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

Comparing AnyEvent/Changes (file contents):
Revision 1.320 by root, Thu Aug 6 13:31:01 2009 UTC vs.
Revision 1.510 by root, Sat Oct 1 21:00:30 2011 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: Danga::Socket backend. 3TODO: document TCP_*** constants
4 4
5TODO: delay tcp_connect 5TODO: AE::logger
64.901 6TODO: add debug/trace logging to some modules
7TODO: poe child watcher hang?
8TODO: general log level "cap"
9TODO: timer increase for arm(?)
10 - INCOMPATIBLE CHANGE: the default log level is now "4" (error
11 and above), and some messages inside AnyEvent have been elevated
12 to higher log levels to print by default.
13 - AnyEvent::log, unlike AnyEvent::Log::log, did not always exit on fatal.
14 - add google public ipv6 dns as fallback servers, so AE::DNS can finally
15 support ipv6-only operation even with defaults (also use ::1 instead
16 of 127.0.0.1).
17 - give better guidelines on how to chose a logging level.
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 - mention some debugging aids in the tutorial.
23 - detect the extremely evil IO::Async::Loop::AnyEvent module and refuse
24 to work with it - the author is well aware of the issues but refuses
25 to fix them. This extreme step has been taken because the module actively
26 attacks the goals of AnyEvent and makes it harder for module authors to use
27 AnyEvent.
28
296.02 Fri Aug 26 20:08:31 CEST 2011
30 - log_to_syslog now accepts facility strings.
31 - log_to_syslog would not allow facilities - "$facility|$level"
32 *literally* means that string. sheesh.
33 - the CPAN parser got confused and thought we provide Sys::Syslog, try
34 to work around.
35 - renamed AnyEvent::Impl::FLTK2 to FLTK.
36
376.01 Fri Aug 26 07:04:11 CEST 2011
38 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
39 the TTL as fourth element - this affects ->resolve, ->request
40 and ->dns_unpack indirectly. Convenience functions such as
41 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
42 AnyEvent::Socket.
43 - INCOMPATIBLE CHANGE: backend authors now should not implement
44 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
45 methods.
46 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
47 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
48 - DEPRECIATION: please change
49 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
50 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
51 in your programs, for future compatibility.
52 - *blush* idle watcher emulation was borked.
53 - AnyEvent::Strict now checks the AE::xxx functions as well.
54 - querying too long domain names would barf deep inside AnyEvent::DNS - since
55 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
56 for too long domain names explicitly and passes an undef to the callback
57 instead (reported by James Bromberger).
58 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
59 - new functions: AnyEvent::postpone, AnyEvent::log.
60 - new module: AnyEvent::Log, for simple logging needs.
61 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
62 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
63 a debug shell.
64 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
65 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
66 - new env variable, PERL_ANYEVENT_HOSTS, to override the place where
67 /etc/hosts can be found.
68 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
69 configuration.
70 - env variables can now be specified using AE_xyz instead of
71 PERL_ANYEVENT_xyz, subject to some rules.
72 - lots of new AnyEvent::Debug functionality.
73 - AnyEvent::Strict now checks for modifications $_ by the callback
74 (by setting it to readonly while executing the callback, if possible).
75 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
76 Invest extra overhead to work around this peculiar behaviour.
77 - split the pure perl backend into a loop module and an impl
78 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
79 - some read types (regex, netstring, json, storable) could
80 errornously keep an AE::Handle object alive on their own
81 (reported by Mohammad Toossi).
82 - AnyEvent::Socket::resolve_sockaddr and all functions using it now
83 supports /etc/hosts overrides.
84 - add more workarounds around child watcher bugs in IO::Async
85 and (newly introduced) in POE.
86 - use glib child watchers and try to work around its limitations,
87 if possible (based on a patch by Kevin Ryde).
88 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
89 - many read types in AnyEvent::Handle would malfunction during
90 transitions to TLS because of a "delete $self->{rbuf}".
91 (reported by Gerald Galster).
92 - improve condvar blocking wait performance (probably only noticable
93 with faster backends), also make it easier for event loops that
94 do not support blocking waits.
95 - PERL_ANYEVENT_MODEL now also allows full module names, although
96 the usefulness of this feature is not yet clear.
97 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
98 - add t/80_ssltext.t.
99 - no longer use AUTOLOAD for the initial loading (for no very good reason).
100 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
101 - do not run event loop detection multiple times when programs call their cached
102 copies of AnyEvent::detect.
103 - parse_hostport now accepts absolute paths als unix domain sockets.
104 - suppress the idiotic warning inside POE by patching the POE::Kernel
105 run flags instead of asking the user to jump through contortions.
106 - add Task::Weaken as dependency (as recommends only at this time).
107 - replace PF_UNSPEC by 0, for good old minix.
108
1096.0 (not officially released, but was on CPAN for about 15 minutes :/)
110
1115.34 Thu May 12 09:59:41 CEST 2011
112 - AE::TLS didn't support both cert_file and key at the same time,
113 due to some copy&paste bug (reported and analyzed by vti).
114 - improve AE::Handle timeout documentation slightly.
115
1165.33 Mon Feb 7 21:51:41 CET 2011
117 - never appeared on CPAN, for unknown reasons.
118 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
119 object when it's already gone (analyzed by Ryan Bullock).
120 - update idna algorithm(s) to UTS#46 version 6.0.0.
121 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
122 uts46 says it's an error, and testsuite makes more sense.
123 - remove leading dots in idn_to_ascii - testsuite says to remove,
124 uts46 says it's an error, and testsuite makes more sense.
125 - fixed a bug in idn ignored character processing.
126 - passes the uts46 6.0.0 testsuite except for three cases which
127 apparently are bugs in the testsuite itself.
128 - passes the uts46 6.0.1 testsuite except for eleven cases which
129 apparently are bugs in the testsuite itself.
130
1315.31 Sun Jan 23 11:49:19 CET 2011
132 - work around a POE bug where POE would first poll for new events and
133 invoke handlers BEFORE handling the message queue.
134 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
135 - stop_read now does just that in TLS mode. see its
136 documentation for caveats.
137 - update to common::sense 3.4 warning set (some warnings in perl impose
138 an arbitrary 32kb limit on regex matches, which affects
139 AnyEvent::Handle).
140 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
141 - new experimental AE::Handle::resettls method.
142
1435.3 Fri Dec 31 05:48:13 CET 2010
144 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
145 issues with IO::Async have been solved in that version, including
146 the availabilty of a default loop!
147 - implement new wbuf_max parameter for AnyEvent::Handle.
148 - added Cocoa::EventLoop backend by Daisuke Murase.
149 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
150 - clarify handle on_prepare documentation.
151 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
152 to avoid leaking them to any child processes.
153 - try to force files into the arch-dependend lib directory, for the
154 half-"turly obsessive compulsive person"(s) that depend on it.
155
1565.29 Sun Dec 5 10:49:21 CET 2010
157 - convert EV backend to EV 4.00 API (so better upgrade EV too).
158 - AE::Handle: implement dynamic read_size adjustment, add
159 max_read_size parameter, reduce default read_size to 2048.
160 - add some FAQ entries for servers and/or on_eof vs. on_error.
161 - work around OS bugs (cygwin again) when lingering in AE::Handle.
162 - reduce memory usage slightly when connect was used in AE::Handle.
163 - wrok around more segfaults in Tk.
164 - document tls_ctx => undef as valid.
165 - detect WNOHANG value at compiletime.
166 - include some probably linux-only support for building AnyEvent
167 as part of the perl core.
168 - improve compatibility of t/08_idna.t to perl 5.8.
169 - make AnyEvent::Strict truly optional (in case somebody builds a
170 minimal perl...).
171
1725.28 Wed Oct 13 04:14:23 CEST 2010
173 - due to a glitch, AnyEvent's internal getprotobyname wasn't
174 actually used in tcp_server.
175 - implement AnyEvent::CondVar->new method, which was documented but
176 missing.
177 - support autoloading of read types in unshift_read.
178 - AnyEvent::Strict: do the test for invalid fh arguments differently,
179 resulting in more informative error messages.
180 - new AnyEvent::FAQ manpage.
181 - clarify that you can actually call push_read etc. in on_read.
182 - put keys and certs in /tmp, not into the current directory,
183 and document this in the SECURITY CONSIDERATIONS sections
184 of AnyEvent::TLS.
185 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
186 requested by Ron Isaacson.
187 - document $AnyEvent::DNS::RESOLVER variable.
188 - incorporated major typo patches by Abhijit Menon-Sen.
189 - AnyEvent::Handle now croaks when negative timeouts are passed.
190 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
191 undocumented, but who knows when they might come in handy.
192
1935.271 Tue Jun 8 12:05:46 CEST 2010
194 - backport to perl 5.8.x.
195
1965.27 Sun Jun 6 12:12:05 CEST 2010
197 - postpone differently in AnyEvent::Socket now, as
198 when not, canceling the connection attempt might fail
199 (found by Felix Antonius Wilhelm Ostmann).
200 - explicitly check for non-stream sockets in AE::Handle, too many
201 clueless people fell into the trap of this somehow working.
202 - simplified and reworked the "OTHER MODULES" section.
203 - better/more condvar examples.
204
2055.261 Wed Apr 28 16:13:36 CEST 2010
206 - AF_INET6 was not properly used from Socket6 during configuration
207 time (found by Andy Grundman).
208
2095.26 Mon Apr 12 04:49:35 CEST 2010
210 - don't generate spurious readyness notifications when select
211 returns EINTR in the pure perl backend (analysed and testcase
212 by Bas Denissen).
213 - give same examples in SYNOPSIS sections of AnyEvent and AE.
214 - provider faster implementations for AE::time/now/now_update for
215 EV and Perl backends.
216 - provide faster AE::cv and AE::time implementations for all
217 backends.
218 - no longer support lower/mixed-case signal names (this was
219 never documented nor universally supported).
220 - some more memory ยต-optimisations, and somewhat less messy
221 code to cope with both AnyEvent and AE APIs.
222
2235.251 Sat Mar 13 00:58:21 CET 2010
224 - make test failed if it was run before make install on perl <=
225 5.8.8, and older versions of perl make our live very difficult,
226 so write constants.pl during Makefile.PL time.
227
2285.25 Sat Mar 13 00:23:14 CET 2010
229 - fix a race condition in AnyEvent::Handle that would cause
230 a "bio_read: unsupported method" error (in _another_
231 ssl connection) after user code threw an exception.
232 - added AnyEvent::Handle->destroyed method.
233 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
234 - the AnyEvent::Util::fh_nonblocking function worked only
235 by lucky accident on win32.
236 - smaller and faster AnyEvent::Util::fh_nonblocking.
237 - when the (required!) Time::HiRes module is missing, AnyEvent
238 did not fall back to built-in time properly.
239 - do not load Fcntl at runtime, saving memory and loading time.
240 - precompile a number of constants and use them instead of runtime
241 detection and eval.
242 - free detection code after detection and similar memory optimisations.
243 - Perl backend timer interval best effort drift has been improved
244 (same algorithm as EV).
245 - update unicode idna mapping table.
246
2475.24 Tue Jan 5 11:39:43 CET 2010
248 - cygwin never reports errors from failed connects. 1.5 just gives
249 you continous readyness and EAGAIN, 1.7 is even more broken and
250 just hangs. work around both issues in a cygwin-specific hack.
251 - improve idle watcher documentation slightly.
252
2535.23 Sun Dec 20 23:48:00 CET 2009
254 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
255 - implement punycode_encode/decode, idn_nameprep,
256 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
257 - provide $AE::VERSION.
258 - removed traces of "no strict 'refs'".
259
2605.22 Sat Dec 5 03:51:13 CET 2009
261 - downgrade-or-fail in AnyEvent::Handle::push_write, to
262 diagnose encoding failures earlier and more succinctly.
263 (this works around bugs in perl, throwing away encoding info
264 when passing scalar data to extensions).
265 - add more examples to AnyEvent::Socket manpage.
266 - upgrade internal warning set to the same as common::sense 2.03.
267 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
268 (John Beppu).
269 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
270 - add google public dns servers to fallback server set and make sure
271 we load-balance properly between the three sets. also add all
272 fallback dns servers, not just a random one, to each dns config.
273
2745.21 Thu Nov 19 02:48:47 CET 2009
275 - fix a problem where socket constants were called with parameters
276 (spotted by David Friedland).
277 - fork_call never use'd POSIX (reported by Daisuke Maki).
278 - improve perl 5.6 compatibility further (but it still won't work
279 unless you rip out everything but the core).
280 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
281 uses 10MB of RAM, but doesn't open a console window. *sigh*.
282
2835.202 Wed Oct 14 22:35:44 CEST 2009
284 - AnyEvent::DNS would unexpectedly clobber $_ under windows
285 (analysed by Matthias Waldorf).
286 - AnyEvent::Handle::run_cmd can now pass the PID of the
287 newly-created process, which is much less useful than it might
288 sound (based on patch by Yann Kerherve).
289
2905.201 Tue Sep 29 12:09:25 CEST 2009
291 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
292 (reported by Torsten Foertsch).
293 - common::sense 2.0 could cause tcp_server to throw an exception
294 (analysed by elmex).
295
2965.2 Mon Sep 14 07:04:49 CEST 2009
297 - INCOMPATIBLE CHANGE: do no longer support register_read_type
298 and register_write_type in AnyEvent::Handle, instead support
299 package names (the facility was mostly abused).
300 - implement "packagename-as-read/write type" support in
301 AnyEvent::Handle.
302 - AnyEvent::Handle: new options "keepalive" and "oobinline".
303 - oobinline set by default to avoid security issues.
304 - the pure-perl event loop backend wrongly detected times() fallback
305 support (spotted by Pavel Boldin).
306
3075.12 Tue Sep 1 20:26:50 CEST 2009
308 - be more lenient when parsing resolv.conf files, as some people
309 use hashmarks as comment indicator inside directives (reported
310 by Michael S. Fischer).
311 - use same set of warnings as common::sense 2.0.
312 - fix a potential 32 bit overflow issue due to perl having problems
313 with large hex constants in the Perl backend.
314 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
315 by default.
316 - fix some prototypes.
317 - (optionally) require Async::Interrupt 1.02.
318 - replace opendns fallbacks by verizon/level3 public dns servers.
319
3205.112 Fri Aug 21 13:59:04 CEST 2009
321 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
322 using starttls with an illegal/uncreatable tls context.
323 - port to broken win32 perls.
324
3255.111 Sun Aug 16 18:44:36 CEST 2009
326 - perl errornously requires sockaddr_un structures returned by the
327 kernel to have a specific length. work around this bug by padding
328 sockaddr_un structures (found to fail on OS X, as perl only uses
329 the correct code for GNU/Linux, but likely to fail on other
330 OSes as well).
331
3325.11 Wed Aug 12 17:49:37 CEST 2009
333 - fixed a bug in run_cmd with input from filename.
334 - create a json coder object when none is given in write direction
335 to reduce dependency on JSON.
336
3375.1 Tue Aug 11 03:17:32 CEST 2009
338 - tcp_connect now reports the correct error in $!, not an unrelated
339 one (was broken since 4.91).
340 - AE::Handle did not properly clear rtimeout/wtimeout watchers
341 when retrying a tcp_connect.
342 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
343 - fixed wrong documentation in AE::Socket.
344
3455.01 Mon Aug 10 03:16:32 CEST 2009
346 - last release broke EV child watchers :/.
347
3485.0 Sun Aug 9 17:32:09 CEST 2009
349 - officially introduce and document the AE API.
350 - lowering the timeout at runtime did not have immediate effect
351 in AnyEvent::Handle.
352 - AE::Handle now has separate and independent read and write timeout
353 settings.
354 - probe for Guard module when AnyEvent::Util is loaded, not at
355 runtime (Event doesn't like eval "use" from callbacks, and it avoids
356 exporting the probe function).
357 - Impl::IOAsync's io watchers did not use the filehandle emulation
358 code to work around it's watcher limits.
359 - work around IO::Async's problems with overloaded objects.
360 - try to disable wearnings inside IO::Async, it's just too buggy.
361 - do no longer use Event by default, only EV or Perl, Event is too buggy.
362 - hackishly provide fast AE:: interface when EV or Perl is the backend.
363 - inherit VERSION from AnyEvent::VERSION in most modules.
364
3654.91 Thu Aug 6 15:42:45 CEST 2009
7 - AE::Handle::starttls could get out of sync when the read buffer 366 - AE::Handle::starttls could get out of sync when the read buffer
8 already contains some TLS handshake. 367 already contains some TLS handshake.
9 - AE::Handle did not properly free the TLS session, causing bigger 368 - AE::Handle did not properly free the TLS session, causing bigger
10 memory leaks in Net::SSLeay than expected (reported by toaster). 369 memory leaks in Net::SSLeay than expected (reported by toaster).
11 - AE::Socket::tcp_connect will now artificially delay invoking 370 - AE::Socket::tcp_connect will now artificially delay invoking
12 the callback to avoid returning after invoking the callback. 371 the callback to avoid returning after invoking the callback.
372 - convert many internal watcher uses to AE API.
13 373
144.9 Sat Aug 1 11:07:01 CEST 2009 3744.9 Sat Aug 1 11:07:01 CEST 2009
15 - Glib silently fails when registering a timeout with 375 - Glib silently fails when registering a timeout with
16 a negative value, so avoid that. 376 a negative value, so avoid that.
17 - call condvar callback immediately when it is set 377 - call condvar callback immediately when it is set
… …
34 - new function: AnyEvent::Socket::format_hostport. 394 - new function: AnyEvent::Socket::format_hostport.
35 395
364.881 Tue Jul 28 12:51:53 CEST 2009 3964.881 Tue Jul 28 12:51:53 CEST 2009
37 - work around a bug in local in pre-5.10 perls, causing 397 - work around a bug in local in pre-5.10 perls, causing
38 AnyEvent::Handle to recurse when it should not 398 AnyEvent::Handle to recurse when it should not
39 (analyzed by elmex). 399 (analysed by elmex).
40 400
414.88 Tue Jul 28 04:04:37 CEST 2009 4014.88 Tue Jul 28 04:04:37 CEST 2009
42 - re-bless the handle into a dummy package after calling 402 - re-bless the handle into a dummy package after calling
43 AnyEvent::Handle::destroy, so the user does not need to check 403 AnyEvent::Handle::destroy, so the user does not need to check
44 for errors after every push_write etc. 404 for errors after every push_write etc.
… …
2664.3 Fri Oct 3 09:18:43 CEST 2008 6264.3 Fri Oct 3 09:18:43 CEST 2008
267 - AnyEvent will now install a no-op signal handler for SIGPIPE, 627 - AnyEvent will now install a no-op signal handler for SIGPIPE,
268 unless one has been installed already. 628 unless one has been installed already.
269 - warn about not loaidng AnyEvent::Impl::POE early enough 629 - warn about not loaidng AnyEvent::Impl::POE early enough
270 (patch by Adam Rosenstein). 630 (patch by Adam Rosenstein).
271 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 631 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
272 - fix a bug in where in SSL connect mode, the 632 - fix a bug in where in SSL connect mode, the
273 client would first wait for some data by the server and 633 client would first wait for some data by the server and
274 otherwise hang. (reported and analyzed in an absolutely 634 otherwise hang. (reported and analysed in an absolutely
275 exemplary manner by Adam Rosenstein). 635 exemplary manner by Adam Rosenstein).
276 - fix a bug in where SSL EOF would not be treated 636 - fix a bug in where SSL EOF would not be treated
277 as stream EOF, putting the connection into a hung state. 637 as stream EOF, putting the connection into a hung state.
278 (reported and analyzed in an absolutely exemplary manner by Adam 638 (reported and analysed in an absolutely exemplary manner by Adam
279 Rosenstein). 639 Rosenstein).
280 - fix a potential segfault when the TLS context 640 - fix a potential segfault when the TLS context
281 would go missing in a read callback (e.g. due to stoptls) - 641 would go missing in a read callback (e.g. due to stoptls) -
282 Net::SSLeay of course makes no type checking whatsoever. 642 Net::SSLeay of course makes no type checking whatsoever.
283 - AnyEvent::Handle will not stop the read watcher in TLS mode. 643 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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