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

Comparing AnyEvent/Changes (file contents):
Revision 1.358 by root, Thu Nov 26 10:30:24 2009 UTC vs.
Revision 1.543 by root, Thu Apr 5 06:14:10 2012 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?
6lib/AnyEvent/DNS.pm: open my $fh, "<", $resolv_conf
7lib/AnyEvent/DNS.pm: if (open my $fh, "ipconfig /all |") {
8lib/AnyEvent/TLS.pm: or croak "$dh_file: failed to open DH parameter file: $!";
9TODO: ssl error messages in 80ssltest
10
116.2
12 - child watchers are broken in POE 1.352 (also many earlier
13 versions)and there seems to be no way to work around it, as POE
14 itself is inherently racy. Document this regression and sleep a
15 second in t/68_poe_03_child.t for the time being.
16 - new module AnyEvent::IO, that is a frontend to either a pure-perl
17 synchronous I/O implementation (AnyEvent::IO::Perl), or to the
18 asynchronous I/O extension IO::AIO (AnyEvent::IO::IOAIO).
19 - load /etc/hosts only when dns fails has no answer.
20 - stat /etc/hosts on every access and reload it if it changed.
21 - load /etc/hosts via AnyEvent::IO - potentially asynchronous.
22 - AnyEvent::Log log_to_file and log_to_path now use AnyEvent::IO. As
23 a side effect, they now use true appending as opposed to libc
24 appending, and the file might not have been opened when the function
25 returns.
26 - the default logging level was not properly documented in a variety of
27 places, this has hopefully been rectified.
28 - updated uts46data.pl for unicode 6.1.0.
29 - made log messages genearted by AnyEvent submodules not
30 include the package name anymore, as it will be logged
31 by default already.
32 - upgrade to the trick used by common-sense 3.5 to work
33 around extra warning torture/breakage under perl 5.15.x.
34 - log messages by anyevent are now ucfirst, are usually full
35 sentences and do no longer include the package name.
36 - the storable read type would sometimes throw an exception instead
37 of causing EBADMSG (analyzed by Maxime Soulé).
38 - work around a bug in openssl 1.0.1 which enforces a minimum rsa
39 keysize (reported by Paul Howarth).
40 - documented Rocco fabricating statements about AnyEvent and me.
41
426.14 Tue Jan 31 20:00:24 CET 2012
43 - AnyEvent::Impl::Tk was broken due to a mysterious "die" inside,
44 probably an editing mistake (reported by Darin McBride).
45
466.13 Thu Jan 12 07:27:01 CET 2012
47 - AnyEvent::Util::fork_call checked for POSIX availability in the wrong way,
48 causing it to fail if POSIX wasn't loaded already (analysed by Rock Power).
49 - AnyEvent::Handle::push_read (line => did pass $1 directly, so
50 regex-matching inside the callback would change the parameter, despite
51 $1 being dynamically scoped per-block. this perl bug is now being
52 worked around (testcase by Cindy Wang).
53
546.12 Mon Dec 12 13:21:10 CET 2011
55 - $! was clobbered by subsequent calls in tcp_connect, due to
56 the postpone not saving/restoring it (analyzed by Richard Garnish).
57
586.11 Tue Nov 22 10:36:05 CET 2011
59 - Tk cannot create windows when tainted unless you set a title, so
60 set a dummy title for AnyEvent's dummy mainwindow. How dummy.
61 - escape any nonprintable/nonascii characters when stringifying
62 backtraces.
63 - log the reason loading an interface module fails to load at level debug.
64
656.1 Tue Oct 4 19:44:30 CEST 2011
66 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
67 and above), and some messages inside AnyEvent have been elevated
68 to higher log levels to print by default.
69 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
70 - work around yet another signal race bug, newly introduced by POE. My
71 how that module sucks.
72 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
73 support ipv6-only operation even with defaults (also use ::1 instead
74 of 127.0.0.1).
75 - give better guidelines on how to chose a logging level.
76 - AnyEvent::Log can now cap (limit) log message priority by context.
77 - AE::log will now save and restore $! and $@, so callers don't have to.
78 - AE::Strict now checks that registered fds are still valid at regular
79 intervals.
80 - skip frequently called 'require' calls if the module appears loaded.
81 - increase signal latency test timer considerably for poor little osx,
82 which can't be bothered with basic posix functionality such as signal
83 delivery.
84 - mention some debugging aids in the tutorial.
85 - new env variable PERL_ANYEVENT_MAX_SIGNAL_LATENCY.
86 - private (%namespace) log contexts in PERL_ANYEVENT_LOG now have the name
87 of the context as title.
88 - AnyEvent::Strict would not restore read-write access to $_ when a callback
89 died - this is a user bug, but since AE::Strict exists to help dbeugging...
90 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
91 to work with it - the author is well aware of the issues but refuses
92 to fix them. This extreme step has been taken because the module actively
93 attacks the goals of AnyEvent and makes it harder for module authors to use
94 AnyEvent.
95
966.02 Fri Aug 26 20:08:31 CEST 2011
97 - log_to_syslog now accepts facility strings.
98 - log_to_syslog would not allow facilities - "$facility|$level"
99 *literally* means that string. sheesh.
100 - the CPAN parser got confused and thought we provide Sys::Syslog, try
101 to work around.
102 - renamed AnyEvent::Impl::FLTK2 to FLTK.
103
1046.01 Fri Aug 26 07:04:11 CEST 2011
105 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
106 the TTL as fourth element - this affects ->resolve, ->request
107 and ->dns_unpack indirectly. Convenience functions such as
108 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
109 AnyEvent::Socket.
110 - INCOMPATIBLE CHANGE: backend authors now should not implement
111 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
112 methods.
113 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
114 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
115 - DEPRECIATION: please change
116 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
117 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
118 in your programs, for future compatibility.
119 - *blush* idle watcher emulation was borked.
120 - AnyEvent::Strict now checks the AE::xxx functions as well.
121 - querying too long domain names would barf deep inside AnyEvent::DNS - since
122 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
123 for too long domain names explicitly and passes an undef to the callback
124 instead (reported by James Bromberger).
125 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
126 - new functions: AnyEvent::postpone, AnyEvent::log.
127 - new module: AnyEvent::Log, for simple logging needs.
128 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
129 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
130 a debug shell.
131 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
132 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
133 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
134 /etc/hosts can be found.
135 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
136 configuration.
137 - env variables can now be specified using AE_xyz instead of
138 PERL_ANYEVENT_xyz, subject to some rules.
139 - lots of new AnyEvent::Debug functionality.
140 - AnyEvent::Strict now checks for modifications of $_ by the callback
141 (by setting it to readonly while executing the callback, if possible).
142 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
143 Invest extra overhead to work around this peculiar behaviour.
144 - split the pure perl backend into a loop module and an impl
145 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
146 - some read types (regex, netstring, json, storable) could
147 errornously keep an AE::Handle object alive on their own
148 (reported by Mohammad Toossi).
149 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
150 supports /etc/hosts overrides.
151 - add more workarounds around child watcher bugs in IO::Async
152 and (newly introduced) in POE.
153 - use glib child watchers and try to work around its limitations,
154 if possible (based on a patch by Kevin Ryde).
155 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
156 - many read types in AnyEvent::Handle would malfunction during
157 transitions to TLS because of a "delete $self->{rbuf}".
158 (reported by Gerald Galster).
159 - improve condvar blocking wait performance (probably only noticable
160 with faster backends), also make it easier for event loops that
161 do not support blocking waits.
162 - PERL_ANYEVENT_MODEL now also allows full module names, although
163 the usefulness of this feature is not yet clear.
164 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
165 - add t/80_ssltext.t.
166 - no longer use AUTOLOAD for the initial loading (for no very good reason).
167 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
168 - do not run event loop detection multiple times when programs call their cached
169 copies of AnyEvent::detect.
170 - parse_hostport now accepts absolute paths als unix domain sockets.
171 - suppress the idiotic warning inside POE by patching the POE::Kernel
172 run flags instead of asking the user to jump through contortions.
173 - add Task::Weaken as dependency (as recommends only at this time).
174 - replace PF_UNSPEC by 0, for good old minix.
175
1766.0 (not officially released, but was on CPAN for about 15 minutes :/)
177
1785.34 Thu May 12 09:59:41 CEST 2011
179 - AE::TLS didn't support both cert_file and key at the same time,
180 due to some copy&paste bug (reported and analyzed by vti).
181 - improve AE::Handle timeout documentation slightly.
182
1835.33 Mon Feb 7 21:51:41 CET 2011
184 - never appeared on CPAN, for unknown reasons.
185 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
186 object when it's already gone (analyzed by Ryan Bullock).
187 - update idna algorithm(s) to UTS#46 version 6.0.0.
188 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
189 uts46 says it's an error, and testsuite makes more sense.
190 - remove leading dots in idn_to_ascii - testsuite says to remove,
191 uts46 says it's an error, and testsuite makes more sense.
192 - fixed a bug in idn ignored character processing.
193 - passes the uts46 6.0.0 testsuite except for three cases which
194 apparently are bugs in the testsuite itself.
195 - passes the uts46 6.0.1 testsuite except for eleven cases which
196 apparently are bugs in the testsuite itself.
197
1985.31 Sun Jan 23 11:49:19 CET 2011
199 - work around a POE bug where POE would first poll for new events and
200 invoke handlers BEFORE handling the message queue.
201 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
202 - stop_read now does just that in TLS mode. see its
203 documentation for caveats.
204 - update to common::sense 3.4 warning set (some warnings in perl impose
205 an arbitrary 32kb limit on regex matches, which affects
206 AnyEvent::Handle).
207 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
208 - new experimental AE::Handle::resettls method.
209
2105.3 Fri Dec 31 05:48:13 CET 2010
211 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
212 issues with IO::Async have been solved in that version, including
213 the availabilty of a default loop!
214 - implement new wbuf_max parameter for AnyEvent::Handle.
215 - added Cocoa::EventLoop backend by Daisuke Murase.
216 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
217 - clarify handle on_prepare documentation.
218 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
219 to avoid leaking them to any child processes.
220 - try to force files into the arch-dependend lib directory, for the
221 half-"turly obsessive compulsive person"(s) that depend on it.
222
2235.29 Sun Dec 5 10:49:21 CET 2010
224 - convert EV backend to EV 4.00 API (so better upgrade EV too).
225 - AE::Handle: implement dynamic read_size adjustment, add
226 max_read_size parameter, reduce default read_size to 2048.
227 - add some FAQ entries for servers and/or on_eof vs. on_error.
228 - work around OS bugs (cygwin again) when lingering in AE::Handle.
229 - reduce memory usage slightly when connect was used in AE::Handle.
230 - wrok around more segfaults in Tk.
231 - document tls_ctx => undef as valid.
232 - detect WNOHANG value at compiletime.
233 - include some probably linux-only support for building AnyEvent
234 as part of the perl core.
235 - improve compatibility of t/08_idna.t to perl 5.8.
236 - make AnyEvent::Strict truly optional (in case somebody builds a
237 minimal perl...).
238
2395.28 Wed Oct 13 04:14:23 CEST 2010
240 - due to a glitch, AnyEvent's internal getprotobyname wasn't
241 actually used in tcp_server.
242 - implement AnyEvent::CondVar->new method, which was documented but
243 missing.
244 - support autoloading of read types in unshift_read.
245 - AnyEvent::Strict: do the test for invalid fh arguments differently,
246 resulting in more informative error messages.
247 - new AnyEvent::FAQ manpage.
248 - clarify that you can actually call push_read etc. in on_read.
249 - put keys and certs in /tmp, not into the current directory,
250 and document this in the SECURITY CONSIDERATIONS sections
251 of AnyEvent::TLS.
252 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
253 requested by Ron Isaacson.
254 - document $AnyEvent::DNS::RESOLVER variable.
255 - incorporated major typo patches by Abhijit Menon-Sen.
256 - AnyEvent::Handle now croaks when negative timeouts are passed.
257 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
258 undocumented, but who knows when they might come in handy.
259
2605.271 Tue Jun 8 12:05:46 CEST 2010
261 - backport to perl 5.8.x.
262
2635.27 Sun Jun 6 12:12:05 CEST 2010
264 - postpone differently in AnyEvent::Socket now, as
265 when not, canceling the connection attempt might fail
266 (found by Felix Antonius Wilhelm Ostmann).
267 - explicitly check for non-stream sockets in AE::Handle, too many
268 clueless people fell into the trap of this somehow working.
269 - simplified and reworked the "OTHER MODULES" section.
270 - better/more condvar examples.
271
2725.261 Wed Apr 28 16:13:36 CEST 2010
273 - AF_INET6 was not properly used from Socket6 during configuration
274 time (found by Andy Grundman).
275
2765.26 Mon Apr 12 04:49:35 CEST 2010
277 - don't generate spurious readyness notifications when select
278 returns EINTR in the pure perl backend (analysed and testcase
279 by Bas Denissen).
280 - give same examples in SYNOPSIS sections of AnyEvent and AE.
281 - provider faster implementations for AE::time/now/now_update for
282 EV and Perl backends.
283 - provide faster AE::cv and AE::time implementations for all
284 backends.
285 - no longer support lower/mixed-case signal names (this was
286 never documented nor universally supported).
287 - some more memory µ-optimisations, and somewhat less messy
288 code to cope with both AnyEvent and AE APIs.
289
2905.251 Sat Mar 13 00:58:21 CET 2010
291 - make test failed if it was run before make install on perl <=
292 5.8.8, and older versions of perl make our live very difficult,
293 so write constants.pl during Makefile.PL time.
294
2955.25 Sat Mar 13 00:23:14 CET 2010
296 - fix a race condition in AnyEvent::Handle that would cause
297 a "bio_read: unsupported method" error (in _another_
298 ssl connection) after user code threw an exception.
299 - added AnyEvent::Handle->destroyed method.
300 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
301 - the AnyEvent::Util::fh_nonblocking function worked only
302 by lucky accident on win32.
303 - smaller and faster AnyEvent::Util::fh_nonblocking.
304 - when the (required!) Time::HiRes module is missing, AnyEvent
305 did not fall back to built-in time properly.
306 - do not load Fcntl at runtime, saving memory and loading time.
307 - precompile a number of constants and use them instead of runtime
308 detection and eval.
309 - free detection code after detection and similar memory optimisations.
310 - Perl backend timer interval best effort drift has been improved
311 (same algorithm as EV).
312 - update unicode idna mapping table.
313
3145.24 Tue Jan 5 11:39:43 CET 2010
315 - cygwin never reports errors from failed connects. 1.5 just gives
316 you continous readyness and EAGAIN, 1.7 is even more broken and
317 just hangs. work around both issues in a cygwin-specific hack.
318 - improve idle watcher documentation slightly.
319
3205.23 Sun Dec 20 23:48:00 CET 2009
321 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
322 - implement punycode_encode/decode, idn_nameprep,
323 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
324 - provide $AE::VERSION.
325 - removed traces of "no strict 'refs'".
326
3275.22 Sat Dec 5 03:51:13 CET 2009
3 - downgrade-or-fail in AnyEvent::Handle::push_write, to 328 - downgrade-or-fail in AnyEvent::Handle::push_write, to
4 diagnose encoding failures earlier and more succinctly. 329 diagnose encoding failures earlier and more succinctly.
330 (this works around bugs in perl, throwing away encoding info
331 when passing scalar data to extensions).
332 - add more examples to AnyEvent::Socket manpage.
333 - upgrade internal warning set to the same as common::sense 2.03.
334 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
335 (John Beppu).
336 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
337 - add google public dns servers to fallback server set and make sure
338 we load-balance properly between the three sets. also add all
339 fallback dns servers, not just a random one, to each dns config.
5 340
65.21 Thu Nov 19 02:48:47 CET 2009 3415.21 Thu Nov 19 02:48:47 CET 2009
7 - fix a problem where socket constants were called with parameters 342 - fix a problem where socket constants were called with parameters
8 (spotted by David Friedland). 343 (spotted by David Friedland).
9 - fork_call never use'd POSIX (reported by Daisuke Maki). 344 - fork_call never use'd POSIX (reported by Daisuke Maki).
… …
3584.3 Fri Oct 3 09:18:43 CEST 2008 6934.3 Fri Oct 3 09:18:43 CEST 2008
359 - AnyEvent will now install a no-op signal handler for SIGPIPE, 694 - AnyEvent will now install a no-op signal handler for SIGPIPE,
360 unless one has been installed already. 695 unless one has been installed already.
361 - warn about not loaidng AnyEvent::Impl::POE early enough 696 - warn about not loaidng AnyEvent::Impl::POE early enough
362 (patch by Adam Rosenstein). 697 (patch by Adam Rosenstein).
363 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 698 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
364 - fix a bug in where in SSL connect mode, the 699 - fix a bug in where in SSL connect mode, the
365 client would first wait for some data by the server and 700 client would first wait for some data by the server and
366 otherwise hang. (reported and analysed in an absolutely 701 otherwise hang. (reported and analysed in an absolutely
367 exemplary manner by Adam Rosenstein). 702 exemplary manner by Adam Rosenstein).
368 - fix a bug in where SSL EOF would not be treated 703 - fix a bug in where SSL EOF would not be treated

Diff Legend

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