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

Comparing AnyEvent/Changes (file contents):
Revision 1.309 by root, Wed Jul 29 13:15:55 2009 UTC vs.
Revision 1.487 by root, Thu Aug 25 00:25:55 2011 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: release a:i 1.0 3TODO: document TCP_*** constants
4
5TODO: AnyEvent::Log
6TODO: ae::log examples
7TODO: consolide slogging, e.g AnyEvent::Debug
8TODO: /etc/hosts, for resolve_sockaddr?
96.01
10 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
11 the TTL as fourth element - this affects ->resolve, ->request
12 and ->dns_unpack indirectly. Convenience functions such as
13 AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither is
14 AnyEvent::Socket.
15 - INCOMPATIBLE CHANGE: backend authors now should not implement
16 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
17 methods.
18 - INCOMPATIBLE CHANGE: verbosity levels have been redefined, see
19 AnyEvent (PERL_ANYEVENT_VERBOSE) and AnyEvent::Log for details.
20 - DEPRECIATION: please change
21 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
22 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
23 in your programs, for future compatibility.
24 - *blush* idle watcher emulation was borked.
25 - AnyEvent::Strict now checks the AE::xxx functions as well.
26 - querying too long domain names would barf deep inside AnyEvent::DNS - since
27 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
28 for too long domain names explicitly and passes an undef to the callback
29 instead (reported by James Bromberger).
30 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
31 - new functions: AnyEvent::postpone, AnyEvent::log.
32 - new module: AnyEvent::Log, for simple logging needs.
33 - new env variable, PERL_ANYEVENT_LOG, for fine-grained logging config.
34 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
35 a debug shell.
36 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
37 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
38 - an empty PERL_ANYEVENT_RESOLV_CONF now selects the default
39 configuration.
40 - env variables can now be specified using AE_xyz instead of
41 PERL_ANYEVENT_xyz, subject to some rules.
42 - lots of new AnyEvent::Debug functionality.
43 - AnyEvent::Strict now checks for modifications $_ by the callback
44 (by setting it to readonly while executing the callback, if possible).
45 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
46 Invest extra overhead to work around this peculiar behaviour.
47 - split the pure perl backend into a loop module and an impl
48 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
49 - some read types (regex, netstring, json, storable) could
50 errornously keep an AE::Handle object alive on their own
51 (reported by Mohammad Toossi).
52 - add more workarounds around child watcher bugs in IO::Async
53 and (newly introduced) in POE.
54 - use glib child watchers and try to work around its limitations,
55 if possible (based on a patch by Kevin Ryde).
56 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
57 - many read types in AnyEvent::Handle would malfunction during
58 transitions to TLS because of a "delete $self->{rbuf}".
59 (reported by Gerald Galster).
60 - improve condvar blocking wait performance (probably only noticable
61 with faster backends), also make it easier for event loops that
62 do not support blocking waits.
63 - PERL_ANYEVENT_MODEL now also allows full module names, although
64 the usefulness of this feature is not yet clear.
65 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
66 - add t/80_ssltext.t.
67 - no longer use AUTOLOAD for the initial loading (for no very good reason).
68 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
69 - do not run event loop detection multiple times when programs call their cached
70 copies of AnyEvent::detect.
71 - parse_hostport now accepts absolute paths als unix domain sockets.
72 - add Task::Weaken as dependency (as recommends only at this time).
73
746.0 (not officially released, but was on CPAN for about 15 minutes :/)
75
765.34 Thu May 12 09:59:41 CEST 2011
77 - AE::TLS didn't support both cert_file and key at the same time,
78 due to some copy&paste bug (reported and analyzed by vti).
79 - improve AE::Handle timeout documentation slightly.
80
815.33 Mon Feb 7 21:51:41 CET 2011
82 - never appeared on CPAN, for unknown reasons.
83 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
84 object when it's already gone (analyzed by Ryan Bullock).
85 - update idna algorithm(s) to UTS#46 version 6.0.0.
86 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
87 uts46 says it's an error, and testsuite makes more sense.
88 - remove leading dots in idn_to_ascii - testsuite says to remove,
89 uts46 says it's an error, and testsuite makes more sense.
90 - fixed a bug in idn ignored character processing.
91 - passes the uts46 6.0.0 testsuite except for three cases which
92 apparently are bugs in the testsuite itself.
93 - passes the uts46 6.0.1 testsuite except for eleven cases which
94 apparently are bugs in the testsuite itself.
95
965.31 Sun Jan 23 11:49:19 CET 2011
97 - work around a POE bug where POE would first poll for new events and
98 invoke handlers BEFORE handling the message queue.
99 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
100 - stop_read now does just that in TLS mode. see its
101 documentation for caveats.
102 - update to common::sense 3.4 warning set (some warnings in perl impose
103 an arbitrary 32kb limit on regex matches, which affects
104 AnyEvent::Handle).
105 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
106 - new experimental AE::Handle::resettls method.
107
1085.3 Fri Dec 31 05:48:13 CET 2010
109 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
110 issues with IO::Async have been solved in that version, including
111 the availabilty of a default loop!
112 - implement new wbuf_max parameter for AnyEvent::Handle.
113 - added Cocoa::EventLoop backend by Daisuke Murase.
114 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
115 - clarify handle on_prepare documentation.
116 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
117 to avoid leaking them to any child processes.
118 - try to force files into the arch-dependend lib directory, for the
119 half-"turly obsessive compulsive person"(s) that depend on it.
120
1215.29 Sun Dec 5 10:49:21 CET 2010
122 - convert EV backend to EV 4.00 API (so better upgrade EV too).
123 - AE::Handle: implement dynamic read_size adjustment, add
124 max_read_size parameter, reduce default read_size to 2048.
125 - add some FAQ entries for servers and/or on_eof vs. on_error.
126 - work around OS bugs (cygwin again) when lingering in AE::Handle.
127 - reduce memory usage slightly when connect was used in AE::Handle.
128 - wrok around more segfaults in Tk.
129 - document tls_ctx => undef as valid.
130 - detect WNOHANG value at compiletime.
131 - include some probably linux-only support for building AnyEvent
132 as part of the perl core.
133 - improve compatibility of t/08_idna.t to perl 5.8.
134 - make AnyEvent::Strict truly optional (in case somebody builds a
135 minimal perl...).
136
1375.28 Wed Oct 13 04:14:23 CEST 2010
138 - due to a glitch, AnyEvent's internal getprotobyname wasn't
139 actually used in tcp_server.
140 - implement AnyEvent::CondVar->new method, which was documented but
141 missing.
142 - support autoloading of read types in unshift_read.
143 - AnyEvent::Strict: do the test for invalid fh arguments differently,
144 resulting in more informative error messages.
145 - new AnyEvent::FAQ manpage.
146 - clarify that you can actually call push_read etc. in on_read.
147 - put keys and certs in /tmp, not into the current directory,
148 and document this in the SECURITY CONSIDERATIONS sections
149 of AnyEvent::TLS.
150 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
151 requested by Ron Isaacson.
152 - document $AnyEvent::DNS::RESOLVER variable.
153 - incorporated major typo patches by Abhijit Menon-Sen.
154 - AnyEvent::Handle now croaks when negative timeouts are passed.
155 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
156 undocumented, but who knows when they might come in handy.
157
1585.271 Tue Jun 8 12:05:46 CEST 2010
159 - backport to perl 5.8.x.
160
1615.27 Sun Jun 6 12:12:05 CEST 2010
162 - postpone differently in AnyEvent::Socket now, as
163 when not, canceling the connection attempt might fail
164 (found by Felix Antonius Wilhelm Ostmann).
165 - explicitly check for non-stream sockets in AE::Handle, too many
166 clueless people fell into the trap of this somehow working.
167 - simplified and reworked the "OTHER MODULES" section.
168 - better/more condvar examples.
169
1705.261 Wed Apr 28 16:13:36 CEST 2010
171 - AF_INET6 was not properly used from Socket6 during configuration
172 time (found by Andy Grundman).
173
1745.26 Mon Apr 12 04:49:35 CEST 2010
175 - don't generate spurious readyness notifications when select
176 returns EINTR in the pure perl backend (analysed and testcase
177 by Bas Denissen).
178 - give same examples in SYNOPSIS sections of AnyEvent and AE.
179 - provider faster implementations for AE::time/now/now_update for
180 EV and Perl backends.
181 - provide faster AE::cv and AE::time implementations for all
182 backends.
183 - no longer support lower/mixed-case signal names (this was
184 never documented nor universally supported).
185 - some more memory ยต-optimisations, and somewhat less messy
186 code to cope with both AnyEvent and AE APIs.
187
1885.251 Sat Mar 13 00:58:21 CET 2010
189 - make test failed if it was run before make install on perl <=
190 5.8.8, and older versions of perl make our live very difficult,
191 so write constants.pl during Makefile.PL time.
192
1935.25 Sat Mar 13 00:23:14 CET 2010
194 - fix a race condition in AnyEvent::Handle that would cause
195 a "bio_read: unsupported method" error (in _another_
196 ssl connection) after user code threw an exception.
197 - added AnyEvent::Handle->destroyed method.
198 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
199 - the AnyEvent::Util::fh_nonblocking function worked only
200 by lucky accident on win32.
201 - smaller and faster AnyEvent::Util::fh_nonblocking.
202 - when the (required!) Time::HiRes module is missing, AnyEvent
203 did not fall back to built-in time properly.
204 - do not load Fcntl at runtime, saving memory and loading time.
205 - precompile a number of constants and use them instead of runtime
206 detection and eval.
207 - free detection code after detection and similar memory optimisations.
208 - Perl backend timer interval best effort drift has been improved
209 (same algorithm as EV).
210 - update unicode idna mapping table.
211
2125.24 Tue Jan 5 11:39:43 CET 2010
213 - cygwin never reports errors from failed connects. 1.5 just gives
214 you continous readyness and EAGAIN, 1.7 is even more broken and
215 just hangs. work around both issues in a cygwin-specific hack.
216 - improve idle watcher documentation slightly.
217
2185.23 Sun Dec 20 23:48:00 CET 2009
219 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
220 - implement punycode_encode/decode, idn_nameprep,
221 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
222 - provide $AE::VERSION.
223 - removed traces of "no strict 'refs'".
224
2255.22 Sat Dec 5 03:51:13 CET 2009
226 - downgrade-or-fail in AnyEvent::Handle::push_write, to
227 diagnose encoding failures earlier and more succinctly.
228 (this works around bugs in perl, throwing away encoding info
229 when passing scalar data to extensions).
230 - add more examples to AnyEvent::Socket manpage.
231 - upgrade internal warning set to the same as common::sense 2.03.
232 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
233 (John Beppu).
234 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
235 - add google public dns servers to fallback server set and make sure
236 we load-balance properly between the three sets. also add all
237 fallback dns servers, not just a random one, to each dns config.
238
2395.21 Thu Nov 19 02:48:47 CET 2009
240 - fix a problem where socket constants were called with parameters
241 (spotted by David Friedland).
242 - fork_call never use'd POSIX (reported by Daisuke Maki).
243 - improve perl 5.6 compatibility further (but it still won't work
244 unless you rip out everything but the core).
245 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
246 uses 10MB of RAM, but doesn't open a console window. *sigh*.
247
2485.202 Wed Oct 14 22:35:44 CEST 2009
249 - AnyEvent::DNS would unexpectedly clobber $_ under windows
250 (analysed by Matthias Waldorf).
251 - AnyEvent::Handle::run_cmd can now pass the PID of the
252 newly-created process, which is much less useful than it might
253 sound (based on patch by Yann Kerherve).
254
2555.201 Tue Sep 29 12:09:25 CEST 2009
256 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
257 (reported by Torsten Foertsch).
258 - common::sense 2.0 could cause tcp_server to throw an exception
259 (analysed by elmex).
260
2615.2 Mon Sep 14 07:04:49 CEST 2009
262 - INCOMPATIBLE CHANGE: do no longer support register_read_type
263 and register_write_type in AnyEvent::Handle, instead support
264 package names (the facility was mostly abused).
265 - implement "packagename-as-read/write type" support in
266 AnyEvent::Handle.
267 - AnyEvent::Handle: new options "keepalive" and "oobinline".
268 - oobinline set by default to avoid security issues.
269 - the pure-perl event loop backend wrongly detected times() fallback
270 support (spotted by Pavel Boldin).
271
2725.12 Tue Sep 1 20:26:50 CEST 2009
273 - be more lenient when parsing resolv.conf files, as some people
274 use hashmarks as comment indicator inside directives (reported
275 by Michael S. Fischer).
276 - use same set of warnings as common::sense 2.0.
277 - fix a potential 32 bit overflow issue due to perl having problems
278 with large hex constants in the Perl backend.
279 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
280 by default.
281 - fix some prototypes.
282 - (optionally) require Async::Interrupt 1.02.
283 - replace opendns fallbacks by verizon/level3 public dns servers.
284
2855.112 Fri Aug 21 13:59:04 CEST 2009
286 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
287 using starttls with an illegal/uncreatable tls context.
288 - port to broken win32 perls.
289
2905.111 Sun Aug 16 18:44:36 CEST 2009
291 - perl errornously requires sockaddr_un structures returned by the
292 kernel to have a specific length. work around this bug by padding
293 sockaddr_un structures (found to fail on OS X, as perl only uses
294 the correct code for GNU/Linux, but likely to fail on other
295 OSes as well).
296
2975.11 Wed Aug 12 17:49:37 CEST 2009
298 - fixed a bug in run_cmd with input from filename.
299 - create a json coder object when none is given in write direction
300 to reduce dependency on JSON.
301
3025.1 Tue Aug 11 03:17:32 CEST 2009
303 - tcp_connect now reports the correct error in $!, not an unrelated
304 one (was broken since 4.91).
305 - AE::Handle did not properly clear rtimeout/wtimeout watchers
306 when retrying a tcp_connect.
307 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
308 - fixed wrong documentation in AE::Socket.
309
3105.01 Mon Aug 10 03:16:32 CEST 2009
311 - last release broke EV child watchers :/.
312
3135.0 Sun Aug 9 17:32:09 CEST 2009
314 - officially introduce and document the AE API.
315 - lowering the timeout at runtime did not have immediate effect
316 in AnyEvent::Handle.
317 - AE::Handle now has separate and independent read and write timeout
318 settings.
319 - probe for Guard module when AnyEvent::Util is loaded, not at
320 runtime (Event doesn't like eval "use" from callbacks, and it avoids
321 exporting the probe function).
322 - Impl::IOAsync's io watchers did not use the filehandle emulation
323 code to work around it's watcher limits.
324 - work around IO::Async's problems with overloaded objects.
325 - try to disable wearnings inside IO::Async, it's just too buggy.
326 - do no longer use Event by default, only EV or Perl, Event is too buggy.
327 - hackishly provide fast AE:: interface when EV or Perl is the backend.
328 - inherit VERSION from AnyEvent::VERSION in most modules.
329
3304.91 Thu Aug 6 15:42:45 CEST 2009
331 - AE::Handle::starttls could get out of sync when the read buffer
332 already contains some TLS handshake.
333 - AE::Handle did not properly free the TLS session, causing bigger
334 memory leaks in Net::SSLeay than expected (reported by toaster).
335 - AE::Socket::tcp_connect will now artificially delay invoking
336 the callback to avoid returning after invoking the callback.
337 - convert many internal watcher uses to AE API.
338
3394.9 Sat Aug 1 11:07:01 CEST 2009
4 - Glib silently fails when registering a timeout with 340 - Glib silently fails when registering a timeout with
5 a negative value, so avoid that. 341 a negative value, so avoid that.
342 - call condvar callback immediately when it is set
343 after the condition is already signalled.
344 - check rbuf_max condition only after trying to consume
345 data.
6 - normalise signal numbers to names when using the ae's 346 - normalise signal numbers to names when using the ae's
7 signal handling, but do not document this (yet). 347 signal handling, but do not document this (yet).
8 - pure perl signal emulation did not properly set nonblocking 348 - pure perl signal emulation did not properly set nonblocking
9 mode on the signal pipe on win32 (this is unlikely to have caused 349 mode on the signal pipe on win32 (this is unlikely to have caused
10 any issues). 350 any issues).
351 - new module: AnyEvent::Debug.
11 - AnyEvent::Strict now checks that a signal was specified 352 - AnyEvent::Strict now checks that a signal was specified
12 by name (not name or number). 353 by name (not name or number).
13 - reduce memory footprint in typical cases by ~50kb by 354 - reduce memory footprint in typical cases by ~50kb by
14 only compiling signal code when necessary. 355 only compiling signal code when necessary.
356 - add AnyEvent::Handle->rbuf_max.
357 - grab the AE namespace for future new API, implement stubs for
358 said future/alternative API.
359 - new function: AnyEvent::Socket::format_hostport.
15 360
164.881 Tue Jul 28 12:51:53 CEST 2009 3614.881 Tue Jul 28 12:51:53 CEST 2009
17 - work around a bug in local in pre-5.10 perls, causing 362 - work around a bug in local in pre-5.10 perls, causing
18 AnyEvent::Handle to recurse when it should not 363 AnyEvent::Handle to recurse when it should not
19 (analyzed by elmex). 364 (analysed by elmex).
20 365
214.88 Tue Jul 28 04:04:37 CEST 2009 3664.88 Tue Jul 28 04:04:37 CEST 2009
22 - re-bless the handle into a dummy package after calling 367 - re-bless the handle into a dummy package after calling
23 AnyEvent::Handle::destroy, so the user does not need to check 368 AnyEvent::Handle::destroy, so the user does not need to check
24 for errors after every push_write etc. 369 for errors after every push_write etc.
… …
2464.3 Fri Oct 3 09:18:43 CEST 2008 5914.3 Fri Oct 3 09:18:43 CEST 2008
247 - AnyEvent will now install a no-op signal handler for SIGPIPE, 592 - AnyEvent will now install a no-op signal handler for SIGPIPE,
248 unless one has been installed already. 593 unless one has been installed already.
249 - warn about not loaidng AnyEvent::Impl::POE early enough 594 - warn about not loaidng AnyEvent::Impl::POE early enough
250 (patch by Adam Rosenstein). 595 (patch by Adam Rosenstein).
251 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 596 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
252 - fix a bug in where in SSL connect mode, the 597 - fix a bug in where in SSL connect mode, the
253 client would first wait for some data by the server and 598 client would first wait for some data by the server and
254 otherwise hang. (reported and analyzed in an absolutely 599 otherwise hang. (reported and analysed in an absolutely
255 exemplary manner by Adam Rosenstein). 600 exemplary manner by Adam Rosenstein).
256 - fix a bug in where SSL EOF would not be treated 601 - fix a bug in where SSL EOF would not be treated
257 as stream EOF, putting the connection into a hung state. 602 as stream EOF, putting the connection into a hung state.
258 (reported and analyzed in an absolutely exemplary manner by Adam 603 (reported and analysed in an absolutely exemplary manner by Adam
259 Rosenstein). 604 Rosenstein).
260 - fix a potential segfault when the TLS context 605 - fix a potential segfault when the TLS context
261 would go missing in a read callback (e.g. due to stoptls) - 606 would go missing in a read callback (e.g. due to stoptls) -
262 Net::SSLeay of course makes no type checking whatsoever. 607 Net::SSLeay of course makes no type checking whatsoever.
263 - AnyEvent::Handle will not stop the read watcher in TLS mode. 608 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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