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

Comparing AnyEvent/Changes (file contents):
Revision 1.235 by root, Wed May 6 03:45:27 2009 UTC vs.
Revision 1.633 by root, Sun May 12 10:32:26 2019 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: more formats for parse_ipv6 (single ipv6 address without port, ...p80, ...#80)
12TODO: inet_aton maybe pack C4? no, add a comment why not
13TODO: socket prepare callbacks, can they easily signal failure as well?
14
15TODO: Олег Г <verdrehung@gmail.com>, empty A record response causes
16AnyEvent::resolve_sockaddr to eventually call pack_sockaddr with undef -
17broken dns server, but maybe improve the response.
18
19 - added some more logging configuration examples.
20
217.15 Tue Feb 26 03:07:42 CET 2019
22 - INCOMPATIBLE CHANGE: AnyEvent::Handle's tls_detect documentation gave
23 separate major and minor versions, while code passed only a single
24 value. This version follows the documentation and now passes separate
25 major and minor values.
26 - work around Net::SSLeay not having been ported to openssl 1.1, but many
27 distributions compiling it against openssl 1.1, which unfortunately
28 succeeds and results in a very broken module.
29 - AnyEvent::DNS::dns_unpack now stores the original DNS packet
30 in the __ member, to allow decoding of undecodable resouce records
31 containing compressed domian names.
32 - AnyEvent::Socket::parse_ipv6 would NOT, as advertised, accept ipv4
33 addresses. It now does and converts them to ipv4 mapped addresses.
34 - support CAA records, based on patch by Steve Atkins.
35 - add freenom and cloudflare nameservers as dns fallback.
36 - AnyEvent::Strict would not properly ward against io watchers
37 on files when the handle passed was a file descriptor.
38 - document "internal" variables used by the dns en-/decoder to allow
39 enterprising users to extend them in a semi-official way.
40
417.14 Sat Jun 24 01:44:19 CEST 2017
42 - fix a crash bug in AnyEvent::Handle with openssl 1.1.0 (patched
43 by Paul Howarth and apparently tracked down by Mike McCauley).
44 - AnyEvent::Handle->keepalive was documented (and defined) twice
45 (found by Matt Merhar).
46 - AnyEvent::Socket::tcp_bind/tcp_server would immediately unlink a unix
47 listening socket unless a guard is used. Change this so that no
48 cleanup will be performed unless a guard is used and document this more
49 clearly.
50 - make tcp_bind/tcp_server error messages more regular.
51 - work around latest perl upstream breakage - this time, perl5porters were
52 apparently bullied by a bogus CVE (CVE-2016-1238), and since distros
53 stupidly apply anything that has "security fix" stamped on it, it's likely
54 going to be a problem in practise (and working around it is trivial on
55 windows and unix).
56 - add TCP_FASTOPEN/MSG_FASTOPEN and MSG_NOSIGNAL/MSG_MORE/MSG_DONTWAIT
57 to constants.
58 - update warnings to common::sense 3.74 standards.
59
607.13 Sat Sep 17 04:31:49 CEST 2016
61 - Only call tlsext_host_name for non-empty common names (reported
62 by Maxime Soulé).
63 - log a (single) notice message if SNI is not supported.
64 - upgrade to UTS-46:9.0.0 draft and switch to non-transitional
65 behaviour, beating thunderbird, ie, edge, chrome and safari to it :)
66 (see also https://bugzilla.mozilla.org/show_bug.cgi?id=1218179)
67 - turns out the UTS-46 IDNA testcase failures were indeed bugs in the
68 testcases and the specification and not in the code - the post-9.0.0
69 unicode files have all known problems fixed, so finally the AnyEvent
70 IDNA implementation can pass the full IDNA testsuite - without needing
71 a single fix :)
72 - guarantee (and document) that condvar callbacks will be removed
73 on invocation - important to avoid circular references.
74
757.12 Wed Jan 27 19:12:26 CET 2016
76 - use common name as hostname for tls connects, if Net::SSLeay
77 supports SNI.
78 - fix documentation of tls_autostart read type in AnyEvent::Handle,
79 analyzed by Felix Ostmann.
80
817.11 Thu Jul 16 14:36:00 CEST 2015
82 - AnyEvent::Socket::parse_ipv6 could accept malformed ipv6
83 addresses (extra "::" at end and similar cases).
84 - add a more explicit warning to AnyEvent::Handle that it doesn't
85 work on files, people keep getting confused.
86 - new function AnyEvent::Socket::tcp_bind.
87 - new functions AnyEvent::fh_block and AnyEvent::fh_unblock.
88 - aligned ipv6 address formatting with RFC 5952 (by not shortening
89 a single :0: to ::).
90 - added stability canary support.
91
927.09 Sat May 2 16:38:53 CEST 2015
93 - AnyEvent::Debug called an internal function (AnyEvent::Log::ft)
94 that was renamed to AnyEvent::Log:format_time. uinder its old name
95 (analyzed by Michael Stovenour).
96 - update AnyEvent::DNS fallback resolver addresses:
97 seems google effectively killed most other free dns resolvers,
98 so remove them, but add cable and wireless (ecrc) since it was
99 stable for 20 years or so, official or not, and there should be
100 an alternative to google.
101 - perl5porters broke windows error codes in 5.20, and mapped
102 WSAEWOULDBLOCK on the (different) EWOULDBLOCK error code, and
103 WSAEINPROGRESS into the incompatible ERINPROGRESS code, probably
104 because they were so cool. They probably broke other error codes
105 for no reason, too, but I didn't care for checking, it's too
106 depressing. This version only works around the WSAEWOULDBLOCK
107 issue, because I don't have a nice way to work around the
108 WSAEINPROGRESS bug.
109
1107.08 Wed Dec 10 05:27:17 CET 2014
111 - work around a newly introduced bug in Socket 2.011 (an
112 errornous sun_length check) (analyzed by Maxime Soulé).
113 - AnyEvent::TLS didn't load (but refer to) AnyEvent::Socket
114 (analyzed by Ben Magistro).
115 - AnyEvent::Strict will now confess, not croak. This is in line with
116 it being a development/debugging tool.
117 - work around a number of libglib bugs (debug builds of libglib enforce
118 certain undocumented behaviour patterns such as not being able to
119 remove a child watch source after it has fired, which we will try
120 to emulate to avoid "criticals". what where they thinking...).
121 - mention json security issues in AnyEvent::Handle, now that Douglas
122 Crockford has foolishly and incompatibly changed JSON.
123 - changed default dns resolver "max_outstanding" value from 1 to 10,
124 the latter beinfg the intended value all along
125 (reported by Ilya Chesnokov).
126 - added new "AnyEvent::Impl::UV" interface module to the UV event lib
127 (written by Mike Lowell).
128
1297.07 Tue Dec 17 17:45:02 CET 2013
130 - the documentation for custom tls verify schemes was wrong. make it agree
131 with the code (reported by Maxime Soulé).
132 - added cbor read and write types to AnyEvent::Handle (using CBOR::XS).
133 - work around an API change in openssl that could cause wrong tls connection
134 aborts, likely on windows only (analyzed by sten).
135 - calling AnyEvent->now_update with AnyEvent::Impl::Perl caused an
136 endless loop (reported by Dietrich Rebmann).
137 - add tlsv1_1 and tlsv1_2 protocols to AnyEvent::TLS
138 (patch by Maxime Soulé).
139 - document AnyEvent::Impl::IOAsync::set_loop and
140 $AnyEvent::Impl::IOAsync::LOOP. Though only documented now, this
141 functionality has _always_ been available.
142 - force a toplevel domain name in t/81_hosts.t (analyzed by
143 David Jack Wange Olrik).
144 - document that AnyEvent::Log uses AnyEvent::IO.
145 - warn about AnyEvent::Filesys::Notify performance.
146 - praise the joys of AnyEvent::Fork::*.
147 - time for an =encoding directive.
148 - do no longer use JSON to create a default json coder, use
149 JSON::XS or JSON::PP directly.
150
1517.05 Wed Aug 21 10:38:08 CEST 2013
152 - uts46data.pl couldn't be found due to wrong naming of the file
153 (reported by Fulko Hew).
154 - handle lone \015's properly in AE::Handle's default line read
155 (reported by various people).
156 - untaint ip addresses found in /etc/hosts (patch by José Micó).
157 - the memleak fix in 7.03 caused resolving via /etc/hosts to always
158 fail on first use (reported and testcase by Andrew Whatson).
159 - expose AnyEvent::Log::format_time, and allow users to redefine it.
160 - expose AnyEvent::Log::default_format, and allow redefinition.
161 - expose AnyEvent::Log::fatal_exit, to allow redefinition.
162 - AnyEvent::Debug shell can now run coro shell commands, if available.
163 - t/63* tests were wrongly in MANIFEST.
164 - kernel.org's finger server went MIA, switch to freebsd.org and
165 icculus.org.
166 - clarify that IO::AIO and AnyEvent::AIO are needed for AnyEvent::IO
167 to function asynchronously (reported by Daniel Carosone).
168 - hard-disable $^W in most tests, it generates too much garbage output.
169 - use a (hopefully) more future-proof method to emulate common::sense.
170 - upgrade to UTS-46:6.2.0.
171 - switch to INSTLIB from INSTLIBDIR, as INSTLIBDIR was wrongly
172 documented. should not affect anything.
173
1747.04 Thu Nov 15 02:23:58 CET 2012
175 - debugging code left in.
176
1777.03 Thu Nov 15 02:16:11 CET 2012
178 - AnyEvent::Socket::inet_aton did not work when dns resolution was
179 used to find the addresses (analyzed and patch by Maxime Soulé).
180 - fix a memory leak in the /etc/hosts lookup code when hosts don't
181 resolve and are not in hosts (analyzed by sten).
182
1837.02 Tue Aug 14 04:27:58 CEST 2012
184 - AnyEvent::Util::run_cmd could block indefinitely (analyzed and test
185 program by Yorhel).
186 - verified that AnyEvent::Socket follows RFC5952.
187 - try to parse "ADDR#PORT" in addition to "ADDR PORT".
188
1897.01 Sun May 13 01:03:17 CEST 2012
190 - fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but
191 not available, instead of throwing an exception.
192 - use File::Spec to get the tmpdir in t/*, to avoid needless
193 failures on (most, not mine :) windows boxes.
194 - new handle read types: tls_detect and tls_autostart.
195
1967.0 Fri Apr 13 06:33:30 CEST 2012
197 - child watchers are broken in POE 1.352 (also many earlier
198 versions) and there seems to be no way to work around it, as POE
199 itself is inherently racy. Document this regression and add a delay
200 in t/68_poe_03_child.t for the time being.
201 - new module AnyEvent::IO, that is a frontend to either a pure-perl
202 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
203 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
204 - load /etc/hosts only when DNS has no answer.
205 - stat /etc/hosts on every access and reload it if it changed.
206 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
207 - fix a buggy croak in the dh parameter reading in AnyEvent::TLS.
208 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
209 a side effect, they now use true appending as opposed to libc
210 appending, and the file might not have been opened when the function
211 returns.
212 - the default logging level was not properly documented in a variety of
213 places, this has hopefully been rectified.
214 - updated uts46data.pl for unicode 6.1.0.
215 - made log messages generated by AnyEvent submodules not
216 include the package name anymore, as it will be logged
217 by default already.
218 - upgrade to the trick used by common-sense 3.5 to work
219 around extra warning torture/breakage under perl 5.15.x.
220 - log messages by anyevent are now ucfirst, are usually full
221 sentences and do no longer include the package name.
222 - the storable read type would sometimes throw an exception instead
223 of causing EBADMSG (analyzed by Maxime Soulé).
224 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
225 keysize (reported by Paul Howarth).
226 - documented Rocco fabricating statements about AnyEvent and me.
227
2286.14 Tue Jan 31 20:00:24 CET 2012
229 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
230 probably an editing mistake (reported by Darin McBride).
231
2326.13 Thu Jan 12 07:27:01 CET 2012
233 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
234 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
235 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
236 regex-matching inside the callback would change the parameter, despite
237 $1 being dynamically scoped per-block. this perl bug is now being
238 worked around (testcase by Cindy Wang).
239
2406.12 Mon Dec 12 13:21:10 CET 2011
241 - $! was clobbered by subsequent calls in tcp_connect, due to
242 the postpone not saving/restoring it (analyzed by Richard Garnish).
243
2446.11 Tue Nov 22 10:36:05 CET 2011
245 - Tk cannot create windows when tainted unless you set a title, so
246 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
247 - escape any nonprintable/nonascii characters when stringifying
248 backtraces.
249 - log the reason loading an interface module fails to load at level debug.
250
2516.1 Tue Oct 4 19:44:30 CEST 2011
252 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
253 and above), and some messages inside AnyEvent have been elevated
254 to higher log levels to print by default.
255 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
256 - work around yet another signal race bug, newly introduced by POE. My
257 how that module sucks.
258 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
259 support ipv6-only operation even with defaults (also use ::1 instead
260 of 127.0.0.1).
261 - give better guidelines on how to chose a logging level.
262 - AnyEvent::Log can now cap (limit) log message priority by context.
263 - AE::log will now save and restore $! and $@, so callers don't have to.
264 - AE::Strict now checks that registered fds are still valid at regular
265 intervals.
266 - skip frequently called 'require' calls if the module appears loaded.
267 - increase signal latency test timer considerably for poor little osx,
268 which can't be bothered with basic posix functionality such as signal
269 delivery.
270 - mention some debugging aids in the tutorial.
271 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
272 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
273 of the context as title.
274 - AnyEvent::Strict would not restore read-write access to $_ when a callback
275 died - this is a user bug, but since AE::Strict exists to help dbeugging...
276 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
277 to work with it - the author is well aware of the issues but refuses
278 to fix them. This extreme step has been taken because the module actively
279 attacks the goals of AnyEvent and makes it harder for module authors to use
280 AnyEvent.
281
2826.02 Fri Aug 26 20:08:31 CEST 2011
283 - log_to_syslog now accepts facility strings.
284 - log_to_syslog would not allow facilities - "$facility|$level"
285 *literally* means that string. sheesh.
286 - the CPAN parser got confused and thought we provide Sys::Syslog, try
287 to work around.
288 - renamed AnyEvent::Impl::FLTK2 to FLTK.
289
2906.01 Fri Aug 26 07:04:11 CEST 2011
291 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
292 the TTL as fourth element - this affects ->resolve, ->request
293 and ->dns_unpack indirectly. Convenience functions such as
294 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
295 AnyEvent::Socket.
296 - INCOMPATIBLE CHANGE: backend authors now should not implement
297 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
298 methods.
299 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
300 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
301 - DEPRECIATION: please change
302 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
303 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
304 in your programs, for future compatibility.
305 - *blush* idle watcher emulation was borked.
306 - AnyEvent::Strict now checks the AE::xxx functions as well.
307 - querying too long domain names would barf deep inside AnyEvent::DNS - since
308 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
309 for too long domain names explicitly and passes an undef to the callback
310 instead (reported by James Bromberger).
311 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
312 - new functions: AnyEvent::postpone, AnyEvent::log.
313 - new module: AnyEvent::Log, for simple logging needs.
314 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
315 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
316 a debug shell.
317 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
318 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
319 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
320 /etc/hosts can be found.
321 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
322 configuration.
323 - env variables can now be specified using AE_xyz instead of
324 PERL_ANYEVENT_xyz, subject to some rules.
325 - lots of new AnyEvent::Debug functionality.
326 - AnyEvent::Strict now checks for modifications of $_ by the callback
327 (by setting it to readonly while executing the callback, if possible).
328 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
329 Invest extra overhead to work around this peculiar behaviour.
330 - split the pure perl backend into a loop module and an impl
331 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
332 - some read types (regex, netstring, json, storable) could
333 errornously keep an AE::Handle object alive on their own
334 (reported by Mohammad Toossi).
335 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
336 supports /etc/hosts overrides.
337 - add more workarounds around child watcher bugs in IO::Async
338 and (newly introduced) in POE.
339 - use glib child watchers and try to work around its limitations,
340 if possible (based on a patch by Kevin Ryde).
341 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
342 - many read types in AnyEvent::Handle would malfunction during
343 transitions to TLS because of a "delete $self->{rbuf}".
344 (reported by Gerald Galster).
345 - improve condvar blocking wait performance (probably only noticable
346 with faster backends), also make it easier for event loops that
347 do not support blocking waits.
348 - PERL_ANYEVENT_MODEL now also allows full module names, although
349 the usefulness of this feature is not yet clear.
350 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
351 - add t/80_ssltext.t.
352 - no longer use AUTOLOAD for the initial loading (for no very good reason).
353 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
354 - do not run event loop detection multiple times when programs call their cached
355 copies of AnyEvent::detect.
356 - parse_hostport now accepts absolute paths als unix domain sockets.
357 - suppress the idiotic warning inside POE by patching the POE::Kernel
358 run flags instead of asking the user to jump through contortions.
359 - add Task::Weaken as dependency (as recommends only at this time).
360 - replace PF_UNSPEC by 0, for good old minix.
361
3626.0 (not officially released, but was on CPAN for about 15 minutes :/)
363
3645.34 Thu May 12 09:59:41 CEST 2011
365 - AE::TLS didn't support both cert_file and key at the same time,
366 due to some copy&paste bug (reported and analyzed by vti).
367 - improve AE::Handle timeout documentation slightly.
368
3695.33 Mon Feb 7 21:51:41 CET 2011
370 - never appeared on CPAN, for unknown reasons.
371 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
372 object when it's already gone (analyzed by Ryan Bullock).
373 - update idna algorithm(s) to UTS#46 version 6.0.0.
374 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
375 uts46 says it's an error, and testsuite makes more sense.
376 - remove leading dots in idn_to_ascii - testsuite says to remove,
377 uts46 says it's an error, and testsuite makes more sense.
378 - fixed a bug in idn ignored character processing.
379 - passes the uts46 6.0.0 testsuite except for three cases which
380 apparently are bugs in the testsuite itself.
381 - passes the uts46 6.0.1 testsuite except for eleven cases which
382 apparently are bugs in the testsuite itself.
383
3845.31 Sun Jan 23 11:49:19 CET 2011
385 - work around a POE bug where POE would first poll for new events and
386 invoke handlers BEFORE handling the message queue.
387 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
388 - stop_read now does just that in TLS mode. see its
389 documentation for caveats.
390 - update to common::sense 3.4 warning set (some warnings in perl impose
391 an arbitrary 32kb limit on regex matches, which affects
392 AnyEvent::Handle).
393 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
394 - new experimental AE::Handle::resettls method.
395
3965.3 Fri Dec 31 05:48:13 CET 2010
397 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
398 issues with IO::Async have been solved in that version, including
399 the availabilty of a default loop!
400 - implement new wbuf_max parameter for AnyEvent::Handle.
401 - added Cocoa::EventLoop backend by Daisuke Murase.
402 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
403 - clarify handle on_prepare documentation.
404 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
405 to avoid leaking them to any child processes.
406 - try to force files into the arch-dependend lib directory, for the
407 half-"turly obsessive compulsive person"(s) that depend on it.
408
4095.29 Sun Dec 5 10:49:21 CET 2010
410 - convert EV backend to EV 4.00 API (so better upgrade EV too).
411 - AE::Handle: implement dynamic read_size adjustment, add
412 max_read_size parameter, reduce default read_size to 2048.
413 - add some FAQ entries for servers and/or on_eof vs. on_error.
414 - work around OS bugs (cygwin again) when lingering in AE::Handle.
415 - reduce memory usage slightly when connect was used in AE::Handle.
416 - wrok around more segfaults in Tk.
417 - document tls_ctx => undef as valid.
418 - detect WNOHANG value at compiletime.
419 - include some probably linux-only support for building AnyEvent
420 as part of the perl core.
421 - improve compatibility of t/08_idna.t to perl 5.8.
422 - make AnyEvent::Strict truly optional (in case somebody builds a
423 minimal perl...).
424
4255.28 Wed Oct 13 04:14:23 CEST 2010
426 - due to a glitch, AnyEvent's internal getprotobyname wasn't
427 actually used in tcp_server.
428 - implement AnyEvent::CondVar->new method, which was documented but
429 missing.
430 - support autoloading of read types in unshift_read.
431 - AnyEvent::Strict: do the test for invalid fh arguments differently,
432 resulting in more informative error messages.
433 - new AnyEvent::FAQ manpage.
434 - clarify that you can actually call push_read etc. in on_read.
435 - put keys and certs in /tmp, not into the current directory,
436 and document this in the SECURITY CONSIDERATIONS sections
437 of AnyEvent::TLS.
438 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
439 requested by Ron Isaacson.
440 - document $AnyEvent::DNS::RESOLVER variable.
441 - incorporated major typo patches by Abhijit Menon-Sen.
442 - AnyEvent::Handle now croaks when negative timeouts are passed.
443 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
444 undocumented, but who knows when they might come in handy.
445
4465.271 Tue Jun 8 12:05:46 CEST 2010
447 - backport to perl 5.8.x.
448
4495.27 Sun Jun 6 12:12:05 CEST 2010
450 - postpone differently in AnyEvent::Socket now, as
451 when not, canceling the connection attempt might fail
452 (found by Felix Antonius Wilhelm Ostmann).
453 - explicitly check for non-stream sockets in AE::Handle, too many
454 clueless people fell into the trap of this somehow working.
455 - simplified and reworked the "OTHER MODULES" section.
456 - better/more condvar examples.
457
4585.261 Wed Apr 28 16:13:36 CEST 2010
459 - AF_INET6 was not properly used from Socket6 during configuration
460 time (found by Andy Grundman).
461
4625.26 Mon Apr 12 04:49:35 CEST 2010
463 - don't generate spurious readyness notifications when select
464 returns EINTR in the pure perl backend (analysed and testcase
465 by Bas Denissen).
466 - give same examples in SYNOPSIS sections of AnyEvent and AE.
467 - provider faster implementations for AE::time/now/now_update for
468 EV and Perl backends.
469 - provide faster AE::cv and AE::time implementations for all
470 backends.
471 - no longer support lower/mixed-case signal names (this was
472 never documented nor universally supported).
473 - some more memory µ-optimisations, and somewhat less messy
474 code to cope with both AnyEvent and AE APIs.
475
4765.251 Sat Mar 13 00:58:21 CET 2010
477 - make test failed if it was run before make install on perl <=
478 5.8.8, and older versions of perl make our live very difficult,
479 so write constants.pl during Makefile.PL time.
480
4815.25 Sat Mar 13 00:23:14 CET 2010
482 - fix a race condition in AnyEvent::Handle that would cause
483 a "bio_read: unsupported method" error (in _another_
484 ssl connection) after user code threw an exception.
485 - added AnyEvent::Handle->destroyed method.
486 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
487 - the AnyEvent::Util::fh_nonblocking function worked only
488 by lucky accident on win32.
489 - smaller and faster AnyEvent::Util::fh_nonblocking.
490 - when the (required!) Time::HiRes module is missing, AnyEvent
491 did not fall back to built-in time properly.
492 - do not load Fcntl at runtime, saving memory and loading time.
493 - precompile a number of constants and use them instead of runtime
494 detection and eval.
495 - free detection code after detection and similar memory optimisations.
496 - Perl backend timer interval best effort drift has been improved
497 (same algorithm as EV).
498 - update unicode idna mapping table.
499
5005.24 Tue Jan 5 11:39:43 CET 2010
501 - cygwin never reports errors from failed connects. 1.5 just gives
502 you continous readyness and EAGAIN, 1.7 is even more broken and
503 just hangs. work around both issues in a cygwin-specific hack.
504 - improve idle watcher documentation slightly.
505
5065.23 Sun Dec 20 23:48:00 CET 2009
507 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
508 - implement punycode_encode/decode, idn_nameprep,
509 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
510 - provide $AE::VERSION.
511 - removed traces of "no strict 'refs'".
512
5135.22 Sat Dec 5 03:51:13 CET 2009
514 - downgrade-or-fail in AnyEvent::Handle::push_write, to
515 diagnose encoding failures earlier and more succinctly.
516 (this works around bugs in perl, throwing away encoding info
517 when passing scalar data to extensions).
518 - add more examples to AnyEvent::Socket manpage.
519 - upgrade internal warning set to the same as common::sense 2.03.
520 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
521 (John Beppu).
522 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
523 - add google public dns servers to fallback server set and make sure
524 we load-balance properly between the three sets. also add all
525 fallback dns servers, not just a random one, to each dns config.
526
5275.21 Thu Nov 19 02:48:47 CET 2009
528 - fix a problem where socket constants were called with parameters
529 (spotted by David Friedland).
530 - fork_call never use'd POSIX (reported by Daisuke Maki).
531 - improve perl 5.6 compatibility further (but it still won't work
532 unless you rip out everything but the core).
533 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
534 uses 10MB of RAM, but doesn't open a console window. *sigh*.
535
5365.202 Wed Oct 14 22:35:44 CEST 2009
537 - AnyEvent::DNS would unexpectedly clobber $_ under windows
538 (analysed by Matthias Waldorf).
539 - AnyEvent::Handle::run_cmd can now pass the PID of the
540 newly-created process, which is much less useful than it might
541 sound (based on patch by Yann Kerherve).
542
5435.201 Tue Sep 29 12:09:25 CEST 2009
544 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
545 (reported by Torsten Foertsch).
546 - common::sense 2.0 could cause tcp_server to throw an exception
547 (analysed by elmex).
548
5495.2 Mon Sep 14 07:04:49 CEST 2009
550 - INCOMPATIBLE CHANGE: do no longer support register_read_type
551 and register_write_type in AnyEvent::Handle, instead support
552 package names (the facility was mostly abused).
553 - implement "packagename-as-read/write type" support in
554 AnyEvent::Handle.
555 - AnyEvent::Handle: new options "keepalive" and "oobinline".
556 - oobinline set by default to avoid security issues.
557 - the pure-perl event loop backend wrongly detected times() fallback
558 support (spotted by Pavel Boldin).
559
5605.12 Tue Sep 1 20:26:50 CEST 2009
561 - be more lenient when parsing resolv.conf files, as some people
562 use hashmarks as comment indicator inside directives (reported
563 by Michael S. Fischer).
564 - use same set of warnings as common::sense 2.0.
565 - fix a potential 32 bit overflow issue due to perl having problems
566 with large hex constants in the Perl backend.
567 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
568 by default.
569 - fix some prototypes.
570 - (optionally) require Async::Interrupt 1.02.
571 - replace opendns fallbacks by verizon/level3 public dns servers.
572
5735.112 Fri Aug 21 13:59:04 CEST 2009
574 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
575 using starttls with an illegal/uncreatable tls context.
576 - port to broken win32 perls.
577
5785.111 Sun Aug 16 18:44:36 CEST 2009
579 - perl errornously requires sockaddr_un structures returned by the
580 kernel to have a specific length. work around this bug by padding
581 sockaddr_un structures (found to fail on OS X, as perl only uses
582 the correct code for GNU/Linux, but likely to fail on other
583 OSes as well).
584
5855.11 Wed Aug 12 17:49:37 CEST 2009
586 - fixed a bug in run_cmd with input from filename.
587 - create a json coder object when none is given in write direction
588 to reduce dependency on JSON.
589
5905.1 Tue Aug 11 03:17:32 CEST 2009
591 - tcp_connect now reports the correct error in $!, not an unrelated
592 one (was broken since 4.91).
593 - AE::Handle did not properly clear rtimeout/wtimeout watchers
594 when retrying a tcp_connect.
595 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
596 - fixed wrong documentation in AE::Socket.
597
5985.01 Mon Aug 10 03:16:32 CEST 2009
599 - last release broke EV child watchers :/.
600
6015.0 Sun Aug 9 17:32:09 CEST 2009
602 - officially introduce and document the AE API.
603 - lowering the timeout at runtime did not have immediate effect
604 in AnyEvent::Handle.
605 - AE::Handle now has separate and independent read and write timeout
606 settings.
607 - probe for Guard module when AnyEvent::Util is loaded, not at
608 runtime (Event doesn't like eval "use" from callbacks, and it avoids
609 exporting the probe function).
610 - Impl::IOAsync's io watchers did not use the filehandle emulation
611 code to work around it's watcher limits.
612 - work around IO::Async's problems with overloaded objects.
613 - try to disable wearnings inside IO::Async, it's just too buggy.
614 - do no longer use Event by default, only EV or Perl, Event is too buggy.
615 - hackishly provide fast AE:: interface when EV or Perl is the backend.
616 - inherit VERSION from AnyEvent::VERSION in most modules.
617
6184.91 Thu Aug 6 15:42:45 CEST 2009
619 - AE::Handle::starttls could get out of sync when the read buffer
620 already contains some TLS handshake.
621 - AE::Handle did not properly free the TLS session, causing bigger
622 memory leaks in Net::SSLeay than expected (reported by toaster).
623 - AE::Socket::tcp_connect will now artificially delay invoking
624 the callback to avoid returning after invoking the callback.
625 - convert many internal watcher uses to AE API.
626
6274.9 Sat Aug 1 11:07:01 CEST 2009
628 - Glib silently fails when registering a timeout with
629 a negative value, so avoid that.
630 - call condvar callback immediately when it is set
631 after the condition is already signalled.
632 - check rbuf_max condition only after trying to consume
633 data.
634 - normalise signal numbers to names when using the ae's
635 signal handling, but do not document this (yet).
636 - pure perl signal emulation did not properly set nonblocking
637 mode on the signal pipe on win32 (this is unlikely to have caused
638 any issues).
639 - new module: AnyEvent::Debug.
640 - AnyEvent::Strict now checks that a signal was specified
641 by name (not name or number).
642 - reduce memory footprint in typical cases by ~50kb by
643 only compiling signal code when necessary.
644 - add AnyEvent::Handle->rbuf_max.
645 - grab the AE namespace for future new API, implement stubs for
646 said future/alternative API.
647 - new function: AnyEvent::Socket::format_hostport.
648
6494.881 Tue Jul 28 12:51:53 CEST 2009
650 - work around a bug in local in pre-5.10 perls, causing
651 AnyEvent::Handle to recurse when it should not
652 (analysed by elmex).
653
6544.88 Tue Jul 28 04:04:37 CEST 2009
655 - re-bless the handle into a dummy package after calling
656 AnyEvent::Handle::destroy, so the user does not need to check
657 for errors after every push_write etc.
658 - do not attempt to run t/02_signals.t on obviously broken
659 platforms.
660
6614.87 Sun Jul 26 02:06:16 CEST 2009
662 - do not attempt to linger when there is no longer a valid fh.
663 - enforce tls mode to be either accept or connect, do not simply
664 segfault in Net::SSLeay.
665 - AnyEvent::Handle can now call tcp_connect itself (new parameters
666 connect, on_prepare, on_connect and on_connect_error). Updated
667 tutorial accordingly.
668 - add AnyEvent::Impl::Irssi backend.
669
6704.86 Mon Jul 20 23:52:29 CEST 2009
671 - since the verbose warning is not enough, explicitly document
672 that versions before 1.33 of Net::SSLeay are not secure.
673 - work around signal handling races in Event and (...) Event::Lib.
674 - try to align signal-race timer to full-second boundaries.
675 - work around Tk not liking negative timeouts.
676 - don't complain of different grades of environmental unfriendlyness
677 in IO::Async.
678
6794.85 Sat Jul 18 06:16:14 CEST 2009
680 - nail the signal race problem in perl once and for all
681 (see $AnyEvent::MAX_SIGNAL_LATENCY).
682 - take advantage of Async::Interrupt if it is available.
683 - load Time::HiRes and Guard modules on demand only.
684 - add optional/recommended modules section to AnyEvent
685 documentation.
686 - reduce memory usage considerably (and reduce startup penalty)
687 by not using "strict", "warnings" and "overload" modules.
688 - work around buggy windows/openbsd perls and provide EBADMSG
689 and EPROTO ourselves when missing.
690 - improve perl 5.6 compatibility of the core event loop.
691 - made Net::SSLeay version 1.33 a soft requirement.
692
6934.83 Fri Jul 17 16:56:26 CEST 2009
694 - implement AnyEvent::Socket::getprotobyname.
695 - AnyEvent::CondVar's will now detect recursive blocking
696 waits and will croak, as too many people fall into
697 this trap.
698 - AnyEvent::Handle will now call ->destroy on itself after
699 executing the on_error callback, instead of doing some
700 half-baked internal shutdown, for fatal errors.
701 - clarify on_eof behaviour w.r.t. the read queue and
702 on_read callbacks.
703 - ignore some possible spurious wake-ups in tcp_connect.
704
7054.82 Sat Jul 11 00:34:55 CEST 2009
706 - POE and Event backends didn't accept some callable objects as
707 callbacks.
708 - use Config module instead of POSIX module to detect signal names
709 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
710 module doesn't even have all POSIX signals :/.
711 - use more workarounds around the many refcnt/corruption bugs in
712 Event::Lib.
713 - work around a race condition in perl's select, causing t/03_child.t
714 to rarely fail.
715
7164.81 Thu Jul 9 10:30:30 CEST 2009
717 - AnyEvent::Handle didn't properly diagnose write errors
718 (it expected -1 from syswrite, how lame... :).
719 - support file descriptors in addition to file handles
720 in AnyEvent->io.
721 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
722 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
723 and PERL_ANYEVENT_CA_PATH.
724 - provide a sensible synopsis section for AnyEvent::TLS.
725 - add a "supported backends" section to the manpage.
726 - added simple io watcher test to testsuite, using a
727 portable_socketpair.
728 - tried to improve the stability of the Event::Lib backend,
729 YMMV.
730
7314.8 Mon Jul 6 23:45:16 CEST 2009
732 - AnyEvent::DNS did not properly follow CNAME records with
733 uppercase targets.
734 - AnyEvent::DNS would errornously return AAAA records
735 with v4 mapped addresses (a faulty record) as ipv4 addresses,
736 causing AnyEvent::Socket to throw an exception.
737 - added new module AnyEvent::TLS for easier SSL/TLS context
738 creation, with many options including hostname verification,
739 secure default configuration, lots of documentation and,
740 predefined diffie-hellman keys for perfect forward security
741 and much more. get it while it's still fresh!
742 - use AnyEvent::TLS in AnyEvent::Handle for context management.
743 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
744 so AnyEvent::Socket users have smaller memory footprint
745 in the common case.
746 - add AnyEvent::Handle->push_shutdown method.
747 - add an additional error message parameter to AnyEvent::Handle's
748 on_error callback (for TLS, $! is still available).
749 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
750 - make AnyEvent::Handle more robust against early connection
751 failures (during new), and return C<undef> in that case
752 from the constructor.
753 - AnyEvent::Handle will now try to load only JSON::XS first,
754 then fall back to JSON.
755 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
756 and because it was documented that way.
757
7584.451 Fri Jul 3 00:28:58 CEST 2009
759 - do not clear rbuf when shutting down an AnyEvent::Handle
760 object - doing so breaks AnyEvent::HTTP.
761
7624.45 Mon Jun 29 22:59:26 CEST 2009
763 - a write error could cause AnyEvent::Handle to create
764 an I/O watcher with an undefined $fh.
765 - special-case mapped ipv4 addresses in both
766 AnyEvent::Socket::format_address and parse_address,
767 to treat them just like ipv4 addresses.
768 - updated and overhauled the AnyEvent::Intro doc.
769 - implement AnyEvent::Socket::format_ipv4/ipv6.
770 - slightly speed up portable_pipe/socketpair.
771 - expand condvar begin/end documentation.
772
7734.42 Fri Jun 26 08:32:18 CEST 2009
774 - preliminary and neccesarily incomplete support for IO::Async.
775 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
776 case it was set to IGNORE, to ensure we can catch child statuses
777 even when the calling env acts stupidly.
778 - updated benchmarks with IO::Async, which performs very well.
779
7804.412 Wed Jun 24 01:35:57 CEST 2009
781 - support an "untaint" attribute for AnyEvent::DNS
782 and set it on the default resolver.
783 - implement and document AnyEvent::Impl::Perl::loop.
784 - remove all anyevent-env variables from %ENV
785 when running in tainted mode.
786 - mention and extend the IO::Lambda benchmark.
787
7884.411 Sun Jun 7 18:48:02 CEST 2009
789 - do not try to use F_SETFD on windows.
790
7914.41 Thu May 14 06:40:11 CEST 2009
792 - work around issues in older perls (5.8.0?) when
793 a signal handler is deleted from the %SIG hash.
794 - use POSIX::_exit in child test, to avoid running
795 destructors.
796 - speed up CHLD handling by relying on SIGCHLD being
797 synchronously delivered, even when we roll our own
798 implementation.
3 - AnyEvent::DNS: add the "dname" resource record name. 799 - AnyEvent::DNS: add the "dname" resource record name
800 for cosmetic reasons.
4 801
54.4 Sun Apr 26 20:12:33 CEST 2009 8024.4 Sun Apr 26 20:12:33 CEST 2009
6 - implemented idle watchers, where applicable. 803 - implemented idle watchers, where applicable.
7 - AnyEvent->time died when Event backend was in use. 804 - AnyEvent->time died when Event backend was in use.
8 - fix a memleak in the Tk backend. 805 - fix a memleak in the Tk backend.
9 - sped up Tk timer handling. 806 - sped up Tk timer handling.
10 - clip negative "after" values to 0 in AnyEvent::Impl::Event 807 - clip negative "after" values to 0 in AnyEvent::Impl::Event
11 to avoid spamming out warning messages. 808 to avoid spamming out warning messages.
12 - fix Qt timers without interval. 809 - fix Qt timers without interval.
13 - avoid Qt zero-timeout specialcase, in old libqt's. 810 - avoid Qt zero-timeout special case, in old libqt's.
14 811
154.352 Mon Apr 20 16:31:11 CEST 2009 8124.352 Mon Apr 20 16:31:11 CEST 2009
16 - fix AnyEvent::Strict error messages for child watchers. 813 - fix AnyEvent::Strict error messages for child watchers.
17 - fix/update Makefile.PL "recommends" versions. 814 - fix/update Makefile.PL "recommends" versions.
18 - add AnyEvent->now_update. 815 - add AnyEvent->now_update.
… …
824.3 Fri Oct 3 09:18:43 CEST 2008 8794.3 Fri Oct 3 09:18:43 CEST 2008
83 - AnyEvent will now install a no-op signal handler for SIGPIPE, 880 - AnyEvent will now install a no-op signal handler for SIGPIPE,
84 unless one has been installed already. 881 unless one has been installed already.
85 - warn about not loaidng AnyEvent::Impl::POE early enough 882 - warn about not loaidng AnyEvent::Impl::POE early enough
86 (patch by Adam Rosenstein). 883 (patch by Adam Rosenstein).
87 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 884 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
88 - fix a bug in where in SSL connect mode, the 885 - fix a bug in where in SSL connect mode, the
89 client would first wait for some data by the server and 886 client would first wait for some data by the server and
90 otherwise hang. (reported and analyzed in an absolutely 887 otherwise hang. (reported and analysed in an absolutely
91 exemplary manner by Adam Rosenstein). 888 exemplary manner by Adam Rosenstein).
92 - fix a bug in where SSL EOF would not be treated 889 - fix a bug in where SSL EOF would not be treated
93 as stream EOF, putting the connection into a hung state. 890 as stream EOF, putting the connection into a hung state.
94 (reported and analyzed in an absolutely exemplary manner by Adam 891 (reported and analysed in an absolutely exemplary manner by Adam
95 Rosenstein). 892 Rosenstein).
96 - fix a potential segfault when the TLS context 893 - fix a potential segfault when the TLS context
97 would go missing in a read callback (e.g. due to stoptls) - 894 would go missing in a read callback (e.g. due to stoptls) -
98 Net::SSLeay of course makes no type checking whatsoever. 895 Net::SSLeay of course makes no type checking whatsoever.
99 - AnyEvent::Handle will not stop the read watcher in TLS mode. 896 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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