1 |
Revision history for Perl extension AnyEvent. |
2 |
|
3 |
TODO: document TCP_*** constants |
4 |
TODO: add debug/trace logging to some modules |
5 |
TODO: use the eval-with-cb-call trick also to speed up JSON? |
6 |
TODO: FAQ: common pitfalls? invoke-before-return |
7 |
TODO: FAQ: scope-ids? hosts file order? |
8 |
TODO: AnyEvent::Handle with AnyEvent::IO? |
9 |
TODO: invalid. and localhost. specialcasing inside AEDNS and not AESocket (rfc6761) |
10 |
TODO: maybe implement env variable to give hosts precedence |
11 |
TODO: hosts always read? that's not expected |
12 |
TODO: more formats for parse_ipv6 (single ipv6 address without port, ...p80, ...#80) |
13 |
|
14 |
- add a more explicit warning to AnyEvent::Handle that it doesn't |
15 |
work on files, people keep getting confused. |
16 |
- new function AnyEvent::Socket::tcp_bind. |
17 |
- new functions AnyEvent::fh_block and AnyEvent::fh_unblock. |
18 |
- aligned ipv6 address formatting with RFC 5952 (by not shortening |
19 |
a single :0: to ::). |
20 |
- added stability canary support. |
21 |
|
22 |
7.09 Sat May 2 16:38:53 CEST 2015 |
23 |
- AnyEvent::Debug called an internal function (AnyEvent::Log::ft) |
24 |
that was renamed to AnyEvent::Log:format_time. uinder its old name |
25 |
(analyzed by Michael Stovenour). |
26 |
- update AnyEvent::DNS fallback resolver addresses: |
27 |
seems google effectively killed most other free dns resolvers, |
28 |
so remove them, but add cable and wireless (ecrc) since it was |
29 |
stable for 20 years or so, official or not, and there should be |
30 |
an alternative to google. |
31 |
- perl5porters broke windows error codes in 5.20, and mapped |
32 |
WSAEWOULDBLOCK on the (different) EWOULDBLOCK error code, and |
33 |
WSAEINPROGRESS into the incompatible ERINPROGRESS code, probably |
34 |
because they were so cool. They probably broke other error codes |
35 |
for no reason, too, but I didn't care for checking, it's too |
36 |
depressing. This version only works around the WSAEWOULDBLOCK |
37 |
issue, because I don't have a nice way to work around the |
38 |
WSAEINPROGRESS bug. |
39 |
|
40 |
7.08 Wed Dec 10 05:27:17 CET 2014 |
41 |
- work around a newly introduced bug in Socket 2.011 (an |
42 |
errornous sun_length check) (analyzed by Maxime Soulé). |
43 |
- AnyEvent::TLS didn't load (but refer to) AnyEvent::Socket |
44 |
(analyzed by Ben Magistro). |
45 |
- AnyEvent::Strict will now confess, not croak. This is in line with |
46 |
it being a development/debugging tool. |
47 |
- work around a number of libglib bugs (debug builds of libglib enforce |
48 |
certain undocumented behaviour patterns such as not being able to |
49 |
remove a child watch source after it has fired, which we will try |
50 |
to emulate to avoid "criticals". what where they thinking...). |
51 |
- mention json security issues in AnyEvent::Handle, now that Douglas |
52 |
Crockford has foolishly and incompatibly changed JSON. |
53 |
- changed default dns resolver "max_outstanding" value from 1 to 10, |
54 |
the latter beinfg the intended value all along |
55 |
(reported by Ilya Chesnokov). |
56 |
- added new "AnyEvent::Impl::UV" interface module to the UV event lib |
57 |
(written by Mike Lowell). |
58 |
|
59 |
7.07 Tue Dec 17 17:45:02 CET 2013 |
60 |
- the documentation for custom tls verify schemes was wrong. make it agree |
61 |
with the code (reported by Maxime Soulé). |
62 |
- added cbor read and write types to AnyEvent::Handle (using CBOR::XS). |
63 |
- work around an API change in openssl that could cause wrong tls connection |
64 |
aborts, likely on windows only (analyzed by sten). |
65 |
- calling AnyEvent->now_update with AnyEvent::Impl::Perl caused an |
66 |
endless loop (reported by Dietrich Rebmann). |
67 |
- add tlsv1_1 and tlsv1_2 protocols to AnyEvent::TLS |
68 |
(patch by Maxime Soulé). |
69 |
- document AnyEvent::Impl::IOAsync::set_loop and |
70 |
$AnyEvent::Impl::IOAsync::LOOP. Though only documented now, this |
71 |
functionality has _always_ been available. |
72 |
- force a toplevel domain name in t/81_hosts.t (analyzed by |
73 |
David Jack Wange Olrik). |
74 |
- document that AnyEvent::Log uses AnyEvent::IO. |
75 |
- warn about AnyEvent::Filesys::Notify performance. |
76 |
- praise the joys of AnyEvent::Fork::*. |
77 |
- time for an =encoding directive. |
78 |
- do no longer use JSON to create a default json coder, use |
79 |
JSON::XS or JSON::PP directly. |
80 |
|
81 |
7.05 Wed Aug 21 10:38:08 CEST 2013 |
82 |
- uts46data.pl couldn't be found due to wrong naming of the file |
83 |
(reported by Fulko Hew). |
84 |
- handle lone \015's properly in AE::Handle's default line read |
85 |
(reported by various people). |
86 |
- untaint ip addresses found in /etc/hosts (patch by José Micó). |
87 |
- the memleak fix in 7.03 caused resolving via /etc/hosts to always |
88 |
fail on first use (reported and testcase by Andrew Whatson). |
89 |
- expose AnyEvent::Log::format_time, and allow users to redefine it. |
90 |
- expose AnyEvent::Log::default_format, and allow redefinition. |
91 |
- expose AnyEvent::Log::fatal_exit, to allow redefinition. |
92 |
- AnyEvent::Debug shell can now run coro shell commands, if available. |
93 |
- t/63* tests were wrongly in MANIFEST. |
94 |
- kernel.org's finger server went MIA, switch to freebsd.org and |
95 |
icculus.org. |
96 |
- clarify that IO::AIO and AnyEvent::AIO are needed for AnyEvent::IO |
97 |
to function asynchronously (reported by Daniel Carosone). |
98 |
- hard-disable $^W in most tests, it generates too much garbage output. |
99 |
- use a (hopefully) more future-proof method to emulate common::sense. |
100 |
- upgrade to UTS-46:6.2.0. |
101 |
- switch to INSTLIB from INSTLIBDIR, as INSTLIBDIR was wrongly |
102 |
documented. should not affect anything. |
103 |
|
104 |
7.04 Thu Nov 15 02:23:58 CET 2012 |
105 |
- debugging code left in. |
106 |
|
107 |
7.03 Thu Nov 15 02:16:11 CET 2012 |
108 |
- AnyEvent::Socket::inet_aton did not work when dns resolution was |
109 |
used to find the addresses (analyzed and patch by Maxime Soulé). |
110 |
- fix a memory leak in the /etc/hosts lookup code when hosts don't |
111 |
resolve and are not in hosts (analyzed by sten). |
112 |
|
113 |
7.02 Tue Aug 14 04:27:58 CEST 2012 |
114 |
- AnyEvent::Util::run_cmd could block indefinitely (analyzed and test |
115 |
program by Yorhel). |
116 |
- verified that AnyEvent::Socket follows RFC5952. |
117 |
- try to parse "ADDR#PORT" in addition to "ADDR PORT". |
118 |
|
119 |
7.01 Sun May 13 01:03:17 CEST 2012 |
120 |
- fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but |
121 |
not available, instead of throwing an exception. |
122 |
- use File::Spec to get the tmpdir in t/*, to avoid needless |
123 |
failures on (most, not mine :) windows boxes. |
124 |
- new handle read types: tls_detect and tls_autostart. |
125 |
|
126 |
7.0 Fri Apr 13 06:33:30 CEST 2012 |
127 |
- child watchers are broken in POE 1.352 (also many earlier |
128 |
versions) and there seems to be no way to work around it, as POE |
129 |
itself is inherently racy. Document this regression and add a delay |
130 |
in t/68_poe_03_child.t for the time being. |
131 |
- new module AnyEvent::IO, that is a frontend to either a pure-perl |
132 |
synchronous I/O implementation (AnyEvent::IO::Perl), or to the |
133 |
asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO). |
134 |
- load /etc/hosts only when DNS has no answer. |
135 |
- stat /etc/hosts on every access and reload it if it changed. |
136 |
- load /etc/hosts via AnyEvent::IO - potentially asynchronous. |
137 |
- fix a buggy croak in the dh parameter reading in AnyEvent::TLS. |
138 |
- AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As |
139 |
a side effect, they now use true appending as opposed to libc |
140 |
appending, and the file might not have been opened when the function |
141 |
returns. |
142 |
- the default logging level was not properly documented in a variety of |
143 |
places, this has hopefully been rectified. |
144 |
- updated uts46data.pl for unicode 6.1.0. |
145 |
- made log messages generated by AnyEvent submodules not |
146 |
include the package name anymore, as it will be logged |
147 |
by default already. |
148 |
- upgrade to the trick used by common-sense 3.5 to work |
149 |
around extra warning torture/breakage under perl 5.15.x. |
150 |
- log messages by anyevent are now ucfirst, are usually full |
151 |
sentences and do no longer include the package name. |
152 |
- the storable read type would sometimes throw an exception instead |
153 |
of causing EBADMSG (analyzed by Maxime Soulé). |
154 |
- work around a bug in openssl 1.0.1 which enforces a minimum rsa |
155 |
keysize (reported by Paul Howarth). |
156 |
- documented Rocco fabricating statements about AnyEvent and me. |
157 |
|
158 |
6.14 Tue Jan 31 20:00:24 CET 2012 |
159 |
- AnyEvent::Impl::Tk was broken due to a mysterious "die" inside, |
160 |
probably an editing mistake (reported by Darin McBride). |
161 |
|
162 |
6.13 Thu Jan 12 07:27:01 CET 2012 |
163 |
- AnyEvent::Util::fork_call checked for POSIX availability in the wrong way, |
164 |
causing it to fail if POSIX wasn't loaded already (analysed by Rock Power). |
165 |
- AnyEvent::Handle::push_read (line => did pass $1 directly, so |
166 |
regex-matching inside the callback would change the parameter, despite |
167 |
$1 being dynamically scoped per-block. this perl bug is now being |
168 |
worked around (testcase by Cindy Wang). |
169 |
|
170 |
6.12 Mon Dec 12 13:21:10 CET 2011 |
171 |
- $! was clobbered by subsequent calls in tcp_connect, due to |
172 |
the postpone not saving/restoring it (analyzed by Richard Garnish). |
173 |
|
174 |
6.11 Tue Nov 22 10:36:05 CET 2011 |
175 |
- Tk cannot create windows when tainted unless you set a title, so |
176 |
set a dummy title for AnyEvent's dummy mainwindow. How dummy. |
177 |
- escape any nonprintable/nonascii characters when stringifying |
178 |
backtraces. |
179 |
- log the reason loading an interface module fails to load at level debug. |
180 |
|
181 |
6.1 Tue Oct 4 19:44:30 CEST 2011 |
182 |
- INCOMPATIBLE CHANGE: the default log level is now "4" (error |
183 |
and above), and some messages inside AnyEvent have been elevated |
184 |
to higher log levels to print by default. |
185 |
- AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal. |
186 |
- work around yet another signal race bug, newly introduced by POE. My |
187 |
how that module sucks. |
188 |
- add google public ipv6 dns as fallback servers, so AE::DNS can finally |
189 |
support ipv6-only operation even with defaults (also use ::1 instead |
190 |
of 127.0.0.1). |
191 |
- give better guidelines on how to chose a logging level. |
192 |
- AnyEvent::Log can now cap (limit) log message priority by context. |
193 |
- AE::log will now save and restore $! and $@, so callers don't have to. |
194 |
- AE::Strict now checks that registered fds are still valid at regular |
195 |
intervals. |
196 |
- skip frequently called 'require' calls if the module appears loaded. |
197 |
- increase signal latency test timer considerably for poor little osx, |
198 |
which can't be bothered with basic posix functionality such as signal |
199 |
delivery. |
200 |
- mention some debugging aids in the tutorial. |
201 |
- new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY. |
202 |
- private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name |
203 |
of the context as title. |
204 |
- AnyEvent::Strict would not restore read-write access to $_ when a callback |
205 |
died - this is a user bug, but since AE::Strict exists to help dbeugging... |
206 |
- detect the extremely evil IO::Async::Loop::AnyEvent module and refuse |
207 |
to work with it - the author is well aware of the issues but refuses |
208 |
to fix them. This extreme step has been taken because the module actively |
209 |
attacks the goals of AnyEvent and makes it harder for module authors to use |
210 |
AnyEvent. |
211 |
|
212 |
6.02 Fri Aug 26 20:08:31 CEST 2011 |
213 |
- log_to_syslog now accepts facility strings. |
214 |
- log_to_syslog would not allow facilities - "$facility|$level" |
215 |
*literally* means that string. sheesh. |
216 |
- the CPAN parser got confused and thought we provide Sys::Syslog, try |
217 |
to work around. |
218 |
- renamed AnyEvent::Impl::FLTK2 to FLTK. |
219 |
|
220 |
6.01 Fri Aug 26 07:04:11 CEST 2011 |
221 |
- INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include |
222 |
the TTL as fourth element - this affects ->resolve, ->request |
223 |
and ->dns_unpack indirectly. Convenience functions such as |
224 |
AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is |
225 |
AnyEvent::Socket. |
226 |
- INCOMPATIBLE CHANGE: backend authors now should not implement |
227 |
one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll |
228 |
methods. |
229 |
- INCOMPATIBLE CHANGE: verbosity levels have been redefined, see |
230 |
AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details. |
231 |
- DEPRECIATION: please change |
232 |
"use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and |
233 |
"AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run" |
234 |
in your programs, for future compatibility. |
235 |
- *blush* idle watcher emulation was borked. |
236 |
- AnyEvent::Strict now checks the AE::xxx functions as well. |
237 |
- querying too long domain names would barf deep inside AnyEvent::DNS - since |
238 |
this is hard to avoid for applications, AnyEvent:.DNS::request now checks for |
239 |
for too long domain names explicitly and passes an undef to the callback |
240 |
instead (reported by James Bromberger). |
241 |
- make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield). |
242 |
- new functions: AnyEvent::postpone, AnyEvent::log. |
243 |
- new module: AnyEvent::Log, for simple logging needs. |
244 |
- new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config. |
245 |
- new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start |
246 |
a debug shell. |
247 |
- new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable |
248 |
watcher instrumentation/wrapping (see AnyEvent::Debug::wrap). |
249 |
- new env variable, PERL_ANYEVENT_HOSTS, to override the place where |
250 |
/etc/hosts can be found. |
251 |
- an empty PERL_ANYEVENT_RESOLV_CONF now selects the default |
252 |
configuration. |
253 |
- env variables can now be specified using AE_xyz instead of |
254 |
PERL_ANYEVENT_xyz, subject to some rules. |
255 |
- lots of new AnyEvent::Debug functionality. |
256 |
- AnyEvent::Strict now checks for modifications of $_ by the callback |
257 |
(by setting it to readonly while executing the callback, if possible). |
258 |
- IO::Async sometimes cancels the wrong timer when you call cancel_timer. |
259 |
Invest extra overhead to work around this peculiar behaviour. |
260 |
- split the pure perl backend into a loop module and an impl |
261 |
module (AnyEvent::Loop and AnyEvent::Impl::Perl). |
262 |
- some read types (regex, netstring, json, storable) could |
263 |
errornously keep an AE::Handle object alive on their own |
264 |
(reported by Mohammad Toossi). |
265 |
- AnyEvent::Socket::resolve_sockaddr and all functions using it now |
266 |
supports /etc/hosts overrides. |
267 |
- add more workarounds around child watcher bugs in IO::Async |
268 |
and (newly introduced) in POE. |
269 |
- use glib child watchers and try to work around its limitations, |
270 |
if possible (based on a patch by Kevin Ryde). |
271 |
- add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2. |
272 |
- many read types in AnyEvent::Handle would malfunction during |
273 |
transitions to TLS because of a "delete $self->{rbuf}". |
274 |
(reported by Gerald Galster). |
275 |
- improve condvar blocking wait performance (probably only noticable |
276 |
with faster backends), also make it easier for event loops that |
277 |
do not support blocking waits. |
278 |
- PERL_ANYEVENT_MODEL now also allows full module names, although |
279 |
the usefulness of this feature is not yet clear. |
280 |
- add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS. |
281 |
- add t/80_ssltext.t. |
282 |
- no longer use AUTOLOAD for the initial loading (for no very good reason). |
283 |
- tcp_server now tries to unlink unix domain sockets when it gets destroyed. |
284 |
- do not run event loop detection multiple times when programs call their cached |
285 |
copies of AnyEvent::detect. |
286 |
- parse_hostport now accepts absolute paths als unix domain sockets. |
287 |
- suppress the idiotic warning inside POE by patching the POE::Kernel |
288 |
run flags instead of asking the user to jump through contortions. |
289 |
- add Task::Weaken as dependency (as recommends only at this time). |
290 |
- replace PF_UNSPEC by 0, for good old minix. |
291 |
|
292 |
6.0 (not officially released, but was on CPAN for about 15 minutes :/) |
293 |
|
294 |
5.34 Thu May 12 09:59:41 CEST 2011 |
295 |
- AE::TLS didn't support both cert_file and key at the same time, |
296 |
due to some copy&paste bug (reported and analyzed by vti). |
297 |
- improve AE::Handle timeout documentation slightly. |
298 |
|
299 |
5.33 Mon Feb 7 21:51:41 CET 2011 |
300 |
- never appeared on CPAN, for unknown reasons. |
301 |
- AnyEvent::Handle on_connect_error - do not try to destroy a handle |
302 |
object when it's already gone (analyzed by Ryan Bullock). |
303 |
- update idna algorithm(s) to UTS#46 version 6.0.0. |
304 |
- preserve trailing dot in idn_to_ascii - testsuite says to preserve, |
305 |
uts46 says it's an error, and testsuite makes more sense. |
306 |
- remove leading dots in idn_to_ascii - testsuite says to remove, |
307 |
uts46 says it's an error, and testsuite makes more sense. |
308 |
- fixed a bug in idn ignored character processing. |
309 |
- passes the uts46 6.0.0 testsuite except for three cases which |
310 |
apparently are bugs in the testsuite itself. |
311 |
- passes the uts46 6.0.1 testsuite except for eleven cases which |
312 |
apparently are bugs in the testsuite itself. |
313 |
|
314 |
5.31 Sun Jan 23 11:49:19 CET 2011 |
315 |
- work around a POE bug where POE would first poll for new events and |
316 |
invoke handlers BEFORE handling the message queue. |
317 |
- fix parsing of "domain" statement in resolv.conf (Johannes Plunien). |
318 |
- stop_read now does just that in TLS mode. see its |
319 |
documentation for caveats. |
320 |
- update to common::sense 3.4 warning set (some warnings in perl impose |
321 |
an arbitrary 32kb limit on regex matches, which affects |
322 |
AnyEvent::Handle). |
323 |
- AE::Handle had a broken rbuf_max method and missing wbuf_max method. |
324 |
- new experimental AE::Handle::resettls method. |
325 |
|
326 |
5.3 Fri Dec 31 05:48:13 CET 2010 |
327 |
- major IOAsync patch by Paul Evans - 0.33 is required, as the major |
328 |
issues with IO::Async have been solved in that version, including |
329 |
the availabilty of a default loop! |
330 |
- implement new wbuf_max parameter for AnyEvent::Handle. |
331 |
- added Cocoa::EventLoop backend by Daisuke Murase. |
332 |
- the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans). |
333 |
- clarify handle on_prepare documentation. |
334 |
- set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd, |
335 |
to avoid leaking them to any child processes. |
336 |
- try to force files into the arch-dependend lib directory, for the |
337 |
half-"turly obsessive compulsive person"(s) that depend on it. |
338 |
|
339 |
5.29 Sun Dec 5 10:49:21 CET 2010 |
340 |
- convert EV backend to EV 4.00 API (so better upgrade EV too). |
341 |
- AE::Handle: implement dynamic read_size adjustment, add |
342 |
max_read_size parameter, reduce default read_size to 2048. |
343 |
- add some FAQ entries for servers and/or on_eof vs. on_error. |
344 |
- work around OS bugs (cygwin again) when lingering in AE::Handle. |
345 |
- reduce memory usage slightly when connect was used in AE::Handle. |
346 |
- wrok around more segfaults in Tk. |
347 |
- document tls_ctx => undef as valid. |
348 |
- detect WNOHANG value at compiletime. |
349 |
- include some probably linux-only support for building AnyEvent |
350 |
as part of the perl core. |
351 |
- improve compatibility of t/08_idna.t to perl 5.8. |
352 |
- make AnyEvent::Strict truly optional (in case somebody builds a |
353 |
minimal perl...). |
354 |
|
355 |
5.28 Wed Oct 13 04:14:23 CEST 2010 |
356 |
- due to a glitch, AnyEvent's internal getprotobyname wasn't |
357 |
actually used in tcp_server. |
358 |
- implement AnyEvent::CondVar->new method, which was documented but |
359 |
missing. |
360 |
- support autoloading of read types in unshift_read. |
361 |
- AnyEvent::Strict: do the test for invalid fh arguments differently, |
362 |
resulting in more informative error messages. |
363 |
- new AnyEvent::FAQ manpage. |
364 |
- clarify that you can actually call push_read etc. in on_read. |
365 |
- put keys and certs in /tmp, not into the current directory, |
366 |
and document this in the SECURITY CONSIDERATIONS sections |
367 |
of AnyEvent::TLS. |
368 |
- support ";" as resolv.conf comment character in AnyEvent::DNS, as |
369 |
requested by Ron Isaacson. |
370 |
- document $AnyEvent::DNS::RESOLVER variable. |
371 |
- incorporated major typo patches by Abhijit Menon-Sen. |
372 |
- AnyEvent::Handle now croaks when negative timeouts are passed. |
373 |
- add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util, |
374 |
undocumented, but who knows when they might come in handy. |
375 |
|
376 |
5.271 Tue Jun 8 12:05:46 CEST 2010 |
377 |
- backport to perl 5.8.x. |
378 |
|
379 |
5.27 Sun Jun 6 12:12:05 CEST 2010 |
380 |
- postpone differently in AnyEvent::Socket now, as |
381 |
when not, canceling the connection attempt might fail |
382 |
(found by Felix Antonius Wilhelm Ostmann). |
383 |
- explicitly check for non-stream sockets in AE::Handle, too many |
384 |
clueless people fell into the trap of this somehow working. |
385 |
- simplified and reworked the "OTHER MODULES" section. |
386 |
- better/more condvar examples. |
387 |
|
388 |
5.261 Wed Apr 28 16:13:36 CEST 2010 |
389 |
- AF_INET6 was not properly used from Socket6 during configuration |
390 |
time (found by Andy Grundman). |
391 |
|
392 |
5.26 Mon Apr 12 04:49:35 CEST 2010 |
393 |
- don't generate spurious readyness notifications when select |
394 |
returns EINTR in the pure perl backend (analysed and testcase |
395 |
by Bas Denissen). |
396 |
- give same examples in SYNOPSIS sections of AnyEvent and AE. |
397 |
- provider faster implementations for AE::time/now/now_update for |
398 |
EV and Perl backends. |
399 |
- provide faster AE::cv and AE::time implementations for all |
400 |
backends. |
401 |
- no longer support lower/mixed-case signal names (this was |
402 |
never documented nor universally supported). |
403 |
- some more memory µ-optimisations, and somewhat less messy |
404 |
code to cope with both AnyEvent and AE APIs. |
405 |
|
406 |
5.251 Sat Mar 13 00:58:21 CET 2010 |
407 |
- make test failed if it was run before make install on perl <= |
408 |
5.8.8, and older versions of perl make our live very difficult, |
409 |
so write constants.pl during Makefile.PL time. |
410 |
|
411 |
5.25 Sat Mar 13 00:23:14 CET 2010 |
412 |
- fix a race condition in AnyEvent::Handle that would cause |
413 |
a "bio_read: unsupported method" error (in _another_ |
414 |
ssl connection) after user code threw an exception. |
415 |
- added AnyEvent::Handle->destroyed method. |
416 |
- speed up AnyEvent::Socket::format_address and ::format_ipv6. |
417 |
- the AnyEvent::Util::fh_nonblocking function worked only |
418 |
by lucky accident on win32. |
419 |
- smaller and faster AnyEvent::Util::fh_nonblocking. |
420 |
- when the (required!) Time::HiRes module is missing, AnyEvent |
421 |
did not fall back to built-in time properly. |
422 |
- do not load Fcntl at runtime, saving memory and loading time. |
423 |
- precompile a number of constants and use them instead of runtime |
424 |
detection and eval. |
425 |
- free detection code after detection and similar memory optimisations. |
426 |
- Perl backend timer interval best effort drift has been improved |
427 |
(same algorithm as EV). |
428 |
- update unicode idna mapping table. |
429 |
|
430 |
5.24 Tue Jan 5 11:39:43 CET 2010 |
431 |
- cygwin never reports errors from failed connects. 1.5 just gives |
432 |
you continous readyness and EAGAIN, 1.7 is even more broken and |
433 |
just hangs. work around both issues in a cygwin-specific hack. |
434 |
- improve idle watcher documentation slightly. |
435 |
|
436 |
5.23 Sun Dec 20 23:48:00 CET 2009 |
437 |
- support IDNs in resolve_sockaddr, and therefore in tcp_connect. |
438 |
- implement punycode_encode/decode, idn_nameprep, |
439 |
idn_to_ascii and idn_to_unicode operations in AnyEvent::Util. |
440 |
- provide $AE::VERSION. |
441 |
- removed traces of "no strict 'refs'". |
442 |
|
443 |
5.22 Sat Dec 5 03:51:13 CET 2009 |
444 |
- downgrade-or-fail in AnyEvent::Handle::push_write, to |
445 |
diagnose encoding failures earlier and more succinctly. |
446 |
(this works around bugs in perl, throwing away encoding info |
447 |
when passing scalar data to extensions). |
448 |
- add more examples to AnyEvent::Socket manpage. |
449 |
- upgrade internal warning set to the same as common::sense 2.03. |
450 |
- use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS |
451 |
(John Beppu). |
452 |
- AnyEvent::Socket::inet_aton now properly supports ipv6, as documented. |
453 |
- add google public dns servers to fallback server set and make sure |
454 |
we load-balance properly between the three sets. also add all |
455 |
fallback dns servers, not just a random one, to each dns config. |
456 |
|
457 |
5.21 Thu Nov 19 02:48:47 CET 2009 |
458 |
- fix a problem where socket constants were called with parameters |
459 |
(spotted by David Friedland). |
460 |
- fork_call never use'd POSIX (reported by Daisuke Maki). |
461 |
- improve perl 5.6 compatibility further (but it still won't work |
462 |
unless you rip out everything but the core). |
463 |
- prefer Net::DNS::Resolver over ipconfig, if installed, on win32. |
464 |
uses 10MB of RAM, but doesn't open a console window. *sigh*. |
465 |
|
466 |
5.202 Wed Oct 14 22:35:44 CEST 2009 |
467 |
- AnyEvent::DNS would unexpectedly clobber $_ under windows |
468 |
(analysed by Matthias Waldorf). |
469 |
- AnyEvent::Handle::run_cmd can now pass the PID of the |
470 |
newly-created process, which is much less useful than it might |
471 |
sound (based on patch by Yann Kerherve). |
472 |
|
473 |
5.201 Tue Sep 29 12:09:25 CEST 2009 |
474 |
- AnyEvent:Handle::on_starttls/on_stoptls methods were broken |
475 |
(reported by Torsten Foertsch). |
476 |
- common::sense 2.0 could cause tcp_server to throw an exception |
477 |
(analysed by elmex). |
478 |
|
479 |
5.2 Mon Sep 14 07:04:49 CEST 2009 |
480 |
- INCOMPATIBLE CHANGE: do no longer support register_read_type |
481 |
and register_write_type in AnyEvent::Handle, instead support |
482 |
package names (the facility was mostly abused). |
483 |
- implement "packagename-as-read/write type" support in |
484 |
AnyEvent::Handle. |
485 |
- AnyEvent::Handle: new options "keepalive" and "oobinline". |
486 |
- oobinline set by default to avoid security issues. |
487 |
- the pure-perl event loop backend wrongly detected times() fallback |
488 |
support (spotted by Pavel Boldin). |
489 |
|
490 |
5.12 Tue Sep 1 20:26:50 CEST 2009 |
491 |
- be more lenient when parsing resolv.conf files, as some people |
492 |
use hashmarks as comment indicator inside directives (reported |
493 |
by Michael S. Fischer). |
494 |
- use same set of warnings as common::sense 2.0. |
495 |
- fix a potential 32 bit overflow issue due to perl having problems |
496 |
with large hex constants in the Perl backend. |
497 |
- do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS |
498 |
by default. |
499 |
- fix some prototypes. |
500 |
- (optionally) require Async::Interrupt 1.02. |
501 |
- replace opendns fallbacks by verizon/level3 public dns servers. |
502 |
|
503 |
5.112 Fri Aug 21 13:59:04 CEST 2009 |
504 |
- AnyEvent::Handle could cause hard crashes in Net::SSLeay when |
505 |
using starttls with an illegal/uncreatable tls context. |
506 |
- port to broken win32 perls. |
507 |
|
508 |
5.111 Sun Aug 16 18:44:36 CEST 2009 |
509 |
- perl errornously requires sockaddr_un structures returned by the |
510 |
kernel to have a specific length. work around this bug by padding |
511 |
sockaddr_un structures (found to fail on OS X, as perl only uses |
512 |
the correct code for GNU/Linux, but likely to fail on other |
513 |
OSes as well). |
514 |
|
515 |
5.11 Wed Aug 12 17:49:37 CEST 2009 |
516 |
- fixed a bug in run_cmd with input from filename. |
517 |
- create a json coder object when none is given in write direction |
518 |
to reduce dependency on JSON. |
519 |
|
520 |
5.1 Tue Aug 11 03:17:32 CEST 2009 |
521 |
- tcp_connect now reports the correct error in $!, not an unrelated |
522 |
one (was broken since 4.91). |
523 |
- AE::Handle did not properly clear rtimeout/wtimeout watchers |
524 |
when retrying a tcp_connect. |
525 |
- new functions: AnyEvent::Util::run_cmd and close_all_fds_except. |
526 |
- fixed wrong documentation in AE::Socket. |
527 |
|
528 |
5.01 Mon Aug 10 03:16:32 CEST 2009 |
529 |
- last release broke EV child watchers :/. |
530 |
|
531 |
5.0 Sun Aug 9 17:32:09 CEST 2009 |
532 |
- officially introduce and document the AE API. |
533 |
- lowering the timeout at runtime did not have immediate effect |
534 |
in AnyEvent::Handle. |
535 |
- AE::Handle now has separate and independent read and write timeout |
536 |
settings. |
537 |
- probe for Guard module when AnyEvent::Util is loaded, not at |
538 |
runtime (Event doesn't like eval "use" from callbacks, and it avoids |
539 |
exporting the probe function). |
540 |
- Impl::IOAsync's io watchers did not use the filehandle emulation |
541 |
code to work around it's watcher limits. |
542 |
- work around IO::Async's problems with overloaded objects. |
543 |
- try to disable wearnings inside IO::Async, it's just too buggy. |
544 |
- do no longer use Event by default, only EV or Perl, Event is too buggy. |
545 |
- hackishly provide fast AE:: interface when EV or Perl is the backend. |
546 |
- inherit VERSION from AnyEvent::VERSION in most modules. |
547 |
|
548 |
4.91 Thu Aug 6 15:42:45 CEST 2009 |
549 |
- AE::Handle::starttls could get out of sync when the read buffer |
550 |
already contains some TLS handshake. |
551 |
- AE::Handle did not properly free the TLS session, causing bigger |
552 |
memory leaks in Net::SSLeay than expected (reported by toaster). |
553 |
- AE::Socket::tcp_connect will now artificially delay invoking |
554 |
the callback to avoid returning after invoking the callback. |
555 |
- convert many internal watcher uses to AE API. |
556 |
|
557 |
4.9 Sat Aug 1 11:07:01 CEST 2009 |
558 |
- Glib silently fails when registering a timeout with |
559 |
a negative value, so avoid that. |
560 |
- call condvar callback immediately when it is set |
561 |
after the condition is already signalled. |
562 |
- check rbuf_max condition only after trying to consume |
563 |
data. |
564 |
- normalise signal numbers to names when using the ae's |
565 |
signal handling, but do not document this (yet). |
566 |
- pure perl signal emulation did not properly set nonblocking |
567 |
mode on the signal pipe on win32 (this is unlikely to have caused |
568 |
any issues). |
569 |
- new module: AnyEvent::Debug. |
570 |
- AnyEvent::Strict now checks that a signal was specified |
571 |
by name (not name or number). |
572 |
- reduce memory footprint in typical cases by ~50kb by |
573 |
only compiling signal code when necessary. |
574 |
- add AnyEvent::Handle->rbuf_max. |
575 |
- grab the AE namespace for future new API, implement stubs for |
576 |
said future/alternative API. |
577 |
- new function: AnyEvent::Socket::format_hostport. |
578 |
|
579 |
4.881 Tue Jul 28 12:51:53 CEST 2009 |
580 |
- work around a bug in local in pre-5.10 perls, causing |
581 |
AnyEvent::Handle to recurse when it should not |
582 |
(analysed by elmex). |
583 |
|
584 |
4.88 Tue Jul 28 04:04:37 CEST 2009 |
585 |
- re-bless the handle into a dummy package after calling |
586 |
AnyEvent::Handle::destroy, so the user does not need to check |
587 |
for errors after every push_write etc. |
588 |
- do not attempt to run t/02_signals.t on obviously broken |
589 |
platforms. |
590 |
|
591 |
4.87 Sun Jul 26 02:06:16 CEST 2009 |
592 |
- do not attempt to linger when there is no longer a valid fh. |
593 |
- enforce tls mode to be either accept or connect, do not simply |
594 |
segfault in Net::SSLeay. |
595 |
- AnyEvent::Handle can now call tcp_connect itself (new parameters |
596 |
connect, on_prepare, on_connect and on_connect_error). Updated |
597 |
tutorial accordingly. |
598 |
- add AnyEvent::Impl::Irssi backend. |
599 |
|
600 |
4.86 Mon Jul 20 23:52:29 CEST 2009 |
601 |
- since the verbose warning is not enough, explicitly document |
602 |
that versions before 1.33 of Net::SSLeay are not secure. |
603 |
- work around signal handling races in Event and (...) Event::Lib. |
604 |
- try to align signal-race timer to full-second boundaries. |
605 |
- work around Tk not liking negative timeouts. |
606 |
- don't complain of different grades of environmental unfriendlyness |
607 |
in IO::Async. |
608 |
|
609 |
4.85 Sat Jul 18 06:16:14 CEST 2009 |
610 |
- nail the signal race problem in perl once and for all |
611 |
(see $AnyEvent::MAX_SIGNAL_LATENCY). |
612 |
- take advantage of Async::Interrupt if it is available. |
613 |
- load Time::HiRes and Guard modules on demand only. |
614 |
- add optional/recommended modules section to AnyEvent |
615 |
documentation. |
616 |
- reduce memory usage considerably (and reduce startup penalty) |
617 |
by not using "strict", "warnings" and "overload" modules. |
618 |
- work around buggy windows/openbsd perls and provide EBADMSG |
619 |
and EPROTO ourselves when missing. |
620 |
- improve perl 5.6 compatibility of the core event loop. |
621 |
- made Net::SSLeay version 1.33 a soft requirement. |
622 |
|
623 |
4.83 Fri Jul 17 16:56:26 CEST 2009 |
624 |
- implement AnyEvent::Socket::getprotobyname. |
625 |
- AnyEvent::CondVar's will now detect recursive blocking |
626 |
waits and will croak, as too many people fall into |
627 |
this trap. |
628 |
- AnyEvent::Handle will now call ->destroy on itself after |
629 |
executing the on_error callback, instead of doing some |
630 |
half-baked internal shutdown, for fatal errors. |
631 |
- clarify on_eof behaviour w.r.t. the read queue and |
632 |
on_read callbacks. |
633 |
- ignore some possible spurious wake-ups in tcp_connect. |
634 |
|
635 |
4.82 Sat Jul 11 00:34:55 CEST 2009 |
636 |
- POE and Event backends didn't accept some callable objects as |
637 |
callbacks. |
638 |
- use Config module instead of POSIX module to detect signal names |
639 |
in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX |
640 |
module doesn't even have all POSIX signals :/. |
641 |
- use more workarounds around the many refcnt/corruption bugs in |
642 |
Event::Lib. |
643 |
- work around a race condition in perl's select, causing t/03_child.t |
644 |
to rarely fail. |
645 |
|
646 |
4.81 Thu Jul 9 10:30:30 CEST 2009 |
647 |
- AnyEvent::Handle didn't properly diagnose write errors |
648 |
(it expected -1 from syswrite, how lame... :). |
649 |
- support file descriptors in addition to file handles |
650 |
in AnyEvent->io. |
651 |
- new env variables: PERL_ANYEVENT_RESOLV_CONF, |
652 |
PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE |
653 |
and PERL_ANYEVENT_CA_PATH. |
654 |
- provide a sensible synopsis section for AnyEvent::TLS. |
655 |
- add a "supported backends" section to the manpage. |
656 |
- added simple io watcher test to testsuite, using a |
657 |
portable_socketpair. |
658 |
- tried to improve the stability of the Event::Lib backend, |
659 |
YMMV. |
660 |
|
661 |
4.8 Mon Jul 6 23:45:16 CEST 2009 |
662 |
- AnyEvent::DNS did not properly follow CNAME records with |
663 |
uppercase targets. |
664 |
- AnyEvent::DNS would errornously return AAAA records |
665 |
with v4 mapped addresses (a faulty record) as ipv4 addresses, |
666 |
causing AnyEvent::Socket to throw an exception. |
667 |
- added new module AnyEvent::TLS for easier SSL/TLS context |
668 |
creation, with many options including hostname verification, |
669 |
secure default configuration, lots of documentation and, |
670 |
predefined diffie-hellman keys for perfect forward security |
671 |
and much more. get it while it's still fresh! |
672 |
- use AnyEvent::TLS in AnyEvent::Handle for context management. |
673 |
- load AnyEvent::Handle only on demand in AnyEvent::DNS, |
674 |
so AnyEvent::Socket users have smaller memory footprint |
675 |
in the common case. |
676 |
- add AnyEvent::Handle->push_shutdown method. |
677 |
- add an additional error message parameter to AnyEvent::Handle's |
678 |
on_error callback (for TLS, $! is still available). |
679 |
- add AnyEvent::Handle on_starttls/on_stoptls callbacks. |
680 |
- make AnyEvent::Handle more robust against early connection |
681 |
failures (during new), and return C<undef> in that case |
682 |
from the constructor. |
683 |
- AnyEvent::Handle will now try to load only JSON::XS first, |
684 |
then fall back to JSON. |
685 |
- format_ipv4/format_ipv6 are now exported by default, for symmetry, |
686 |
and because it was documented that way. |
687 |
|
688 |
4.451 Fri Jul 3 00:28:58 CEST 2009 |
689 |
- do not clear rbuf when shutting down an AnyEvent::Handle |
690 |
object - doing so breaks AnyEvent::HTTP. |
691 |
|
692 |
4.45 Mon Jun 29 22:59:26 CEST 2009 |
693 |
- a write error could cause AnyEvent::Handle to create |
694 |
an I/O watcher with an undefined $fh. |
695 |
- special-case mapped ipv4 addresses in both |
696 |
AnyEvent::Socket::format_address and parse_address, |
697 |
to treat them just like ipv4 addresses. |
698 |
- updated and overhauled the AnyEvent::Intro doc. |
699 |
- implement AnyEvent::Socket::format_ipv4/ipv6. |
700 |
- slightly speed up portable_pipe/socketpair. |
701 |
- expand condvar begin/end documentation. |
702 |
|
703 |
4.42 Fri Jun 26 08:32:18 CEST 2009 |
704 |
- preliminary and neccesarily incomplete support for IO::Async. |
705 |
- reset SIGCHLD to DEFAULT when AnyEvent is loaded, in |
706 |
case it was set to IGNORE, to ensure we can catch child statuses |
707 |
even when the calling env acts stupidly. |
708 |
- updated benchmarks with IO::Async, which performs very well. |
709 |
|
710 |
4.412 Wed Jun 24 01:35:57 CEST 2009 |
711 |
- support an "untaint" attribute for AnyEvent::DNS |
712 |
and set it on the default resolver. |
713 |
- implement and document AnyEvent::Impl::Perl::loop. |
714 |
- remove all anyevent-env variables from %ENV |
715 |
when running in tainted mode. |
716 |
- mention and extend the IO::Lambda benchmark. |
717 |
|
718 |
4.411 Sun Jun 7 18:48:02 CEST 2009 |
719 |
- do not try to use F_SETFD on windows. |
720 |
|
721 |
4.41 Thu May 14 06:40:11 CEST 2009 |
722 |
- work around issues in older perls (5.8.0?) when |
723 |
a signal handler is deleted from the %SIG hash. |
724 |
- use POSIX::_exit in child test, to avoid running |
725 |
destructors. |
726 |
- speed up CHLD handling by relying on SIGCHLD being |
727 |
synchronously delivered, even when we roll our own |
728 |
implementation. |
729 |
- AnyEvent::DNS: add the "dname" resource record name |
730 |
for cosmetic reasons. |
731 |
|
732 |
4.4 Sun Apr 26 20:12:33 CEST 2009 |
733 |
- implemented idle watchers, where applicable. |
734 |
- AnyEvent->time died when Event backend was in use. |
735 |
- fix a memleak in the Tk backend. |
736 |
- sped up Tk timer handling. |
737 |
- clip negative "after" values to 0 in AnyEvent::Impl::Event |
738 |
to avoid spamming out warning messages. |
739 |
- fix Qt timers without interval. |
740 |
- avoid Qt zero-timeout special case, in old libqt's. |
741 |
|
742 |
4.352 Mon Apr 20 16:31:11 CEST 2009 |
743 |
- fix AnyEvent::Strict error messages for child watchers. |
744 |
- fix/update Makefile.PL "recommends" versions. |
745 |
- add AnyEvent->now_update. |
746 |
|
747 |
4.351 Sat Apr 11 07:56:14 CEST 2009 |
748 |
- actually make the signal pipe work under win32. |
749 |
- localise $! in signal handler to avoid changing $!, although |
750 |
perl itself does not seem to save/restore errno either. |
751 |
- set the cloexec flag on the signal pipe (normally set by perl too). |
752 |
|
753 |
4.35 Fri Mar 27 11:48:20 CET 2009 |
754 |
- event models relying on AnyEvent's signal watcher emulation did |
755 |
invoke the callback asynchronously, contradicting documentation |
756 |
and causing signals to get lost (this includes AnyEvent's own |
757 |
event loop). AnyEvent now uses the standard pipe trick to make |
758 |
callback execution synchronous to the event loop. |
759 |
- AnyEvent::Handle didn't free TLS context data on DESTROY |
760 |
(patch by Pavel Shaydo). |
761 |
- work around the ever-incompatibly-changing API of MakeMaker. |
762 |
- document that changing global variables without restoring them |
763 |
is a bad idea in Perl (noted by Adam Rosenstein). |
764 |
- AnyEvent::Strict now barfs if ->io is passed a file. |
765 |
|
766 |
4.34 Thu Feb 12 18:32:45 CET 2009 |
767 |
- separately buffer TLS read data, as otherwise the read queue could |
768 |
deadlock as receiving data is not expected while draining the |
769 |
read queue (which cna only happen with TLS). |
770 |
- raise EBADMSG error on JSON decoding errors. |
771 |
- fix some minor manpage bugs (reported by Maximilian Gaß). |
772 |
- speed up select bitmask parsing quite a bit in the pure perl |
773 |
backend. |
774 |
- use CORE::select instead of just select in AnyEvent::Impl::Perl. |
775 |
|
776 |
4.331 Tue Jan 6 21:07:25 CET 2009 |
777 |
- socketpair fails on many vista machines because vista has |
778 |
completely broken accept/getpeername and getsockname functions, |
779 |
so we provide our own socketpair emulation that kind of works |
780 |
(AnyEvent::Util::portable_pipe). |
781 |
- new function: AnyEvent::Util::portable_socketpair. |
782 |
- take advantage of the Guard module if it exists. |
783 |
|
784 |
4.33 Fri Nov 21 02:35:40 CET 2008 |
785 |
- AnyEvent::Strict did errornously flag a fileno of 0 as |
786 |
illegal. |
787 |
- reduce memory usage and slightly speed up the pure perl backend |
788 |
by only storing the file descriptor, not the file handle. |
789 |
- add missing autocork method to AnyEvent::Handle (reported |
790 |
by Adam Rosenstein). |
791 |
- AnyEvent::DNS->resolve errornously documented an $rcode result |
792 |
argument, but there isn't (spotted by Henrik Krohns). |
793 |
- the naked truth about Tk - it's basically unmaintained. |
794 |
|
795 |
4.32 Mon Nov 3 22:46:32 CET 2008 |
796 |
- fix AnyEvent::Socket::resolve_sockaddr to properly support |
797 |
unix sockets again and choose a proper default for |
798 |
the socket type. |
799 |
- fix call to resolve_sockaddr in tcp_connect. |
800 |
|
801 |
4.31 Thu Oct 30 04:41:48 CET 2008 |
802 |
- implemented AnyEvent::Handle->destroy method. |
803 |
- hint about unexpected effects in TLS mode. |
804 |
- speed up AnyEvent::Impl::Perl by using more arrays |
805 |
instead of hashes without the slightest loss of readability :) |
806 |
- work around a perl argument refcounting bug. |
807 |
- with some perl patches applied, AnyEvent now seems leak-free. |
808 |
|
809 |
4.3 Fri Oct 3 09:18:43 CEST 2008 |
810 |
- AnyEvent will now install a no-op signal handler for SIGPIPE, |
811 |
unless one has been installed already. |
812 |
- warn about not loaidng AnyEvent::Impl::POE early enough |
813 |
(patch by Adam Rosenstein). |
814 |
- fixed a great number of bugs and corner cases in AnyEvent::Handle. |
815 |
- fix a bug in where in SSL connect mode, the |
816 |
client would first wait for some data by the server and |
817 |
otherwise hang. (reported and analysed in an absolutely |
818 |
exemplary manner by Adam Rosenstein). |
819 |
- fix a bug in where SSL EOF would not be treated |
820 |
as stream EOF, putting the connection into a hung state. |
821 |
(reported and analysed in an absolutely exemplary manner by Adam |
822 |
Rosenstein). |
823 |
- fix a potential segfault when the TLS context |
824 |
would go missing in a read callback (e.g. due to stoptls) - |
825 |
Net::SSLeay of course makes no type checking whatsoever. |
826 |
- AnyEvent::Handle will not stop the read watcher in TLS mode. |
827 |
- AnyEvent::Handle->stoptls will send a TLS close notify |
828 |
instead of simply destroying the stream now. |
829 |
- fix a bug where large blocks of data written in TLS mode |
830 |
would not be sent unless triggered by receives. |
831 |
- on_drain will now take the tls write buffer into account. |
832 |
- SSL operations have been streamlined a bit and should be faster |
833 |
now, more to come, though. |
834 |
- added a FAQ section, will document nontrivial issues in there. |
835 |
|
836 |
4.234 Mon Sep 29 04:08:13 CEST 2008 |
837 |
- fix child watcher documentation: only child exits will be reported, |
838 |
no trace events. |
839 |
- mention SIGPIPE in AnyEvent::Handle. |
840 |
- perl backend will now use POSIX::times as monotonic clock |
841 |
when available and the monotonic clock option is not. |
842 |
- run even without Time::HiRes in some cases now, but do not |
843 |
provide sub-second accuracy (all *supported* perl versions have |
844 |
Time::HiRes). |
845 |
|
846 |
4.233 Fri Aug 22 01:48:05 CEST 2008 |
847 |
- fix a bug in the testsuite, causing a hang. |
848 |
- clarified lots of AnyEvent::Handle constructor arguments. |
849 |
|
850 |
4.232 Thu Aug 21 20:44:25 CEST 2008 |
851 |
- fix a bug in AnyEvent::Handle that could cause two fatal error |
852 |
callbacks on EOF, which caused AnyEvent::HTTP to signal success twice |
853 |
(testcase by Leon Brocard). |
854 |
- clarified on_eof/on_error documentation a bit. |
855 |
|
856 |
4.231 Tue Jul 29 13:12:15 CEST 2008 |
857 |
- remove some debugging code left in AnyEvent::Util::fork_call |
858 |
(and no, it's impossible to implement with the broken windows |
859 |
perls without resource leaks or worse). |
860 |
|
861 |
4.23 Tue Jul 29 12:19:59 CEST 2008 |
862 |
- document the first parameter passed to condvar callbacks to be |
863 |
the callback. |
864 |
- add AnyEvent::Socket::{ntoa,aton} aliases. |
865 |
- optimize the AE::Handle->push_read (line) for the default |
866 |
eol marker. |
867 |
- optimize push_read (packstring|storable) for small packets. |
868 |
- invoke on_error callback when no on_eof callback is set. |
869 |
- fix a bug in push_read (storable) of unknown impact. |
870 |
|
871 |
4.22 Sun Jul 20 16:34:13 CEST 2008 |
872 |
- new function AnyEvent::Socket::parse_hostport. |
873 |
- as the bulkheads at microsoft can't even get getprotobyname reliably |
874 |
working on their shitty fucking broken os we need to hardcode |
875 |
some common protocol numbers in AnyEvent::Socket. How can |
876 |
people even bother with such a pile of shit as windows. |
877 |
|
878 |
4.21 Thu Jul 17 14:40:05 CEST 2008 |
879 |
- INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype |
880 |
which is incompatible to previous invocation syntax. |
881 |
- work around more windows perl bugs in fork_call. I cannot imagine |
882 |
why anybody would bother implementing fork in such an obviously |
883 |
unusable way. |
884 |
- avoid calling $do_retry if the request has already finished in |
885 |
AnyEvent::DNS, thus avoiding an exception (reported by Anatoly |
886 |
K. Sharifulin). |
887 |
|
888 |
4.2 Sat Jul 12 22:42:11 CEST 2008 |
889 |
- support an "interval" argument for repeating timers. |
890 |
- fix a bug in the pure perl event loop that caused timers to |
891 |
be delayed under some circumstances. |
892 |
- implement optional argument checking via PERL_ANYEVENT_STRICT. |
893 |
- not importing anything from Event causes it to malfunction |
894 |
and not use Time::HiRes, work around this bug. |
895 |
- more bug workarounds against the endlessly broken Tk module, |
896 |
seems to work now on 32 bit machines, but less so on 64 bit. |
897 |
- minor optimisations applied to most backends. |
898 |
|
899 |
4.161 Fri Jul 4 14:24:48 CEST 2008 |
900 |
- fixed the t/handle/02_write.t, it read after write, with the |
901 |
assumption that all data could be written before read |
902 |
(reported by Jim Razmus). |
903 |
|
904 |
4.160 Thu Jul 3 04:02:21 CEST 2008 |
905 |
- re-registering signal watchers after unregistering one did |
906 |
not work if a backend falls back on AnyEvent's default |
907 |
implementation (also affected child watchers). |
908 |
- new AnyEvent::Handle options: autocork and no_delay. |
909 |
|
910 |
4.152 Sun Jun 22 14:15:44 CEST 2008 |
911 |
- allow for 32-bit perls that implement shifts differently |
912 |
on different architectures in parse_ipv4 (reported and |
913 |
analysed by Keiichi DAIBA). |
914 |
|
915 |
4.151 Fri Jun 6 17:34:24 CEST 2008 |
916 |
- make sure specifying _only_ on_read and never pushing reads |
917 |
works. |
918 |
|
919 |
4.15 Fri Jun 6 13:00:46 CEST 2008 |
920 |
- the pure perl backend would keep some watchers alive when more than |
921 |
one watcher was registered for the same fd. |
922 |
- new "packstring" and "storable" read and write types |
923 |
for AnyEvent::Handle. |
924 |
- allow on_eof handler to be called after on_error with EPIPE returns. |
925 |
- do not immediately call on_read callback in handle constructor. |
926 |
|
927 |
4.14 Thu Jun 5 20:29:31 CEST 2008 |
928 |
- Fixed a bug in DNS SRV priority sorting. |
929 |
- AnyEvent::Util::guard now reports runtime errors while |
930 |
executing the guard block as warnings. |
931 |
- handle 0-byte-reads just before EOF correctly in AnyEvent::Handle. |
932 |
|
933 |
4.13 Thu Jun 5 00:47:59 CEST 2008 |
934 |
- AnyEvent::DNS only followed cname chains with length 2, |
935 |
contrary to documentation. bumped it up to 10, thanks to |
936 |
microsoft, the current broken-dns-config-king. |
937 |
- AnyEvent::DNS didn't check for socket return status |
938 |
but instead relied on perl not creating filehandles in that |
939 |
case - too bad it gives you a bogus file handle (reported |
940 |
and analysed by Vladimir Timofeev). |
941 |
- fix queue management logic in AnyEvent::Handle: |
942 |
when on_read was registered, the queue was empty |
943 |
and no progress could be made AnyEvent::Handle would |
944 |
enter an endless loop. |
945 |
- correctly start reading again when the handle became |
946 |
busy again after idling. |
947 |
- correctly treat tls shut-downs as EOF condition |
948 |
(for the time being, would be nice to have a callback for that). |
949 |
- correctly call eof callback under all conditions |
950 |
(hopefully). |
951 |
- the timeout callback did not expect that $self can go |
952 |
away any time. |
953 |
|
954 |
4.12 Tue Jun 3 10:58:04 CEST 2008 |
955 |
- include AnyEvent::Intro, a tutorial for anyevent, |
956 |
anyevent::socket and anyevent::handle. |
957 |
- allow more options in on_error. |
958 |
|
959 |
4.11 Fri May 30 23:42:25 CEST 2008 |
960 |
- INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide |
961 |
reverse_lookup and reverse_verify to replace it, support |
962 |
v4mapped and v4compat addresses. |
963 |
- provide more documentation for the resolver class. |
964 |
- really replace longest run of :0: by :: in format_address, |
965 |
also properly convert :: and ::1 again. |
966 |
- support NAPTR record name and decode it. |
967 |
- implement random weight sampling for SRV records, as per |
968 |
rfc 2782. |
969 |
- correctly abort on srv-record targets of ".". |
970 |
- added AnyEvent::DNS::wait_for_slot. |
971 |
- in the unlikely event of a virtual circuit connection |
972 |
being invalidated by a delayed udp reply, AnyEvent::DNS |
973 |
could die. this has been fixed. |
974 |
- plug a probable memleak in the DNS vc code. |
975 |
- use configured timeout also for VC request phase. |
976 |
- implement timeout and max_outstanding methods |
977 |
for resolver class. |
978 |
- update version numbers in all modules. |
979 |
|
980 |
4.1 Thu May 29 05:45:40 CEST 2008 |
981 |
- INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address |
982 |
which mirrors their purpose better (old names still available |
983 |
for a while). Also moved AnyEvent::DNS::addr to |
984 |
AnyEvent::Socket::resolve_sockaddr. |
985 |
- implement AnyEvent->time and AnyEvent->now. |
986 |
- fix IPv6 support in pack_sockaddr. |
987 |
- officially un-experimentalise all newly introduced code. |
988 |
- support unix domain sockets everywhere by specifying |
989 |
a host of "unix/" and the pathname as service. |
990 |
- implement an activity timeout in AnyEvent::Handle. |
991 |
- added a regex read type to AnyEvent::Handle. |
992 |
- added a json read/write type to AnyEvent::Handle. |
993 |
- always croak in on_error in AnyEvent::Handle. |
994 |
- document how to subclass AnyEvent::Handle. |
995 |
- implement AnyEvent::Util::fork_call. |
996 |
- add support for IPv6 nameservers and nameserver statements. |
997 |
- work around _yet_ _another_ windows perl bug in where empty |
998 |
select masks cause errors in the pure perl backend (this |
999 |
workaround cannot be implemented for the other event loops, |
1000 |
but EV also works around this). |
1001 |
- supply AnyEvent::Socket::address_type and make good use of it. |
1002 |
- clarify and fix the Handle documentation w.r.t $self vs. $handle |
1003 |
vs. callback arguments. |
1004 |
- add some recommends to the META.yml. |
1005 |
|
1006 |
4.05 Mon May 26 19:44:06 CEST 2008 |
1007 |
- some platforms ignore the rfc and prepend an extra sa_len member to |
1008 |
the sockaddr structure, cater for those. also use sockaddr_family, |
1009 |
didn't know it was there, but comes in handy for the workaround. |
1010 |
- undo the import WIN32 hack. |
1011 |
|
1012 |
4.04 Mon May 26 08:03:31 CEST 2008 |
1013 |
- try to work around yet another windows bug: failed connects are reported |
1014 |
as if it were out-of-band data. windows users: you suck. Only |
1015 |
EV, Glib, Event and the pure perl backend can handle this condition. |
1016 |
- optimize write algorithm in AnyEvent::Handle. |
1017 |
- properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly. |
1018 |
- tcp_connect now properly iterates through all targets. |
1019 |
- check wether IPv6 sockets can actually be created, otherwise, disable ipv6 |
1020 |
support entirely. also hardcode some AF_INET6 constants for the |
1021 |
sake of perl 5.8. |
1022 |
- vastly improved nameserver/suffix detection algorithm for windows |
1023 |
(still a hack...) |
1024 |
- try to use a fallback dns server if no dns servers could be found. |
1025 |
- splatter around some local $SIG{__DIE__}'s to work around bad |
1026 |
die handlers. |
1027 |
|
1028 |
4.03 Sun May 25 05:05:57 CEST 2008 |
1029 |
- work around perl bugs on windows, where perl returns undocumented |
1030 |
error codes for sysread, syswrite etc. |
1031 |
- fix AnyEvent::Handle::unshift_read to really unshift and not push. |
1032 |
- AnyEvent::Handle could sometimes stumble over 0-byte writes. |
1033 |
- fix unpack_sockaddr, now works for AF_INET6 sockaddr's. |
1034 |
- updated documentation. |
1035 |
- lots of workarounds for perl 5.8 quirks. |
1036 |
- implement a more flexible type system for both reads and writes |
1037 |
in AnyEvent::Handle. |
1038 |
- support IPv6 in tcp_server. |
1039 |
- pass local hostname and port to the prepare callback in tcp_server. |
1040 |
- make types extendable in AnyEvent::Handle. |
1041 |
- croak, not die, in AnyEvent::Handle; |
1042 |
|
1043 |
4.0 Sat May 24 19:58:08 CEST 2008 |
1044 |
- added (experimental) AnyEvent::DNS module. |
1045 |
- added (experimental) AnyEvent::Socket module. |
1046 |
- added (experiemntal) support for TLS/SSL in AnyEvent::Handle. |
1047 |
- make condvar->end call ->send as documented. |
1048 |
- make condvar constructor support the documented cb => argument. |
1049 |
- overload condvars so you can use them instead of code references. |
1050 |
- support more of the POSIX forms instead of a pure dotted quad for |
1051 |
verbatim ip addresses. |
1052 |
- AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. |
1053 |
- try very hard not to freeze on broken windows perls. |
1054 |
- make on_eof optional in AnyEvent::Handle. |
1055 |
- support service names in tcp_connect and tcp_server. |
1056 |
- work around netbsd/openbsd bug causing unix sockets not to |
1057 |
become writable until empty. |
1058 |
- only make a dummy read for error detection when getpeername |
1059 |
returns ENOTCONN. |
1060 |
- defend AnyEvent::Util::inet_aton against Coro::LWP's brutal |
1061 |
override. |
1062 |
- new AnyEvent::Guard::cancel. |
1063 |
|
1064 |
3.5 Sat May 17 23:17:14 CEST 2008 |
1065 |
- gracefully deal with systems not defining CLOCK_MONOTONIC |
1066 |
(as opposed to not supporting the clock itself, which worked in 3.41). |
1067 |
- added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server |
1068 |
helpers, to replace the ill-designed AnyEvent::Socket class. |
1069 |
- added some examples (eg/connect, eg/handle). |
1070 |
- added AnyEvent::Util::guard. |
1071 |
- AnyEvent::Util now exports most of its functions. |
1072 |
- fixed a bug in *_read_chunk, where the callback was called with $self as |
1073 |
first and second argument. |
1074 |
- simplified and fixed regex handling in AnyEvent::Handle::*_read_line. |
1075 |
- fix manpage condvars examples to use new syntax. |
1076 |
- considerably improve inet_aton with EV::ADNS backend to support |
1077 |
cname chains (common dns configuration bug, e.g. with www.google.de) |
1078 |
and aaaa records if no a record could be found. |
1079 |
|
1080 |
3.41 Sun May 11 19:53:13 CEST 2008 |
1081 |
- work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. |
1082 |
|
1083 |
3.4 Sun May 11 00:29:25 CEST 2008 |
1084 |
- complete re-implementation of condvars with a hopefully much more useful |
1085 |
API (backward compatible functions exist but are undocumented). |
1086 |
- AnyEvent::Handle fully rewritten with a hopefully more powerful API |
1087 |
(still subject to change). |
1088 |
- take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
1089 |
- provide AnyEvent::post_detect and @AnyEvent::post_detect, which |
1090 |
allows module authors to avoid forcing event loop detection. |
1091 |
(used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example). |
1092 |
- remove coro backends: Coro now provides generic support for AnyEvent |
1093 |
via Coro::AnyEvent. |
1094 |
|
1095 |
3.3 Mon Apr 28 09:51:06 CEST 2008 |
1096 |
- added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. |
1097 |
- fix a bug in the pure perl backend that kept watchers alive |
1098 |
when multiple watchers were registered for the same fh/poll combo. |
1099 |
- add a benchmark section showing AnyEvent overhead and comparing |
1100 |
the different event loops with each other. |
1101 |
- prefer pure perl over tk when autoprobing, as it's about as fast, |
1102 |
but doesn't crash with many watchers. |
1103 |
- declare Qt support non-experimental. |
1104 |
- clarify the confusing section about the file descriptor being |
1105 |
kept alive. |
1106 |
- document the race between loading of an event module and |
1107 |
child processes exiting. |
1108 |
- support POE as "backend" (with some caveats, POE is not generic |
1109 |
enough, and darn slow). |
1110 |
- support Wx and Prima through POE. |
1111 |
- optimise perl backend to use 20% less memory and take advantage |
1112 |
of typical timeout behaviour. It can now compete with |
1113 |
select/poll-based C event loops in most cases (it is usually |
1114 |
faster than Event and Glib :). |
1115 |
- roughly cut EV memory use in half and increase its speed by 30%, |
1116 |
by removing undocumented functionality. Did something similar |
1117 |
to other event loops where possible. |
1118 |
|
1119 |
3.2 Thu Apr 24 10:10:40 CEST 2008 |
1120 |
- do not die when anyevent watchers are destroyed while running callbacks |
1121 |
in the pureperl backend (could only happen when two watchers are registered |
1122 |
for the same fh/poll combo). |
1123 |
- support autoloading for child watchers, was broken in all versions. |
1124 |
- implement PERL_ANYEVENT_MODEL env variable. |
1125 |
- (experimental) implement interface to the (very crashy/buggy) Event::Lib module. |
1126 |
- (experimental) implement interface to the Qt module (cannot be autoprobed). |
1127 |
- this release is IO:AnyEvent-proof. |
1128 |
|
1129 |
3.12 Tue Apr 22 07:11:46 CEST 2008 |
1130 |
- reinstate AUTHOR section that got lost somehow. |
1131 |
- do not hang in the testsuite with badly broken perls (activestate, |
1132 |
strawberry...), but instead diagnose the problem and continue. |
1133 |
- use INT instead of CHLD in an attempt to support broken windows perls |
1134 |
better (this decreases the test quality, unfortunately). |
1135 |
- do not send the signal to the process group |
1136 |
(no problem for CHLD, bad for INT :). |
1137 |
|
1138 |
3.11 Sat Apr 19 06:57:31 CEST 2008 |
1139 |
- major documentation rework. |
1140 |
- document the fact that child watchers only watch for zombies. |
1141 |
- fix the child watcher example. |
1142 |
|
1143 |
3.1 Wed Apr 16 17:09:01 CEST 2008 |
1144 |
- work around recurring bugs in Tk by dup'ing filehandles, the |
1145 |
only method with good success chances on Tk (the bugs apparently |
1146 |
don't get fixed anytime soon). |
1147 |
- lift the restriction of only one watcher per fh direction |
1148 |
(as the Tk bug workaround also lifts it and only Tk imposed |
1149 |
such strong limits). |
1150 |
- changed probe order to prefer coro adaptors. |
1151 |
- explain why recursion into the event loop is not supported |
1152 |
unless the backend supports it (only Coro::EV does without |
1153 |
any restrictions...). |
1154 |
- add simple manpages for all backend modules. |
1155 |
|
1156 |
3.0 Mon Apr 7 21:30:23 CEST 2008 |
1157 |
- Coro::Signal changed semantics, roll our own, also cleaning |
1158 |
up the Coro implementation in general. |
1159 |
- rename Coro backend to CoroEvent. |
1160 |
- add some decision helping paragraph to the manpage that should |
1161 |
help people to decide whether AnyEvent is the right thing for them. |
1162 |
|
1163 |
2.9 Mon Jan 28 13:31:54 CET 2008 |
1164 |
- update for EV 3.0 API changes. |
1165 |
|
1166 |
2.8 Sun Nov 25 15:06:03 CET 2007 |
1167 |
- waitpid can validly return 0. accept this fact of life |
1168 |
instead of reporting it to any watchers. |
1169 |
|
1170 |
2.7 Fri Nov 23 11:41:14 CET 2007 |
1171 |
- force use of AnyEvent::Impl::Perl in testsuite, there is |
1172 |
too much breakage outside AnyEvent. |
1173 |
- deliver signals synchronously in AnyEvent::Impl::Perl |
1174 |
or any other event loop that relies on AnyEvents child |
1175 |
watcher emulation. *Could* help with hanging testsuite |
1176 |
(except when Event or EV are installed). |
1177 |
|
1178 |
2.6 Fri Nov 9 20:36:35 CET 2007 |
1179 |
- fix bug in testsuite. |
1180 |
- move EV adaptor modules to AnyEvent. |
1181 |
- add Coro+EV adaptor module. |
1182 |
|
1183 |
2.55 Tue Nov 6 17:41:32 CET 2007 |
1184 |
- add EV to the list of supported event models. |
1185 |
- do not auto-reset pid watchers, pass pid and status to them. |
1186 |
- allow a pid of zero to watch for all children in child watchers. |
1187 |
|
1188 |
2.54 Wed Jul 18 17:36:23 CEST 2007 |
1189 |
- work around a perl bug that results in |
1190 |
BEGIN not safe after errors--compilation aborted without |
1191 |
any discernible reason or error message by once |
1192 |
again not use'ing strict. |
1193 |
|
1194 |
2.53 Sun Jul 8 10:51:53 CEST 2007 |
1195 |
- make Glib interface work again (spotted by elmex). |
1196 |
|
1197 |
2.52 Wed Mar 7 18:36:16 CET 2007 |
1198 |
- child watchers stopped working permanently when |
1199 |
all current child watchers were destroyed. |
1200 |
|
1201 |
2.51 Mon Dec 11 21:33:24 CET 2006 |
1202 |
- work around bugs in perl where eval "require Module" returns |
1203 |
true even if the module couldn't be loaded :/. |
1204 |
|
1205 |
2.5 Mon Dec 11 02:15:28 CET 2006 |
1206 |
- avoid $AUTOLOAD because many perls corrupt it. |
1207 |
- AnyEvent::detect forces autodetection. |
1208 |
- implement signal watchers (experimental) |
1209 |
(Tk does not support async signals (<= 804.027 at least), |
1210 |
so they might get delayed indefinitely on Tk). |
1211 |
- implement child watchers (experimental). |
1212 |
- moved default condvar implementation into base module, simplifying |
1213 |
most implementation modules. |
1214 |
|
1215 |
2.1 Fri Nov 24 15:50:48 CET 2006 |
1216 |
- better docs. |
1217 |
- simple testsuite. |
1218 |
- added AnyEvent->one_event method for special purposes. |
1219 |
|
1220 |
2.0 Wed Nov 1 02:21:30 CET 2006 |
1221 |
- INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both. |
1222 |
- INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore. |
1223 |
- EXPERIMENTAL: |
1224 |
new pure-perl model added, AnyEvent now always finds a suitable |
1225 |
event model. |
1226 |
- improved documentation. |
1227 |
- reduced cpu overhead. |
1228 |
|
1229 |
1.02 Fri Jan 13 14:15:40 CET 2006 |
1230 |
- add COPYING to clarify license. |
1231 |
|
1232 |
1.01 Sun Jan 8 05:48:09 CET 2006 |
1233 |
- correctly cancel Event timers, avoid memory leaks. |
1234 |
|
1235 |
1.0 Sun Jan 8 05:40:58 CET 2006 |
1236 |
- allow third-party interfaces to be added at runtime. |
1237 |
- fix version number. |
1238 |
|
1239 |
0.9 Fri Dec 30 02:25:41 CET 2005 |
1240 |
- fixed errors in the documentation. |
1241 |
- going back to using $VERSION as "load check", |
1242 |
as suggested by Jörn Reder. Let's see why I stopped using |
1243 |
this method. |
1244 |
- add 'hup' to the list of events to receive for glib |
1245 |
and generate 'rw' events for it, as some versions of |
1246 |
glib seem to require this, while this breaks other |
1247 |
(older) versions of glib. |
1248 |
- implemented PERL_ANYEVENT_VERBOSE env variable. |
1249 |
|
1250 |
0.3 Sun Dec 4 10:43:33 CET 2005 |
1251 |
- no changes to the code. |
1252 |
- fixed documentation. |
1253 |
- clarified some corner cases. |
1254 |
- vastly improved documentation with a long example. |
1255 |
|
1256 |
0.2 Thu Dec 1 22:18:49 CET 2005 |
1257 |
- check whether a valid method was called, to avoid endless loops. |
1258 |
- move condvars into their own class (cleaner, safer). |
1259 |
- don't die when some modules are not available. |
1260 |
- Tk is too broken to allow for words. Will work once Tk works. |
1261 |
|
1262 |
0.1 Thu Dec 1 19:44:57 CET 2005 |
1263 |
- spiced up and cleaned up for release. |
1264 |
|
1265 |
0.01 Tue Sep 21 18:12:09 CEST 2004 |
1266 |
- original version; copied from Convert::Scalar. |
1267 |
|