE API. |
157 |
- lowering the timeout at runtime did not have immediate effect |
158 |
in AnyEvent::Handle. |
159 |
- AE::Handle now has separate and independent read and write timeout |
160 |
settings. |
161 |
- probe for Guard module when AnyEvent::Util is loaded, not at |
162 |
runtime (Event doesn't like eval "use" from callbacks, and it avoids |
163 |
exporting the probe function). |
164 |
- Impl::IOAsync's io watchers did not use the filehandle emulation |
165 |
code to work around it's watcher limits. |
166 |
- work around IO::Async's problems with overloaded objects. |
167 |
- try to disable wearnings inside IO::Async, it's just too buggy. |
168 |
- do no longer use Event by default, only EV or Perl, Event is too buggy. |
169 |
- hackishly provide fast AE:: interface when EV or Perl is the backend. |
170 |
- inherit VERSION from AnyEvent::VERSION in most modules. |
171 |
|
172 |
4.91 Thu Aug 6 15:42:45 CEST 2009 |
173 |
- AE::Handle::starttls could get out of sync when the read buffer |
174 |
already contains some TLS handshake. |
175 |
- AE::Handle did not properly free the TLS session, causing bigger |
176 |
memory leaks in Net::SSLeay than expected (reported by toaster). |
177 |
- AE::Socket::tcp_connect will now artificially delay invoking |
178 |
the callback to avoid returning after invoking the callback. |
179 |
- convert many internal watcher uses to AE API. |
180 |
|
181 |
4.9 Sat Aug 1 11:07:01 CEST 2009 |
182 |
- Glib silently fails when registering a timeout with |
183 |
a negative value, so avoid that. |
184 |
- call condvar callback immediately when it is set |
185 |
after the condition is already signalled. |
186 |
- check rbuf_max condition only after trying to consume |
187 |
data. |
188 |
- normalise signal numbers to names when using the ae's |
189 |
signal handling, but do not document this (yet). |
190 |
- pure perl signal emulation did not properly set nonblocking |
191 |
mode on the signal pipe on win32 (this is unlikely to have caused |
192 |
any issues). |
193 |
- new module: AnyEvent::Debug. |
194 |
- AnyEvent::Strict now checks that a signal was specified |
195 |
by name (not name or number). |
196 |
- reduce memory footprint in typical cases by ~50kb by |
197 |
only compiling signal code when necessary. |
198 |
- add AnyEvent::Handle->rbuf_max. |
199 |
- grab the AE namespace for future new API, implement stubs for |
200 |
said future/alternative API. |
201 |
- new function: AnyEvent::Socket::format_hostport. |
202 |
|
203 |
4.881 Tue Jul 28 12:51:53 CEST 2009 |
204 |
- work around a bug in local in pre-5.10 perls, causing |
205 |
AnyEvent::Handle to recurse when it should not |
206 |
(analysed by elmex). |
207 |
|
208 |
4.88 Tue Jul 28 04:04:37 CEST 2009 |
209 |
- re-bless the handle into a dummy package after calling |
210 |
AnyEvent::Handle::destroy, so the user does not need to check |
211 |
for errors after every push_write etc. |
212 |
- do not attempt to run t/02_signals.t on obviously broken |
213 |
platforms. |
214 |
|
215 |
4.87 Sun Jul 26 02:06:16 CEST 2009 |
216 |
- do not attempt to linger when there is no longer a valid fh. |
217 |
- enforce tls mode to be either accept or connect, do not simply |
218 |
segfault in Net::SSLeay. |
219 |
- AnyEvent::Handle can now call tcp_connect itself (new parameters |
220 |
connect, on_prepare, on_connect and on_connect_error). Updated |
221 |
tutorial accordingly. |
222 |
- add AnyEvent::Impl::Irssi backend. |
223 |
|
224 |
4.86 Mon Jul 20 23:52:29 CEST 2009 |
225 |
- since the verbose warning is not enough, explicitly document |
226 |
that versions before 1.33 of Net::SSLeay are not secure. |
227 |
- work around signal handling races in Event and (...) Event::Lib. |
228 |
- try to align signal-race timer to full-second boundaries. |
229 |
- work around Tk not liking negative timeouts. |
230 |
- don't complain of different grades of environmental unfriendlyness |
231 |
in IO::Async. |
232 |
|
233 |
4.85 Sat Jul 18 06:16:14 CEST 2009 |
234 |
- nail the signal race problem in perl once and for all |
235 |
(see $AnyEvent::MAX_SIGNAL_LATENCY). |
236 |
- take advantage of Async::Interrupt if it is available. |
237 |
- load Time::HiRes and Guard modules on demand only. |
238 |
- add optional/recommended modules section to AnyEvent |
239 |
documentation. |
240 |
- reduce memory usage considerably (and reduce startup penalty) |
241 |
by not using "strict", "warnings" and "overload" modules. |
242 |
- work around buggy windows/openbsd perls and provide EBADMSG |
243 |
and EPROTO ourselves when missing. |
244 |
- improve perl 5.6 compatibility of the core event loop. |
245 |
- made Net::SSLeay version 1.33 a soft requirement. |
246 |
|
247 |
4.83 Fri Jul 17 16:56:26 CEST 2009 |
248 |
- implement AnyEvent::Socket::getprotobyname. |
249 |
- AnyEvent::CondVar's will now detect recursive blocking |
250 |
waits and will croak, as too many people fall into |
251 |
this trap. |
252 |
- AnyEvent::Handle will now call ->destroy on itself after |
253 |
executing the on_error callback, instead of doing some |
254 |
half-baked internal shutdown, for fatal errors. |
255 |
- clarify on_eof behaviour w.r.t. the read queue and |
256 |
on_read callbacks. |
257 |
- ignore some possible spurious wake-ups in tcp_connect. |
258 |
|
259 |
4.82 Sat Jul 11 00:34:55 CEST 2009 |
260 |
- POE and Event backends didn't accept some callable objects as |
261 |
callbacks. |
262 |
- use Config module instead of POSIX module to detect signal names |
263 |
in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX |
264 |
module doesn't even have all POSIX signals :/. |
265 |
- use more workarounds around the many refcnt/corruption bugs in |
266 |
Event::Lib. |
267 |
- work around a race condition in perl's select, causing t/03_child.t |
268 |
to rarely fail. |
269 |
|
270 |
4.81 Thu Jul 9 10:30:30 CEST 2009 |
271 |
- AnyEvent::Handle didn't properly diagnose write errors |
272 |
(it expected -1 from syswrite, how lame... :). |
273 |
- support file descriptors in addition to file handles |
274 |
in AnyEvent->io. |
275 |
- new env variables: PERL_ANYEVENT_RESOLV_CONF, |
276 |
PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE |
277 |
and PERL_ANYEVENT_CA_PATH. |
278 |
- provide a sensible synopsis section for AnyEvent::TLS. |
279 |
- add a "supported backends" section to the manpage. |
280 |
- added simple io watcher test to testsuite, using a |
281 |
portable_socketpair. |
282 |
- tried to improve the stability of the Event::Lib backend, |
283 |
YMMV. |
284 |
|
285 |
4.8 Mon Jul 6 23:45:16 CEST 2009 |
286 |
- AnyEvent::DNS did not properly follow CNAME records with |
287 |
uppercase targets. |
288 |
- AnyEvent::DNS would errornously return AAAA records |
289 |
with v4 mapped addresses (a faulty record) as ipv4 addresses, |
290 |
causing AnyEvent::Socket to throw an exception. |
291 |
- added new module AnyEvent::TLS for easier SSL/TLS context |
292 |
creation, with many options including hostname verification, |
293 |
secure default configuration, lots of documentation and, |
294 |
predefined diffie-hellman keys for perfect forward security |
295 |
and much more. get it while it's still fresh! |
296 |
- use AnyEvent::TLS in AnyEvent::Handle for context management. |
297 |
- load AnyEvent::Handle only on demand in AnyEvent::DNS, |
298 |
so AnyEvent::Socket users have smaller memory footprint |
299 |
in the common case. |
300 |
- add AnyEvent::Handle->push_shutdown method. |
301 |
- add an additional error message parameter to AnyEvent::Handle's |
302 |
on_error callback (for TLS, $! is still available). |
303 |
- add AnyEvent::Handle on_starttls/on_stoptls callbacks. |
304 |
- make AnyEvent::Handle more robust against early conenction |
305 |
failures (during new), and return C<undef> in that case |
306 |
from the constructor. |
307 |
- AnyEvent::Handle will now try to load only JSON::XS first, |
308 |
then fall back to JSON. |
309 |
- format_ipv4/format_ipv6 are now exported by default, for symmetry, |
310 |
and because it was documented that way. |
311 |
|
312 |
4.451 Fri Jul 3 00:28:58 CEST 2009 |
313 |
- do not clear rbuf when shutting down an AnyEvent::Handle |
314 |
object - doing so breaks AnyEvent::HTTP. |
315 |
|
316 |
4.45 Mon Jun 29 22:59:26 CEST 2009 |
317 |
- a write error could cause AnyEvent::Handle to create |
318 |
an I/O watcher with an undefined $fh. |
319 |
- special-case mapped ipv4 addresses in both |
320 |
AnyEvent::Socket::format_address and parse_address, |
321 |
to treat them just like ipv4 addresses. |
322 |
- updated and overhauled the AnyEvent::Intro doc. |
323 |
- implement AnyEvent::Socket::format_ipv4/ipv6. |
324 |
- slightly speed up portable_pipe/socketpair. |
325 |
- expand condvar begin/end documentation. |
326 |
|
327 |
4.42 Fri Jun 26 08:32:18 CEST 2009 |
328 |
- preliminary and neccesarily incomplete support for IO::Async. |
329 |
- reset SIGCHLD to DEFAULT when AnyEvent is loaded, in |
330 |
case it was set to IGNORE, to ensure we can catch child statuses |
331 |
even when the calling env acts stupidly. |
332 |
- updated benchmarks with IO::Async, which performs very well. |
333 |
|
334 |
4.412 Wed Jun 24 01:35:57 CEST 2009 |
335 |
- support an "untaint" attribute for AnyEvent::DNS |
336 |
and set it on the default resolver. |
337 |
- implement and document AnyEvent::Impl::Perl::loop. |
338 |
- remove all anyevent-env variables from %ENV |
339 |
when running in tainted mode. |
340 |
- mention and extend the IO::Lambda benchmark. |
341 |
|
342 |
4.411 Sun Jun 7 18:48:02 CEST 2009 |
343 |
- do not try to use F_SETFD on windows. |
344 |
|
345 |
4.41 Thu May 14 06:40:11 CEST 2009 |
346 |
- work around issues in older perls (5.8.0?) when |
347 |
a signal handler is deleted from the %SIG hash. |
348 |
- use POSIX::_exit in child test, to avoid running |
349 |
destructors. |
350 |
- speed up CHLD handling by relying on SIGCHLD being |
351 |
synchronously delivered, even when we roll our own |
352 |
implementation. |
353 |
- AnyEvent::DNS: add the "dname" resource record name |
354 |
for cosmetic reasons. |
355 |
|
356 |
4.4 Sun Apr 26 20:12:33 CEST 2009 |
357 |
- implemented idle watchers, where applicable. |
358 |
- AnyEvent->time died when Event backend was in use. |
359 |
- fix a memleak in the Tk backend. |
360 |
- sped up Tk timer handling. |
361 |
- clip negative "after" values to 0 in AnyEvent::Impl::Event |
362 |
to avoid spamming out warning messages. |
363 |
- fix Qt timers without interval. |
364 |
- avoid Qt zero-timeout special case, in old libqt's. |
365 |
|
366 |
4.352 Mon Apr 20 16:31:11 CEST 2009 |
367 |
- fix AnyEvent::Strict error messages for child watchers. |
368 |
- fix/update Makefile.PL "recommends" versions. |
369 |
- add AnyEvent->now_update. |
370 |
|
371 |
4.351 Sat Apr 11 07:56:14 CEST 2009 |
372 |
- actually make the signal pipe work under win32. |
373 |
- localise $! in signal handler to avoid changing $!, although |
374 |
perl itself does not seem to save/restore errno either. |
375 |
- set the cloexec flag on the signal pipe (normally set by perl too). |
376 |
|
377 |
4.35 Fri Mar 27 11:48:20 CET 2009 |
378 |
- event models relying on AnyEvent's signal watcher emulation did |
379 |
invoke the callback asynchronously, contradicting documentation |
380 |
and causing signals to get lost (this includes AnyEvent's own |
381 |
event loop). AnyEvent now uses the standard pipe trick to make |
382 |
callback execution synchronous to the event loop. |
383 |
- AnyEvent::Handle didn't free TLS context data on DESTROY |
384 |
(patch by Pavel Shaydo). |
385 |
- work around the ever-incompatibly-changing API of MakeMaker. |
386 |
- document that changing global variables without restoring them |
387 |
is a bad idea in Perl (noted by Adam Rosenstein). |
388 |
- AnyEvent::Strict now barfs if ->io is passed a file. |
389 |
|
390 |
4.34 Thu Feb 12 18:32:45 CET 2009 |
391 |
- separately buffer TLS read data, as otherwise the read queue could |
392 |
deadlock as receiving data is not expected while draining the |
393 |
read queue (which cna only happen with TLS). |
394 |
- raise EBADMSG error on JSON decoding errors. |
395 |
- fix some minor manpage bugs (reported by Maximilian Gaß). |
396 |
- speed up select bitmask parsing quite a bit in the pure perl |
397 |
backend. |
398 |
- use CORE::select instead of just select in AnyEvent::Impl::Perl. |
399 |
|
400 |
4.331 Tue Jan 6 21:07:25 CET 2009 |
401 |
- socketpair fails on many vista machines because vista has |
402 |
completely broken accept/getpeername and getsockname functions, |
403 |
so we provide our own socketpair emulation that kind of works |
404 |
(AnyEvent::Util::portable_pipe). |
405 |
- new function: AnyEvent::Util::portable_socketpair. |
406 |
- take advantage of the Guard module if it exists. |
407 |
|
408 |
4.33 Fri Nov 21 02:35:40 CET 2008 |
409 |
- AnyEvent::Strict did errornously flag a fileno of 0 as |
410 |
illegal. |
411 |
- reduce memory usage and slightly speed up the pure perl backend |
412 |
by only storing the file descriptor, not the file handle. |
413 |
- add missing autocork method to AnyEvent::Handle (reported |
414 |
by Adam Rosenstein). |
415 |
- AnyEvent::DNS->resolve errornously documented an $rcode result |
416 |
argument, but there isn't (spotted by Henrik Krohns). |
417 |
- the naked truth about Tk - it's basically unmaintained. |
418 |
|
419 |
4.32 Mon Nov 3 22:46:32 CET 2008 |
420 |
- fix AnyEvent::Socket::resolve_sockaddr to properly support |
421 |
unix sockets again and choose a proper default for |
422 |
the socket type. |
423 |
- fix call to resolve_sockaddr in tcp_connect. |
424 |
|
425 |
4.31 Thu Oct 30 04:41:48 CET 2008 |
426 |
- implemented AnyEvent::Handle->destroy method. |
427 |
- hint about unexpected effects in TLS mode. |
428 |
- speed up AnyEvent::Impl::Perl by using more arrays |
429 |
instead of hashes without the slightest loss of readability :) |
430 |
- work around a perl argument refcounting bug. |
431 |
- with some perl patches applied, AnyEvent now seems leak-free. |
432 |
|
433 |
4.3 Fri Oct 3 09:18:43 CEST 2008 |
434 |
- AnyEvent will now install a no-op signal handler for SIGPIPE, |
435 |
unless one has been installed already. |
436 |
- warn about not loaidng AnyEvent::Impl::POE early enough |
437 |
(patch by Adam Rosenstein). |
438 |
- fixed a great number of bugs and corner cases in AnyEvent::Handle. |
439 |
- fix a bug in where in SSL connect mode, the |
440 |
client would first wait for some data by the server and |
441 |
otherwise hang. (reported and analysed in an absolutely |
442 |
exemplary manner by Adam Rosenstein). |
443 |
- fix a bug in where SSL EOF would not be treated |
444 |
as stream EOF, putting the connection into a hung state. |
445 |
(reported and analysed in an absolutely exemplary manner by Adam |
446 |
Rosenstein). |
447 |
- fix a potential segfault when the TLS context |
448 |
would go missing in a read callback (e.g. due to stoptls) - |
449 |
Net::SSLeay of course makes no type checking whatsoever. |
450 |
- AnyEvent::Handle will not stop the read watcher in TLS mode. |
451 |
- AnyEvent::Handle->stoptls will send a TLS close notify |
452 |
instead of simply destroying the stream now. |
453 |
- fix a bug where large blocks of data written in TLS mode |
454 |
would not be sent unless triggered by receives. |
455 |
- on_drain will now take the tls write buffer into account. |
456 |
- SSL operations have been streamlined a bit and should be faster |
457 |
now, more to come, though. |
458 |
- added a FAQ section, will document nontrivial issues in there. |
459 |
|
460 |
4.234 Mon Sep 29 04:08:13 CEST 2008 |
461 |
- fix child watcher documentation: only child exits will be reported, |
462 |
no trace events. |
463 |
- mention SIGPIPE in AnyEvent::Handle. |
464 |
- perl backend will now use POSIX::times as monotonic clock |
465 |
when available and the monotonic clock option is not. |
466 |
- run even without Time::HiRes in some cases now, but do not |
467 |
provide sub-second accuracy (all *supported* perl versions have |
468 |
Time::HiRes). |
469 |
|
470 |
4.233 Fri Aug 22 01:48:05 CEST 2008 |
471 |
- fix a bug in the testsuite, causing a hang. |
472 |
- clarified lots of AnyEvent::Handle constructor arguments. |
473 |
|
474 |
4.232 Thu Aug 21 20:44:25 CEST 2008 |
475 |
- fix a bug in AnyEvent::Handle that could cause two fatal error |
476 |
callbacks on EOF, which caused AnyEvent::HTTP to signal success twice |
477 |
(testcase by Leon Brocard). |
478 |
- clarified on_eof/on_error documentation a bit. |
479 |
|
480 |
4.231 Tue Jul 29 13:12:15 CEST 2008 |
481 |
- remove some debugging code left in AnyEvent::Util::fork_call |
482 |
(and no, it's impossible to implement with the broken windows |
483 |
perls without resource leaks or worse). |
484 |
|
485 |
4.23 Tue Jul 29 12:19:59 CEST 2008 |
486 |
- document the first parameter passed to condvar callbacks to be |
487 |
the callback. |
488 |
- add AnyEvent::Socket::{ntoa,aton} aliases. |
489 |
- optimize the AE::Handle->push_read (line) for the default |
490 |
eol marker. |
491 |
- optimize push_read (packstring|storable) for small packets. |
492 |
- invoke on_error callback when no on_eof callback is set. |
493 |
- fix a bug in push_read (storable) of unknown impact. |
494 |
|
495 |
4.22 Sun Jul 20 16:34:13 CEST 2008 |
496 |
- new function AnyEvent::Socket::parse_hostport. |
497 |
- as the bulkheads at microsoft can't even get getprotobyname reliably |
498 |
working on their shitty fucking broken os we need to hardcode |
499 |
some common protocol numbers in AnyEvent::Socket. How can |
500 |
people even bother with such a pile of shit as windows. |
501 |
|
502 |
4.21 Thu Jul 17 14:40:05 CEST 2008 |
503 |
- INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype |
504 |
which is incompatible to previous invocation syntax. |
505 |
- work around more windows perl bugs in fork_call. I cannot imagine |
506 |
why anybody would bother implementing fork in such an obviously |
507 |
unusable way. |
508 |
- avoid calling $do_retry if the request has already finished in |
509 |
AnyEvent::DNS, thus avoiding an exception (reported by Anatoly |
510 |
K. Sharifulin). |
511 |
|
512 |
4.2 Sat Jul 12 22:42:11 CEST 2008 |
513 |
- support an "interval" argument for repeating timers. |
514 |
- fix a bug in the pure perl event loop that caused timers to |
515 |
be delayed under some circumstances. |
516 |
- implement optional argument checking via PERL_ANYEVENT_STRICT. |
517 |
- not importing anything from Event causes it to malfunction |
518 |
and not use Time::HiRes, work around this bug. |
519 |
- more bug workarounds against the endlessly broken Tk module, |
520 |
seems to work now on 32 bit machines, but less so on 64 bit. |
521 |
- minor optimisations applied to most backends. |
522 |
|
523 |
4.161 Fri Jul 4 14:24:48 CEST 2008 |
524 |
- fixed the t/handle/02_write.t, it read after write, with the |
525 |
assumption that all data could be written before read |
526 |
(reported by Jim Razmus). |
527 |
|
528 |
4.160 Thu Jul 3 04:02:21 CEST 2008 |
529 |
- re-registering signal watchers after unregistering one did |
530 |
not work if a backend falls back on AnyEvent's default |
531 |
implementation (also affected child watchers). |
532 |
- new AnyEvent::Handle options: autocork and no_delay. |
533 |
|
534 |
4.152 Sun Jun 22 14:15:44 CEST 2008 |
535 |
- allow for 32-bit perls that implement shifts differently |
536 |
on different architectures in parse_ipv4 (reported and |
537 |
analysed by Keiichi DAIBA). |
538 |
|
539 |
4.151 Fri Jun 6 17:34:24 CEST 2008 |
540 |
- make sure specifying _only_ on_read and never pushing reads |
541 |
works. |
542 |
|
543 |
4.15 Fri Jun 6 13:00:46 CEST 2008 |
544 |
- the pure perl backend would keep some watchers alive when more than |
545 |
one watcher was registered for the same fd. |
546 |
- new "packstring" and "storable" read and write types |
547 |
for AnyEvent::Handle. |
548 |
- allow on_eof handler to be called after on_error with EPIPE returns. |
549 |
- do not immediately call on_read callback in handle constructor. |
550 |
|
551 |
4.14 Thu Jun 5 20:29:31 CEST 2008 |
552 |
- Fixed a bug in DNS SRV priority sorting. |
553 |
- AnyEvent::Util::guard now reports runtime errors while |
554 |
executing the guard block as warnings. |
555 |
- handle 0-byte-reads just before EOF correctly in AnyEvent::Handle. |
556 |
|
557 |
4.13 Thu Jun 5 00:47:59 CEST 2008 |
558 |
- AnyEvent::DNS only followed cname chains with length 2, |
559 |
contrary to documentation. bumped it up to 10, thanks to |
560 |
microsoft, the current broken-dns-config-king. |
561 |
- AnyEvent::DNS didn't check for socket return status |
562 |
but instead relied on perl not creating filehandles in that |
563 |
case - too bad it gives you a bogus file handle (reported |
564 |
and analysed by Vladimir Timofeev). |
565 |
- fix queue management logic in AnyEvent::Handle: |
566 |
when on_read was registered, the queue was empty |
567 |
and no progress could be made AnyEvent::Handle would |
568 |
enter an endless loop. |
569 |
- correctly start reading again when the handle became |
570 |
busy again after idling. |
571 |
- correctly treat tls shut-downs as EOF condition |
572 |
(for the time being, would be nice to have a callback for that). |
573 |
- correctly call eof callback under all conditions |
574 |
(hopefully). |
575 |
- the timeout callback did not expect that $self can go |
576 |
away any time. |
577 |
|
578 |
4.12 Tue Jun 3 10:58:04 CEST 2008 |
579 |
- include AnyEvent::Intro, a tutorial for anyevent, |
580 |
anyevent::socket and anyevent::handle. |
581 |
- allow more options in on_error. |
582 |
|
583 |
4.11 Fri May 30 23:42:25 CEST 2008 |
584 |
- INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide |
585 |
reverse_lookup and reverse_verify to replace it, support |
586 |
v4mapped and v4compat addresses. |
587 |
- provide more documentation for the resolver class. |
588 |
- really replace longest run of :0: by :: in format_address, |
589 |
also properly convert :: and ::1 again. |
590 |
- support NAPTR record name and decode it. |
591 |
- implement random weight sampling for SRV records, as per |
592 |
rfc 2782. |
593 |
- correctly abort on srv-record targets of ".". |
594 |
- added AnyEvent::DNS::wait_for_slot. |
595 |
- in the unlikely event of a virtual circuit connection |
596 |
being invalidated by a delayed udp reply, AnyEvent::DNS |
597 |
could die. this has been fixed. |
598 |
- plug a probable memleak in the DNS vc code. |
599 |
- use configured timeout also for VC request phase. |
600 |
- implement timeout and max_outstanding methods |
601 |
for resolver class. |
602 |
- update version numbers in all modules. |
603 |
|
604 |
4.1 Thu May 29 05:45:40 CEST 2008 |
605 |
- INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address |
606 |
which mirrors their purpose better (old names still available |
607 |
for a while). Also moved AnyEvent::DNS::addr to |
608 |
AnyEvent::Socket::resolve_sockaddr. |
609 |
- implement AnyEvent->time and AnyEvent->now. |
610 |
- fix IPv6 support in pack_sockaddr. |
611 |
- officially un-experimentalise all newly introduced code. |
612 |
- support unix domain sockets everywhere by specifying |
613 |
a host of "unix/" and the pathname as service. |
614 |
- implement an activity timeout in AnyEvent::Handle. |
615 |
- added a regex read type to AnyEvent::Handle. |
616 |
- added a json read/write type to AnyEvent::Handle. |
617 |
- always croak in on_error in AnyEvent::Handle. |
618 |
- document how to subclass AnyEvent::Handle. |
619 |
- implement AnyEvent::Util::fork_call. |
620 |
- add support for IPv6 nameservers and nameserver statements. |
621 |
- work around _yet_ _another_ windows perl bug in where empty |
622 |
select masks cause errors in the pure perl backend (this |
623 |
workaround cannot be implemented for the other event loops, |
624 |
but EV also works around this). |
625 |
- supply AnyEvent::Socket::address_type and make good use of it. |
626 |
- clarify and fix the Handle documentation w.r.t $self vs. $handle |
627 |
vs. callback arguments. |
628 |
- add some recommends to the META.yml. |
629 |
|
630 |
4.05 Mon May 26 19:44:06 CEST 2008 |
631 |
- some platforms ignore the rfc and prepend an extra sa_len member to |
632 |
the sockaddr structure, cater for those. also use sockaddr_family, |
633 |
didn't know it was there, but comes in handy for the workaround. |
634 |
- undo the import WIN32 hack. |
635 |
|
636 |
4.04 Mon May 26 08:03:31 CEST 2008 |
637 |
- try to work around yet another windows bug: failed connects are reported |
638 |
as if it were out-of-band data. windows users: you suck. Only |
639 |
EV, Glib, Event and the pure perl backend can handle this condition. |
640 |
- optimize write algorithm in AnyEvent::Handle. |
641 |
- properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly. |
642 |
- tcp_connect now properly iterates through all targets. |
643 |
- check wether IPv6 sockets can actually be created, otherwise, disable ipv6 |
644 |
support entirely. also hardcode some AF_INET6 constants for the |
645 |
sake of perl 5.8. |
646 |
- vastly improved nameserver/suffix detection algorithm for windows |
647 |
(still a hack...) |
648 |
- try to use a fallback dns server if no dns servers could be found. |
649 |
- splatter around some local $SIG{__DIE__}'s to work around bad |
650 |
die handlers. |
651 |
|
652 |
4.03 Sun May 25 05:05:57 CEST 2008 |
653 |
- work around perl bugs on windows, where perl returns undocumented |
654 |
error codes for sysread, syswrite etc. |
655 |
- fix AnyEvent::Handle::unshift_read to really unshift and not push. |
656 |
- AnyEvent::Handle could sometimes stumble over 0-byte writes. |
657 |
- fix unpack_sockaddr, now works for AF_INET6 sockaddr's. |
658 |
- updated documentation. |
659 |
- lots of workarounds for perl 5.8 quirks. |
660 |
- implement a more flexible type system for both reads and writes |
661 |
in AnyEvent::Handle. |
662 |
- support IPv6 in tcp_server. |
663 |
- pass local hostname and port to the prepare callback in tcp_server. |
664 |
- make types extendable in AnyEvent::Handle. |
665 |
- croak, not die, in AnyEvent::Handle; |
666 |
|
667 |
4.0 Sat May 24 19:58:08 CEST 2008 |
668 |
- added (experimental) AnyEvent::DNS module. |
669 |
- added (experimental) AnyEvent::Socket module. |
670 |
- added (experiemntal) support for TLS/SSL in AnyEvent::Handle. |
671 |
- make condvar->end call ->send as documented. |
672 |
- make condvar constructor support the documented cb => argument. |
673 |
- overload condvars so you can use them instead of code references. |
674 |
- support more of the POSIX forms instead of a pure dotted quad for |
675 |
verbatim ip addresses. |
676 |
- AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. |
677 |
- try very hard not to freeze on broken windows perls. |
678 |
- make on_eof optional in AnyEvent::Handle. |
679 |
- support service names in tcp_connect and tcp_server. |
680 |
- work around netbsd/openbsd bug causing unix sockets not to |
681 |
become writable until empty. |
682 |
- only make a dummy read for error detection when getpeername |
683 |
returns ENOTCONN. |
684 |
- defend AnyEvent::Util::inet_aton against Coro::LWP's brutal |
685 |
override. |
686 |
- new AnyEvent::Guard::cancel. |
687 |
|
688 |
3.5 Sat May 17 23:17:14 CEST 2008 |
689 |
- gracefully deal with systems not defining CLOCK_MONOTONIC |
690 |
(as opposed to not supporting the clock itself, which worked in 3.41). |
691 |
- added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server |
692 |
helpers, to replace the ill-designed AnyEvent::Socket class. |
693 |
- added some examples (eg/connect, eg/handle). |
694 |
- added AnyEvent::Util::guard. |
695 |
- AnyEvent::Util now exports most of its functions. |
696 |
- fixed a bug in *_read_chunk, where the callback was called with $self as |
697 |
first and second argument. |
698 |
- simplified and fixed regex handling in AnyEvent::Handle::*_read_line. |
699 |
- fix manpage condvars examples to use new syntax. |
700 |
- considerably improve inet_aton with EV::ADNS backend to support |
701 |
cname chains (common dns configuration bug, e.g. with www.google.de) |
702 |
and aaaa records if no a record could be found. |
703 |
|
704 |
3.41 Sun May 11 19:53:13 CEST 2008 |
705 |
- work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. |
706 |
|
707 |
3.4 Sun May 11 00:29:25 CEST 2008 |
708 |
- complete re-implementation of condvars with a hopefully much more useful |
709 |
API (backward compatible functions exist but are undocumented). |
710 |
- AnyEvent::Handle fully rewritten with a hopefully more powerful API |
711 |
(still subject to change). |
712 |
- take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
713 |
- provide AnyEvent::post_detect and @AnyEvent::post_detect, which |
714 |
allows module authors to avoid forcing event loop detection. |
715 |
(used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example). |
716 |
- remove coro backends: Coro now provides generic support for AnyEvent |
717 |
via Coro::AnyEvent. |
718 |
|
719 |
3.3 Mon Apr 28 09:51:06 CEST 2008 |
720 |
- added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. |
721 |
- fix a bug in the pure perl backend that kept watchers alive |
722 |
when multiple watchers were registered for the same fh/poll combo. |
723 |
- add a benchmark section showing AnyEvent overhead and comparing |
724 |
the different event loops with each other. |
725 |
- prefer pure perl over tk when autoprobing, as it's about as fast, |
726 |
but doesn't crash with many watchers. |
727 |
- declare Qt support non-experimental. |
728 |
- clarify the confusing section about the file descriptor being |
729 |
kept alive. |
730 |
- document the race between loading of an event module and |
731 |
child processes exiting. |
732 |
- support POE as "backend" (with some caveats, POE is not generic |
733 |
enough, and darn slow). |
734 |
- support Wx and Prima through POE. |
735 |
- optimise perl backend to use 20% less memory and take advantage |
736 |
of typical timeout behaviour. It can now compete with |
737 |
select/poll-based C event loops in most cases (it is usually |
738 |
faster than Event and Glib :). |
739 |
- roughly cut EV memory use in half and increase its speed by 30%, |
740 |
by removing undocumented functionality. Did something similar |
741 |
to other event loops where possible. |
742 |
|
743 |
3.2 Thu Apr 24 10:10:40 CEST 2008 |
744 |
- do not die when anyevent watchers are destroyed while running callbacks |
745 |
in the pureperl backend (could only happen when two watchers are registered |
746 |
for the same fh/poll combo). |
747 |
- support autoloading for child watchers, was broken in all versions. |
748 |
- implement PERL_ANYEVENT_MODEL env variable. |
749 |
- (experimental) implement interface to the (very crashy/buggy) Event::Lib module. |
750 |
- (experimental) implement interface to the Qt module (cannot be autoprobed). |
751 |
- this release is IO:AnyEvent-proof. |
752 |
|
753 |
3.12 Tue Apr 22 07:11:46 CEST 2008 |
754 |
- reinstate AUTHOR section that got lost somehow. |
755 |
- do not hang in the testsuite with badly broken perls (activestate, |
756 |
strawberry...), but instead diagnose the problem and continue. |
757 |
- use INT instead of CHLD in an attempt to support broken windows perls |
758 |
better (this decreases the test quality, unfortunately). |
759 |
- do not send the signal to the process group |
760 |
(no problem for CHLD, bad for INT :). |
761 |
|
762 |
3.11 Sat Apr 19 06:57:31 CEST 2008 |
763 |
- major documentation rework. |
764 |
- document the fact that child watchers only watch for zombies. |
765 |
- fix the child watcher example. |
766 |
|
767 |
3.1 Wed Apr 16 17:09:01 CEST 2008 |
768 |
- work around recurring bugs in Tk by dup'ing filehandles, the |
769 |
only method with good success chances on Tk (the bugs apparently |
770 |
don't get fixed anytime soon). |
771 |
- lift the restriction of only one watcher per fh direction |
772 |
(as the Tk bug workaround also lifts it and only Tk imposed |
773 |
such strong limits). |
774 |
- changed probe order to prefer coro adaptors. |
775 |
- explain why recursion into the event loop is not supported |
776 |
unless the backend supports it (only Coro::EV does without |
777 |
any restrictions...). |
778 |
- add simple manpages for all backend modules. |
779 |
|
780 |
3.0 Mon Apr 7 21:30:23 CEST 2008 |
781 |
- Coro::Signal changed semantics, roll our own, also cleaning |
782 |
up the Coro implementation in general. |
783 |
- rename Coro backend to CoroEvent. |
784 |
- add some decision helping paragraph to the manpage that should |
785 |
help people to decide whether AnyEvent is the right thing for them. |
786 |
|
787 |
2.9 Mon Jan 28 13:31:54 CET 2008 |
788 |
- update for EV 3.0 API changes. |
789 |
|
790 |
2.8 Sun Nov 25 15:06:03 CET 2007 |
791 |
- waitpid can validly return 0. accept this fact of life |
792 |
instead of reporting it to any watchers. |
793 |
|
794 |
2.7 Fri Nov 23 11:41:14 CET 2007 |
795 |
- force use of AnyEvent::Impl::Perl in testsuite, there is |
796 |
too much breakage outside AnyEvent. |
797 |
- deliver signals synchronously in AnyEvent::Impl::Perl |
798 |
or any other event loop that relies on AnyEvents child |
799 |
watcher emulation. *Could* help with hanging testsuite |
800 |
(except when Event or EV are installed). |
801 |
|
802 |
2.6 Fri Nov 9 20:36:35 CET 2007 |
803 |
- fix bug in testsuite. |
804 |
- move EV adaptor modules to AnyEvent. |
805 |
- add Coro+EV adaptor module. |
806 |
|
807 |
2.55 Tue Nov 6 17:41:32 CET 2007 |
808 |
- add EV to the list of supported event models. |
809 |
- do not auto-reset pid watchers, pass pid and status to them. |
810 |
- allow a pid of zero to watch for all children in child watchers. |
811 |
|
812 |
2.54 Wed Jul 18 17:36:23 CEST 2007 |
813 |
- work around a perl bug that results in |
814 |
BEGIN not safe after errors--compilation aborted without |
815 |
any discernible reason or error message by once |
816 |
again not use'ing strict. |
817 |
|
818 |
2.53 Sun Jul 8 10:51:53 CEST 2007 |
819 |
- make Glib interface work again (spotted by elmex). |
820 |
|
821 |
2.52 Wed Mar 7 18:36:16 CET 2007 |
822 |
- child watchers stopped working permanently when |
823 |
all current child watchers were destroyed. |
824 |
|
825 |
2.51 Mon Dec 11 21:33:24 CET 2006 |
826 |
- work around bugs in perl where eval "require Module" returns |
827 |
true even if the module couldn't be loaded :/. |
828 |
|
829 |
2.5 Mon Dec 11 02:15:28 CET 2006 |
830 |
- avoid $AUTOLOAD because many perls corrupt it. |
831 |
- AnyEvent::detect forces autodetection. |
832 |
- implement signal watchers (experimental) |
833 |
(Tk does not support async signals (<= 804.027 at least), |
834 |
so they might get delayed indefinitely on Tk). |
835 |
- implement child watchers (experimental). |
836 |
- moved default condvar implementation into base module, simplifying |
837 |
most implementation modules. |
838 |
|
839 |
2.1 Fri Nov 24 15:50:48 CET 2006 |
840 |
- better docs. |
841 |
- simple testsuite. |
842 |
- added AnyEvent->one_event method for special purposes. |
843 |
|
844 |
2.0 Wed Nov 1 02:21:30 CET 2006 |
845 |
- INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both. |
846 |
- INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore. |
847 |
- EXPERIMENTAL: |
848 |
new pure-perl model added, AnyEvent now always finds a suitable |
849 |
event model. |
850 |
- improved documentation. |
851 |
- reduced cpu overhead. |
852 |
|
853 |
1.02 Fri Jan 13 14:15:40 CET 2006 |
854 |
- add COPYING to clarify license. |
855 |
|
856 |
1.01 Sun Jan 8 05:48:09 CET 2006 |
857 |
- correctly cancel Event timers, avoid memory leaks. |
858 |
|
859 |
1.0 Sun Jan 8 05:40:58 CET 2006 |
860 |
- allow third-party interfaces to be added at runtime. |
861 |
- fix version number. |
862 |
|
863 |
0.9 Fri Dec 30 02:25:41 CET 2005 |
864 |
- fixed errors in the documentation. |
865 |
- going back to using $VERSION as "load check", |
866 |
as suggested by Jörn Reder. Let's see why I stopped using |
867 |
this method. |
868 |
- add 'hup' to the list of events to receive for glib |
869 |
and generate 'rw' events for it, as some versions of |
870 |
glib seem to require this, while this breaks other |
871 |
(older) versions of glib. |
872 |
- implemented PERL_ANYEVENT_VERBOSE env variable. |
873 |
|
874 |
0.3 Sun Dec 4 10:43:33 CET 2005 |
875 |
- no changes to the code. |
876 |
- fixed documentation. |
877 |
- clarified some corner cases. |
878 |
- vastly improved documentation with a long example. |
879 |
|
880 |
0.2 Thu Dec 1 22:18:49 CET 2005 |
881 |
- check whether a valid method was called, to avoid endless loops. |
882 |
- move condvars into their own class (cleaner, safer). |
883 |
- don't die when some modules are not available. |
884 |
- Tk is too broken to allow for words. Will work once Tk works. |
885 |
|
886 |
0.1 Thu Dec 1 19:44:57 CET 2005 |
887 |
- spiced up and cleaned up for release. |
888 |
|
889 |
0.01 Tue Sep 21 18:12:09 CEST 2004 |
890 |
- original version; copied from Convert::Scalar. |
891 |
|