ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent/Changes
(Generate patch)

Comparing AnyEvent/Changes (file contents):
Revision 1.189 by root, Sat Sep 6 10:54:32 2008 UTC vs.
Revision 1.603 by root, Tue Jun 16 03:22:19 2015 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: document TCP_*** constants
4TODO: add debug/trace logging to some modules
5TODO: use the eval-with-cb-call trick also to speed up JSON?
6TODO: FAQ: common pitfalls? invoke-before-return
7TODO: FAQ: scope-ids? hosts file order?
8TODO: AnyEvent::Handle with AnyEvent::IO?
9TODO: invalid. and localhost. specialcasing inside AEDNS and not AESocket (rfc6761)
10TODO: maybe implement env variable to give hosts precedence
11TODO: hosts always read? that's not expected
12TODO: more formats for parse_ipv6 (single ipv6 address without port, ...p80, ...#80)
13
14 - new function AnyEvent::Socket::tcp_bind.
15 - new functions AnyEvent::fh_block and AnyEvent::fh_unblock.
16 - aligned ipv6 address formatting with RFC 5952 (by not shortening
17 a single :0: to ::).
18 - added stability canary support.
19
207.09 Sat May 2 16:38:53 CEST 2015
21 - AnyEvent::Debug called an internal function (AnyEvent::Log::ft)
22 that was renamed to AnyEvent::Log:format_time. uinder its old name
23 (analyzed by Michael Stovenour).
24 - update AnyEvent::DNS fallback resolver addresses:
25 seems google effectively killed most other free dns resolvers,
26 so remove them, but add cable and wireless (ecrc) since it was
27 stable for 20 years or so, official or not, and there should be
28 an alternative to google.
29 - perl5porters broke windows error codes in 5.20, and mapped
30 WSAEWOULDBLOCK on the (different) EWOULDBLOCK error code, and
31 WSAEINPROGRESS into the incompatible ERINPROGRESS code, probably
32 because they were so cool. They probably broke other error codes
33 for no reason, too, but I didn't care for checking, it's too
34 depressing. This version only works around the WSAEWOULDBLOCK
35 issue, because I don't have a nice way to work around the
36 WSAEINPROGRESS bug.
37
387.08 Wed Dec 10 05:27:17 CET 2014
39 - work around a newly introduced bug in Socket 2.011 (an
40 errornous sun_length check) (analyzed by Maxime Soulé).
41 - AnyEvent::TLS didn't load (but refer to) AnyEvent::Socket
42 (analyzed by Ben Magistro).
43 - AnyEvent::Strict will now confess, not croak. This is in line with
44 it being a development/debugging tool.
45 - work around a number of libglib bugs (debug builds of libglib enforce
46 certain undocumented behaviour patterns such as not being able to
47 remove a child watch source after it has fired, which we will try
48 to emulate to avoid "criticals". what where they thinking...).
49 - mention json security issues in AnyEvent::Handle, now that Douglas
50 Crockford has foolishly and incompatibly changed JSON.
51 - changed default dns resolver "max_outstanding" value from 1 to 10,
52 the latter beinfg the intended value all along
53 (reported by Ilya Chesnokov).
54 - added new "AnyEvent::Impl::UV" interface module to the UV event lib
55 (written by Mike Lowell).
56
577.07 Tue Dec 17 17:45:02 CET 2013
58 - the documentation for custom tls verify schemes was wrong. make it agree
59 with the code (reported by Maxime Soulé).
60 - added cbor read and write types to AnyEvent::Handle (using CBOR::XS).
61 - work around an API change in openssl that could cause wrong tls connection
62 aborts, likely on windows only (analyzed by sten).
63 - calling AnyEvent->now_update with AnyEvent::Impl::Perl caused an
64 endless loop (reported by Dietrich Rebmann).
65 - add tlsv1_1 and tlsv1_2 protocols to AnyEvent::TLS
66 (patch by Maxime Soulé).
67 - document AnyEvent::Impl::IOAsync::set_loop and
68 $AnyEvent::Impl::IOAsync::LOOP. Though only documented now, this
69 functionality has _always_ been available.
70 - force a toplevel domain name in t/81_hosts.t (analyzed by
71 David Jack Wange Olrik).
72 - document that AnyEvent::Log uses AnyEvent::IO.
73 - warn about AnyEvent::Filesys::Notify performance.
74 - praise the joys of AnyEvent::Fork::*.
75 - time for an =encoding directive.
76 - do no longer use JSON to create a default json coder, use
77 JSON::XS or JSON::PP directly.
78
797.05 Wed Aug 21 10:38:08 CEST 2013
80 - uts46data.pl couldn't be found due to wrong naming of the file
81 (reported by Fulko Hew).
82 - handle lone \015's properly in AE::Handle's default line read
83 (reported by various people).
84 - untaint ip addresses found in /etc/hosts (patch by José Micó).
85 - the memleak fix in 7.03 caused resolving via /etc/hosts to always
86 fail on first use (reported and testcase by Andrew Whatson).
87 - expose AnyEvent::Log::format_time, and allow users to redefine it.
88 - expose AnyEvent::Log::default_format, and allow redefinition.
89 - expose AnyEvent::Log::fatal_exit, to allow redefinition.
90 - AnyEvent::Debug shell can now run coro shell commands, if available.
91 - t/63* tests were wrongly in MANIFEST.
92 - kernel.org's finger server went MIA, switch to freebsd.org and
93 icculus.org.
94 - clarify that IO::AIO and AnyEvent::AIO are needed for AnyEvent::IO
95 to function asynchronously (reported by Daniel Carosone).
96 - hard-disable $^W in most tests, it generates too much garbage output.
97 - use a (hopefully) more future-proof method to emulate common::sense.
98 - upgrade to UTS-46:6.2.0.
99 - switch to INSTLIB from INSTLIBDIR, as INSTLIBDIR was wrongly
100 documented. should not affect anything.
101
1027.04 Thu Nov 15 02:23:58 CET 2012
103 - debugging code left in.
104
1057.03 Thu Nov 15 02:16:11 CET 2012
106 - AnyEvent::Socket::inet_aton did not work when dns resolution was
107 used to find the addresses (analyzed and patch by Maxime Soulé).
108 - fix a memory leak in the /etc/hosts lookup code when hosts don't
109 resolve and are not in hosts (analyzed by sten).
110
1117.02 Tue Aug 14 04:27:58 CEST 2012
112 - AnyEvent::Util::run_cmd could block indefinitely (analyzed and test
113 program by Yorhel).
114 - verified that AnyEvent::Socket follows RFC5952.
115 - try to parse "ADDR#PORT" in addition to "ADDR PORT".
116
1177.01 Sun May 13 01:03:17 CEST 2012
118 - fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but
119 not available, instead of throwing an exception.
120 - use File::Spec to get the tmpdir in t/*, to avoid needless
121 failures on (most, not mine :) windows boxes.
122 - new handle read types: tls_detect and tls_autostart.
123
1247.0 Fri Apr 13 06:33:30 CEST 2012
125 - child watchers are broken in POE 1.352 (also many earlier
126 versions) and there seems to be no way to work around it, as POE
127 itself is inherently racy. Document this regression and add a delay
128 in t/68_poe_03_child.t for the time being.
129 - new module AnyEvent::IO, that is a frontend to either a pure-perl
130 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
131 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
132 - load /etc/hosts only when DNS has no answer.
133 - stat /etc/hosts on every access and reload it if it changed.
134 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
135 - fix a buggy croak in the dh parameter reading in AnyEvent::TLS.
136 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
137 a side effect, they now use true appending as opposed to libc
138 appending, and the file might not have been opened when the function
139 returns.
140 - the default logging level was not properly documented in a variety of
141 places, this has hopefully been rectified.
142 - updated uts46data.pl for unicode 6.1.0.
143 - made log messages generated by AnyEvent submodules not
144 include the package name anymore, as it will be logged
145 by default already.
146 - upgrade to the trick used by common-sense 3.5 to work
147 around extra warning torture/breakage under perl 5.15.x.
148 - log messages by anyevent are now ucfirst, are usually full
149 sentences and do no longer include the package name.
150 - the storable read type would sometimes throw an exception instead
151 of causing EBADMSG (analyzed by Maxime Soulé).
152 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
153 keysize (reported by Paul Howarth).
154 - documented Rocco fabricating statements about AnyEvent and me.
155
1566.14 Tue Jan 31 20:00:24 CET 2012
157 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
158 probably an editing mistake (reported by Darin McBride).
159
1606.13 Thu Jan 12 07:27:01 CET 2012
161 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
162 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
163 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
164 regex-matching inside the callback would change the parameter, despite
165 $1 being dynamically scoped per-block. this perl bug is now being
166 worked around (testcase by Cindy Wang).
167
1686.12 Mon Dec 12 13:21:10 CET 2011
169 - $! was clobbered by subsequent calls in tcp_connect, due to
170 the postpone not saving/restoring it (analyzed by Richard Garnish).
171
1726.11 Tue Nov 22 10:36:05 CET 2011
173 - Tk cannot create windows when tainted unless you set a title, so
174 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
175 - escape any nonprintable/nonascii characters when stringifying
176 backtraces.
177 - log the reason loading an interface module fails to load at level debug.
178
1796.1 Tue Oct 4 19:44:30 CEST 2011
180 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
181 and above), and some messages inside AnyEvent have been elevated
182 to higher log levels to print by default.
183 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
184 - work around yet another signal race bug, newly introduced by POE. My
185 how that module sucks.
186 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
187 support ipv6-only operation even with defaults (also use ::1 instead
188 of 127.0.0.1).
189 - give better guidelines on how to chose a logging level.
190 - AnyEvent::Log can now cap (limit) log message priority by context.
191 - AE::log will now save and restore $! and $@, so callers don't have to.
192 - AE::Strict now checks that registered fds are still valid at regular
193 intervals.
194 - skip frequently called 'require' calls if the module appears loaded.
195 - increase signal latency test timer considerably for poor little osx,
196 which can't be bothered with basic posix functionality such as signal
197 delivery.
198 - mention some debugging aids in the tutorial.
199 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
200 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
201 of the context as title.
202 - AnyEvent::Strict would not restore read-write access to $_ when a callback
203 died - this is a user bug, but since AE::Strict exists to help dbeugging...
204 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
205 to work with it - the author is well aware of the issues but refuses
206 to fix them. This extreme step has been taken because the module actively
207 attacks the goals of AnyEvent and makes it harder for module authors to use
208 AnyEvent.
209
2106.02 Fri Aug 26 20:08:31 CEST 2011
211 - log_to_syslog now accepts facility strings.
212 - log_to_syslog would not allow facilities - "$facility|$level"
213 *literally* means that string. sheesh.
214 - the CPAN parser got confused and thought we provide Sys::Syslog, try
215 to work around.
216 - renamed AnyEvent::Impl::FLTK2 to FLTK.
217
2186.01 Fri Aug 26 07:04:11 CEST 2011
219 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
220 the TTL as fourth element - this affects ->resolve, ->request
221 and ->dns_unpack indirectly. Convenience functions such as
222 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
223 AnyEvent::Socket.
224 - INCOMPATIBLE CHANGE: backend authors now should not implement
225 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
226 methods.
227 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
228 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
229 - DEPRECIATION: please change
230 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
231 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
232 in your programs, for future compatibility.
233 - *blush* idle watcher emulation was borked.
234 - AnyEvent::Strict now checks the AE::xxx functions as well.
235 - querying too long domain names would barf deep inside AnyEvent::DNS - since
236 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
237 for too long domain names explicitly and passes an undef to the callback
238 instead (reported by James Bromberger).
239 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
240 - new functions: AnyEvent::postpone, AnyEvent::log.
241 - new module: AnyEvent::Log, for simple logging needs.
242 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
243 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
244 a debug shell.
245 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
246 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
247 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
248 /etc/hosts can be found.
249 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
250 configuration.
251 - env variables can now be specified using AE_xyz instead of
252 PERL_ANYEVENT_xyz, subject to some rules.
253 - lots of new AnyEvent::Debug functionality.
254 - AnyEvent::Strict now checks for modifications of $_ by the callback
255 (by setting it to readonly while executing the callback, if possible).
256 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
257 Invest extra overhead to work around this peculiar behaviour.
258 - split the pure perl backend into a loop module and an impl
259 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
260 - some read types (regex, netstring, json, storable) could
261 errornously keep an AE::Handle object alive on their own
262 (reported by Mohammad Toossi).
263 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
264 supports /etc/hosts overrides.
265 - add more workarounds around child watcher bugs in IO::Async
266 and (newly introduced) in POE.
267 - use glib child watchers and try to work around its limitations,
268 if possible (based on a patch by Kevin Ryde).
269 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
270 - many read types in AnyEvent::Handle would malfunction during
271 transitions to TLS because of a "delete $self->{rbuf}".
272 (reported by Gerald Galster).
273 - improve condvar blocking wait performance (probably only noticable
274 with faster backends), also make it easier for event loops that
275 do not support blocking waits.
276 - PERL_ANYEVENT_MODEL now also allows full module names, although
277 the usefulness of this feature is not yet clear.
278 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
279 - add t/80_ssltext.t.
280 - no longer use AUTOLOAD for the initial loading (for no very good reason).
281 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
282 - do not run event loop detection multiple times when programs call their cached
283 copies of AnyEvent::detect.
284 - parse_hostport now accepts absolute paths als unix domain sockets.
285 - suppress the idiotic warning inside POE by patching the POE::Kernel
286 run flags instead of asking the user to jump through contortions.
287 - add Task::Weaken as dependency (as recommends only at this time).
288 - replace PF_UNSPEC by 0, for good old minix.
289
2906.0 (not officially released, but was on CPAN for about 15 minutes :/)
291
2925.34 Thu May 12 09:59:41 CEST 2011
293 - AE::TLS didn't support both cert_file and key at the same time,
294 due to some copy&paste bug (reported and analyzed by vti).
295 - improve AE::Handle timeout documentation slightly.
296
2975.33 Mon Feb 7 21:51:41 CET 2011
298 - never appeared on CPAN, for unknown reasons.
299 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
300 object when it's already gone (analyzed by Ryan Bullock).
301 - update idna algorithm(s) to UTS#46 version 6.0.0.
302 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
303 uts46 says it's an error, and testsuite makes more sense.
304 - remove leading dots in idn_to_ascii - testsuite says to remove,
305 uts46 says it's an error, and testsuite makes more sense.
306 - fixed a bug in idn ignored character processing.
307 - passes the uts46 6.0.0 testsuite except for three cases which
308 apparently are bugs in the testsuite itself.
309 - passes the uts46 6.0.1 testsuite except for eleven cases which
310 apparently are bugs in the testsuite itself.
311
3125.31 Sun Jan 23 11:49:19 CET 2011
313 - work around a POE bug where POE would first poll for new events and
314 invoke handlers BEFORE handling the message queue.
315 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
316 - stop_read now does just that in TLS mode. see its
317 documentation for caveats.
318 - update to common::sense 3.4 warning set (some warnings in perl impose
319 an arbitrary 32kb limit on regex matches, which affects
320 AnyEvent::Handle).
321 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
322 - new experimental AE::Handle::resettls method.
323
3245.3 Fri Dec 31 05:48:13 CET 2010
325 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
326 issues with IO::Async have been solved in that version, including
327 the availabilty of a default loop!
328 - implement new wbuf_max parameter for AnyEvent::Handle.
329 - added Cocoa::EventLoop backend by Daisuke Murase.
330 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
331 - clarify handle on_prepare documentation.
332 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
333 to avoid leaking them to any child processes.
334 - try to force files into the arch-dependend lib directory, for the
335 half-"turly obsessive compulsive person"(s) that depend on it.
336
3375.29 Sun Dec 5 10:49:21 CET 2010
338 - convert EV backend to EV 4.00 API (so better upgrade EV too).
339 - AE::Handle: implement dynamic read_size adjustment, add
340 max_read_size parameter, reduce default read_size to 2048.
341 - add some FAQ entries for servers and/or on_eof vs. on_error.
342 - work around OS bugs (cygwin again) when lingering in AE::Handle.
343 - reduce memory usage slightly when connect was used in AE::Handle.
344 - wrok around more segfaults in Tk.
345 - document tls_ctx => undef as valid.
346 - detect WNOHANG value at compiletime.
347 - include some probably linux-only support for building AnyEvent
348 as part of the perl core.
349 - improve compatibility of t/08_idna.t to perl 5.8.
350 - make AnyEvent::Strict truly optional (in case somebody builds a
351 minimal perl...).
352
3535.28 Wed Oct 13 04:14:23 CEST 2010
354 - due to a glitch, AnyEvent's internal getprotobyname wasn't
355 actually used in tcp_server.
356 - implement AnyEvent::CondVar->new method, which was documented but
357 missing.
358 - support autoloading of read types in unshift_read.
359 - AnyEvent::Strict: do the test for invalid fh arguments differently,
360 resulting in more informative error messages.
361 - new AnyEvent::FAQ manpage.
362 - clarify that you can actually call push_read etc. in on_read.
363 - put keys and certs in /tmp, not into the current directory,
364 and document this in the SECURITY CONSIDERATIONS sections
365 of AnyEvent::TLS.
366 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
367 requested by Ron Isaacson.
368 - document $AnyEvent::DNS::RESOLVER variable.
369 - incorporated major typo patches by Abhijit Menon-Sen.
370 - AnyEvent::Handle now croaks when negative timeouts are passed.
371 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
372 undocumented, but who knows when they might come in handy.
373
3745.271 Tue Jun 8 12:05:46 CEST 2010
375 - backport to perl 5.8.x.
376
3775.27 Sun Jun 6 12:12:05 CEST 2010
378 - postpone differently in AnyEvent::Socket now, as
379 when not, canceling the connection attempt might fail
380 (found by Felix Antonius Wilhelm Ostmann).
381 - explicitly check for non-stream sockets in AE::Handle, too many
382 clueless people fell into the trap of this somehow working.
383 - simplified and reworked the "OTHER MODULES" section.
384 - better/more condvar examples.
385
3865.261 Wed Apr 28 16:13:36 CEST 2010
387 - AF_INET6 was not properly used from Socket6 during configuration
388 time (found by Andy Grundman).
389
3905.26 Mon Apr 12 04:49:35 CEST 2010
391 - don't generate spurious readyness notifications when select
392 returns EINTR in the pure perl backend (analysed and testcase
393 by Bas Denissen).
394 - give same examples in SYNOPSIS sections of AnyEvent and AE.
395 - provider faster implementations for AE::time/now/now_update for
396 EV and Perl backends.
397 - provide faster AE::cv and AE::time implementations for all
398 backends.
399 - no longer support lower/mixed-case signal names (this was
400 never documented nor universally supported).
401 - some more memory µ-optimisations, and somewhat less messy
402 code to cope with both AnyEvent and AE APIs.
403
4045.251 Sat Mar 13 00:58:21 CET 2010
405 - make test failed if it was run before make install on perl <=
406 5.8.8, and older versions of perl make our live very difficult,
407 so write constants.pl during Makefile.PL time.
408
4095.25 Sat Mar 13 00:23:14 CET 2010
410 - fix a race condition in AnyEvent::Handle that would cause
411 a "bio_read: unsupported method" error (in _another_
412 ssl connection) after user code threw an exception.
413 - added AnyEvent::Handle->destroyed method.
414 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
415 - the AnyEvent::Util::fh_nonblocking function worked only
416 by lucky accident on win32.
417 - smaller and faster AnyEvent::Util::fh_nonblocking.
418 - when the (required!) Time::HiRes module is missing, AnyEvent
419 did not fall back to built-in time properly.
420 - do not load Fcntl at runtime, saving memory and loading time.
421 - precompile a number of constants and use them instead of runtime
422 detection and eval.
423 - free detection code after detection and similar memory optimisations.
424 - Perl backend timer interval best effort drift has been improved
425 (same algorithm as EV).
426 - update unicode idna mapping table.
427
4285.24 Tue Jan 5 11:39:43 CET 2010
429 - cygwin never reports errors from failed connects. 1.5 just gives
430 you continous readyness and EAGAIN, 1.7 is even more broken and
431 just hangs. work around both issues in a cygwin-specific hack.
432 - improve idle watcher documentation slightly.
433
4345.23 Sun Dec 20 23:48:00 CET 2009
435 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
436 - implement punycode_encode/decode, idn_nameprep,
437 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
438 - provide $AE::VERSION.
439 - removed traces of "no strict 'refs'".
440
4415.22 Sat Dec 5 03:51:13 CET 2009
442 - downgrade-or-fail in AnyEvent::Handle::push_write, to
443 diagnose encoding failures earlier and more succinctly.
444 (this works around bugs in perl, throwing away encoding info
445 when passing scalar data to extensions).
446 - add more examples to AnyEvent::Socket manpage.
447 - upgrade internal warning set to the same as common::sense 2.03.
448 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
449 (John Beppu).
450 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
451 - add google public dns servers to fallback server set and make sure
452 we load-balance properly between the three sets. also add all
453 fallback dns servers, not just a random one, to each dns config.
454
4555.21 Thu Nov 19 02:48:47 CET 2009
456 - fix a problem where socket constants were called with parameters
457 (spotted by David Friedland).
458 - fork_call never use'd POSIX (reported by Daisuke Maki).
459 - improve perl 5.6 compatibility further (but it still won't work
460 unless you rip out everything but the core).
461 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
462 uses 10MB of RAM, but doesn't open a console window. *sigh*.
463
4645.202 Wed Oct 14 22:35:44 CEST 2009
465 - AnyEvent::DNS would unexpectedly clobber $_ under windows
466 (analysed by Matthias Waldorf).
467 - AnyEvent::Handle::run_cmd can now pass the PID of the
468 newly-created process, which is much less useful than it might
469 sound (based on patch by Yann Kerherve).
470
4715.201 Tue Sep 29 12:09:25 CEST 2009
472 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
473 (reported by Torsten Foertsch).
474 - common::sense 2.0 could cause tcp_server to throw an exception
475 (analysed by elmex).
476
4775.2 Mon Sep 14 07:04:49 CEST 2009
478 - INCOMPATIBLE CHANGE: do no longer support register_read_type
479 and register_write_type in AnyEvent::Handle, instead support
480 package names (the facility was mostly abused).
481 - implement "packagename-as-read/write type" support in
482 AnyEvent::Handle.
483 - AnyEvent::Handle: new options "keepalive" and "oobinline".
484 - oobinline set by default to avoid security issues.
485 - the pure-perl event loop backend wrongly detected times() fallback
486 support (spotted by Pavel Boldin).
487
4885.12 Tue Sep 1 20:26:50 CEST 2009
489 - be more lenient when parsing resolv.conf files, as some people
490 use hashmarks as comment indicator inside directives (reported
491 by Michael S. Fischer).
492 - use same set of warnings as common::sense 2.0.
493 - fix a potential 32 bit overflow issue due to perl having problems
494 with large hex constants in the Perl backend.
495 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
496 by default.
497 - fix some prototypes.
498 - (optionally) require Async::Interrupt 1.02.
499 - replace opendns fallbacks by verizon/level3 public dns servers.
500
5015.112 Fri Aug 21 13:59:04 CEST 2009
502 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
503 using starttls with an illegal/uncreatable tls context.
504 - port to broken win32 perls.
505
5065.111 Sun Aug 16 18:44:36 CEST 2009
507 - perl errornously requires sockaddr_un structures returned by the
508 kernel to have a specific length. work around this bug by padding
509 sockaddr_un structures (found to fail on OS X, as perl only uses
510 the correct code for GNU/Linux, but likely to fail on other
511 OSes as well).
512
5135.11 Wed Aug 12 17:49:37 CEST 2009
514 - fixed a bug in run_cmd with input from filename.
515 - create a json coder object when none is given in write direction
516 to reduce dependency on JSON.
517
5185.1 Tue Aug 11 03:17:32 CEST 2009
519 - tcp_connect now reports the correct error in $!, not an unrelated
520 one (was broken since 4.91).
521 - AE::Handle did not properly clear rtimeout/wtimeout watchers
522 when retrying a tcp_connect.
523 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
524 - fixed wrong documentation in AE::Socket.
525
5265.01 Mon Aug 10 03:16:32 CEST 2009
527 - last release broke EV child watchers :/.
528
5295.0 Sun Aug 9 17:32:09 CEST 2009
530 - officially introduce and document the AE API.
531 - lowering the timeout at runtime did not have immediate effect
532 in AnyEvent::Handle.
533 - AE::Handle now has separate and independent read and write timeout
534 settings.
535 - probe for Guard module when AnyEvent::Util is loaded, not at
536 runtime (Event doesn't like eval "use" from callbacks, and it avoids
537 exporting the probe function).
538 - Impl::IOAsync's io watchers did not use the filehandle emulation
539 code to work around it's watcher limits.
540 - work around IO::Async's problems with overloaded objects.
541 - try to disable wearnings inside IO::Async, it's just too buggy.
542 - do no longer use Event by default, only EV or Perl, Event is too buggy.
543 - hackishly provide fast AE:: interface when EV or Perl is the backend.
544 - inherit VERSION from AnyEvent::VERSION in most modules.
545
5464.91 Thu Aug 6 15:42:45 CEST 2009
547 - AE::Handle::starttls could get out of sync when the read buffer
548 already contains some TLS handshake.
549 - AE::Handle did not properly free the TLS session, causing bigger
550 memory leaks in Net::SSLeay than expected (reported by toaster).
551 - AE::Socket::tcp_connect will now artificially delay invoking
552 the callback to avoid returning after invoking the callback.
553 - convert many internal watcher uses to AE API.
554
5554.9 Sat Aug 1 11:07:01 CEST 2009
556 - Glib silently fails when registering a timeout with
557 a negative value, so avoid that.
558 - call condvar callback immediately when it is set
559 after the condition is already signalled.
560 - check rbuf_max condition only after trying to consume
561 data.
562 - normalise signal numbers to names when using the ae's
563 signal handling, but do not document this (yet).
564 - pure perl signal emulation did not properly set nonblocking
565 mode on the signal pipe on win32 (this is unlikely to have caused
566 any issues).
567 - new module: AnyEvent::Debug.
568 - AnyEvent::Strict now checks that a signal was specified
569 by name (not name or number).
570 - reduce memory footprint in typical cases by ~50kb by
571 only compiling signal code when necessary.
572 - add AnyEvent::Handle->rbuf_max.
573 - grab the AE namespace for future new API, implement stubs for
574 said future/alternative API.
575 - new function: AnyEvent::Socket::format_hostport.
576
5774.881 Tue Jul 28 12:51:53 CEST 2009
578 - work around a bug in local in pre-5.10 perls, causing
579 AnyEvent::Handle to recurse when it should not
580 (analysed by elmex).
581
5824.88 Tue Jul 28 04:04:37 CEST 2009
583 - re-bless the handle into a dummy package after calling
584 AnyEvent::Handle::destroy, so the user does not need to check
585 for errors after every push_write etc.
586 - do not attempt to run t/02_signals.t on obviously broken
587 platforms.
588
5894.87 Sun Jul 26 02:06:16 CEST 2009
590 - do not attempt to linger when there is no longer a valid fh.
591 - enforce tls mode to be either accept or connect, do not simply
592 segfault in Net::SSLeay.
593 - AnyEvent::Handle can now call tcp_connect itself (new parameters
594 connect, on_prepare, on_connect and on_connect_error). Updated
595 tutorial accordingly.
596 - add AnyEvent::Impl::Irssi backend.
597
5984.86 Mon Jul 20 23:52:29 CEST 2009
599 - since the verbose warning is not enough, explicitly document
600 that versions before 1.33 of Net::SSLeay are not secure.
601 - work around signal handling races in Event and (...) Event::Lib.
602 - try to align signal-race timer to full-second boundaries.
603 - work around Tk not liking negative timeouts.
604 - don't complain of different grades of environmental unfriendlyness
605 in IO::Async.
606
6074.85 Sat Jul 18 06:16:14 CEST 2009
608 - nail the signal race problem in perl once and for all
609 (see $AnyEvent::MAX_SIGNAL_LATENCY).
610 - take advantage of Async::Interrupt if it is available.
611 - load Time::HiRes and Guard modules on demand only.
612 - add optional/recommended modules section to AnyEvent
613 documentation.
614 - reduce memory usage considerably (and reduce startup penalty)
615 by not using "strict", "warnings" and "overload" modules.
616 - work around buggy windows/openbsd perls and provide EBADMSG
617 and EPROTO ourselves when missing.
618 - improve perl 5.6 compatibility of the core event loop.
619 - made Net::SSLeay version 1.33 a soft requirement.
620
6214.83 Fri Jul 17 16:56:26 CEST 2009
622 - implement AnyEvent::Socket::getprotobyname.
623 - AnyEvent::CondVar's will now detect recursive blocking
624 waits and will croak, as too many people fall into
625 this trap.
626 - AnyEvent::Handle will now call ->destroy on itself after
627 executing the on_error callback, instead of doing some
628 half-baked internal shutdown, for fatal errors.
629 - clarify on_eof behaviour w.r.t. the read queue and
630 on_read callbacks.
631 - ignore some possible spurious wake-ups in tcp_connect.
632
6334.82 Sat Jul 11 00:34:55 CEST 2009
634 - POE and Event backends didn't accept some callable objects as
635 callbacks.
636 - use Config module instead of POSIX module to detect signal names
637 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
638 module doesn't even have all POSIX signals :/.
639 - use more workarounds around the many refcnt/corruption bugs in
640 Event::Lib.
641 - work around a race condition in perl's select, causing t/03_child.t
642 to rarely fail.
643
6444.81 Thu Jul 9 10:30:30 CEST 2009
645 - AnyEvent::Handle didn't properly diagnose write errors
646 (it expected -1 from syswrite, how lame... :).
647 - support file descriptors in addition to file handles
648 in AnyEvent->io.
649 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
650 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
651 and PERL_ANYEVENT_CA_PATH.
652 - provide a sensible synopsis section for AnyEvent::TLS.
653 - add a "supported backends" section to the manpage.
654 - added simple io watcher test to testsuite, using a
655 portable_socketpair.
656 - tried to improve the stability of the Event::Lib backend,
657 YMMV.
658
6594.8 Mon Jul 6 23:45:16 CEST 2009
660 - AnyEvent::DNS did not properly follow CNAME records with
661 uppercase targets.
662 - AnyEvent::DNS would errornously return AAAA records
663 with v4 mapped addresses (a faulty record) as ipv4 addresses,
664 causing AnyEvent::Socket to throw an exception.
665 - added new module AnyEvent::TLS for easier SSL/TLS context
666 creation, with many options including hostname verification,
667 secure default configuration, lots of documentation and,
668 predefined diffie-hellman keys for perfect forward security
669 and much more. get it while it's still fresh!
670 - use AnyEvent::TLS in AnyEvent::Handle for context management.
671 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
672 so AnyEvent::Socket users have smaller memory footprint
673 in the common case.
674 - add AnyEvent::Handle->push_shutdown method.
675 - add an additional error message parameter to AnyEvent::Handle's
676 on_error callback (for TLS, $! is still available).
677 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
678 - make AnyEvent::Handle more robust against early connection
679 failures (during new), and return C<undef> in that case
680 from the constructor.
681 - AnyEvent::Handle will now try to load only JSON::XS first,
682 then fall back to JSON.
683 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
684 and because it was documented that way.
685
6864.451 Fri Jul 3 00:28:58 CEST 2009
687 - do not clear rbuf when shutting down an AnyEvent::Handle
688 object - doing so breaks AnyEvent::HTTP.
689
6904.45 Mon Jun 29 22:59:26 CEST 2009
691 - a write error could cause AnyEvent::Handle to create
692 an I/O watcher with an undefined $fh.
693 - special-case mapped ipv4 addresses in both
694 AnyEvent::Socket::format_address and parse_address,
695 to treat them just like ipv4 addresses.
696 - updated and overhauled the AnyEvent::Intro doc.
697 - implement AnyEvent::Socket::format_ipv4/ipv6.
698 - slightly speed up portable_pipe/socketpair.
699 - expand condvar begin/end documentation.
700
7014.42 Fri Jun 26 08:32:18 CEST 2009
702 - preliminary and neccesarily incomplete support for IO::Async.
703 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
704 case it was set to IGNORE, to ensure we can catch child statuses
705 even when the calling env acts stupidly.
706 - updated benchmarks with IO::Async, which performs very well.
707
7084.412 Wed Jun 24 01:35:57 CEST 2009
709 - support an "untaint" attribute for AnyEvent::DNS
710 and set it on the default resolver.
711 - implement and document AnyEvent::Impl::Perl::loop.
712 - remove all anyevent-env variables from %ENV
713 when running in tainted mode.
714 - mention and extend the IO::Lambda benchmark.
715
7164.411 Sun Jun 7 18:48:02 CEST 2009
717 - do not try to use F_SETFD on windows.
718
7194.41 Thu May 14 06:40:11 CEST 2009
720 - work around issues in older perls (5.8.0?) when
721 a signal handler is deleted from the %SIG hash.
722 - use POSIX::_exit in child test, to avoid running
723 destructors.
724 - speed up CHLD handling by relying on SIGCHLD being
725 synchronously delivered, even when we roll our own
726 implementation.
727 - AnyEvent::DNS: add the "dname" resource record name
728 for cosmetic reasons.
729
7304.4 Sun Apr 26 20:12:33 CEST 2009
731 - implemented idle watchers, where applicable.
732 - AnyEvent->time died when Event backend was in use.
733 - fix a memleak in the Tk backend.
734 - sped up Tk timer handling.
735 - clip negative "after" values to 0 in AnyEvent::Impl::Event
736 to avoid spamming out warning messages.
737 - fix Qt timers without interval.
738 - avoid Qt zero-timeout special case, in old libqt's.
739
7404.352 Mon Apr 20 16:31:11 CEST 2009
741 - fix AnyEvent::Strict error messages for child watchers.
742 - fix/update Makefile.PL "recommends" versions.
743 - add AnyEvent->now_update.
744
7454.351 Sat Apr 11 07:56:14 CEST 2009
746 - actually make the signal pipe work under win32.
747 - localise $! in signal handler to avoid changing $!, although
748 perl itself does not seem to save/restore errno either.
749 - set the cloexec flag on the signal pipe (normally set by perl too).
750
7514.35 Fri Mar 27 11:48:20 CET 2009
752 - event models relying on AnyEvent's signal watcher emulation did
753 invoke the callback asynchronously, contradicting documentation
754 and causing signals to get lost (this includes AnyEvent's own
755 event loop). AnyEvent now uses the standard pipe trick to make
756 callback execution synchronous to the event loop.
757 - AnyEvent::Handle didn't free TLS context data on DESTROY
758 (patch by Pavel Shaydo).
759 - work around the ever-incompatibly-changing API of MakeMaker.
760 - document that changing global variables without restoring them
761 is a bad idea in Perl (noted by Adam Rosenstein).
762 - AnyEvent::Strict now barfs if ->io is passed a file.
763
7644.34 Thu Feb 12 18:32:45 CET 2009
765 - separately buffer TLS read data, as otherwise the read queue could
766 deadlock as receiving data is not expected while draining the
767 read queue (which cna only happen with TLS).
768 - raise EBADMSG error on JSON decoding errors.
769 - fix some minor manpage bugs (reported by Maximilian Gaß).
770 - speed up select bitmask parsing quite a bit in the pure perl
771 backend.
772 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
773
7744.331 Tue Jan 6 21:07:25 CET 2009
775 - socketpair fails on many vista machines because vista has
776 completely broken accept/getpeername and getsockname functions,
777 so we provide our own socketpair emulation that kind of works
778 (AnyEvent::Util::portable_pipe).
779 - new function: AnyEvent::Util::portable_socketpair.
780 - take advantage of the Guard module if it exists.
781
7824.33 Fri Nov 21 02:35:40 CET 2008
783 - AnyEvent::Strict did errornously flag a fileno of 0 as
784 illegal.
785 - reduce memory usage and slightly speed up the pure perl backend
786 by only storing the file descriptor, not the file handle.
787 - add missing autocork method to AnyEvent::Handle (reported
788 by Adam Rosenstein).
789 - AnyEvent::DNS->resolve errornously documented an $rcode result
790 argument, but there isn't (spotted by Henrik Krohns).
791 - the naked truth about Tk - it's basically unmaintained.
792
7934.32 Mon Nov 3 22:46:32 CET 2008
794 - fix AnyEvent::Socket::resolve_sockaddr to properly support
795 unix sockets again and choose a proper default for
796 the socket type.
797 - fix call to resolve_sockaddr in tcp_connect.
798
7994.31 Thu Oct 30 04:41:48 CET 2008
800 - implemented AnyEvent::Handle->destroy method.
801 - hint about unexpected effects in TLS mode.
802 - speed up AnyEvent::Impl::Perl by using more arrays
803 instead of hashes without the slightest loss of readability :)
804 - work around a perl argument refcounting bug.
805 - with some perl patches applied, AnyEvent now seems leak-free.
806
8074.3 Fri Oct 3 09:18:43 CEST 2008
808 - AnyEvent will now install a no-op signal handler for SIGPIPE,
809 unless one has been installed already.
810 - warn about not loaidng AnyEvent::Impl::POE early enough
811 (patch by Adam Rosenstein).
812 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
813 - fix a bug in where in SSL connect mode, the
814 client would first wait for some data by the server and
815 otherwise hang. (reported and analysed in an absolutely
816 exemplary manner by Adam Rosenstein).
817 - fix a bug in where SSL EOF would not be treated
818 as stream EOF, putting the connection into a hung state.
819 (reported and analysed in an absolutely exemplary manner by Adam
820 Rosenstein).
821 - fix a potential segfault when the TLS context
822 would go missing in a read callback (e.g. due to stoptls) -
823 Net::SSLeay of course makes no type checking whatsoever.
824 - AnyEvent::Handle will not stop the read watcher in TLS mode.
825 - AnyEvent::Handle->stoptls will send a TLS close notify
826 instead of simply destroying the stream now.
827 - fix a bug where large blocks of data written in TLS mode
828 would not be sent unless triggered by receives.
829 - on_drain will now take the tls write buffer into account.
830 - SSL operations have been streamlined a bit and should be faster
831 now, more to come, though.
832 - added a FAQ section, will document nontrivial issues in there.
833
8344.234 Mon Sep 29 04:08:13 CEST 2008
3 - fix child watcher documentation: only child exits will be reported, 835 - fix child watcher documentation: only child exits will be reported,
4 no trace events. 836 no trace events.
837 - mention SIGPIPE in AnyEvent::Handle.
5 - perl backend will not use POSIX::times as monotonic clock 838 - perl backend will now use POSIX::times as monotonic clock
6 when available and the monotonic clock option is not. 839 when available and the monotonic clock option is not.
7 - run even without Time::HiRes in some cases now, but do not 840 - run even without Time::HiRes in some cases now, but do not
8 provide sub-second accuracy (all supported perl versions have 841 provide sub-second accuracy (all *supported* perl versions have
9 Time::HiRes). 842 Time::HiRes).
10 843
114.233 Fri Aug 22 01:48:05 CEST 2008 8444.233 Fri Aug 22 01:48:05 CEST 2008
12 - fix a bug in the testsuite, causing a hang. 845 - fix a bug in the testsuite, causing a hang.
13 - clarified lots of AnyEvent::Handle constructor arguments. 846 - clarified lots of AnyEvent::Handle constructor arguments.

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines