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

Comparing AnyEvent/Changes (file contents):
Revision 1.38 by root, Mon Apr 7 19:23:58 2008 UTC vs.
Revision 1.647 by root, Sat Jul 25 22:20:29 2020 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
33.0 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?
14TODO: Олег Г <verdrehung@gmail.com>, empty A record response causes
15AnyEvent::resolve_sockaddr to eventually call pack_sockaddr with undef -
16broken dns server, but maybe improve the response.
17TODO: AnyEvent::Loop should not use POSIIX when ti isn't used.
18
19TODO: AnyEvent::Log contexts should have a way to inject log messages received form elsewhere, using timestamp etc.
20TODO: add some cheta-sheet to log synopsis regarding AE_LOG=AnyEvent::MP=+
21TODO: tcp_connect option for AnyEvent::Handle?
22TODO: hosts files on windows probably contain \r, which is not stripped
23 - AnyEvent::Loop no longer loads the POSIX module if Time::HiRes
24 has monotonic clock support (suggested by Felipe Gasper).
25 - since Net::SSLeay (up to 1.88-1) is _still not ported to openssl 1.1
26 and many distributions therefore will have a completely broken
27 Net::SSLeay module for the forseeable future, AnyEvent::TLS
28 will now log a warning (that is not seen by default unless you
29 increase log verbosity).
30 - add tlsv1_3 disable switch to AnyEvent::TLS.
31 - add experimental "tls" method support for AnyEvent::TLS.
32 - tls methods were lowercase, but documented to be uppercase.
33 fix documentation.
34 - EV backend now uses EV::break when waiting for condvars,
35 making waiting for condvars more efficient. This might
36 "break" already-broken programs that recursively enter
37 the event loop even more.
38 - add a dependency on Coro 6.57 if the installed version is older,
39 to avoid an incompatibility.
40
417.17 Wed Sep 18 03:04:49 CEST 2019
42 - work around antique openssl version in RHEL 7 by formatting
43 dh parameters differently (reported by several people).
44 - add t/13_weaken.t.
45
467.16 Fri Jul 19 18:00:21 CEST 2019
47 - add ffdhe group dh parameters from RFC 7919, and use ffdhe3072
48 as new default, instead of schmorp1539.
49 - AnyEvent::Log did not reassess logging status of
50 AnyEvent::Log::logger's when contexts were changed with
51 ->attach/detach/slaves, causing them to not log even though a
52 recent attach should have caused them to log.
53 - added some more logging configuration examples.
54 - mention RFC 8482 in AnyEvent::DNS.
55
567.15 Tue Feb 26 03:07:42 CET 2019
57 - INCOMPATIBLE CHANGE: AnyEvent::Handle's tls_detect documentation gave
58 separate major and minor versions, while code passed only a single
59 value. This version follows the documentation and now passes separate
60 major and minor values.
61 - work around Net::SSLeay not having been ported to openssl 1.1, but many
62 distributions compiling it against openssl 1.1, which unfortunately
63 succeeds and results in a very broken module.
64 - AnyEvent::DNS::dns_unpack now stores the original DNS packet
65 in the __ member, to allow decoding of undecodable resouce records
66 containing compressed domian names.
67 - AnyEvent::Socket::parse_ipv6 would NOT, as advertised, accept ipv4
68 addresses. It now does and converts them to ipv4 mapped addresses.
69 - support CAA records, based on patch by Steve Atkins.
70 - add freenom and cloudflare nameservers as dns fallback.
71 - AnyEvent::Strict would not properly ward against io watchers
72 on files when the handle passed was a file descriptor.
73 - document "internal" variables used by the dns en-/decoder to allow
74 enterprising users to extend them in a semi-official way.
75
767.14 Sat Jun 24 01:44:19 CEST 2017
77 - fix a crash bug in AnyEvent::Handle with openssl 1.1.0 (patched
78 by Paul Howarth and apparently tracked down by Mike McCauley).
79 - AnyEvent::Handle->keepalive was documented (and defined) twice
80 (found by Matt Merhar).
81 - AnyEvent::Socket::tcp_bind/tcp_server would immediately unlink a unix
82 listening socket unless a guard is used. Change this so that no
83 cleanup will be performed unless a guard is used and document this more
84 clearly.
85 - make tcp_bind/tcp_server error messages more regular.
86 - work around latest perl upstream breakage - this time, perl5porters were
87 apparently bullied by a bogus CVE (CVE-2016-1238), and since distros
88 stupidly apply anything that has "security fix" stamped on it, it's likely
89 going to be a problem in practise (and working around it is trivial on
90 windows and unix).
91 - add TCP_FASTOPEN/MSG_FASTOPEN and MSG_NOSIGNAL/MSG_MORE/MSG_DONTWAIT
92 to constants.
93 - update warnings to common::sense 3.74 standards.
94
957.13 Sat Sep 17 04:31:49 CEST 2016
96 - Only call tlsext_host_name for non-empty common names (reported
97 by Maxime Soulé).
98 - log a (single) notice message if SNI is not supported.
99 - upgrade to UTS-46:9.0.0 draft and switch to non-transitional
100 behaviour, beating thunderbird, ie, edge, chrome and safari to it :)
101 (see also https://bugzilla.mozilla.org/show_bug.cgi?id=1218179)
102 - turns out the UTS-46 IDNA testcase failures were indeed bugs in the
103 testcases and the specification and not in the code - the post-9.0.0
104 unicode files have all known problems fixed, so finally the AnyEvent
105 IDNA implementation can pass the full IDNA testsuite - without needing
106 a single fix :)
107 - guarantee (and document) that condvar callbacks will be removed
108 on invocation - important to avoid circular references.
109
1107.12 Wed Jan 27 19:12:26 CET 2016
111 - use common name as hostname for tls connects, if Net::SSLeay
112 supports SNI.
113 - fix documentation of tls_autostart read type in AnyEvent::Handle,
114 analyzed by Felix Ostmann.
115
1167.11 Thu Jul 16 14:36:00 CEST 2015
117 - AnyEvent::Socket::parse_ipv6 could accept malformed ipv6
118 addresses (extra "::" at end and similar cases).
119 - add a more explicit warning to AnyEvent::Handle that it doesn't
120 work on files, people keep getting confused.
121 - new function AnyEvent::Socket::tcp_bind.
122 - new functions AnyEvent::fh_block and AnyEvent::fh_unblock.
123 - aligned ipv6 address formatting with RFC 5952 (by not shortening
124 a single :0: to ::).
125 - added stability canary support.
126
1277.09 Sat May 2 16:38:53 CEST 2015
128 - AnyEvent::Debug called an internal function (AnyEvent::Log::ft)
129 that was renamed to AnyEvent::Log:format_time. uinder its old name
130 (analyzed by Michael Stovenour).
131 - update AnyEvent::DNS fallback resolver addresses:
132 seems google effectively killed most other free dns resolvers,
133 so remove them, but add cable and wireless (ecrc) since it was
134 stable for 20 years or so, official or not, and there should be
135 an alternative to google.
136 - perl5porters broke windows error codes in 5.20, and mapped
137 WSAEWOULDBLOCK on the (different) EWOULDBLOCK error code, and
138 WSAEINPROGRESS into the incompatible ERINPROGRESS code, probably
139 because they were so cool. They probably broke other error codes
140 for no reason, too, but I didn't care for checking, it's too
141 depressing. This version only works around the WSAEWOULDBLOCK
142 issue, because I don't have a nice way to work around the
143 WSAEINPROGRESS bug.
144
1457.08 Wed Dec 10 05:27:17 CET 2014
146 - work around a newly introduced bug in Socket 2.011 (an
147 errornous sun_length check) (analyzed by Maxime Soulé).
148 - AnyEvent::TLS didn't load (but refer to) AnyEvent::Socket
149 (analyzed by Ben Magistro).
150 - AnyEvent::Strict will now confess, not croak. This is in line with
151 it being a development/debugging tool.
152 - work around a number of libglib bugs (debug builds of libglib enforce
153 certain undocumented behaviour patterns such as not being able to
154 remove a child watch source after it has fired, which we will try
155 to emulate to avoid "criticals". what where they thinking...).
156 - mention json security issues in AnyEvent::Handle, now that Douglas
157 Crockford has foolishly and incompatibly changed JSON.
158 - changed default dns resolver "max_outstanding" value from 1 to 10,
159 the latter beinfg the intended value all along
160 (reported by Ilya Chesnokov).
161 - added new "AnyEvent::Impl::UV" interface module to the UV event lib
162 (written by Mike Lowell).
163
1647.07 Tue Dec 17 17:45:02 CET 2013
165 - the documentation for custom tls verify schemes was wrong. make it agree
166 with the code (reported by Maxime Soulé).
167 - added cbor read and write types to AnyEvent::Handle (using CBOR::XS).
168 - work around an API change in openssl that could cause wrong tls connection
169 aborts, likely on windows only (analyzed by sten).
170 - calling AnyEvent->now_update with AnyEvent::Impl::Perl caused an
171 endless loop (reported by Dietrich Rebmann).
172 - add tlsv1_1 and tlsv1_2 protocols to AnyEvent::TLS
173 (patch by Maxime Soulé).
174 - document AnyEvent::Impl::IOAsync::set_loop and
175 $AnyEvent::Impl::IOAsync::LOOP. Though only documented now, this
176 functionality has _always_ been available.
177 - force a toplevel domain name in t/81_hosts.t (analyzed by
178 David Jack Wange Olrik).
179 - document that AnyEvent::Log uses AnyEvent::IO.
180 - warn about AnyEvent::Filesys::Notify performance.
181 - praise the joys of AnyEvent::Fork::*.
182 - time for an =encoding directive.
183 - do no longer use JSON to create a default json coder, use
184 JSON::XS or JSON::PP directly.
185
1867.05 Wed Aug 21 10:38:08 CEST 2013
187 - uts46data.pl couldn't be found due to wrong naming of the file
188 (reported by Fulko Hew).
189 - handle lone \015's properly in AE::Handle's default line read
190 (reported by various people).
191 - untaint ip addresses found in /etc/hosts (patch by José Micó).
192 - the memleak fix in 7.03 caused resolving via /etc/hosts to always
193 fail on first use (reported and testcase by Andrew Whatson).
194 - expose AnyEvent::Log::format_time, and allow users to redefine it.
195 - expose AnyEvent::Log::default_format, and allow redefinition.
196 - expose AnyEvent::Log::fatal_exit, to allow redefinition.
197 - AnyEvent::Debug shell can now run coro shell commands, if available.
198 - t/63* tests were wrongly in MANIFEST.
199 - kernel.org's finger server went MIA, switch to freebsd.org and
200 icculus.org.
201 - clarify that IO::AIO and AnyEvent::AIO are needed for AnyEvent::IO
202 to function asynchronously (reported by Daniel Carosone).
203 - hard-disable $^W in most tests, it generates too much garbage output.
204 - use a (hopefully) more future-proof method to emulate common::sense.
205 - upgrade to UTS-46:6.2.0.
206 - switch to INSTLIB from INSTLIBDIR, as INSTLIBDIR was wrongly
207 documented. should not affect anything.
208
2097.04 Thu Nov 15 02:23:58 CET 2012
210 - debugging code left in.
211
2127.03 Thu Nov 15 02:16:11 CET 2012
213 - AnyEvent::Socket::inet_aton did not work when dns resolution was
214 used to find the addresses (analyzed and patch by Maxime Soulé).
215 - fix a memory leak in the /etc/hosts lookup code when hosts don't
216 resolve and are not in hosts (analyzed by sten).
217
2187.02 Tue Aug 14 04:27:58 CEST 2012
219 - AnyEvent::Util::run_cmd could block indefinitely (analyzed and test
220 program by Yorhel).
221 - verified that AnyEvent::Socket follows RFC5952.
222 - try to parse "ADDR#PORT" in addition to "ADDR PORT".
223
2247.01 Sun May 13 01:03:17 CEST 2012
225 - fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but
226 not available, instead of throwing an exception.
227 - use File::Spec to get the tmpdir in t/*, to avoid needless
228 failures on (most, not mine :) windows boxes.
229 - new handle read types: tls_detect and tls_autostart.
230
2317.0 Fri Apr 13 06:33:30 CEST 2012
232 - child watchers are broken in POE 1.352 (also many earlier
233 versions) and there seems to be no way to work around it, as POE
234 itself is inherently racy. Document this regression and add a delay
235 in t/68_poe_03_child.t for the time being.
236 - new module AnyEvent::IO, that is a frontend to either a pure-perl
237 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
238 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
239 - load /etc/hosts only when DNS has no answer.
240 - stat /etc/hosts on every access and reload it if it changed.
241 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
242 - fix a buggy croak in the dh parameter reading in AnyEvent::TLS.
243 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
244 a side effect, they now use true appending as opposed to libc
245 appending, and the file might not have been opened when the function
246 returns.
247 - the default logging level was not properly documented in a variety of
248 places, this has hopefully been rectified.
249 - updated uts46data.pl for unicode 6.1.0.
250 - made log messages generated by AnyEvent submodules not
251 include the package name anymore, as it will be logged
252 by default already.
253 - upgrade to the trick used by common-sense 3.5 to work
254 around extra warning torture/breakage under perl 5.15.x.
255 - log messages by anyevent are now ucfirst, are usually full
256 sentences and do no longer include the package name.
257 - the storable read type would sometimes throw an exception instead
258 of causing EBADMSG (analyzed by Maxime Soulé).
259 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
260 keysize (reported by Paul Howarth).
261 - documented Rocco fabricating statements about AnyEvent and me.
262
2636.14 Tue Jan 31 20:00:24 CET 2012
264 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
265 probably an editing mistake (reported by Darin McBride).
266
2676.13 Thu Jan 12 07:27:01 CET 2012
268 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
269 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
270 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
271 regex-matching inside the callback would change the parameter, despite
272 $1 being dynamically scoped per-block. this perl bug is now being
273 worked around (testcase by Cindy Wang).
274
2756.12 Mon Dec 12 13:21:10 CET 2011
276 - $! was clobbered by subsequent calls in tcp_connect, due to
277 the postpone not saving/restoring it (analyzed by Richard Garnish).
278
2796.11 Tue Nov 22 10:36:05 CET 2011
280 - Tk cannot create windows when tainted unless you set a title, so
281 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
282 - escape any nonprintable/nonascii characters when stringifying
283 backtraces.
284 - log the reason loading an interface module fails to load at level debug.
285
2866.1 Tue Oct 4 19:44:30 CEST 2011
287 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
288 and above), and some messages inside AnyEvent have been elevated
289 to higher log levels to print by default.
290 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
291 - work around yet another signal race bug, newly introduced by POE. My
292 how that module sucks.
293 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
294 support ipv6-only operation even with defaults (also use ::1 instead
295 of 127.0.0.1).
296 - give better guidelines on how to chose a logging level.
297 - AnyEvent::Log can now cap (limit) log message priority by context.
298 - AE::log will now save and restore $! and $@, so callers don't have to.
299 - AE::Strict now checks that registered fds are still valid at regular
300 intervals.
301 - skip frequently called 'require' calls if the module appears loaded.
302 - increase signal latency test timer considerably for poor little osx,
303 which can't be bothered with basic posix functionality such as signal
304 delivery.
305 - mention some debugging aids in the tutorial.
306 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
307 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
308 of the context as title.
309 - AnyEvent::Strict would not restore read-write access to $_ when a callback
310 died - this is a user bug, but since AE::Strict exists to help dbeugging...
311 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
312 to work with it - the author is well aware of the issues but refuses
313 to fix them. This extreme step has been taken because the module actively
314 attacks the goals of AnyEvent and makes it harder for module authors to use
315 AnyEvent.
316
3176.02 Fri Aug 26 20:08:31 CEST 2011
318 - log_to_syslog now accepts facility strings.
319 - log_to_syslog would not allow facilities - "$facility|$level"
320 *literally* means that string. sheesh.
321 - the CPAN parser got confused and thought we provide Sys::Syslog, try
322 to work around.
323 - renamed AnyEvent::Impl::FLTK2 to FLTK.
324
3256.01 Fri Aug 26 07:04:11 CEST 2011
326 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
327 the TTL as fourth element - this affects ->resolve, ->request
328 and ->dns_unpack indirectly. Convenience functions such as
329 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
330 AnyEvent::Socket.
331 - INCOMPATIBLE CHANGE: backend authors now should not implement
332 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
333 methods.
334 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
335 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
336 - DEPRECIATION: please change
337 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
338 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
339 in your programs, for future compatibility.
340 - *blush* idle watcher emulation was borked.
341 - AnyEvent::Strict now checks the AE::xxx functions as well.
342 - querying too long domain names would barf deep inside AnyEvent::DNS - since
343 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
344 for too long domain names explicitly and passes an undef to the callback
345 instead (reported by James Bromberger).
346 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
347 - new functions: AnyEvent::postpone, AnyEvent::log.
348 - new module: AnyEvent::Log, for simple logging needs.
349 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
350 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
351 a debug shell.
352 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
353 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
354 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
355 /etc/hosts can be found.
356 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
357 configuration.
358 - env variables can now be specified using AE_xyz instead of
359 PERL_ANYEVENT_xyz, subject to some rules.
360 - lots of new AnyEvent::Debug functionality.
361 - AnyEvent::Strict now checks for modifications of $_ by the callback
362 (by setting it to readonly while executing the callback, if possible).
363 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
364 Invest extra overhead to work around this peculiar behaviour.
365 - split the pure perl backend into a loop module and an impl
366 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
367 - some read types (regex, netstring, json, storable) could
368 errornously keep an AE::Handle object alive on their own
369 (reported by Mohammad Toossi).
370 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
371 supports /etc/hosts overrides.
372 - add more workarounds around child watcher bugs in IO::Async
373 and (newly introduced) in POE.
374 - use glib child watchers and try to work around its limitations,
375 if possible (based on a patch by Kevin Ryde).
376 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
377 - many read types in AnyEvent::Handle would malfunction during
378 transitions to TLS because of a "delete $self->{rbuf}".
379 (reported by Gerald Galster).
380 - improve condvar blocking wait performance (probably only noticable
381 with faster backends), also make it easier for event loops that
382 do not support blocking waits.
383 - PERL_ANYEVENT_MODEL now also allows full module names, although
384 the usefulness of this feature is not yet clear.
385 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
386 - add t/80_ssltext.t.
387 - no longer use AUTOLOAD for the initial loading (for no very good reason).
388 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
389 - do not run event loop detection multiple times when programs call their cached
390 copies of AnyEvent::detect.
391 - parse_hostport now accepts absolute paths als unix domain sockets.
392 - suppress the idiotic warning inside POE by patching the POE::Kernel
393 run flags instead of asking the user to jump through contortions.
394 - add Task::Weaken as dependency (as recommends only at this time).
395 - replace PF_UNSPEC by 0, for good old minix.
396
3976.0 (not officially released, but was on CPAN for about 15 minutes :/)
398
3995.34 Thu May 12 09:59:41 CEST 2011
400 - AE::TLS didn't support both cert_file and key at the same time,
401 due to some copy&paste bug (reported and analyzed by vti).
402 - improve AE::Handle timeout documentation slightly.
403
4045.33 Mon Feb 7 21:51:41 CET 2011
405 - never appeared on CPAN, for unknown reasons.
406 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
407 object when it's already gone (analyzed by Ryan Bullock).
408 - update idna algorithm(s) to UTS#46 version 6.0.0.
409 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
410 uts46 says it's an error, and testsuite makes more sense.
411 - remove leading dots in idn_to_ascii - testsuite says to remove,
412 uts46 says it's an error, and testsuite makes more sense.
413 - fixed a bug in idn ignored character processing.
414 - passes the uts46 6.0.0 testsuite except for three cases which
415 apparently are bugs in the testsuite itself.
416 - passes the uts46 6.0.1 testsuite except for eleven cases which
417 apparently are bugs in the testsuite itself.
418
4195.31 Sun Jan 23 11:49:19 CET 2011
420 - work around a POE bug where POE would first poll for new events and
421 invoke handlers BEFORE handling the message queue.
422 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
423 - stop_read now does just that in TLS mode. see its
424 documentation for caveats.
425 - update to common::sense 3.4 warning set (some warnings in perl impose
426 an arbitrary 32kb limit on regex matches, which affects
427 AnyEvent::Handle).
428 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
429 - new experimental AE::Handle::resettls method.
430
4315.3 Fri Dec 31 05:48:13 CET 2010
432 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
433 issues with IO::Async have been solved in that version, including
434 the availabilty of a default loop!
435 - implement new wbuf_max parameter for AnyEvent::Handle.
436 - added Cocoa::EventLoop backend by Daisuke Murase.
437 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
438 - clarify handle on_prepare documentation.
439 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
440 to avoid leaking them to any child processes.
441 - try to force files into the arch-dependend lib directory, for the
442 half-"turly obsessive compulsive person"(s) that depend on it.
443
4445.29 Sun Dec 5 10:49:21 CET 2010
445 - convert EV backend to EV 4.00 API (so better upgrade EV too).
446 - AE::Handle: implement dynamic read_size adjustment, add
447 max_read_size parameter, reduce default read_size to 2048.
448 - add some FAQ entries for servers and/or on_eof vs. on_error.
449 - work around OS bugs (cygwin again) when lingering in AE::Handle.
450 - reduce memory usage slightly when connect was used in AE::Handle.
451 - wrok around more segfaults in Tk.
452 - document tls_ctx => undef as valid.
453 - detect WNOHANG value at compiletime.
454 - include some probably linux-only support for building AnyEvent
455 as part of the perl core.
456 - improve compatibility of t/08_idna.t to perl 5.8.
457 - make AnyEvent::Strict truly optional (in case somebody builds a
458 minimal perl...).
459
4605.28 Wed Oct 13 04:14:23 CEST 2010
461 - due to a glitch, AnyEvent's internal getprotobyname wasn't
462 actually used in tcp_server.
463 - implement AnyEvent::CondVar->new method, which was documented but
464 missing.
465 - support autoloading of read types in unshift_read.
466 - AnyEvent::Strict: do the test for invalid fh arguments differently,
467 resulting in more informative error messages.
468 - new AnyEvent::FAQ manpage.
469 - clarify that you can actually call push_read etc. in on_read.
470 - put keys and certs in /tmp, not into the current directory,
471 and document this in the SECURITY CONSIDERATIONS sections
472 of AnyEvent::TLS.
473 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
474 requested by Ron Isaacson.
475 - document $AnyEvent::DNS::RESOLVER variable.
476 - incorporated major typo patches by Abhijit Menon-Sen.
477 - AnyEvent::Handle now croaks when negative timeouts are passed.
478 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
479 undocumented, but who knows when they might come in handy.
480
4815.271 Tue Jun 8 12:05:46 CEST 2010
482 - backport to perl 5.8.x.
483
4845.27 Sun Jun 6 12:12:05 CEST 2010
485 - postpone differently in AnyEvent::Socket now, as
486 when not, canceling the connection attempt might fail
487 (found by Felix Antonius Wilhelm Ostmann).
488 - explicitly check for non-stream sockets in AE::Handle, too many
489 clueless people fell into the trap of this somehow working.
490 - simplified and reworked the "OTHER MODULES" section.
491 - better/more condvar examples.
492
4935.261 Wed Apr 28 16:13:36 CEST 2010
494 - AF_INET6 was not properly used from Socket6 during configuration
495 time (found by Andy Grundman).
496
4975.26 Mon Apr 12 04:49:35 CEST 2010
498 - don't generate spurious readyness notifications when select
499 returns EINTR in the pure perl backend (analysed and testcase
500 by Bas Denissen).
501 - give same examples in SYNOPSIS sections of AnyEvent and AE.
502 - provider faster implementations for AE::time/now/now_update for
503 EV and Perl backends.
504 - provide faster AE::cv and AE::time implementations for all
505 backends.
506 - no longer support lower/mixed-case signal names (this was
507 never documented nor universally supported).
508 - some more memory µ-optimisations, and somewhat less messy
509 code to cope with both AnyEvent and AE APIs.
510
5115.251 Sat Mar 13 00:58:21 CET 2010
512 - make test failed if it was run before make install on perl <=
513 5.8.8, and older versions of perl make our live very difficult,
514 so write constants.pl during Makefile.PL time.
515
5165.25 Sat Mar 13 00:23:14 CET 2010
517 - fix a race condition in AnyEvent::Handle that would cause
518 a "bio_read: unsupported method" error (in _another_
519 ssl connection) after user code threw an exception.
520 - added AnyEvent::Handle->destroyed method.
521 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
522 - the AnyEvent::Util::fh_nonblocking function worked only
523 by lucky accident on win32.
524 - smaller and faster AnyEvent::Util::fh_nonblocking.
525 - when the (required!) Time::HiRes module is missing, AnyEvent
526 did not fall back to built-in time properly.
527 - do not load Fcntl at runtime, saving memory and loading time.
528 - precompile a number of constants and use them instead of runtime
529 detection and eval.
530 - free detection code after detection and similar memory optimisations.
531 - Perl backend timer interval best effort drift has been improved
532 (same algorithm as EV).
533 - update unicode idna mapping table.
534
5355.24 Tue Jan 5 11:39:43 CET 2010
536 - cygwin never reports errors from failed connects. 1.5 just gives
537 you continous readyness and EAGAIN, 1.7 is even more broken and
538 just hangs. work around both issues in a cygwin-specific hack.
539 - improve idle watcher documentation slightly.
540
5415.23 Sun Dec 20 23:48:00 CET 2009
542 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
543 - implement punycode_encode/decode, idn_nameprep,
544 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
545 - provide $AE::VERSION.
546 - removed traces of "no strict 'refs'".
547
5485.22 Sat Dec 5 03:51:13 CET 2009
549 - downgrade-or-fail in AnyEvent::Handle::push_write, to
550 diagnose encoding failures earlier and more succinctly.
551 (this works around bugs in perl, throwing away encoding info
552 when passing scalar data to extensions).
553 - add more examples to AnyEvent::Socket manpage.
554 - upgrade internal warning set to the same as common::sense 2.03.
555 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
556 (John Beppu).
557 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
558 - add google public dns servers to fallback server set and make sure
559 we load-balance properly between the three sets. also add all
560 fallback dns servers, not just a random one, to each dns config.
561
5625.21 Thu Nov 19 02:48:47 CET 2009
563 - fix a problem where socket constants were called with parameters
564 (spotted by David Friedland).
565 - fork_call never use'd POSIX (reported by Daisuke Maki).
566 - improve perl 5.6 compatibility further (but it still won't work
567 unless you rip out everything but the core).
568 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
569 uses 10MB of RAM, but doesn't open a console window. *sigh*.
570
5715.202 Wed Oct 14 22:35:44 CEST 2009
572 - AnyEvent::DNS would unexpectedly clobber $_ under windows
573 (analysed by Matthias Waldorf).
574 - AnyEvent::Handle::run_cmd can now pass the PID of the
575 newly-created process, which is much less useful than it might
576 sound (based on patch by Yann Kerherve).
577
5785.201 Tue Sep 29 12:09:25 CEST 2009
579 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
580 (reported by Torsten Foertsch).
581 - common::sense 2.0 could cause tcp_server to throw an exception
582 (analysed by elmex).
583
5845.2 Mon Sep 14 07:04:49 CEST 2009
585 - INCOMPATIBLE CHANGE: do no longer support register_read_type
586 and register_write_type in AnyEvent::Handle, instead support
587 package names (the facility was mostly abused).
588 - implement "packagename-as-read/write type" support in
589 AnyEvent::Handle.
590 - AnyEvent::Handle: new options "keepalive" and "oobinline".
591 - oobinline set by default to avoid security issues.
592 - the pure-perl event loop backend wrongly detected times() fallback
593 support (spotted by Pavel Boldin).
594
5955.12 Tue Sep 1 20:26:50 CEST 2009
596 - be more lenient when parsing resolv.conf files, as some people
597 use hashmarks as comment indicator inside directives (reported
598 by Michael S. Fischer).
599 - use same set of warnings as common::sense 2.0.
600 - fix a potential 32 bit overflow issue due to perl having problems
601 with large hex constants in the Perl backend.
602 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
603 by default.
604 - fix some prototypes.
605 - (optionally) require Async::Interrupt 1.02.
606 - replace opendns fallbacks by verizon/level3 public dns servers.
607
6085.112 Fri Aug 21 13:59:04 CEST 2009
609 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
610 using starttls with an illegal/uncreatable tls context.
611 - port to broken win32 perls.
612
6135.111 Sun Aug 16 18:44:36 CEST 2009
614 - perl errornously requires sockaddr_un structures returned by the
615 kernel to have a specific length. work around this bug by padding
616 sockaddr_un structures (found to fail on OS X, as perl only uses
617 the correct code for GNU/Linux, but likely to fail on other
618 OSes as well).
619
6205.11 Wed Aug 12 17:49:37 CEST 2009
621 - fixed a bug in run_cmd with input from filename.
622 - create a json coder object when none is given in write direction
623 to reduce dependency on JSON.
624
6255.1 Tue Aug 11 03:17:32 CEST 2009
626 - tcp_connect now reports the correct error in $!, not an unrelated
627 one (was broken since 4.91).
628 - AE::Handle did not properly clear rtimeout/wtimeout watchers
629 when retrying a tcp_connect.
630 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
631 - fixed wrong documentation in AE::Socket.
632
6335.01 Mon Aug 10 03:16:32 CEST 2009
634 - last release broke EV child watchers :/.
635
6365.0 Sun Aug 9 17:32:09 CEST 2009
637 - officially introduce and document the AE API.
638 - lowering the timeout at runtime did not have immediate effect
639 in AnyEvent::Handle.
640 - AE::Handle now has separate and independent read and write timeout
641 settings.
642 - probe for Guard module when AnyEvent::Util is loaded, not at
643 runtime (Event doesn't like eval "use" from callbacks, and it avoids
644 exporting the probe function).
645 - Impl::IOAsync's io watchers did not use the filehandle emulation
646 code to work around it's watcher limits.
647 - work around IO::Async's problems with overloaded objects.
648 - try to disable wearnings inside IO::Async, it's just too buggy.
649 - do no longer use Event by default, only EV or Perl, Event is too buggy.
650 - hackishly provide fast AE:: interface when EV or Perl is the backend.
651 - inherit VERSION from AnyEvent::VERSION in most modules.
652
6534.91 Thu Aug 6 15:42:45 CEST 2009
654 - AE::Handle::starttls could get out of sync when the read buffer
655 already contains some TLS handshake.
656 - AE::Handle did not properly free the TLS session, causing bigger
657 memory leaks in Net::SSLeay than expected (reported by toaster).
658 - AE::Socket::tcp_connect will now artificially delay invoking
659 the callback to avoid returning after invoking the callback.
660 - convert many internal watcher uses to AE API.
661
6624.9 Sat Aug 1 11:07:01 CEST 2009
663 - Glib silently fails when registering a timeout with
664 a negative value, so avoid that.
665 - call condvar callback immediately when it is set
666 after the condition is already signalled.
667 - check rbuf_max condition only after trying to consume
668 data.
669 - normalise signal numbers to names when using the ae's
670 signal handling, but do not document this (yet).
671 - pure perl signal emulation did not properly set nonblocking
672 mode on the signal pipe on win32 (this is unlikely to have caused
673 any issues).
674 - new module: AnyEvent::Debug.
675 - AnyEvent::Strict now checks that a signal was specified
676 by name (not name or number).
677 - reduce memory footprint in typical cases by ~50kb by
678 only compiling signal code when necessary.
679 - add AnyEvent::Handle->rbuf_max.
680 - grab the AE namespace for future new API, implement stubs for
681 said future/alternative API.
682 - new function: AnyEvent::Socket::format_hostport.
683
6844.881 Tue Jul 28 12:51:53 CEST 2009
685 - work around a bug in local in pre-5.10 perls, causing
686 AnyEvent::Handle to recurse when it should not
687 (analysed by elmex).
688
6894.88 Tue Jul 28 04:04:37 CEST 2009
690 - re-bless the handle into a dummy package after calling
691 AnyEvent::Handle::destroy, so the user does not need to check
692 for errors after every push_write etc.
693 - do not attempt to run t/02_signals.t on obviously broken
694 platforms.
695
6964.87 Sun Jul 26 02:06:16 CEST 2009
697 - do not attempt to linger when there is no longer a valid fh.
698 - enforce tls mode to be either accept or connect, do not simply
699 segfault in Net::SSLeay.
700 - AnyEvent::Handle can now call tcp_connect itself (new parameters
701 connect, on_prepare, on_connect and on_connect_error). Updated
702 tutorial accordingly.
703 - add AnyEvent::Impl::Irssi backend.
704
7054.86 Mon Jul 20 23:52:29 CEST 2009
706 - since the verbose warning is not enough, explicitly document
707 that versions before 1.33 of Net::SSLeay are not secure.
708 - work around signal handling races in Event and (...) Event::Lib.
709 - try to align signal-race timer to full-second boundaries.
710 - work around Tk not liking negative timeouts.
711 - don't complain of different grades of environmental unfriendlyness
712 in IO::Async.
713
7144.85 Sat Jul 18 06:16:14 CEST 2009
715 - nail the signal race problem in perl once and for all
716 (see $AnyEvent::MAX_SIGNAL_LATENCY).
717 - take advantage of Async::Interrupt if it is available.
718 - load Time::HiRes and Guard modules on demand only.
719 - add optional/recommended modules section to AnyEvent
720 documentation.
721 - reduce memory usage considerably (and reduce startup penalty)
722 by not using "strict", "warnings" and "overload" modules.
723 - work around buggy windows/openbsd perls and provide EBADMSG
724 and EPROTO ourselves when missing.
725 - improve perl 5.6 compatibility of the core event loop.
726 - made Net::SSLeay version 1.33 a soft requirement.
727
7284.83 Fri Jul 17 16:56:26 CEST 2009
729 - implement AnyEvent::Socket::getprotobyname.
730 - AnyEvent::CondVar's will now detect recursive blocking
731 waits and will croak, as too many people fall into
732 this trap.
733 - AnyEvent::Handle will now call ->destroy on itself after
734 executing the on_error callback, instead of doing some
735 half-baked internal shutdown, for fatal errors.
736 - clarify on_eof behaviour w.r.t. the read queue and
737 on_read callbacks.
738 - ignore some possible spurious wake-ups in tcp_connect.
739
7404.82 Sat Jul 11 00:34:55 CEST 2009
741 - POE and Event backends didn't accept some callable objects as
742 callbacks.
743 - use Config module instead of POSIX module to detect signal names
744 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
745 module doesn't even have all POSIX signals :/.
746 - use more workarounds around the many refcnt/corruption bugs in
747 Event::Lib.
748 - work around a race condition in perl's select, causing t/03_child.t
749 to rarely fail.
750
7514.81 Thu Jul 9 10:30:30 CEST 2009
752 - AnyEvent::Handle didn't properly diagnose write errors
753 (it expected -1 from syswrite, how lame... :).
754 - support file descriptors in addition to file handles
755 in AnyEvent->io.
756 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
757 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
758 and PERL_ANYEVENT_CA_PATH.
759 - provide a sensible synopsis section for AnyEvent::TLS.
760 - add a "supported backends" section to the manpage.
761 - added simple io watcher test to testsuite, using a
762 portable_socketpair.
763 - tried to improve the stability of the Event::Lib backend,
764 YMMV.
765
7664.8 Mon Jul 6 23:45:16 CEST 2009
767 - AnyEvent::DNS did not properly follow CNAME records with
768 uppercase targets.
769 - AnyEvent::DNS would errornously return AAAA records
770 with v4 mapped addresses (a faulty record) as ipv4 addresses,
771 causing AnyEvent::Socket to throw an exception.
772 - added new module AnyEvent::TLS for easier SSL/TLS context
773 creation, with many options including hostname verification,
774 secure default configuration, lots of documentation and,
775 predefined diffie-hellman keys for perfect forward security
776 and much more. get it while it's still fresh!
777 - use AnyEvent::TLS in AnyEvent::Handle for context management.
778 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
779 so AnyEvent::Socket users have smaller memory footprint
780 in the common case.
781 - add AnyEvent::Handle->push_shutdown method.
782 - add an additional error message parameter to AnyEvent::Handle's
783 on_error callback (for TLS, $! is still available).
784 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
785 - make AnyEvent::Handle more robust against early connection
786 failures (during new), and return C<undef> in that case
787 from the constructor.
788 - AnyEvent::Handle will now try to load only JSON::XS first,
789 then fall back to JSON.
790 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
791 and because it was documented that way.
792
7934.451 Fri Jul 3 00:28:58 CEST 2009
794 - do not clear rbuf when shutting down an AnyEvent::Handle
795 object - doing so breaks AnyEvent::HTTP.
796
7974.45 Mon Jun 29 22:59:26 CEST 2009
798 - a write error could cause AnyEvent::Handle to create
799 an I/O watcher with an undefined $fh.
800 - special-case mapped ipv4 addresses in both
801 AnyEvent::Socket::format_address and parse_address,
802 to treat them just like ipv4 addresses.
803 - updated and overhauled the AnyEvent::Intro doc.
804 - implement AnyEvent::Socket::format_ipv4/ipv6.
805 - slightly speed up portable_pipe/socketpair.
806 - expand condvar begin/end documentation.
807
8084.42 Fri Jun 26 08:32:18 CEST 2009
809 - preliminary and neccesarily incomplete support for IO::Async.
810 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
811 case it was set to IGNORE, to ensure we can catch child statuses
812 even when the calling env acts stupidly.
813 - updated benchmarks with IO::Async, which performs very well.
814
8154.412 Wed Jun 24 01:35:57 CEST 2009
816 - support an "untaint" attribute for AnyEvent::DNS
817 and set it on the default resolver.
818 - implement and document AnyEvent::Impl::Perl::loop.
819 - remove all anyevent-env variables from %ENV
820 when running in tainted mode.
821 - mention and extend the IO::Lambda benchmark.
822
8234.411 Sun Jun 7 18:48:02 CEST 2009
824 - do not try to use F_SETFD on windows.
825
8264.41 Thu May 14 06:40:11 CEST 2009
827 - work around issues in older perls (5.8.0?) when
828 a signal handler is deleted from the %SIG hash.
829 - use POSIX::_exit in child test, to avoid running
830 destructors.
831 - speed up CHLD handling by relying on SIGCHLD being
832 synchronously delivered, even when we roll our own
833 implementation.
834 - AnyEvent::DNS: add the "dname" resource record name
835 for cosmetic reasons.
836
8374.4 Sun Apr 26 20:12:33 CEST 2009
838 - implemented idle watchers, where applicable.
839 - AnyEvent->time died when Event backend was in use.
840 - fix a memleak in the Tk backend.
841 - sped up Tk timer handling.
842 - clip negative "after" values to 0 in AnyEvent::Impl::Event
843 to avoid spamming out warning messages.
844 - fix Qt timers without interval.
845 - avoid Qt zero-timeout special case, in old libqt's.
846
8474.352 Mon Apr 20 16:31:11 CEST 2009
848 - fix AnyEvent::Strict error messages for child watchers.
849 - fix/update Makefile.PL "recommends" versions.
850 - add AnyEvent->now_update.
851
8524.351 Sat Apr 11 07:56:14 CEST 2009
853 - actually make the signal pipe work under win32.
854 - localise $! in signal handler to avoid changing $!, although
855 perl itself does not seem to save/restore errno either.
856 - set the cloexec flag on the signal pipe (normally set by perl too).
857
8584.35 Fri Mar 27 11:48:20 CET 2009
859 - event models relying on AnyEvent's signal watcher emulation did
860 invoke the callback asynchronously, contradicting documentation
861 and causing signals to get lost (this includes AnyEvent's own
862 event loop). AnyEvent now uses the standard pipe trick to make
863 callback execution synchronous to the event loop.
864 - AnyEvent::Handle didn't free TLS context data on DESTROY
865 (patch by Pavel Shaydo).
866 - work around the ever-incompatibly-changing API of MakeMaker.
867 - document that changing global variables without restoring them
868 is a bad idea in Perl (noted by Adam Rosenstein).
869 - AnyEvent::Strict now barfs if ->io is passed a file.
870
8714.34 Thu Feb 12 18:32:45 CET 2009
872 - separately buffer TLS read data, as otherwise the read queue could
873 deadlock as receiving data is not expected while draining the
874 read queue (which cna only happen with TLS).
875 - raise EBADMSG error on JSON decoding errors.
876 - fix some minor manpage bugs (reported by Maximilian Gaß).
877 - speed up select bitmask parsing quite a bit in the pure perl
878 backend.
879 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
880
8814.331 Tue Jan 6 21:07:25 CET 2009
882 - socketpair fails on many vista machines because vista has
883 completely broken accept/getpeername and getsockname functions,
884 so we provide our own socketpair emulation that kind of works
885 (AnyEvent::Util::portable_pipe).
886 - new function: AnyEvent::Util::portable_socketpair.
887 - take advantage of the Guard module if it exists.
888
8894.33 Fri Nov 21 02:35:40 CET 2008
890 - AnyEvent::Strict did errornously flag a fileno of 0 as
891 illegal.
892 - reduce memory usage and slightly speed up the pure perl backend
893 by only storing the file descriptor, not the file handle.
894 - add missing autocork method to AnyEvent::Handle (reported
895 by Adam Rosenstein).
896 - AnyEvent::DNS->resolve errornously documented an $rcode result
897 argument, but there isn't (spotted by Henrik Krohns).
898 - the naked truth about Tk - it's basically unmaintained.
899
9004.32 Mon Nov 3 22:46:32 CET 2008
901 - fix AnyEvent::Socket::resolve_sockaddr to properly support
902 unix sockets again and choose a proper default for
903 the socket type.
904 - fix call to resolve_sockaddr in tcp_connect.
905
9064.31 Thu Oct 30 04:41:48 CET 2008
907 - implemented AnyEvent::Handle->destroy method.
908 - hint about unexpected effects in TLS mode.
909 - speed up AnyEvent::Impl::Perl by using more arrays
910 instead of hashes without the slightest loss of readability :)
911 - work around a perl argument refcounting bug.
912 - with some perl patches applied, AnyEvent now seems leak-free.
913
9144.3 Fri Oct 3 09:18:43 CEST 2008
915 - AnyEvent will now install a no-op signal handler for SIGPIPE,
916 unless one has been installed already.
917 - warn about not loaidng AnyEvent::Impl::POE early enough
918 (patch by Adam Rosenstein).
919 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
920 - fix a bug in where in SSL connect mode, the
921 client would first wait for some data by the server and
922 otherwise hang. (reported and analysed in an absolutely
923 exemplary manner by Adam Rosenstein).
924 - fix a bug in where SSL EOF would not be treated
925 as stream EOF, putting the connection into a hung state.
926 (reported and analysed in an absolutely exemplary manner by Adam
927 Rosenstein).
928 - fix a potential segfault when the TLS context
929 would go missing in a read callback (e.g. due to stoptls) -
930 Net::SSLeay of course makes no type checking whatsoever.
931 - AnyEvent::Handle will not stop the read watcher in TLS mode.
932 - AnyEvent::Handle->stoptls will send a TLS close notify
933 instead of simply destroying the stream now.
934 - fix a bug where large blocks of data written in TLS mode
935 would not be sent unless triggered by receives.
936 - on_drain will now take the tls write buffer into account.
937 - SSL operations have been streamlined a bit and should be faster
938 now, more to come, though.
939 - added a FAQ section, will document nontrivial issues in there.
940
9414.234 Mon Sep 29 04:08:13 CEST 2008
942 - fix child watcher documentation: only child exits will be reported,
943 no trace events.
944 - mention SIGPIPE in AnyEvent::Handle.
945 - perl backend will now use POSIX::times as monotonic clock
946 when available and the monotonic clock option is not.
947 - run even without Time::HiRes in some cases now, but do not
948 provide sub-second accuracy (all *supported* perl versions have
949 Time::HiRes).
950
9514.233 Fri Aug 22 01:48:05 CEST 2008
952 - fix a bug in the testsuite, causing a hang.
953 - clarified lots of AnyEvent::Handle constructor arguments.
954
9554.232 Thu Aug 21 20:44:25 CEST 2008
956 - fix a bug in AnyEvent::Handle that could cause two fatal error
957 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
958 (testcase by Leon Brocard).
959 - clarified on_eof/on_error documentation a bit.
960
9614.231 Tue Jul 29 13:12:15 CEST 2008
962 - remove some debugging code left in AnyEvent::Util::fork_call
963 (and no, it's impossible to implement with the broken windows
964 perls without resource leaks or worse).
965
9664.23 Tue Jul 29 12:19:59 CEST 2008
967 - document the first parameter passed to condvar callbacks to be
968 the callback.
969 - add AnyEvent::Socket::{ntoa,aton} aliases.
970 - optimize the AE::Handle->push_read (line) for the default
971 eol marker.
972 - optimize push_read (packstring|storable) for small packets.
973 - invoke on_error callback when no on_eof callback is set.
974 - fix a bug in push_read (storable) of unknown impact.
975
9764.22 Sun Jul 20 16:34:13 CEST 2008
977 - new function AnyEvent::Socket::parse_hostport.
978 - as the bulkheads at microsoft can't even get getprotobyname reliably
979 working on their shitty fucking broken os we need to hardcode
980 some common protocol numbers in AnyEvent::Socket. How can
981 people even bother with such a pile of shit as windows.
982
9834.21 Thu Jul 17 14:40:05 CEST 2008
984 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
985 which is incompatible to previous invocation syntax.
986 - work around more windows perl bugs in fork_call. I cannot imagine
987 why anybody would bother implementing fork in such an obviously
988 unusable way.
989 - avoid calling $do_retry if the request has already finished in
990 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
991 K. Sharifulin).
992
9934.2 Sat Jul 12 22:42:11 CEST 2008
994 - support an "interval" argument for repeating timers.
995 - fix a bug in the pure perl event loop that caused timers to
996 be delayed under some circumstances.
997 - implement optional argument checking via PERL_ANYEVENT_STRICT.
998 - not importing anything from Event causes it to malfunction
999 and not use Time::HiRes, work around this bug.
1000 - more bug workarounds against the endlessly broken Tk module,
1001 seems to work now on 32 bit machines, but less so on 64 bit.
1002 - minor optimisations applied to most backends.
1003
10044.161 Fri Jul 4 14:24:48 CEST 2008
1005 - fixed the t/handle/02_write.t, it read after write, with the
1006 assumption that all data could be written before read
1007 (reported by Jim Razmus).
1008
10094.160 Thu Jul 3 04:02:21 CEST 2008
1010 - re-registering signal watchers after unregistering one did
1011 not work if a backend falls back on AnyEvent's default
1012 implementation (also affected child watchers).
1013 - new AnyEvent::Handle options: autocork and no_delay.
1014
10154.152 Sun Jun 22 14:15:44 CEST 2008
1016 - allow for 32-bit perls that implement shifts differently
1017 on different architectures in parse_ipv4 (reported and
1018 analysed by Keiichi DAIBA).
1019
10204.151 Fri Jun 6 17:34:24 CEST 2008
1021 - make sure specifying _only_ on_read and never pushing reads
1022 works.
1023
10244.15 Fri Jun 6 13:00:46 CEST 2008
1025 - the pure perl backend would keep some watchers alive when more than
1026 one watcher was registered for the same fd.
1027 - new "packstring" and "storable" read and write types
1028 for AnyEvent::Handle.
1029 - allow on_eof handler to be called after on_error with EPIPE returns.
1030 - do not immediately call on_read callback in handle constructor.
1031
10324.14 Thu Jun 5 20:29:31 CEST 2008
1033 - Fixed a bug in DNS SRV priority sorting.
1034 - AnyEvent::Util::guard now reports runtime errors while
1035 executing the guard block as warnings.
1036 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
1037
10384.13 Thu Jun 5 00:47:59 CEST 2008
1039 - AnyEvent::DNS only followed cname chains with length 2,
1040 contrary to documentation. bumped it up to 10, thanks to
1041 microsoft, the current broken-dns-config-king.
1042 - AnyEvent::DNS didn't check for socket return status
1043 but instead relied on perl not creating filehandles in that
1044 case - too bad it gives you a bogus file handle (reported
1045 and analysed by Vladimir Timofeev).
1046 - fix queue management logic in AnyEvent::Handle:
1047 when on_read was registered, the queue was empty
1048 and no progress could be made AnyEvent::Handle would
1049 enter an endless loop.
1050 - correctly start reading again when the handle became
1051 busy again after idling.
1052 - correctly treat tls shut-downs as EOF condition
1053 (for the time being, would be nice to have a callback for that).
1054 - correctly call eof callback under all conditions
1055 (hopefully).
1056 - the timeout callback did not expect that $self can go
1057 away any time.
1058
10594.12 Tue Jun 3 10:58:04 CEST 2008
1060 - include AnyEvent::Intro, a tutorial for anyevent,
1061 anyevent::socket and anyevent::handle.
1062 - allow more options in on_error.
1063
10644.11 Fri May 30 23:42:25 CEST 2008
1065 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
1066 reverse_lookup and reverse_verify to replace it, support
1067 v4mapped and v4compat addresses.
1068 - provide more documentation for the resolver class.
1069 - really replace longest run of :0: by :: in format_address,
1070 also properly convert :: and ::1 again.
1071 - support NAPTR record name and decode it.
1072 - implement random weight sampling for SRV records, as per
1073 rfc 2782.
1074 - correctly abort on srv-record targets of ".".
1075 - added AnyEvent::DNS::wait_for_slot.
1076 - in the unlikely event of a virtual circuit connection
1077 being invalidated by a delayed udp reply, AnyEvent::DNS
1078 could die. this has been fixed.
1079 - plug a probable memleak in the DNS vc code.
1080 - use configured timeout also for VC request phase.
1081 - implement timeout and max_outstanding methods
1082 for resolver class.
1083 - update version numbers in all modules.
1084
10854.1 Thu May 29 05:45:40 CEST 2008
1086 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
1087 which mirrors their purpose better (old names still available
1088 for a while). Also moved AnyEvent::DNS::addr to
1089 AnyEvent::Socket::resolve_sockaddr.
1090 - implement AnyEvent->time and AnyEvent->now.
1091 - fix IPv6 support in pack_sockaddr.
1092 - officially un-experimentalise all newly introduced code.
1093 - support unix domain sockets everywhere by specifying
1094 a host of "unix/" and the pathname as service.
1095 - implement an activity timeout in AnyEvent::Handle.
1096 - added a regex read type to AnyEvent::Handle.
1097 - added a json read/write type to AnyEvent::Handle.
1098 - always croak in on_error in AnyEvent::Handle.
1099 - document how to subclass AnyEvent::Handle.
1100 - implement AnyEvent::Util::fork_call.
1101 - add support for IPv6 nameservers and nameserver statements.
1102 - work around _yet_ _another_ windows perl bug in where empty
1103 select masks cause errors in the pure perl backend (this
1104 workaround cannot be implemented for the other event loops,
1105 but EV also works around this).
1106 - supply AnyEvent::Socket::address_type and make good use of it.
1107 - clarify and fix the Handle documentation w.r.t $self vs. $handle
1108 vs. callback arguments.
1109 - add some recommends to the META.yml.
1110
11114.05 Mon May 26 19:44:06 CEST 2008
1112 - some platforms ignore the rfc and prepend an extra sa_len member to
1113 the sockaddr structure, cater for those. also use sockaddr_family,
1114 didn't know it was there, but comes in handy for the workaround.
1115 - undo the import WIN32 hack.
1116
11174.04 Mon May 26 08:03:31 CEST 2008
1118 - try to work around yet another windows bug: failed connects are reported
1119 as if it were out-of-band data. windows users: you suck. Only
1120 EV, Glib, Event and the pure perl backend can handle this condition.
1121 - optimize write algorithm in AnyEvent::Handle.
1122 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
1123 - tcp_connect now properly iterates through all targets.
1124 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
1125 support entirely. also hardcode some AF_INET6 constants for the
1126 sake of perl 5.8.
1127 - vastly improved nameserver/suffix detection algorithm for windows
1128 (still a hack...)
1129 - try to use a fallback dns server if no dns servers could be found.
1130 - splatter around some local $SIG{__DIE__}'s to work around bad
1131 die handlers.
1132
11334.03 Sun May 25 05:05:57 CEST 2008
1134 - work around perl bugs on windows, where perl returns undocumented
1135 error codes for sysread, syswrite etc.
1136 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
1137 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
1138 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
1139 - updated documentation.
1140 - lots of workarounds for perl 5.8 quirks.
1141 - implement a more flexible type system for both reads and writes
1142 in AnyEvent::Handle.
1143 - support IPv6 in tcp_server.
1144 - pass local hostname and port to the prepare callback in tcp_server.
1145 - make types extendable in AnyEvent::Handle.
1146 - croak, not die, in AnyEvent::Handle;
1147
11484.0 Sat May 24 19:58:08 CEST 2008
1149 - added (experimental) AnyEvent::DNS module.
1150 - added (experimental) AnyEvent::Socket module.
1151 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
1152 - make condvar->end call ->send as documented.
1153 - make condvar constructor support the documented cb => argument.
1154 - overload condvars so you can use them instead of code references.
1155 - support more of the POSIX forms instead of a pure dotted quad for
1156 verbatim ip addresses.
1157 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
1158 - try very hard not to freeze on broken windows perls.
1159 - make on_eof optional in AnyEvent::Handle.
1160 - support service names in tcp_connect and tcp_server.
1161 - work around netbsd/openbsd bug causing unix sockets not to
1162 become writable until empty.
1163 - only make a dummy read for error detection when getpeername
1164 returns ENOTCONN.
1165 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
1166 override.
1167 - new AnyEvent::Guard::cancel.
1168
11693.5 Sat May 17 23:17:14 CEST 2008
1170 - gracefully deal with systems not defining CLOCK_MONOTONIC
1171 (as opposed to not supporting the clock itself, which worked in 3.41).
1172 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
1173 helpers, to replace the ill-designed AnyEvent::Socket class.
1174 - added some examples (eg/connect, eg/handle).
1175 - added AnyEvent::Util::guard.
1176 - AnyEvent::Util now exports most of its functions.
1177 - fixed a bug in *_read_chunk, where the callback was called with $self as
1178 first and second argument.
1179 - simplified and fixed regex handling in AnyEvent::Handle::*_read_line.
1180 - fix manpage condvars examples to use new syntax.
1181 - considerably improve inet_aton with EV::ADNS backend to support
1182 cname chains (common dns configuration bug, e.g. with www.google.de)
1183 and aaaa records if no a record could be found.
1184
11853.41 Sun May 11 19:53:13 CEST 2008
1186 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
1187
11883.4 Sun May 11 00:29:25 CEST 2008
1189 - complete re-implementation of condvars with a hopefully much more useful
1190 API (backward compatible functions exist but are undocumented).
1191 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
1192 (still subject to change).
1193 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
1194 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which
1195 allows module authors to avoid forcing event loop detection.
1196 (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example).
1197 - remove coro backends: Coro now provides generic support for AnyEvent
1198 via Coro::AnyEvent.
1199
12003.3 Mon Apr 28 09:51:06 CEST 2008
1201 - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules.
1202 - fix a bug in the pure perl backend that kept watchers alive
1203 when multiple watchers were registered for the same fh/poll combo.
1204 - add a benchmark section showing AnyEvent overhead and comparing
1205 the different event loops with each other.
1206 - prefer pure perl over tk when autoprobing, as it's about as fast,
1207 but doesn't crash with many watchers.
1208 - declare Qt support non-experimental.
1209 - clarify the confusing section about the file descriptor being
1210 kept alive.
1211 - document the race between loading of an event module and
1212 child processes exiting.
1213 - support POE as "backend" (with some caveats, POE is not generic
1214 enough, and darn slow).
1215 - support Wx and Prima through POE.
1216 - optimise perl backend to use 20% less memory and take advantage
1217 of typical timeout behaviour. It can now compete with
1218 select/poll-based C event loops in most cases (it is usually
1219 faster than Event and Glib :).
1220 - roughly cut EV memory use in half and increase its speed by 30%,
1221 by removing undocumented functionality. Did something similar
1222 to other event loops where possible.
1223
12243.2 Thu Apr 24 10:10:40 CEST 2008
1225 - do not die when anyevent watchers are destroyed while running callbacks
1226 in the pureperl backend (could only happen when two watchers are registered
1227 for the same fh/poll combo).
1228 - support autoloading for child watchers, was broken in all versions.
1229 - implement PERL_ANYEVENT_MODEL env variable.
1230 - (experimental) implement interface to the (very crashy/buggy) Event::Lib module.
1231 - (experimental) implement interface to the Qt module (cannot be autoprobed).
1232 - this release is IO:AnyEvent-proof.
1233
12343.12 Tue Apr 22 07:11:46 CEST 2008
1235 - reinstate AUTHOR section that got lost somehow.
1236 - do not hang in the testsuite with badly broken perls (activestate,
1237 strawberry...), but instead diagnose the problem and continue.
1238 - use INT instead of CHLD in an attempt to support broken windows perls
1239 better (this decreases the test quality, unfortunately).
1240 - do not send the signal to the process group
1241 (no problem for CHLD, bad for INT :).
1242
12433.11 Sat Apr 19 06:57:31 CEST 2008
1244 - major documentation rework.
1245 - document the fact that child watchers only watch for zombies.
1246 - fix the child watcher example.
1247
12483.1 Wed Apr 16 17:09:01 CEST 2008
1249 - work around recurring bugs in Tk by dup'ing filehandles, the
1250 only method with good success chances on Tk (the bugs apparently
1251 don't get fixed anytime soon).
1252 - lift the restriction of only one watcher per fh direction
1253 (as the Tk bug workaround also lifts it and only Tk imposed
1254 such strong limits).
1255 - changed probe order to prefer coro adaptors.
1256 - explain why recursion into the event loop is not supported
1257 unless the backend supports it (only Coro::EV does without
1258 any restrictions...).
1259 - add simple manpages for all backend modules.
1260
12613.0 Mon Apr 7 21:30:23 CEST 2008
4 - Coro::Signal changed semantics, roll our own, also cleaning 1262 - Coro::Signal changed semantics, roll our own, also cleaning
5 up the Coro implementation in general. 1263 up the Coro implementation in general.
6 - rename Coro backend to CoroEvent. 1264 - rename Coro backend to CoroEvent.
7 - add some decision helping paragraph to the manpage that should 1265 - add some decision helping paragraph to the manpage that should
8 help people to decide wether AnyEvent is the right thing for them. 1266 help people to decide whether AnyEvent is the right thing for them.
9 1267
102.9 Mon Jan 28 13:31:54 CET 2008 12682.9 Mon Jan 28 13:31:54 CET 2008
11 - update for EV 3.0 API changes. 1269 - update for EV 3.0 API changes.
12 1270
132.8 Sun Nov 25 15:06:03 CET 2007 12712.8 Sun Nov 25 15:06:03 CET 2007
… …
99 - fixed documentation. 1357 - fixed documentation.
100 - clarified some corner cases. 1358 - clarified some corner cases.
101 - vastly improved documentation with a long example. 1359 - vastly improved documentation with a long example.
102 1360
1030.2 Thu Dec 1 22:18:49 CET 2005 13610.2 Thu Dec 1 22:18:49 CET 2005
104 - check wether a valid method was called, to avoid endless loops. 1362 - check whether a valid method was called, to avoid endless loops.
105 - move condvars into their own class (cleaner, safer). 1363 - move condvars into their own class (cleaner, safer).
106 - don't die when some modules are not available. 1364 - don't die when some modules are not available.
107 - Tk is too broken to allow for words. Will work once Tk works. 1365 - Tk is too broken to allow for words. Will work once Tk works.
108 1366
1090.1 Thu Dec 1 19:44:57 CET 2005 13670.1 Thu Dec 1 19:44:57 CET 2005

Diff Legend

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