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

Comparing AnyEvent/Changes (file contents):
Revision 1.314 by root, Fri Jul 31 07:57:43 2009 UTC vs.
Revision 1.516 by root, Tue Oct 4 15:29:00 2011 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
34.9 3TODO: document TCP_*** constants
4
5TODO: add debug/trace logging to some modules
6 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
7 and above), and some messages inside AnyEvent have been elevated
8 to higher log levels to print by default.
9 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
10 - work around yet another signal race bug, newly introduced by POE. My
11 how that module sucks.
12 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
13 support ipv6-only operation even with defaults (also use ::1 instead
14 of 127.0.0.1).
15 - give better guidelines on how to chose a logging level.
16 - AnyEvent::Log can now cap (limit) log message priority by context.
17 - AE::log will now save and restore $! and $@, so callers don't have to.
18 - AE::Strict now checks that registered fds are still valid at regular
19 intervals.
20 - skip frequently called 'require' calls if the module appears loaded.
21 - increase signal latency test timer considerably for poor little osx,
22 which can't be bothered with basic posix functionality such as signal
23 delivery.
24 - mention some debugging aids in the tutorial.
25 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
26 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
27 of the context as title.
28 - AnyEvent::Strict would not restore read-write access to $_ when a callback
29 died - this is a user bug, but since AE::Strict exists to help dbeugging...
30 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
31 to work with it - the author is well aware of the issues but refuses
32 to fix them. This extreme step has been taken because the module actively
33 attacks the goals of AnyEvent and makes it harder for module authors to use
34 AnyEvent.
35
366.02 Fri Aug 26 20:08:31 CEST 2011
37 - log_to_syslog now accepts facility strings.
38 - log_to_syslog would not allow facilities - "$facility|$level"
39 *literally* means that string. sheesh.
40 - the CPAN parser got confused and thought we provide Sys::Syslog, try
41 to work around.
42 - renamed AnyEvent::Impl::FLTK2 to FLTK.
43
446.01 Fri Aug 26 07:04:11 CEST 2011
45 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
46 the TTL as fourth element - this affects ->resolve, ->request
47 and ->dns_unpack indirectly. Convenience functions such as
48 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
49 AnyEvent::Socket.
50 - INCOMPATIBLE CHANGE: backend authors now should not implement
51 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
52 methods.
53 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
54 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
55 - DEPRECIATION: please change
56 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
57 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
58 in your programs, for future compatibility.
59 - *blush* idle watcher emulation was borked.
60 - AnyEvent::Strict now checks the AE::xxx functions as well.
61 - querying too long domain names would barf deep inside AnyEvent::DNS - since
62 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
63 for too long domain names explicitly and passes an undef to the callback
64 instead (reported by James Bromberger).
65 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
66 - new functions: AnyEvent::postpone, AnyEvent::log.
67 - new module: AnyEvent::Log, for simple logging needs.
68 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
69 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
70 a debug shell.
71 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
72 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
73 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
74 /etc/hosts can be found.
75 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
76 configuration.
77 - env variables can now be specified using AE_xyz instead of
78 PERL_ANYEVENT_xyz, subject to some rules.
79 - lots of new AnyEvent::Debug functionality.
80 - AnyEvent::Strict now checks for modifications $_ by the callback
81 (by setting it to readonly while executing the callback, if possible).
82 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
83 Invest extra overhead to work around this peculiar behaviour.
84 - split the pure perl backend into a loop module and an impl
85 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
86 - some read types (regex, netstring, json, storable) could
87 errornously keep an AE::Handle object alive on their own
88 (reported by Mohammad Toossi).
89 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
90 supports /etc/hosts overrides.
91 - add more workarounds around child watcher bugs in IO::Async
92 and (newly introduced) in POE.
93 - use glib child watchers and try to work around its limitations,
94 if possible (based on a patch by Kevin Ryde).
95 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
96 - many read types in AnyEvent::Handle would malfunction during
97 transitions to TLS because of a "delete $self->{rbuf}".
98 (reported by Gerald Galster).
99 - improve condvar blocking wait performance (probably only noticable
100 with faster backends), also make it easier for event loops that
101 do not support blocking waits.
102 - PERL_ANYEVENT_MODEL now also allows full module names, although
103 the usefulness of this feature is not yet clear.
104 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
105 - add t/80_ssltext.t.
106 - no longer use AUTOLOAD for the initial loading (for no very good reason).
107 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
108 - do not run event loop detection multiple times when programs call their cached
109 copies of AnyEvent::detect.
110 - parse_hostport now accepts absolute paths als unix domain sockets.
111 - suppress the idiotic warning inside POE by patching the POE::Kernel
112 run flags instead of asking the user to jump through contortions.
113 - add Task::Weaken as dependency (as recommends only at this time).
114 - replace PF_UNSPEC by 0, for good old minix.
115
1166.0 (not officially released, but was on CPAN for about 15 minutes :/)
117
1185.34 Thu May 12 09:59:41 CEST 2011
119 - AE::TLS didn't support both cert_file and key at the same time,
120 due to some copy&paste bug (reported and analyzed by vti).
121 - improve AE::Handle timeout documentation slightly.
122
1235.33 Mon Feb 7 21:51:41 CET 2011
124 - never appeared on CPAN, for unknown reasons.
125 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
126 object when it's already gone (analyzed by Ryan Bullock).
127 - update idna algorithm(s) to UTS#46 version 6.0.0.
128 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
129 uts46 says it's an error, and testsuite makes more sense.
130 - remove leading dots in idn_to_ascii - testsuite says to remove,
131 uts46 says it's an error, and testsuite makes more sense.
132 - fixed a bug in idn ignored character processing.
133 - passes the uts46 6.0.0 testsuite except for three cases which
134 apparently are bugs in the testsuite itself.
135 - passes the uts46 6.0.1 testsuite except for eleven cases which
136 apparently are bugs in the testsuite itself.
137
1385.31 Sun Jan 23 11:49:19 CET 2011
139 - work around a POE bug where POE would first poll for new events and
140 invoke handlers BEFORE handling the message queue.
141 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
142 - stop_read now does just that in TLS mode. see its
143 documentation for caveats.
144 - update to common::sense 3.4 warning set (some warnings in perl impose
145 an arbitrary 32kb limit on regex matches, which affects
146 AnyEvent::Handle).
147 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
148 - new experimental AE::Handle::resettls method.
149
1505.3 Fri Dec 31 05:48:13 CET 2010
151 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
152 issues with IO::Async have been solved in that version, including
153 the availabilty of a default loop!
154 - implement new wbuf_max parameter for AnyEvent::Handle.
155 - added Cocoa::EventLoop backend by Daisuke Murase.
156 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
157 - clarify handle on_prepare documentation.
158 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
159 to avoid leaking them to any child processes.
160 - try to force files into the arch-dependend lib directory, for the
161 half-"turly obsessive compulsive person"(s) that depend on it.
162
1635.29 Sun Dec 5 10:49:21 CET 2010
164 - convert EV backend to EV 4.00 API (so better upgrade EV too).
165 - AE::Handle: implement dynamic read_size adjustment, add
166 max_read_size parameter, reduce default read_size to 2048.
167 - add some FAQ entries for servers and/or on_eof vs. on_error.
168 - work around OS bugs (cygwin again) when lingering in AE::Handle.
169 - reduce memory usage slightly when connect was used in AE::Handle.
170 - wrok around more segfaults in Tk.
171 - document tls_ctx => undef as valid.
172 - detect WNOHANG value at compiletime.
173 - include some probably linux-only support for building AnyEvent
174 as part of the perl core.
175 - improve compatibility of t/08_idna.t to perl 5.8.
176 - make AnyEvent::Strict truly optional (in case somebody builds a
177 minimal perl...).
178
1795.28 Wed Oct 13 04:14:23 CEST 2010
180 - due to a glitch, AnyEvent's internal getprotobyname wasn't
181 actually used in tcp_server.
182 - implement AnyEvent::CondVar->new method, which was documented but
183 missing.
184 - support autoloading of read types in unshift_read.
185 - AnyEvent::Strict: do the test for invalid fh arguments differently,
186 resulting in more informative error messages.
187 - new AnyEvent::FAQ manpage.
188 - clarify that you can actually call push_read etc. in on_read.
189 - put keys and certs in /tmp, not into the current directory,
190 and document this in the SECURITY CONSIDERATIONS sections
191 of AnyEvent::TLS.
192 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
193 requested by Ron Isaacson.
194 - document $AnyEvent::DNS::RESOLVER variable.
195 - incorporated major typo patches by Abhijit Menon-Sen.
196 - AnyEvent::Handle now croaks when negative timeouts are passed.
197 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
198 undocumented, but who knows when they might come in handy.
199
2005.271 Tue Jun 8 12:05:46 CEST 2010
201 - backport to perl 5.8.x.
202
2035.27 Sun Jun 6 12:12:05 CEST 2010
204 - postpone differently in AnyEvent::Socket now, as
205 when not, canceling the connection attempt might fail
206 (found by Felix Antonius Wilhelm Ostmann).
207 - explicitly check for non-stream sockets in AE::Handle, too many
208 clueless people fell into the trap of this somehow working.
209 - simplified and reworked the "OTHER MODULES" section.
210 - better/more condvar examples.
211
2125.261 Wed Apr 28 16:13:36 CEST 2010
213 - AF_INET6 was not properly used from Socket6 during configuration
214 time (found by Andy Grundman).
215
2165.26 Mon Apr 12 04:49:35 CEST 2010
217 - don't generate spurious readyness notifications when select
218 returns EINTR in the pure perl backend (analysed and testcase
219 by Bas Denissen).
220 - give same examples in SYNOPSIS sections of AnyEvent and AE.
221 - provider faster implementations for AE::time/now/now_update for
222 EV and Perl backends.
223 - provide faster AE::cv and AE::time implementations for all
224 backends.
225 - no longer support lower/mixed-case signal names (this was
226 never documented nor universally supported).
227 - some more memory ยต-optimisations, and somewhat less messy
228 code to cope with both AnyEvent and AE APIs.
229
2305.251 Sat Mar 13 00:58:21 CET 2010
231 - make test failed if it was run before make install on perl <=
232 5.8.8, and older versions of perl make our live very difficult,
233 so write constants.pl during Makefile.PL time.
234
2355.25 Sat Mar 13 00:23:14 CET 2010
236 - fix a race condition in AnyEvent::Handle that would cause
237 a "bio_read: unsupported method" error (in _another_
238 ssl connection) after user code threw an exception.
239 - added AnyEvent::Handle->destroyed method.
240 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
241 - the AnyEvent::Util::fh_nonblocking function worked only
242 by lucky accident on win32.
243 - smaller and faster AnyEvent::Util::fh_nonblocking.
244 - when the (required!) Time::HiRes module is missing, AnyEvent
245 did not fall back to built-in time properly.
246 - do not load Fcntl at runtime, saving memory and loading time.
247 - precompile a number of constants and use them instead of runtime
248 detection and eval.
249 - free detection code after detection and similar memory optimisations.
250 - Perl backend timer interval best effort drift has been improved
251 (same algorithm as EV).
252 - update unicode idna mapping table.
253
2545.24 Tue Jan 5 11:39:43 CET 2010
255 - cygwin never reports errors from failed connects. 1.5 just gives
256 you continous readyness and EAGAIN, 1.7 is even more broken and
257 just hangs. work around both issues in a cygwin-specific hack.
258 - improve idle watcher documentation slightly.
259
2605.23 Sun Dec 20 23:48:00 CET 2009
261 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
262 - implement punycode_encode/decode, idn_nameprep,
263 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
264 - provide $AE::VERSION.
265 - removed traces of "no strict 'refs'".
266
2675.22 Sat Dec 5 03:51:13 CET 2009
268 - downgrade-or-fail in AnyEvent::Handle::push_write, to
269 diagnose encoding failures earlier and more succinctly.
270 (this works around bugs in perl, throwing away encoding info
271 when passing scalar data to extensions).
272 - add more examples to AnyEvent::Socket manpage.
273 - upgrade internal warning set to the same as common::sense 2.03.
274 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
275 (John Beppu).
276 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
277 - add google public dns servers to fallback server set and make sure
278 we load-balance properly between the three sets. also add all
279 fallback dns servers, not just a random one, to each dns config.
280
2815.21 Thu Nov 19 02:48:47 CET 2009
282 - fix a problem where socket constants were called with parameters
283 (spotted by David Friedland).
284 - fork_call never use'd POSIX (reported by Daisuke Maki).
285 - improve perl 5.6 compatibility further (but it still won't work
286 unless you rip out everything but the core).
287 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
288 uses 10MB of RAM, but doesn't open a console window. *sigh*.
289
2905.202 Wed Oct 14 22:35:44 CEST 2009
291 - AnyEvent::DNS would unexpectedly clobber $_ under windows
292 (analysed by Matthias Waldorf).
293 - AnyEvent::Handle::run_cmd can now pass the PID of the
294 newly-created process, which is much less useful than it might
295 sound (based on patch by Yann Kerherve).
296
2975.201 Tue Sep 29 12:09:25 CEST 2009
298 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
299 (reported by Torsten Foertsch).
300 - common::sense 2.0 could cause tcp_server to throw an exception
301 (analysed by elmex).
302
3035.2 Mon Sep 14 07:04:49 CEST 2009
304 - INCOMPATIBLE CHANGE: do no longer support register_read_type
305 and register_write_type in AnyEvent::Handle, instead support
306 package names (the facility was mostly abused).
307 - implement "packagename-as-read/write type" support in
308 AnyEvent::Handle.
309 - AnyEvent::Handle: new options "keepalive" and "oobinline".
310 - oobinline set by default to avoid security issues.
311 - the pure-perl event loop backend wrongly detected times() fallback
312 support (spotted by Pavel Boldin).
313
3145.12 Tue Sep 1 20:26:50 CEST 2009
315 - be more lenient when parsing resolv.conf files, as some people
316 use hashmarks as comment indicator inside directives (reported
317 by Michael S. Fischer).
318 - use same set of warnings as common::sense 2.0.
319 - fix a potential 32 bit overflow issue due to perl having problems
320 with large hex constants in the Perl backend.
321 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
322 by default.
323 - fix some prototypes.
324 - (optionally) require Async::Interrupt 1.02.
325 - replace opendns fallbacks by verizon/level3 public dns servers.
326
3275.112 Fri Aug 21 13:59:04 CEST 2009
328 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
329 using starttls with an illegal/uncreatable tls context.
330 - port to broken win32 perls.
331
3325.111 Sun Aug 16 18:44:36 CEST 2009
333 - perl errornously requires sockaddr_un structures returned by the
334 kernel to have a specific length. work around this bug by padding
335 sockaddr_un structures (found to fail on OS X, as perl only uses
336 the correct code for GNU/Linux, but likely to fail on other
337 OSes as well).
338
3395.11 Wed Aug 12 17:49:37 CEST 2009
340 - fixed a bug in run_cmd with input from filename.
341 - create a json coder object when none is given in write direction
342 to reduce dependency on JSON.
343
3445.1 Tue Aug 11 03:17:32 CEST 2009
345 - tcp_connect now reports the correct error in $!, not an unrelated
346 one (was broken since 4.91).
347 - AE::Handle did not properly clear rtimeout/wtimeout watchers
348 when retrying a tcp_connect.
349 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
350 - fixed wrong documentation in AE::Socket.
351
3525.01 Mon Aug 10 03:16:32 CEST 2009
353 - last release broke EV child watchers :/.
354
3555.0 Sun Aug 9 17:32:09 CEST 2009
356 - officially introduce and document the AE API.
357 - lowering the timeout at runtime did not have immediate effect
358 in AnyEvent::Handle.
359 - AE::Handle now has separate and independent read and write timeout
360 settings.
361 - probe for Guard module when AnyEvent::Util is loaded, not at
362 runtime (Event doesn't like eval "use" from callbacks, and it avoids
363 exporting the probe function).
364 - Impl::IOAsync's io watchers did not use the filehandle emulation
365 code to work around it's watcher limits.
366 - work around IO::Async's problems with overloaded objects.
367 - try to disable wearnings inside IO::Async, it's just too buggy.
368 - do no longer use Event by default, only EV or Perl, Event is too buggy.
369 - hackishly provide fast AE:: interface when EV or Perl is the backend.
370 - inherit VERSION from AnyEvent::VERSION in most modules.
371
3724.91 Thu Aug 6 15:42:45 CEST 2009
373 - AE::Handle::starttls could get out of sync when the read buffer
374 already contains some TLS handshake.
375 - AE::Handle did not properly free the TLS session, causing bigger
376 memory leaks in Net::SSLeay than expected (reported by toaster).
377 - AE::Socket::tcp_connect will now artificially delay invoking
378 the callback to avoid returning after invoking the callback.
379 - convert many internal watcher uses to AE API.
380
3814.9 Sat Aug 1 11:07:01 CEST 2009
4 - Glib silently fails when registering a timeout with 382 - Glib silently fails when registering a timeout with
5 a negative value, so avoid that. 383 a negative value, so avoid that.
384 - call condvar callback immediately when it is set
385 after the condition is already signalled.
6 - check rbuf_max condition only after trying to consume 386 - check rbuf_max condition only after trying to consume
7 data. 387 data.
8 - normalise signal numbers to names when using the ae's 388 - normalise signal numbers to names when using the ae's
9 signal handling, but do not document this (yet). 389 signal handling, but do not document this (yet).
10 - pure perl signal emulation did not properly set nonblocking 390 - pure perl signal emulation did not properly set nonblocking
… …
16 - reduce memory footprint in typical cases by ~50kb by 396 - reduce memory footprint in typical cases by ~50kb by
17 only compiling signal code when necessary. 397 only compiling signal code when necessary.
18 - add AnyEvent::Handle->rbuf_max. 398 - add AnyEvent::Handle->rbuf_max.
19 - grab the AE namespace for future new API, implement stubs for 399 - grab the AE namespace for future new API, implement stubs for
20 said future/alternative API. 400 said future/alternative API.
401 - new function: AnyEvent::Socket::format_hostport.
21 402
224.881 Tue Jul 28 12:51:53 CEST 2009 4034.881 Tue Jul 28 12:51:53 CEST 2009
23 - work around a bug in local in pre-5.10 perls, causing 404 - work around a bug in local in pre-5.10 perls, causing
24 AnyEvent::Handle to recurse when it should not 405 AnyEvent::Handle to recurse when it should not
25 (analyzed by elmex). 406 (analysed by elmex).
26 407
274.88 Tue Jul 28 04:04:37 CEST 2009 4084.88 Tue Jul 28 04:04:37 CEST 2009
28 - re-bless the handle into a dummy package after calling 409 - re-bless the handle into a dummy package after calling
29 AnyEvent::Handle::destroy, so the user does not need to check 410 AnyEvent::Handle::destroy, so the user does not need to check
30 for errors after every push_write etc. 411 for errors after every push_write etc.
… …
2524.3 Fri Oct 3 09:18:43 CEST 2008 6334.3 Fri Oct 3 09:18:43 CEST 2008
253 - AnyEvent will now install a no-op signal handler for SIGPIPE, 634 - AnyEvent will now install a no-op signal handler for SIGPIPE,
254 unless one has been installed already. 635 unless one has been installed already.
255 - warn about not loaidng AnyEvent::Impl::POE early enough 636 - warn about not loaidng AnyEvent::Impl::POE early enough
256 (patch by Adam Rosenstein). 637 (patch by Adam Rosenstein).
257 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 638 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
258 - fix a bug in where in SSL connect mode, the 639 - fix a bug in where in SSL connect mode, the
259 client would first wait for some data by the server and 640 client would first wait for some data by the server and
260 otherwise hang. (reported and analyzed in an absolutely 641 otherwise hang. (reported and analysed in an absolutely
261 exemplary manner by Adam Rosenstein). 642 exemplary manner by Adam Rosenstein).
262 - fix a bug in where SSL EOF would not be treated 643 - fix a bug in where SSL EOF would not be treated
263 as stream EOF, putting the connection into a hung state. 644 as stream EOF, putting the connection into a hung state.
264 (reported and analyzed in an absolutely exemplary manner by Adam 645 (reported and analysed in an absolutely exemplary manner by Adam
265 Rosenstein). 646 Rosenstein).
266 - fix a potential segfault when the TLS context 647 - fix a potential segfault when the TLS context
267 would go missing in a read callback (e.g. due to stoptls) - 648 would go missing in a read callback (e.g. due to stoptls) -
268 Net::SSLeay of course makes no type checking whatsoever. 649 Net::SSLeay of course makes no type checking whatsoever.
269 - AnyEvent::Handle will not stop the read watcher in TLS mode. 650 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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