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

Comparing AnyEvent/Changes (file contents):
Revision 1.233 by root, Sat Apr 25 20:27:47 2009 UTC vs.
Revision 1.639 by root, Wed Nov 27 19:59:20 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?
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.
20
21 - AnyEvent::Loop no longer loads the POSIX module if Time::HiRes
22 has monotonic clock support.
23
247.17 Wed Sep 18 03:04:49 CEST 2019
25 - work around antique openssl version in RHEL 7 by formatting
26 dh parameters differently (reported by several people).
27 - add t/13_weaken.t.
28
297.16 Fri Jul 19 18:00:21 CEST 2019
30 - add ffdhe group dh parameters from RFC 7919, and use ffdhe3072
31 as new default, instead of schmorp1539.
32 - AnyEvent::Log did not reassess logging status of
33 AnyEvent::Log::logger's when contexts were changed with
34 ->attach/detach/slaves, causing them to not log even though a
35 recent attach should have caused them to log.
36 - added some more logging configuration examples.
37 - mention RFC 8482 in AnyEvent::DNS.
38
397.15 Tue Feb 26 03:07:42 CET 2019
40 - INCOMPATIBLE CHANGE: AnyEvent::Handle's tls_detect documentation gave
41 separate major and minor versions, while code passed only a single
42 value. This version follows the documentation and now passes separate
43 major and minor values.
44 - work around Net::SSLeay not having been ported to openssl 1.1, but many
45 distributions compiling it against openssl 1.1, which unfortunately
46 succeeds and results in a very broken module.
47 - AnyEvent::DNS::dns_unpack now stores the original DNS packet
48 in the __ member, to allow decoding of undecodable resouce records
49 containing compressed domian names.
50 - AnyEvent::Socket::parse_ipv6 would NOT, as advertised, accept ipv4
51 addresses. It now does and converts them to ipv4 mapped addresses.
52 - support CAA records, based on patch by Steve Atkins.
53 - add freenom and cloudflare nameservers as dns fallback.
54 - AnyEvent::Strict would not properly ward against io watchers
55 on files when the handle passed was a file descriptor.
56 - document "internal" variables used by the dns en-/decoder to allow
57 enterprising users to extend them in a semi-official way.
58
597.14 Sat Jun 24 01:44:19 CEST 2017
60 - fix a crash bug in AnyEvent::Handle with openssl 1.1.0 (patched
61 by Paul Howarth and apparently tracked down by Mike McCauley).
62 - AnyEvent::Handle->keepalive was documented (and defined) twice
63 (found by Matt Merhar).
64 - AnyEvent::Socket::tcp_bind/tcp_server would immediately unlink a unix
65 listening socket unless a guard is used. Change this so that no
66 cleanup will be performed unless a guard is used and document this more
67 clearly.
68 - make tcp_bind/tcp_server error messages more regular.
69 - work around latest perl upstream breakage - this time, perl5porters were
70 apparently bullied by a bogus CVE (CVE-2016-1238), and since distros
71 stupidly apply anything that has "security fix" stamped on it, it's likely
72 going to be a problem in practise (and working around it is trivial on
73 windows and unix).
74 - add TCP_FASTOPEN/MSG_FASTOPEN and MSG_NOSIGNAL/MSG_MORE/MSG_DONTWAIT
75 to constants.
76 - update warnings to common::sense 3.74 standards.
77
787.13 Sat Sep 17 04:31:49 CEST 2016
79 - Only call tlsext_host_name for non-empty common names (reported
80 by Maxime Soulé).
81 - log a (single) notice message if SNI is not supported.
82 - upgrade to UTS-46:9.0.0 draft and switch to non-transitional
83 behaviour, beating thunderbird, ie, edge, chrome and safari to it :)
84 (see also https://bugzilla.mozilla.org/show_bug.cgi?id=1218179)
85 - turns out the UTS-46 IDNA testcase failures were indeed bugs in the
86 testcases and the specification and not in the code - the post-9.0.0
87 unicode files have all known problems fixed, so finally the AnyEvent
88 IDNA implementation can pass the full IDNA testsuite - without needing
89 a single fix :)
90 - guarantee (and document) that condvar callbacks will be removed
91 on invocation - important to avoid circular references.
92
937.12 Wed Jan 27 19:12:26 CET 2016
94 - use common name as hostname for tls connects, if Net::SSLeay
95 supports SNI.
96 - fix documentation of tls_autostart read type in AnyEvent::Handle,
97 analyzed by Felix Ostmann.
98
997.11 Thu Jul 16 14:36:00 CEST 2015
100 - AnyEvent::Socket::parse_ipv6 could accept malformed ipv6
101 addresses (extra "::" at end and similar cases).
102 - add a more explicit warning to AnyEvent::Handle that it doesn't
103 work on files, people keep getting confused.
104 - new function AnyEvent::Socket::tcp_bind.
105 - new functions AnyEvent::fh_block and AnyEvent::fh_unblock.
106 - aligned ipv6 address formatting with RFC 5952 (by not shortening
107 a single :0: to ::).
108 - added stability canary support.
109
1107.09 Sat May 2 16:38:53 CEST 2015
111 - AnyEvent::Debug called an internal function (AnyEvent::Log::ft)
112 that was renamed to AnyEvent::Log:format_time. uinder its old name
113 (analyzed by Michael Stovenour).
114 - update AnyEvent::DNS fallback resolver addresses:
115 seems google effectively killed most other free dns resolvers,
116 so remove them, but add cable and wireless (ecrc) since it was
117 stable for 20 years or so, official or not, and there should be
118 an alternative to google.
119 - perl5porters broke windows error codes in 5.20, and mapped
120 WSAEWOULDBLOCK on the (different) EWOULDBLOCK error code, and
121 WSAEINPROGRESS into the incompatible ERINPROGRESS code, probably
122 because they were so cool. They probably broke other error codes
123 for no reason, too, but I didn't care for checking, it's too
124 depressing. This version only works around the WSAEWOULDBLOCK
125 issue, because I don't have a nice way to work around the
126 WSAEINPROGRESS bug.
127
1287.08 Wed Dec 10 05:27:17 CET 2014
129 - work around a newly introduced bug in Socket 2.011 (an
130 errornous sun_length check) (analyzed by Maxime Soulé).
131 - AnyEvent::TLS didn't load (but refer to) AnyEvent::Socket
132 (analyzed by Ben Magistro).
133 - AnyEvent::Strict will now confess, not croak. This is in line with
134 it being a development/debugging tool.
135 - work around a number of libglib bugs (debug builds of libglib enforce
136 certain undocumented behaviour patterns such as not being able to
137 remove a child watch source after it has fired, which we will try
138 to emulate to avoid "criticals". what where they thinking...).
139 - mention json security issues in AnyEvent::Handle, now that Douglas
140 Crockford has foolishly and incompatibly changed JSON.
141 - changed default dns resolver "max_outstanding" value from 1 to 10,
142 the latter beinfg the intended value all along
143 (reported by Ilya Chesnokov).
144 - added new "AnyEvent::Impl::UV" interface module to the UV event lib
145 (written by Mike Lowell).
146
1477.07 Tue Dec 17 17:45:02 CET 2013
148 - the documentation for custom tls verify schemes was wrong. make it agree
149 with the code (reported by Maxime Soulé).
150 - added cbor read and write types to AnyEvent::Handle (using CBOR::XS).
151 - work around an API change in openssl that could cause wrong tls connection
152 aborts, likely on windows only (analyzed by sten).
153 - calling AnyEvent->now_update with AnyEvent::Impl::Perl caused an
154 endless loop (reported by Dietrich Rebmann).
155 - add tlsv1_1 and tlsv1_2 protocols to AnyEvent::TLS
156 (patch by Maxime Soulé).
157 - document AnyEvent::Impl::IOAsync::set_loop and
158 $AnyEvent::Impl::IOAsync::LOOP. Though only documented now, this
159 functionality has _always_ been available.
160 - force a toplevel domain name in t/81_hosts.t (analyzed by
161 David Jack Wange Olrik).
162 - document that AnyEvent::Log uses AnyEvent::IO.
163 - warn about AnyEvent::Filesys::Notify performance.
164 - praise the joys of AnyEvent::Fork::*.
165 - time for an =encoding directive.
166 - do no longer use JSON to create a default json coder, use
167 JSON::XS or JSON::PP directly.
168
1697.05 Wed Aug 21 10:38:08 CEST 2013
170 - uts46data.pl couldn't be found due to wrong naming of the file
171 (reported by Fulko Hew).
172 - handle lone \015's properly in AE::Handle's default line read
173 (reported by various people).
174 - untaint ip addresses found in /etc/hosts (patch by José Micó).
175 - the memleak fix in 7.03 caused resolving via /etc/hosts to always
176 fail on first use (reported and testcase by Andrew Whatson).
177 - expose AnyEvent::Log::format_time, and allow users to redefine it.
178 - expose AnyEvent::Log::default_format, and allow redefinition.
179 - expose AnyEvent::Log::fatal_exit, to allow redefinition.
180 - AnyEvent::Debug shell can now run coro shell commands, if available.
181 - t/63* tests were wrongly in MANIFEST.
182 - kernel.org's finger server went MIA, switch to freebsd.org and
183 icculus.org.
184 - clarify that IO::AIO and AnyEvent::AIO are needed for AnyEvent::IO
185 to function asynchronously (reported by Daniel Carosone).
186 - hard-disable $^W in most tests, it generates too much garbage output.
187 - use a (hopefully) more future-proof method to emulate common::sense.
188 - upgrade to UTS-46:6.2.0.
189 - switch to INSTLIB from INSTLIBDIR, as INSTLIBDIR was wrongly
190 documented. should not affect anything.
191
1927.04 Thu Nov 15 02:23:58 CET 2012
193 - debugging code left in.
194
1957.03 Thu Nov 15 02:16:11 CET 2012
196 - AnyEvent::Socket::inet_aton did not work when dns resolution was
197 used to find the addresses (analyzed and patch by Maxime Soulé).
198 - fix a memory leak in the /etc/hosts lookup code when hosts don't
199 resolve and are not in hosts (analyzed by sten).
200
2017.02 Tue Aug 14 04:27:58 CEST 2012
202 - AnyEvent::Util::run_cmd could block indefinitely (analyzed and test
203 program by Yorhel).
204 - verified that AnyEvent::Socket follows RFC5952.
205 - try to parse "ADDR#PORT" in addition to "ADDR PORT".
206
2077.01 Sun May 13 01:03:17 CEST 2012
208 - fail with EPROTO in AnyEvent::Handle wqhen TLS is requested but
209 not available, instead of throwing an exception.
210 - use File::Spec to get the tmpdir in t/*, to avoid needless
211 failures on (most, not mine :) windows boxes.
212 - new handle read types: tls_detect and tls_autostart.
213
2147.0 Fri Apr 13 06:33:30 CEST 2012
215 - child watchers are broken in POE 1.352 (also many earlier
216 versions) and there seems to be no way to work around it, as POE
217 itself is inherently racy. Document this regression and add a delay
218 in t/68_poe_03_child.t for the time being.
219 - new module AnyEvent::IO, that is a frontend to either a pure-perl
220 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
221 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
222 - load /etc/hosts only when DNS has no answer.
223 - stat /etc/hosts on every access and reload it if it changed.
224 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
225 - fix a buggy croak in the dh parameter reading in AnyEvent::TLS.
226 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
227 a side effect, they now use true appending as opposed to libc
228 appending, and the file might not have been opened when the function
229 returns.
230 - the default logging level was not properly documented in a variety of
231 places, this has hopefully been rectified.
232 - updated uts46data.pl for unicode 6.1.0.
233 - made log messages generated by AnyEvent submodules not
234 include the package name anymore, as it will be logged
235 by default already.
236 - upgrade to the trick used by common-sense 3.5 to work
237 around extra warning torture/breakage under perl 5.15.x.
238 - log messages by anyevent are now ucfirst, are usually full
239 sentences and do no longer include the package name.
240 - the storable read type would sometimes throw an exception instead
241 of causing EBADMSG (analyzed by Maxime Soulé).
242 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
243 keysize (reported by Paul Howarth).
244 - documented Rocco fabricating statements about AnyEvent and me.
245
2466.14 Tue Jan 31 20:00:24 CET 2012
247 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
248 probably an editing mistake (reported by Darin McBride).
249
2506.13 Thu Jan 12 07:27:01 CET 2012
251 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
252 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
253 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
254 regex-matching inside the callback would change the parameter, despite
255 $1 being dynamically scoped per-block. this perl bug is now being
256 worked around (testcase by Cindy Wang).
257
2586.12 Mon Dec 12 13:21:10 CET 2011
259 - $! was clobbered by subsequent calls in tcp_connect, due to
260 the postpone not saving/restoring it (analyzed by Richard Garnish).
261
2626.11 Tue Nov 22 10:36:05 CET 2011
263 - Tk cannot create windows when tainted unless you set a title, so
264 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
265 - escape any nonprintable/nonascii characters when stringifying
266 backtraces.
267 - log the reason loading an interface module fails to load at level debug.
268
2696.1 Tue Oct 4 19:44:30 CEST 2011
270 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
271 and above), and some messages inside AnyEvent have been elevated
272 to higher log levels to print by default.
273 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
274 - work around yet another signal race bug, newly introduced by POE. My
275 how that module sucks.
276 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
277 support ipv6-only operation even with defaults (also use ::1 instead
278 of 127.0.0.1).
279 - give better guidelines on how to chose a logging level.
280 - AnyEvent::Log can now cap (limit) log message priority by context.
281 - AE::log will now save and restore $! and $@, so callers don't have to.
282 - AE::Strict now checks that registered fds are still valid at regular
283 intervals.
284 - skip frequently called 'require' calls if the module appears loaded.
285 - increase signal latency test timer considerably for poor little osx,
286 which can't be bothered with basic posix functionality such as signal
287 delivery.
288 - mention some debugging aids in the tutorial.
289 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
290 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
291 of the context as title.
292 - AnyEvent::Strict would not restore read-write access to $_ when a callback
293 died - this is a user bug, but since AE::Strict exists to help dbeugging...
294 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
295 to work with it - the author is well aware of the issues but refuses
296 to fix them. This extreme step has been taken because the module actively
297 attacks the goals of AnyEvent and makes it harder for module authors to use
298 AnyEvent.
299
3006.02 Fri Aug 26 20:08:31 CEST 2011
301 - log_to_syslog now accepts facility strings.
302 - log_to_syslog would not allow facilities - "$facility|$level"
303 *literally* means that string. sheesh.
304 - the CPAN parser got confused and thought we provide Sys::Syslog, try
305 to work around.
306 - renamed AnyEvent::Impl::FLTK2 to FLTK.
307
3086.01 Fri Aug 26 07:04:11 CEST 2011
309 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
310 the TTL as fourth element - this affects ->resolve, ->request
311 and ->dns_unpack indirectly. Convenience functions such as
312 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
313 AnyEvent::Socket.
314 - INCOMPATIBLE CHANGE: backend authors now should not implement
315 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
316 methods.
317 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
318 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
319 - DEPRECIATION: please change
320 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
321 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
322 in your programs, for future compatibility.
323 - *blush* idle watcher emulation was borked.
324 - AnyEvent::Strict now checks the AE::xxx functions as well.
325 - querying too long domain names would barf deep inside AnyEvent::DNS - since
326 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
327 for too long domain names explicitly and passes an undef to the callback
328 instead (reported by James Bromberger).
329 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
330 - new functions: AnyEvent::postpone, AnyEvent::log.
331 - new module: AnyEvent::Log, for simple logging needs.
332 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
333 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
334 a debug shell.
335 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
336 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
337 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
338 /etc/hosts can be found.
339 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
340 configuration.
341 - env variables can now be specified using AE_xyz instead of
342 PERL_ANYEVENT_xyz, subject to some rules.
343 - lots of new AnyEvent::Debug functionality.
344 - AnyEvent::Strict now checks for modifications of $_ by the callback
345 (by setting it to readonly while executing the callback, if possible).
346 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
347 Invest extra overhead to work around this peculiar behaviour.
348 - split the pure perl backend into a loop module and an impl
349 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
350 - some read types (regex, netstring, json, storable) could
351 errornously keep an AE::Handle object alive on their own
352 (reported by Mohammad Toossi).
353 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
354 supports /etc/hosts overrides.
355 - add more workarounds around child watcher bugs in IO::Async
356 and (newly introduced) in POE.
357 - use glib child watchers and try to work around its limitations,
358 if possible (based on a patch by Kevin Ryde).
359 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
360 - many read types in AnyEvent::Handle would malfunction during
361 transitions to TLS because of a "delete $self->{rbuf}".
362 (reported by Gerald Galster).
363 - improve condvar blocking wait performance (probably only noticable
364 with faster backends), also make it easier for event loops that
365 do not support blocking waits.
366 - PERL_ANYEVENT_MODEL now also allows full module names, although
367 the usefulness of this feature is not yet clear.
368 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
369 - add t/80_ssltext.t.
370 - no longer use AUTOLOAD for the initial loading (for no very good reason).
371 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
372 - do not run event loop detection multiple times when programs call their cached
373 copies of AnyEvent::detect.
374 - parse_hostport now accepts absolute paths als unix domain sockets.
375 - suppress the idiotic warning inside POE by patching the POE::Kernel
376 run flags instead of asking the user to jump through contortions.
377 - add Task::Weaken as dependency (as recommends only at this time).
378 - replace PF_UNSPEC by 0, for good old minix.
379
3806.0 (not officially released, but was on CPAN for about 15 minutes :/)
381
3825.34 Thu May 12 09:59:41 CEST 2011
383 - AE::TLS didn't support both cert_file and key at the same time,
384 due to some copy&paste bug (reported and analyzed by vti).
385 - improve AE::Handle timeout documentation slightly.
386
3875.33 Mon Feb 7 21:51:41 CET 2011
388 - never appeared on CPAN, for unknown reasons.
389 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
390 object when it's already gone (analyzed by Ryan Bullock).
391 - update idna algorithm(s) to UTS#46 version 6.0.0.
392 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
393 uts46 says it's an error, and testsuite makes more sense.
394 - remove leading dots in idn_to_ascii - testsuite says to remove,
395 uts46 says it's an error, and testsuite makes more sense.
396 - fixed a bug in idn ignored character processing.
397 - passes the uts46 6.0.0 testsuite except for three cases which
398 apparently are bugs in the testsuite itself.
399 - passes the uts46 6.0.1 testsuite except for eleven cases which
400 apparently are bugs in the testsuite itself.
401
4025.31 Sun Jan 23 11:49:19 CET 2011
403 - work around a POE bug where POE would first poll for new events and
404 invoke handlers BEFORE handling the message queue.
405 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
406 - stop_read now does just that in TLS mode. see its
407 documentation for caveats.
408 - update to common::sense 3.4 warning set (some warnings in perl impose
409 an arbitrary 32kb limit on regex matches, which affects
410 AnyEvent::Handle).
411 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
412 - new experimental AE::Handle::resettls method.
413
4145.3 Fri Dec 31 05:48:13 CET 2010
415 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
416 issues with IO::Async have been solved in that version, including
417 the availabilty of a default loop!
418 - implement new wbuf_max parameter for AnyEvent::Handle.
419 - added Cocoa::EventLoop backend by Daisuke Murase.
420 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
421 - clarify handle on_prepare documentation.
422 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
423 to avoid leaking them to any child processes.
424 - try to force files into the arch-dependend lib directory, for the
425 half-"turly obsessive compulsive person"(s) that depend on it.
426
4275.29 Sun Dec 5 10:49:21 CET 2010
428 - convert EV backend to EV 4.00 API (so better upgrade EV too).
429 - AE::Handle: implement dynamic read_size adjustment, add
430 max_read_size parameter, reduce default read_size to 2048.
431 - add some FAQ entries for servers and/or on_eof vs. on_error.
432 - work around OS bugs (cygwin again) when lingering in AE::Handle.
433 - reduce memory usage slightly when connect was used in AE::Handle.
434 - wrok around more segfaults in Tk.
435 - document tls_ctx => undef as valid.
436 - detect WNOHANG value at compiletime.
437 - include some probably linux-only support for building AnyEvent
438 as part of the perl core.
439 - improve compatibility of t/08_idna.t to perl 5.8.
440 - make AnyEvent::Strict truly optional (in case somebody builds a
441 minimal perl...).
442
4435.28 Wed Oct 13 04:14:23 CEST 2010
444 - due to a glitch, AnyEvent's internal getprotobyname wasn't
445 actually used in tcp_server.
446 - implement AnyEvent::CondVar->new method, which was documented but
447 missing.
448 - support autoloading of read types in unshift_read.
449 - AnyEvent::Strict: do the test for invalid fh arguments differently,
450 resulting in more informative error messages.
451 - new AnyEvent::FAQ manpage.
452 - clarify that you can actually call push_read etc. in on_read.
453 - put keys and certs in /tmp, not into the current directory,
454 and document this in the SECURITY CONSIDERATIONS sections
455 of AnyEvent::TLS.
456 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
457 requested by Ron Isaacson.
458 - document $AnyEvent::DNS::RESOLVER variable.
459 - incorporated major typo patches by Abhijit Menon-Sen.
460 - AnyEvent::Handle now croaks when negative timeouts are passed.
461 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
462 undocumented, but who knows when they might come in handy.
463
4645.271 Tue Jun 8 12:05:46 CEST 2010
465 - backport to perl 5.8.x.
466
4675.27 Sun Jun 6 12:12:05 CEST 2010
468 - postpone differently in AnyEvent::Socket now, as
469 when not, canceling the connection attempt might fail
470 (found by Felix Antonius Wilhelm Ostmann).
471 - explicitly check for non-stream sockets in AE::Handle, too many
472 clueless people fell into the trap of this somehow working.
473 - simplified and reworked the "OTHER MODULES" section.
474 - better/more condvar examples.
475
4765.261 Wed Apr 28 16:13:36 CEST 2010
477 - AF_INET6 was not properly used from Socket6 during configuration
478 time (found by Andy Grundman).
479
4805.26 Mon Apr 12 04:49:35 CEST 2010
481 - don't generate spurious readyness notifications when select
482 returns EINTR in the pure perl backend (analysed and testcase
483 by Bas Denissen).
484 - give same examples in SYNOPSIS sections of AnyEvent and AE.
485 - provider faster implementations for AE::time/now/now_update for
486 EV and Perl backends.
487 - provide faster AE::cv and AE::time implementations for all
488 backends.
489 - no longer support lower/mixed-case signal names (this was
490 never documented nor universally supported).
491 - some more memory µ-optimisations, and somewhat less messy
492 code to cope with both AnyEvent and AE APIs.
493
4945.251 Sat Mar 13 00:58:21 CET 2010
495 - make test failed if it was run before make install on perl <=
496 5.8.8, and older versions of perl make our live very difficult,
497 so write constants.pl during Makefile.PL time.
498
4995.25 Sat Mar 13 00:23:14 CET 2010
500 - fix a race condition in AnyEvent::Handle that would cause
501 a "bio_read: unsupported method" error (in _another_
502 ssl connection) after user code threw an exception.
503 - added AnyEvent::Handle->destroyed method.
504 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
505 - the AnyEvent::Util::fh_nonblocking function worked only
506 by lucky accident on win32.
507 - smaller and faster AnyEvent::Util::fh_nonblocking.
508 - when the (required!) Time::HiRes module is missing, AnyEvent
509 did not fall back to built-in time properly.
510 - do not load Fcntl at runtime, saving memory and loading time.
511 - precompile a number of constants and use them instead of runtime
512 detection and eval.
513 - free detection code after detection and similar memory optimisations.
514 - Perl backend timer interval best effort drift has been improved
515 (same algorithm as EV).
516 - update unicode idna mapping table.
517
5185.24 Tue Jan 5 11:39:43 CET 2010
519 - cygwin never reports errors from failed connects. 1.5 just gives
520 you continous readyness and EAGAIN, 1.7 is even more broken and
521 just hangs. work around both issues in a cygwin-specific hack.
522 - improve idle watcher documentation slightly.
523
5245.23 Sun Dec 20 23:48:00 CET 2009
525 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
526 - implement punycode_encode/decode, idn_nameprep,
527 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
528 - provide $AE::VERSION.
529 - removed traces of "no strict 'refs'".
530
5315.22 Sat Dec 5 03:51:13 CET 2009
532 - downgrade-or-fail in AnyEvent::Handle::push_write, to
533 diagnose encoding failures earlier and more succinctly.
534 (this works around bugs in perl, throwing away encoding info
535 when passing scalar data to extensions).
536 - add more examples to AnyEvent::Socket manpage.
537 - upgrade internal warning set to the same as common::sense 2.03.
538 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
539 (John Beppu).
540 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
541 - add google public dns servers to fallback server set and make sure
542 we load-balance properly between the three sets. also add all
543 fallback dns servers, not just a random one, to each dns config.
544
5455.21 Thu Nov 19 02:48:47 CET 2009
546 - fix a problem where socket constants were called with parameters
547 (spotted by David Friedland).
548 - fork_call never use'd POSIX (reported by Daisuke Maki).
549 - improve perl 5.6 compatibility further (but it still won't work
550 unless you rip out everything but the core).
551 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
552 uses 10MB of RAM, but doesn't open a console window. *sigh*.
553
5545.202 Wed Oct 14 22:35:44 CEST 2009
555 - AnyEvent::DNS would unexpectedly clobber $_ under windows
556 (analysed by Matthias Waldorf).
557 - AnyEvent::Handle::run_cmd can now pass the PID of the
558 newly-created process, which is much less useful than it might
559 sound (based on patch by Yann Kerherve).
560
5615.201 Tue Sep 29 12:09:25 CEST 2009
562 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
563 (reported by Torsten Foertsch).
564 - common::sense 2.0 could cause tcp_server to throw an exception
565 (analysed by elmex).
566
5675.2 Mon Sep 14 07:04:49 CEST 2009
568 - INCOMPATIBLE CHANGE: do no longer support register_read_type
569 and register_write_type in AnyEvent::Handle, instead support
570 package names (the facility was mostly abused).
571 - implement "packagename-as-read/write type" support in
572 AnyEvent::Handle.
573 - AnyEvent::Handle: new options "keepalive" and "oobinline".
574 - oobinline set by default to avoid security issues.
575 - the pure-perl event loop backend wrongly detected times() fallback
576 support (spotted by Pavel Boldin).
577
5785.12 Tue Sep 1 20:26:50 CEST 2009
579 - be more lenient when parsing resolv.conf files, as some people
580 use hashmarks as comment indicator inside directives (reported
581 by Michael S. Fischer).
582 - use same set of warnings as common::sense 2.0.
583 - fix a potential 32 bit overflow issue due to perl having problems
584 with large hex constants in the Perl backend.
585 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
586 by default.
587 - fix some prototypes.
588 - (optionally) require Async::Interrupt 1.02.
589 - replace opendns fallbacks by verizon/level3 public dns servers.
590
5915.112 Fri Aug 21 13:59:04 CEST 2009
592 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
593 using starttls with an illegal/uncreatable tls context.
594 - port to broken win32 perls.
595
5965.111 Sun Aug 16 18:44:36 CEST 2009
597 - perl errornously requires sockaddr_un structures returned by the
598 kernel to have a specific length. work around this bug by padding
599 sockaddr_un structures (found to fail on OS X, as perl only uses
600 the correct code for GNU/Linux, but likely to fail on other
601 OSes as well).
602
6035.11 Wed Aug 12 17:49:37 CEST 2009
604 - fixed a bug in run_cmd with input from filename.
605 - create a json coder object when none is given in write direction
606 to reduce dependency on JSON.
607
6085.1 Tue Aug 11 03:17:32 CEST 2009
609 - tcp_connect now reports the correct error in $!, not an unrelated
610 one (was broken since 4.91).
611 - AE::Handle did not properly clear rtimeout/wtimeout watchers
612 when retrying a tcp_connect.
613 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
614 - fixed wrong documentation in AE::Socket.
615
6165.01 Mon Aug 10 03:16:32 CEST 2009
617 - last release broke EV child watchers :/.
618
6195.0 Sun Aug 9 17:32:09 CEST 2009
620 - officially introduce and document the AE API.
621 - lowering the timeout at runtime did not have immediate effect
622 in AnyEvent::Handle.
623 - AE::Handle now has separate and independent read and write timeout
624 settings.
625 - probe for Guard module when AnyEvent::Util is loaded, not at
626 runtime (Event doesn't like eval "use" from callbacks, and it avoids
627 exporting the probe function).
628 - Impl::IOAsync's io watchers did not use the filehandle emulation
629 code to work around it's watcher limits.
630 - work around IO::Async's problems with overloaded objects.
631 - try to disable wearnings inside IO::Async, it's just too buggy.
632 - do no longer use Event by default, only EV or Perl, Event is too buggy.
633 - hackishly provide fast AE:: interface when EV or Perl is the backend.
634 - inherit VERSION from AnyEvent::VERSION in most modules.
635
6364.91 Thu Aug 6 15:42:45 CEST 2009
637 - AE::Handle::starttls could get out of sync when the read buffer
638 already contains some TLS handshake.
639 - AE::Handle did not properly free the TLS session, causing bigger
640 memory leaks in Net::SSLeay than expected (reported by toaster).
641 - AE::Socket::tcp_connect will now artificially delay invoking
642 the callback to avoid returning after invoking the callback.
643 - convert many internal watcher uses to AE API.
644
6454.9 Sat Aug 1 11:07:01 CEST 2009
646 - Glib silently fails when registering a timeout with
647 a negative value, so avoid that.
648 - call condvar callback immediately when it is set
649 after the condition is already signalled.
650 - check rbuf_max condition only after trying to consume
651 data.
652 - normalise signal numbers to names when using the ae's
653 signal handling, but do not document this (yet).
654 - pure perl signal emulation did not properly set nonblocking
655 mode on the signal pipe on win32 (this is unlikely to have caused
656 any issues).
657 - new module: AnyEvent::Debug.
658 - AnyEvent::Strict now checks that a signal was specified
659 by name (not name or number).
660 - reduce memory footprint in typical cases by ~50kb by
661 only compiling signal code when necessary.
662 - add AnyEvent::Handle->rbuf_max.
663 - grab the AE namespace for future new API, implement stubs for
664 said future/alternative API.
665 - new function: AnyEvent::Socket::format_hostport.
666
6674.881 Tue Jul 28 12:51:53 CEST 2009
668 - work around a bug in local in pre-5.10 perls, causing
669 AnyEvent::Handle to recurse when it should not
670 (analysed by elmex).
671
6724.88 Tue Jul 28 04:04:37 CEST 2009
673 - re-bless the handle into a dummy package after calling
674 AnyEvent::Handle::destroy, so the user does not need to check
675 for errors after every push_write etc.
676 - do not attempt to run t/02_signals.t on obviously broken
677 platforms.
678
6794.87 Sun Jul 26 02:06:16 CEST 2009
680 - do not attempt to linger when there is no longer a valid fh.
681 - enforce tls mode to be either accept or connect, do not simply
682 segfault in Net::SSLeay.
683 - AnyEvent::Handle can now call tcp_connect itself (new parameters
684 connect, on_prepare, on_connect and on_connect_error). Updated
685 tutorial accordingly.
686 - add AnyEvent::Impl::Irssi backend.
687
6884.86 Mon Jul 20 23:52:29 CEST 2009
689 - since the verbose warning is not enough, explicitly document
690 that versions before 1.33 of Net::SSLeay are not secure.
691 - work around signal handling races in Event and (...) Event::Lib.
692 - try to align signal-race timer to full-second boundaries.
693 - work around Tk not liking negative timeouts.
694 - don't complain of different grades of environmental unfriendlyness
695 in IO::Async.
696
6974.85 Sat Jul 18 06:16:14 CEST 2009
698 - nail the signal race problem in perl once and for all
699 (see $AnyEvent::MAX_SIGNAL_LATENCY).
700 - take advantage of Async::Interrupt if it is available.
701 - load Time::HiRes and Guard modules on demand only.
702 - add optional/recommended modules section to AnyEvent
703 documentation.
704 - reduce memory usage considerably (and reduce startup penalty)
705 by not using "strict", "warnings" and "overload" modules.
706 - work around buggy windows/openbsd perls and provide EBADMSG
707 and EPROTO ourselves when missing.
708 - improve perl 5.6 compatibility of the core event loop.
709 - made Net::SSLeay version 1.33 a soft requirement.
710
7114.83 Fri Jul 17 16:56:26 CEST 2009
712 - implement AnyEvent::Socket::getprotobyname.
713 - AnyEvent::CondVar's will now detect recursive blocking
714 waits and will croak, as too many people fall into
715 this trap.
716 - AnyEvent::Handle will now call ->destroy on itself after
717 executing the on_error callback, instead of doing some
718 half-baked internal shutdown, for fatal errors.
719 - clarify on_eof behaviour w.r.t. the read queue and
720 on_read callbacks.
721 - ignore some possible spurious wake-ups in tcp_connect.
722
7234.82 Sat Jul 11 00:34:55 CEST 2009
724 - POE and Event backends didn't accept some callable objects as
725 callbacks.
726 - use Config module instead of POSIX module to detect signal names
727 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
728 module doesn't even have all POSIX signals :/.
729 - use more workarounds around the many refcnt/corruption bugs in
730 Event::Lib.
731 - work around a race condition in perl's select, causing t/03_child.t
732 to rarely fail.
733
7344.81 Thu Jul 9 10:30:30 CEST 2009
735 - AnyEvent::Handle didn't properly diagnose write errors
736 (it expected -1 from syswrite, how lame... :).
737 - support file descriptors in addition to file handles
738 in AnyEvent->io.
739 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
740 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
741 and PERL_ANYEVENT_CA_PATH.
742 - provide a sensible synopsis section for AnyEvent::TLS.
743 - add a "supported backends" section to the manpage.
744 - added simple io watcher test to testsuite, using a
745 portable_socketpair.
746 - tried to improve the stability of the Event::Lib backend,
747 YMMV.
748
7494.8 Mon Jul 6 23:45:16 CEST 2009
750 - AnyEvent::DNS did not properly follow CNAME records with
751 uppercase targets.
752 - AnyEvent::DNS would errornously return AAAA records
753 with v4 mapped addresses (a faulty record) as ipv4 addresses,
754 causing AnyEvent::Socket to throw an exception.
755 - added new module AnyEvent::TLS for easier SSL/TLS context
756 creation, with many options including hostname verification,
757 secure default configuration, lots of documentation and,
758 predefined diffie-hellman keys for perfect forward security
759 and much more. get it while it's still fresh!
760 - use AnyEvent::TLS in AnyEvent::Handle for context management.
761 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
762 so AnyEvent::Socket users have smaller memory footprint
763 in the common case.
764 - add AnyEvent::Handle->push_shutdown method.
765 - add an additional error message parameter to AnyEvent::Handle's
766 on_error callback (for TLS, $! is still available).
767 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
768 - make AnyEvent::Handle more robust against early connection
769 failures (during new), and return C<undef> in that case
770 from the constructor.
771 - AnyEvent::Handle will now try to load only JSON::XS first,
772 then fall back to JSON.
773 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
774 and because it was documented that way.
775
7764.451 Fri Jul 3 00:28:58 CEST 2009
777 - do not clear rbuf when shutting down an AnyEvent::Handle
778 object - doing so breaks AnyEvent::HTTP.
779
7804.45 Mon Jun 29 22:59:26 CEST 2009
781 - a write error could cause AnyEvent::Handle to create
782 an I/O watcher with an undefined $fh.
783 - special-case mapped ipv4 addresses in both
784 AnyEvent::Socket::format_address and parse_address,
785 to treat them just like ipv4 addresses.
786 - updated and overhauled the AnyEvent::Intro doc.
787 - implement AnyEvent::Socket::format_ipv4/ipv6.
788 - slightly speed up portable_pipe/socketpair.
789 - expand condvar begin/end documentation.
790
7914.42 Fri Jun 26 08:32:18 CEST 2009
792 - preliminary and neccesarily incomplete support for IO::Async.
793 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
794 case it was set to IGNORE, to ensure we can catch child statuses
795 even when the calling env acts stupidly.
796 - updated benchmarks with IO::Async, which performs very well.
797
7984.412 Wed Jun 24 01:35:57 CEST 2009
799 - support an "untaint" attribute for AnyEvent::DNS
800 and set it on the default resolver.
801 - implement and document AnyEvent::Impl::Perl::loop.
802 - remove all anyevent-env variables from %ENV
803 when running in tainted mode.
804 - mention and extend the IO::Lambda benchmark.
805
8064.411 Sun Jun 7 18:48:02 CEST 2009
807 - do not try to use F_SETFD on windows.
808
8094.41 Thu May 14 06:40:11 CEST 2009
810 - work around issues in older perls (5.8.0?) when
811 a signal handler is deleted from the %SIG hash.
812 - use POSIX::_exit in child test, to avoid running
813 destructors.
814 - speed up CHLD handling by relying on SIGCHLD being
815 synchronously delivered, even when we roll our own
816 implementation.
817 - AnyEvent::DNS: add the "dname" resource record name
818 for cosmetic reasons.
819
8204.4 Sun Apr 26 20:12:33 CEST 2009
3TODO: - implemented idle watchers, where applicable. 821 - implemented idle watchers, where applicable.
4 - AnyEvent->time died when Event backend was in use. 822 - AnyEvent->time died when Event backend was in use.
5 - fix a memleak in the Tk backend. 823 - fix a memleak in the Tk backend.
6 - sped up Tk timer handling. 824 - sped up Tk timer handling.
7 - clip negative "after" values to 0 in AnyEvent::Impl::Event 825 - clip negative "after" values to 0 in AnyEvent::Impl::Event
8 to avoid spamming out warning messages. 826 to avoid spamming out warning messages.
9 - fix Qt timers without interval. 827 - fix Qt timers without interval.
10 - avoid Qt zero-timeout specialcase, in old libqt's. 828 - avoid Qt zero-timeout special case, in old libqt's.
11 829
124.352 Mon Apr 20 16:31:11 CEST 2009 8304.352 Mon Apr 20 16:31:11 CEST 2009
13 - fix AnyEvent::Strict error messages for child watchers. 831 - fix AnyEvent::Strict error messages for child watchers.
14 - fix/update Makefile.PL "recommends" versions. 832 - fix/update Makefile.PL "recommends" versions.
15 - add AnyEvent->now_update. 833 - add AnyEvent->now_update.
… …
794.3 Fri Oct 3 09:18:43 CEST 2008 8974.3 Fri Oct 3 09:18:43 CEST 2008
80 - AnyEvent will now install a no-op signal handler for SIGPIPE, 898 - AnyEvent will now install a no-op signal handler for SIGPIPE,
81 unless one has been installed already. 899 unless one has been installed already.
82 - warn about not loaidng AnyEvent::Impl::POE early enough 900 - warn about not loaidng AnyEvent::Impl::POE early enough
83 (patch by Adam Rosenstein). 901 (patch by Adam Rosenstein).
84 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 902 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
85 - fix a bug in where in SSL connect mode, the 903 - fix a bug in where in SSL connect mode, the
86 client would first wait for some data by the server and 904 client would first wait for some data by the server and
87 otherwise hang. (reported and analyzed in an absolutely 905 otherwise hang. (reported and analysed in an absolutely
88 exemplary manner by Adam Rosenstein). 906 exemplary manner by Adam Rosenstein).
89 - fix a bug in where SSL EOF would not be treated 907 - fix a bug in where SSL EOF would not be treated
90 as stream EOF, putting the connection into a hung state. 908 as stream EOF, putting the connection into a hung state.
91 (reported and analyzed in an absolutely exemplary manner by Adam 909 (reported and analysed in an absolutely exemplary manner by Adam
92 Rosenstein). 910 Rosenstein).
93 - fix a potential segfault when the TLS context 911 - fix a potential segfault when the TLS context
94 would go missing in a read callback (e.g. due to stoptls) - 912 would go missing in a read callback (e.g. due to stoptls) -
95 Net::SSLeay of course makes no type checking whatsoever. 913 Net::SSLeay of course makes no type checking whatsoever.
96 - AnyEvent::Handle will not stop the read watcher in TLS mode. 914 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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