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

Comparing AnyEvent/Changes (file contents):
Revision 1.90 by root, Fri May 23 22:52:31 2008 UTC vs.
Revision 1.502 by root, Thu Sep 1 22:09:25 2011 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
33.51 3TODO: document TCP_*** constants
4
5TODO: ae::stritc, regulalry check for invalid fds?
6TODO: critical messages log by default?
7TODO: document- boosted default logging level to critical.
8TOOD: stop internals abusing
9TODO: set loglevel to warn, even, as thats recommended?
10TODO: document default loglevel in table
11TODO: AE_LOG in tutorial?
12TODO: add google ipv6 public dns
13TODO: #TODO in Intro
14 - increase signal latency test timer considerably for poor little osx.
15
166.02 Fri Aug 26 20:08:31 CEST 2011
17 - log_to_syslog now accepts facility strings.
18 - log_to_syslog would not allow facilities - "$facility|$level"
19 *literally* means that string. sheesh.
20 - the CPAN parser got confused and thought we provide Sys::Syslog, try
21 to work around.
22 - renamed AnyEvent::Impl::FLTK2 to FLTK.
23
246.01 Fri Aug 26 07:04:11 CEST 2011
25 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
26 the TTL as fourth element - this affects ->resolve, ->request
27 and ->dns_unpack indirectly. Convenience functions such as
28 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
29 AnyEvent::Socket.
30 - INCOMPATIBLE CHANGE: backend authors now should not implement
31 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
32 methods.
33 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
34 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
35 - DEPRECIATION: please change
36 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
37 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
38 in your programs, for future compatibility.
39 - *blush* idle watcher emulation was borked.
40 - AnyEvent::Strict now checks the AE::xxx functions as well.
41 - querying too long domain names would barf deep inside AnyEvent::DNS - since
42 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
43 for too long domain names explicitly and passes an undef to the callback
44 instead (reported by James Bromberger).
45 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
46 - new functions: AnyEvent::postpone, AnyEvent::log.
47 - new module: AnyEvent::Log, for simple logging needs.
48 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
49 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
50 a debug shell.
51 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
52 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
53 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
54 F</etc/hosts> can be found.
55 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
56 configuration.
57 - env variables can now be specified using AE_xyz instead of
58 PERL_ANYEVENT_xyz, subject to some rules.
59 - lots of new AnyEvent::Debug functionality.
60 - AnyEvent::Strict now checks for modifications $_ by the callback
61 (by setting it to readonly while executing the callback, if possible).
62 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
63 Invest extra overhead to work around this peculiar behaviour.
64 - split the pure perl backend into a loop module and an impl
65 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
66 - some read types (regex, netstring, json, storable) could
67 errornously keep an AE::Handle object alive on their own
68 (reported by Mohammad Toossi).
69 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
70 supports /etc/hosts overrides.
71 - add more workarounds around child watcher bugs in IO::Async
72 and (newly introduced) in POE.
73 - use glib child watchers and try to work around its limitations,
74 if possible (based on a patch by Kevin Ryde).
75 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
76 - many read types in AnyEvent::Handle would malfunction during
77 transitions to TLS because of a "delete $self->{rbuf}".
78 (reported by Gerald Galster).
79 - improve condvar blocking wait performance (probably only noticable
80 with faster backends), also make it easier for event loops that
81 do not support blocking waits.
82 - PERL_ANYEVENT_MODEL now also allows full module names, although
83 the usefulness of this feature is not yet clear.
84 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
85 - add t/80_ssltext.t.
86 - no longer use AUTOLOAD for the initial loading (for no very good reason).
87 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
88 - do not run event loop detection multiple times when programs call their cached
89 copies of AnyEvent::detect.
90 - parse_hostport now accepts absolute paths als unix domain sockets.
91 - suppress the idiotic warning inside POE by patching the POE::Kernel
92 run flags instead of asking the user to jump through contortions.
93 - add Task::Weaken as dependency (as recommends only at this time).
94 - replace PF_UNSPEC by 0, for good old minix.
95
966.0 (not officially released, but was on CPAN for about 15 minutes :/)
97
985.34 Thu May 12 09:59:41 CEST 2011
99 - AE::TLS didn't support both cert_file and key at the same time,
100 due to some copy&paste bug (reported and analyzed by vti).
101 - improve AE::Handle timeout documentation slightly.
102
1035.33 Mon Feb 7 21:51:41 CET 2011
104 - never appeared on CPAN, for unknown reasons.
105 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
106 object when it's already gone (analyzed by Ryan Bullock).
107 - update idna algorithm(s) to UTS#46 version 6.0.0.
108 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
109 uts46 says it's an error, and testsuite makes more sense.
110 - remove leading dots in idn_to_ascii - testsuite says to remove,
111 uts46 says it's an error, and testsuite makes more sense.
112 - fixed a bug in idn ignored character processing.
113 - passes the uts46 6.0.0 testsuite except for three cases which
114 apparently are bugs in the testsuite itself.
115 - passes the uts46 6.0.1 testsuite except for eleven cases which
116 apparently are bugs in the testsuite itself.
117
1185.31 Sun Jan 23 11:49:19 CET 2011
119 - work around a POE bug where POE would first poll for new events and
120 invoke handlers BEFORE handling the message queue.
121 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
122 - stop_read now does just that in TLS mode. see its
123 documentation for caveats.
124 - update to common::sense 3.4 warning set (some warnings in perl impose
125 an arbitrary 32kb limit on regex matches, which affects
126 AnyEvent::Handle).
127 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
128 - new experimental AE::Handle::resettls method.
129
1305.3 Fri Dec 31 05:48:13 CET 2010
131 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
132 issues with IO::Async have been solved in that version, including
133 the availabilty of a default loop!
134 - implement new wbuf_max parameter for AnyEvent::Handle.
135 - added Cocoa::EventLoop backend by Daisuke Murase.
136 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
137 - clarify handle on_prepare documentation.
138 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
139 to avoid leaking them to any child processes.
140 - try to force files into the arch-dependend lib directory, for the
141 half-"turly obsessive compulsive person"(s) that depend on it.
142
1435.29 Sun Dec 5 10:49:21 CET 2010
144 - convert EV backend to EV 4.00 API (so better upgrade EV too).
145 - AE::Handle: implement dynamic read_size adjustment, add
146 max_read_size parameter, reduce default read_size to 2048.
147 - add some FAQ entries for servers and/or on_eof vs. on_error.
148 - work around OS bugs (cygwin again) when lingering in AE::Handle.
149 - reduce memory usage slightly when connect was used in AE::Handle.
150 - wrok around more segfaults in Tk.
151 - document tls_ctx => undef as valid.
152 - detect WNOHANG value at compiletime.
153 - include some probably linux-only support for building AnyEvent
154 as part of the perl core.
155 - improve compatibility of t/08_idna.t to perl 5.8.
156 - make AnyEvent::Strict truly optional (in case somebody builds a
157 minimal perl...).
158
1595.28 Wed Oct 13 04:14:23 CEST 2010
160 - due to a glitch, AnyEvent's internal getprotobyname wasn't
161 actually used in tcp_server.
162 - implement AnyEvent::CondVar->new method, which was documented but
163 missing.
164 - support autoloading of read types in unshift_read.
165 - AnyEvent::Strict: do the test for invalid fh arguments differently,
166 resulting in more informative error messages.
167 - new AnyEvent::FAQ manpage.
168 - clarify that you can actually call push_read etc. in on_read.
169 - put keys and certs in /tmp, not into the current directory,
170 and document this in the SECURITY CONSIDERATIONS sections
171 of AnyEvent::TLS.
172 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
173 requested by Ron Isaacson.
174 - document $AnyEvent::DNS::RESOLVER variable.
175 - incorporated major typo patches by Abhijit Menon-Sen.
176 - AnyEvent::Handle now croaks when negative timeouts are passed.
177 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
178 undocumented, but who knows when they might come in handy.
179
1805.271 Tue Jun 8 12:05:46 CEST 2010
181 - backport to perl 5.8.x.
182
1835.27 Sun Jun 6 12:12:05 CEST 2010
184 - postpone differently in AnyEvent::Socket now, as
185 when not, canceling the connection attempt might fail
186 (found by Felix Antonius Wilhelm Ostmann).
187 - explicitly check for non-stream sockets in AE::Handle, too many
188 clueless people fell into the trap of this somehow working.
189 - simplified and reworked the "OTHER MODULES" section.
190 - better/more condvar examples.
191
1925.261 Wed Apr 28 16:13:36 CEST 2010
193 - AF_INET6 was not properly used from Socket6 during configuration
194 time (found by Andy Grundman).
195
1965.26 Mon Apr 12 04:49:35 CEST 2010
197 - don't generate spurious readyness notifications when select
198 returns EINTR in the pure perl backend (analysed and testcase
199 by Bas Denissen).
200 - give same examples in SYNOPSIS sections of AnyEvent and AE.
201 - provider faster implementations for AE::time/now/now_update for
202 EV and Perl backends.
203 - provide faster AE::cv and AE::time implementations for all
204 backends.
205 - no longer support lower/mixed-case signal names (this was
206 never documented nor universally supported).
207 - some more memory µ-optimisations, and somewhat less messy
208 code to cope with both AnyEvent and AE APIs.
209
2105.251 Sat Mar 13 00:58:21 CET 2010
211 - make test failed if it was run before make install on perl <=
212 5.8.8, and older versions of perl make our live very difficult,
213 so write constants.pl during Makefile.PL time.
214
2155.25 Sat Mar 13 00:23:14 CET 2010
216 - fix a race condition in AnyEvent::Handle that would cause
217 a "bio_read: unsupported method" error (in _another_
218 ssl connection) after user code threw an exception.
219 - added AnyEvent::Handle->destroyed method.
220 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
221 - the AnyEvent::Util::fh_nonblocking function worked only
222 by lucky accident on win32.
223 - smaller and faster AnyEvent::Util::fh_nonblocking.
224 - when the (required!) Time::HiRes module is missing, AnyEvent
225 did not fall back to built-in time properly.
226 - do not load Fcntl at runtime, saving memory and loading time.
227 - precompile a number of constants and use them instead of runtime
228 detection and eval.
229 - free detection code after detection and similar memory optimisations.
230 - Perl backend timer interval best effort drift has been improved
231 (same algorithm as EV).
232 - update unicode idna mapping table.
233
2345.24 Tue Jan 5 11:39:43 CET 2010
235 - cygwin never reports errors from failed connects. 1.5 just gives
236 you continous readyness and EAGAIN, 1.7 is even more broken and
237 just hangs. work around both issues in a cygwin-specific hack.
238 - improve idle watcher documentation slightly.
239
2405.23 Sun Dec 20 23:48:00 CET 2009
241 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
242 - implement punycode_encode/decode, idn_nameprep,
243 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
244 - provide $AE::VERSION.
245 - removed traces of "no strict 'refs'".
246
2475.22 Sat Dec 5 03:51:13 CET 2009
248 - downgrade-or-fail in AnyEvent::Handle::push_write, to
249 diagnose encoding failures earlier and more succinctly.
250 (this works around bugs in perl, throwing away encoding info
251 when passing scalar data to extensions).
252 - add more examples to AnyEvent::Socket manpage.
253 - upgrade internal warning set to the same as common::sense 2.03.
254 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
255 (John Beppu).
256 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
257 - add google public dns servers to fallback server set and make sure
258 we load-balance properly between the three sets. also add all
259 fallback dns servers, not just a random one, to each dns config.
260
2615.21 Thu Nov 19 02:48:47 CET 2009
262 - fix a problem where socket constants were called with parameters
263 (spotted by David Friedland).
264 - fork_call never use'd POSIX (reported by Daisuke Maki).
265 - improve perl 5.6 compatibility further (but it still won't work
266 unless you rip out everything but the core).
267 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
268 uses 10MB of RAM, but doesn't open a console window. *sigh*.
269
2705.202 Wed Oct 14 22:35:44 CEST 2009
271 - AnyEvent::DNS would unexpectedly clobber $_ under windows
272 (analysed by Matthias Waldorf).
273 - AnyEvent::Handle::run_cmd can now pass the PID of the
274 newly-created process, which is much less useful than it might
275 sound (based on patch by Yann Kerherve).
276
2775.201 Tue Sep 29 12:09:25 CEST 2009
278 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
279 (reported by Torsten Foertsch).
280 - common::sense 2.0 could cause tcp_server to throw an exception
281 (analysed by elmex).
282
2835.2 Mon Sep 14 07:04:49 CEST 2009
284 - INCOMPATIBLE CHANGE: do no longer support register_read_type
285 and register_write_type in AnyEvent::Handle, instead support
286 package names (the facility was mostly abused).
287 - implement "packagename-as-read/write type" support in
288 AnyEvent::Handle.
289 - AnyEvent::Handle: new options "keepalive" and "oobinline".
290 - oobinline set by default to avoid security issues.
291 - the pure-perl event loop backend wrongly detected times() fallback
292 support (spotted by Pavel Boldin).
293
2945.12 Tue Sep 1 20:26:50 CEST 2009
295 - be more lenient when parsing resolv.conf files, as some people
296 use hashmarks as comment indicator inside directives (reported
297 by Michael S. Fischer).
298 - use same set of warnings as common::sense 2.0.
299 - fix a potential 32 bit overflow issue due to perl having problems
300 with large hex constants in the Perl backend.
301 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
302 by default.
303 - fix some prototypes.
304 - (optionally) require Async::Interrupt 1.02.
305 - replace opendns fallbacks by verizon/level3 public dns servers.
306
3075.112 Fri Aug 21 13:59:04 CEST 2009
308 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
309 using starttls with an illegal/uncreatable tls context.
310 - port to broken win32 perls.
311
3125.111 Sun Aug 16 18:44:36 CEST 2009
313 - perl errornously requires sockaddr_un structures returned by the
314 kernel to have a specific length. work around this bug by padding
315 sockaddr_un structures (found to fail on OS X, as perl only uses
316 the correct code for GNU/Linux, but likely to fail on other
317 OSes as well).
318
3195.11 Wed Aug 12 17:49:37 CEST 2009
320 - fixed a bug in run_cmd with input from filename.
321 - create a json coder object when none is given in write direction
322 to reduce dependency on JSON.
323
3245.1 Tue Aug 11 03:17:32 CEST 2009
325 - tcp_connect now reports the correct error in $!, not an unrelated
326 one (was broken since 4.91).
327 - AE::Handle did not properly clear rtimeout/wtimeout watchers
328 when retrying a tcp_connect.
329 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
330 - fixed wrong documentation in AE::Socket.
331
3325.01 Mon Aug 10 03:16:32 CEST 2009
333 - last release broke EV child watchers :/.
334
3355.0 Sun Aug 9 17:32:09 CEST 2009
336 - officially introduce and document the AE API.
337 - lowering the timeout at runtime did not have immediate effect
338 in AnyEvent::Handle.
339 - AE::Handle now has separate and independent read and write timeout
340 settings.
341 - probe for Guard module when AnyEvent::Util is loaded, not at
342 runtime (Event doesn't like eval "use" from callbacks, and it avoids
343 exporting the probe function).
344 - Impl::IOAsync's io watchers did not use the filehandle emulation
345 code to work around it's watcher limits.
346 - work around IO::Async's problems with overloaded objects.
347 - try to disable wearnings inside IO::Async, it's just too buggy.
348 - do no longer use Event by default, only EV or Perl, Event is too buggy.
349 - hackishly provide fast AE:: interface when EV or Perl is the backend.
350 - inherit VERSION from AnyEvent::VERSION in most modules.
351
3524.91 Thu Aug 6 15:42:45 CEST 2009
353 - AE::Handle::starttls could get out of sync when the read buffer
354 already contains some TLS handshake.
355 - AE::Handle did not properly free the TLS session, causing bigger
356 memory leaks in Net::SSLeay than expected (reported by toaster).
357 - AE::Socket::tcp_connect will now artificially delay invoking
358 the callback to avoid returning after invoking the callback.
359 - convert many internal watcher uses to AE API.
360
3614.9 Sat Aug 1 11:07:01 CEST 2009
362 - Glib silently fails when registering a timeout with
363 a negative value, so avoid that.
364 - call condvar callback immediately when it is set
365 after the condition is already signalled.
366 - check rbuf_max condition only after trying to consume
367 data.
368 - normalise signal numbers to names when using the ae's
369 signal handling, but do not document this (yet).
370 - pure perl signal emulation did not properly set nonblocking
371 mode on the signal pipe on win32 (this is unlikely to have caused
372 any issues).
373 - new module: AnyEvent::Debug.
374 - AnyEvent::Strict now checks that a signal was specified
375 by name (not name or number).
376 - reduce memory footprint in typical cases by ~50kb by
377 only compiling signal code when necessary.
378 - add AnyEvent::Handle->rbuf_max.
379 - grab the AE namespace for future new API, implement stubs for
380 said future/alternative API.
381 - new function: AnyEvent::Socket::format_hostport.
382
3834.881 Tue Jul 28 12:51:53 CEST 2009
384 - work around a bug in local in pre-5.10 perls, causing
385 AnyEvent::Handle to recurse when it should not
386 (analysed by elmex).
387
3884.88 Tue Jul 28 04:04:37 CEST 2009
389 - re-bless the handle into a dummy package after calling
390 AnyEvent::Handle::destroy, so the user does not need to check
391 for errors after every push_write etc.
392 - do not attempt to run t/02_signals.t on obviously broken
393 platforms.
394
3954.87 Sun Jul 26 02:06:16 CEST 2009
396 - do not attempt to linger when there is no longer a valid fh.
397 - enforce tls mode to be either accept or connect, do not simply
398 segfault in Net::SSLeay.
399 - AnyEvent::Handle can now call tcp_connect itself (new parameters
400 connect, on_prepare, on_connect and on_connect_error). Updated
401 tutorial accordingly.
402 - add AnyEvent::Impl::Irssi backend.
403
4044.86 Mon Jul 20 23:52:29 CEST 2009
405 - since the verbose warning is not enough, explicitly document
406 that versions before 1.33 of Net::SSLeay are not secure.
407 - work around signal handling races in Event and (...) Event::Lib.
408 - try to align signal-race timer to full-second boundaries.
409 - work around Tk not liking negative timeouts.
410 - don't complain of different grades of environmental unfriendlyness
411 in IO::Async.
412
4134.85 Sat Jul 18 06:16:14 CEST 2009
414 - nail the signal race problem in perl once and for all
415 (see $AnyEvent::MAX_SIGNAL_LATENCY).
416 - take advantage of Async::Interrupt if it is available.
417 - load Time::HiRes and Guard modules on demand only.
418 - add optional/recommended modules section to AnyEvent
419 documentation.
420 - reduce memory usage considerably (and reduce startup penalty)
421 by not using "strict", "warnings" and "overload" modules.
422 - work around buggy windows/openbsd perls and provide EBADMSG
423 and EPROTO ourselves when missing.
424 - improve perl 5.6 compatibility of the core event loop.
425 - made Net::SSLeay version 1.33 a soft requirement.
426
4274.83 Fri Jul 17 16:56:26 CEST 2009
428 - implement AnyEvent::Socket::getprotobyname.
429 - AnyEvent::CondVar's will now detect recursive blocking
430 waits and will croak, as too many people fall into
431 this trap.
432 - AnyEvent::Handle will now call ->destroy on itself after
433 executing the on_error callback, instead of doing some
434 half-baked internal shutdown, for fatal errors.
435 - clarify on_eof behaviour w.r.t. the read queue and
436 on_read callbacks.
437 - ignore some possible spurious wake-ups in tcp_connect.
438
4394.82 Sat Jul 11 00:34:55 CEST 2009
440 - POE and Event backends didn't accept some callable objects as
441 callbacks.
442 - use Config module instead of POSIX module to detect signal names
443 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
444 module doesn't even have all POSIX signals :/.
445 - use more workarounds around the many refcnt/corruption bugs in
446 Event::Lib.
447 - work around a race condition in perl's select, causing t/03_child.t
448 to rarely fail.
449
4504.81 Thu Jul 9 10:30:30 CEST 2009
451 - AnyEvent::Handle didn't properly diagnose write errors
452 (it expected -1 from syswrite, how lame... :).
453 - support file descriptors in addition to file handles
454 in AnyEvent->io.
455 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
456 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
457 and PERL_ANYEVENT_CA_PATH.
458 - provide a sensible synopsis section for AnyEvent::TLS.
459 - add a "supported backends" section to the manpage.
460 - added simple io watcher test to testsuite, using a
461 portable_socketpair.
462 - tried to improve the stability of the Event::Lib backend,
463 YMMV.
464
4654.8 Mon Jul 6 23:45:16 CEST 2009
466 - AnyEvent::DNS did not properly follow CNAME records with
467 uppercase targets.
468 - AnyEvent::DNS would errornously return AAAA records
469 with v4 mapped addresses (a faulty record) as ipv4 addresses,
470 causing AnyEvent::Socket to throw an exception.
471 - added new module AnyEvent::TLS for easier SSL/TLS context
472 creation, with many options including hostname verification,
473 secure default configuration, lots of documentation and,
474 predefined diffie-hellman keys for perfect forward security
475 and much more. get it while it's still fresh!
476 - use AnyEvent::TLS in AnyEvent::Handle for context management.
477 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
478 so AnyEvent::Socket users have smaller memory footprint
479 in the common case.
480 - add AnyEvent::Handle->push_shutdown method.
481 - add an additional error message parameter to AnyEvent::Handle's
482 on_error callback (for TLS, $! is still available).
483 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
484 - make AnyEvent::Handle more robust against early conenction
485 failures (during new), and return C<undef> in that case
486 from the constructor.
487 - AnyEvent::Handle will now try to load only JSON::XS first,
488 then fall back to JSON.
489 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
490 and because it was documented that way.
491
4924.451 Fri Jul 3 00:28:58 CEST 2009
493 - do not clear rbuf when shutting down an AnyEvent::Handle
494 object - doing so breaks AnyEvent::HTTP.
495
4964.45 Mon Jun 29 22:59:26 CEST 2009
497 - a write error could cause AnyEvent::Handle to create
498 an I/O watcher with an undefined $fh.
499 - special-case mapped ipv4 addresses in both
500 AnyEvent::Socket::format_address and parse_address,
501 to treat them just like ipv4 addresses.
502 - updated and overhauled the AnyEvent::Intro doc.
503 - implement AnyEvent::Socket::format_ipv4/ipv6.
504 - slightly speed up portable_pipe/socketpair.
505 - expand condvar begin/end documentation.
506
5074.42 Fri Jun 26 08:32:18 CEST 2009
508 - preliminary and neccesarily incomplete support for IO::Async.
509 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
510 case it was set to IGNORE, to ensure we can catch child statuses
511 even when the calling env acts stupidly.
512 - updated benchmarks with IO::Async, which performs very well.
513
5144.412 Wed Jun 24 01:35:57 CEST 2009
515 - support an "untaint" attribute for AnyEvent::DNS
516 and set it on the default resolver.
517 - implement and document AnyEvent::Impl::Perl::loop.
518 - remove all anyevent-env variables from %ENV
519 when running in tainted mode.
520 - mention and extend the IO::Lambda benchmark.
521
5224.411 Sun Jun 7 18:48:02 CEST 2009
523 - do not try to use F_SETFD on windows.
524
5254.41 Thu May 14 06:40:11 CEST 2009
526 - work around issues in older perls (5.8.0?) when
527 a signal handler is deleted from the %SIG hash.
528 - use POSIX::_exit in child test, to avoid running
529 destructors.
530 - speed up CHLD handling by relying on SIGCHLD being
531 synchronously delivered, even when we roll our own
532 implementation.
533 - AnyEvent::DNS: add the "dname" resource record name
534 for cosmetic reasons.
535
5364.4 Sun Apr 26 20:12:33 CEST 2009
537 - implemented idle watchers, where applicable.
538 - AnyEvent->time died when Event backend was in use.
539 - fix a memleak in the Tk backend.
540 - sped up Tk timer handling.
541 - clip negative "after" values to 0 in AnyEvent::Impl::Event
542 to avoid spamming out warning messages.
543 - fix Qt timers without interval.
544 - avoid Qt zero-timeout special case, in old libqt's.
545
5464.352 Mon Apr 20 16:31:11 CEST 2009
547 - fix AnyEvent::Strict error messages for child watchers.
548 - fix/update Makefile.PL "recommends" versions.
549 - add AnyEvent->now_update.
550
5514.351 Sat Apr 11 07:56:14 CEST 2009
552 - actually make the signal pipe work under win32.
553 - localise $! in signal handler to avoid changing $!, although
554 perl itself does not seem to save/restore errno either.
555 - set the cloexec flag on the signal pipe (normally set by perl too).
556
5574.35 Fri Mar 27 11:48:20 CET 2009
558 - event models relying on AnyEvent's signal watcher emulation did
559 invoke the callback asynchronously, contradicting documentation
560 and causing signals to get lost (this includes AnyEvent's own
561 event loop). AnyEvent now uses the standard pipe trick to make
562 callback execution synchronous to the event loop.
563 - AnyEvent::Handle didn't free TLS context data on DESTROY
564 (patch by Pavel Shaydo).
565 - work around the ever-incompatibly-changing API of MakeMaker.
566 - document that changing global variables without restoring them
567 is a bad idea in Perl (noted by Adam Rosenstein).
568 - AnyEvent::Strict now barfs if ->io is passed a file.
569
5704.34 Thu Feb 12 18:32:45 CET 2009
571 - separately buffer TLS read data, as otherwise the read queue could
572 deadlock as receiving data is not expected while draining the
573 read queue (which cna only happen with TLS).
574 - raise EBADMSG error on JSON decoding errors.
575 - fix some minor manpage bugs (reported by Maximilian Gaß).
576 - speed up select bitmask parsing quite a bit in the pure perl
577 backend.
578 - use CORE::select instead of just select in AnyEvent::Impl::Perl.
579
5804.331 Tue Jan 6 21:07:25 CET 2009
581 - socketpair fails on many vista machines because vista has
582 completely broken accept/getpeername and getsockname functions,
583 so we provide our own socketpair emulation that kind of works
584 (AnyEvent::Util::portable_pipe).
585 - new function: AnyEvent::Util::portable_socketpair.
586 - take advantage of the Guard module if it exists.
587
5884.33 Fri Nov 21 02:35:40 CET 2008
589 - AnyEvent::Strict did errornously flag a fileno of 0 as
590 illegal.
591 - reduce memory usage and slightly speed up the pure perl backend
592 by only storing the file descriptor, not the file handle.
593 - add missing autocork method to AnyEvent::Handle (reported
594 by Adam Rosenstein).
595 - AnyEvent::DNS->resolve errornously documented an $rcode result
596 argument, but there isn't (spotted by Henrik Krohns).
597 - the naked truth about Tk - it's basically unmaintained.
598
5994.32 Mon Nov 3 22:46:32 CET 2008
600 - fix AnyEvent::Socket::resolve_sockaddr to properly support
601 unix sockets again and choose a proper default for
602 the socket type.
603 - fix call to resolve_sockaddr in tcp_connect.
604
6054.31 Thu Oct 30 04:41:48 CET 2008
606 - implemented AnyEvent::Handle->destroy method.
607 - hint about unexpected effects in TLS mode.
608 - speed up AnyEvent::Impl::Perl by using more arrays
609 instead of hashes without the slightest loss of readability :)
610 - work around a perl argument refcounting bug.
611 - with some perl patches applied, AnyEvent now seems leak-free.
612
6134.3 Fri Oct 3 09:18:43 CEST 2008
614 - AnyEvent will now install a no-op signal handler for SIGPIPE,
615 unless one has been installed already.
616 - warn about not loaidng AnyEvent::Impl::POE early enough
617 (patch by Adam Rosenstein).
618 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
619 - fix a bug in where in SSL connect mode, the
620 client would first wait for some data by the server and
621 otherwise hang. (reported and analysed in an absolutely
622 exemplary manner by Adam Rosenstein).
623 - fix a bug in where SSL EOF would not be treated
624 as stream EOF, putting the connection into a hung state.
625 (reported and analysed in an absolutely exemplary manner by Adam
626 Rosenstein).
627 - fix a potential segfault when the TLS context
628 would go missing in a read callback (e.g. due to stoptls) -
629 Net::SSLeay of course makes no type checking whatsoever.
630 - AnyEvent::Handle will not stop the read watcher in TLS mode.
631 - AnyEvent::Handle->stoptls will send a TLS close notify
632 instead of simply destroying the stream now.
633 - fix a bug where large blocks of data written in TLS mode
634 would not be sent unless triggered by receives.
635 - on_drain will now take the tls write buffer into account.
636 - SSL operations have been streamlined a bit and should be faster
637 now, more to come, though.
638 - added a FAQ section, will document nontrivial issues in there.
639
6404.234 Mon Sep 29 04:08:13 CEST 2008
641 - fix child watcher documentation: only child exits will be reported,
642 no trace events.
643 - mention SIGPIPE in AnyEvent::Handle.
644 - perl backend will now use POSIX::times as monotonic clock
645 when available and the monotonic clock option is not.
646 - run even without Time::HiRes in some cases now, but do not
647 provide sub-second accuracy (all *supported* perl versions have
648 Time::HiRes).
649
6504.233 Fri Aug 22 01:48:05 CEST 2008
651 - fix a bug in the testsuite, causing a hang.
652 - clarified lots of AnyEvent::Handle constructor arguments.
653
6544.232 Thu Aug 21 20:44:25 CEST 2008
655 - fix a bug in AnyEvent::Handle that could cause two fatal error
656 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
657 (testcase by Leon Brocard).
658 - clarified on_eof/on_error documentation a bit.
659
6604.231 Tue Jul 29 13:12:15 CEST 2008
661 - remove some debugging code left in AnyEvent::Util::fork_call
662 (and no, it's impossible to implement with the broken windows
663 perls without resource leaks or worse).
664
6654.23 Tue Jul 29 12:19:59 CEST 2008
666 - document the first parameter passed to condvar callbacks to be
667 the callback.
668 - add AnyEvent::Socket::{ntoa,aton} aliases.
669 - optimize the AE::Handle->push_read (line) for the default
670 eol marker.
671 - optimize push_read (packstring|storable) for small packets.
672 - invoke on_error callback when no on_eof callback is set.
673 - fix a bug in push_read (storable) of unknown impact.
674
6754.22 Sun Jul 20 16:34:13 CEST 2008
676 - new function AnyEvent::Socket::parse_hostport.
677 - as the bulkheads at microsoft can't even get getprotobyname reliably
678 working on their shitty fucking broken os we need to hardcode
679 some common protocol numbers in AnyEvent::Socket. How can
680 people even bother with such a pile of shit as windows.
681
6824.21 Thu Jul 17 14:40:05 CEST 2008
683 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
684 which is incompatible to previous invocation syntax.
685 - work around more windows perl bugs in fork_call. I cannot imagine
686 why anybody would bother implementing fork in such an obviously
687 unusable way.
688 - avoid calling $do_retry if the request has already finished in
689 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
690 K. Sharifulin).
691
6924.2 Sat Jul 12 22:42:11 CEST 2008
693 - support an "interval" argument for repeating timers.
694 - fix a bug in the pure perl event loop that caused timers to
695 be delayed under some circumstances.
696 - implement optional argument checking via PERL_ANYEVENT_STRICT.
697 - not importing anything from Event causes it to malfunction
698 and not use Time::HiRes, work around this bug.
699 - more bug workarounds against the endlessly broken Tk module,
700 seems to work now on 32 bit machines, but less so on 64 bit.
701 - minor optimisations applied to most backends.
702
7034.161 Fri Jul 4 14:24:48 CEST 2008
704 - fixed the t/handle/02_write.t, it read after write, with the
705 assumption that all data could be written before read
706 (reported by Jim Razmus).
707
7084.160 Thu Jul 3 04:02:21 CEST 2008
709 - re-registering signal watchers after unregistering one did
710 not work if a backend falls back on AnyEvent's default
711 implementation (also affected child watchers).
712 - new AnyEvent::Handle options: autocork and no_delay.
713
7144.152 Sun Jun 22 14:15:44 CEST 2008
715 - allow for 32-bit perls that implement shifts differently
716 on different architectures in parse_ipv4 (reported and
717 analysed by Keiichi DAIBA).
718
7194.151 Fri Jun 6 17:34:24 CEST 2008
720 - make sure specifying _only_ on_read and never pushing reads
721 works.
722
7234.15 Fri Jun 6 13:00:46 CEST 2008
724 - the pure perl backend would keep some watchers alive when more than
725 one watcher was registered for the same fd.
726 - new "packstring" and "storable" read and write types
727 for AnyEvent::Handle.
728 - allow on_eof handler to be called after on_error with EPIPE returns.
729 - do not immediately call on_read callback in handle constructor.
730
7314.14 Thu Jun 5 20:29:31 CEST 2008
732 - Fixed a bug in DNS SRV priority sorting.
733 - AnyEvent::Util::guard now reports runtime errors while
734 executing the guard block as warnings.
735 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
736
7374.13 Thu Jun 5 00:47:59 CEST 2008
738 - AnyEvent::DNS only followed cname chains with length 2,
739 contrary to documentation. bumped it up to 10, thanks to
740 microsoft, the current broken-dns-config-king.
741 - AnyEvent::DNS didn't check for socket return status
742 but instead relied on perl not creating filehandles in that
743 case - too bad it gives you a bogus file handle (reported
744 and analysed by Vladimir Timofeev).
745 - fix queue management logic in AnyEvent::Handle:
746 when on_read was registered, the queue was empty
747 and no progress could be made AnyEvent::Handle would
748 enter an endless loop.
749 - correctly start reading again when the handle became
750 busy again after idling.
751 - correctly treat tls shut-downs as EOF condition
752 (for the time being, would be nice to have a callback for that).
753 - correctly call eof callback under all conditions
754 (hopefully).
755 - the timeout callback did not expect that $self can go
756 away any time.
757
7584.12 Tue Jun 3 10:58:04 CEST 2008
759 - include AnyEvent::Intro, a tutorial for anyevent,
760 anyevent::socket and anyevent::handle.
761 - allow more options in on_error.
762
7634.11 Fri May 30 23:42:25 CEST 2008
764 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
765 reverse_lookup and reverse_verify to replace it, support
766 v4mapped and v4compat addresses.
767 - provide more documentation for the resolver class.
768 - really replace longest run of :0: by :: in format_address,
769 also properly convert :: and ::1 again.
770 - support NAPTR record name and decode it.
771 - implement random weight sampling for SRV records, as per
772 rfc 2782.
773 - correctly abort on srv-record targets of ".".
774 - added AnyEvent::DNS::wait_for_slot.
775 - in the unlikely event of a virtual circuit connection
776 being invalidated by a delayed udp reply, AnyEvent::DNS
777 could die. this has been fixed.
778 - plug a probable memleak in the DNS vc code.
779 - use configured timeout also for VC request phase.
780 - implement timeout and max_outstanding methods
781 for resolver class.
782 - update version numbers in all modules.
783
7844.1 Thu May 29 05:45:40 CEST 2008
785 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
786 which mirrors their purpose better (old names still available
787 for a while). Also moved AnyEvent::DNS::addr to
788 AnyEvent::Socket::resolve_sockaddr.
789 - implement AnyEvent->time and AnyEvent->now.
790 - fix IPv6 support in pack_sockaddr.
791 - officially un-experimentalise all newly introduced code.
792 - support unix domain sockets everywhere by specifying
793 a host of "unix/" and the pathname as service.
794 - implement an activity timeout in AnyEvent::Handle.
795 - added a regex read type to AnyEvent::Handle.
796 - added a json read/write type to AnyEvent::Handle.
797 - always croak in on_error in AnyEvent::Handle.
798 - document how to subclass AnyEvent::Handle.
799 - implement AnyEvent::Util::fork_call.
800 - add support for IPv6 nameservers and nameserver statements.
801 - work around _yet_ _another_ windows perl bug in where empty
802 select masks cause errors in the pure perl backend (this
803 workaround cannot be implemented for the other event loops,
804 but EV also works around this).
805 - supply AnyEvent::Socket::address_type and make good use of it.
806 - clarify and fix the Handle documentation w.r.t $self vs. $handle
807 vs. callback arguments.
808 - add some recommends to the META.yml.
809
8104.05 Mon May 26 19:44:06 CEST 2008
811 - some platforms ignore the rfc and prepend an extra sa_len member to
812 the sockaddr structure, cater for those. also use sockaddr_family,
813 didn't know it was there, but comes in handy for the workaround.
814 - undo the import WIN32 hack.
815
8164.04 Mon May 26 08:03:31 CEST 2008
817 - try to work around yet another windows bug: failed connects are reported
818 as if it were out-of-band data. windows users: you suck. Only
819 EV, Glib, Event and the pure perl backend can handle this condition.
820 - optimize write algorithm in AnyEvent::Handle.
821 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
822 - tcp_connect now properly iterates through all targets.
823 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
824 support entirely. also hardcode some AF_INET6 constants for the
825 sake of perl 5.8.
826 - vastly improved nameserver/suffix detection algorithm for windows
827 (still a hack...)
828 - try to use a fallback dns server if no dns servers could be found.
829 - splatter around some local $SIG{__DIE__}'s to work around bad
830 die handlers.
831
8324.03 Sun May 25 05:05:57 CEST 2008
833 - work around perl bugs on windows, where perl returns undocumented
834 error codes for sysread, syswrite etc.
835 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
836 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
837 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
838 - updated documentation.
839 - lots of workarounds for perl 5.8 quirks.
840 - implement a more flexible type system for both reads and writes
841 in AnyEvent::Handle.
842 - support IPv6 in tcp_server.
843 - pass local hostname and port to the prepare callback in tcp_server.
844 - make types extendable in AnyEvent::Handle.
845 - croak, not die, in AnyEvent::Handle;
846
8474.0 Sat May 24 19:58:08 CEST 2008
4 - added (experimental) AnyEvent::DNS module. 848 - added (experimental) AnyEvent::DNS module.
5 - added (experimental) AnyEvent::Socket module. 849 - added (experimental) AnyEvent::Socket module.
850 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
6 - make condvar->end call ->send as documented. 851 - make condvar->end call ->send as documented.
7 - make condvar constructor support the documented cb => argument. 852 - make condvar constructor support the documented cb => argument.
853 - overload condvars so you can use them instead of code references.
8 - support more of the POSIX forms instead of a pure dooted quad for 854 - support more of the POSIX forms instead of a pure dotted quad for
9 verbatim ip addresses. 855 verbatim ip addresses.
10 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. 856 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
11 - try very hard not to freeze on broken windows perls. 857 - try very hard not to freeze on broken windows perls.
12 - make on_eof optional in AnyEvent::Handle. 858 - make on_eof optional in AnyEvent::Handle.
13 - support service names in tcp_connect and tcp_server. 859 - support service names in tcp_connect and tcp_server.
14 - work around netbsd/openbsd bug causing unix sockets not to 860 - work around netbsd/openbsd bug causing unix sockets not to
15 become writable untill empty. 861 become writable until empty.
16 - only make a dummy read for error detection when getpeername 862 - only make a dummy read for error detection when getpeername
17 returns ENOTCONN. 863 returns ENOTCONN.
18 - defend AnyEvent::Util::inet_aton against Coro::LWP brutal 864 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
19 override. 865 override.
866 - new AnyEvent::Guard::cancel.
20 867
213.5 Sat May 17 23:17:14 CEST 2008 8683.5 Sat May 17 23:17:14 CEST 2008
22 - gracefully deal with systems not defining CLOCK_MONOTONIC 869 - gracefully deal with systems not defining CLOCK_MONOTONIC
23 (as opposed to not supporting the clock itself, which worked in 3.41). 870 (as opposed to not supporting the clock itself, which worked in 3.41).
24 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server 871 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
… …
36 883
373.41 Sun May 11 19:53:13 CEST 2008 8843.41 Sun May 11 19:53:13 CEST 2008
38 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. 885 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
39 886
403.4 Sun May 11 00:29:25 CEST 2008 8873.4 Sun May 11 00:29:25 CEST 2008
41 - complete re-implementation of condvars with a hopefully much mroe useful 888 - complete re-implementation of condvars with a hopefully much more useful
42 API (backward compatible functions exist but are undocumented). 889 API (backward compatible functions exist but are undocumented).
43 - AnyEvent::Handle fully rewritten with a hopefully more powerful API 890 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
44 (still subject to change). 891 (still subject to change).
45 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. 892 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
46 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which 893 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which

Diff Legend

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