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

Comparing AnyEvent/Changes (file contents):
Revision 1.296 by root, Wed Jul 22 23:07:13 2009 UTC vs.
Revision 1.461 by root, Sat Aug 13 02:20:29 2011 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: maybe support "peerip" verificstion, i.e. getpeername diretcly as 3TODO: document TCP_*** constants
4additional peername? 4TODO: catch AE::xxx for anyevent strict and anyevent debug
5TODO: anyevent::trace? or debug tracing all watchers?
6TODO: anyevent strict checking for $_ changes?
5 7
8TODO: redesign Impl
9TODO: make ae::run mroe robust against eintr, Andrew Suffield
10TODO: IO::Async sometimes cancels the wrong timer when you call cancel_timer -wait for IO::Async release.
11TODO: wrap
12TODO: shell on env
13TODO: FLTK docs
14 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
15 the TTL as fourth element - this affects ->resolve directly,
16 and ->request and ->dns_unpack indirectly. Convenience functions
17 such as AnyEvent::DNS::{a,txt,mx,srv} are not affected.
18 - INCOMPATIBLE CHANGE: backend authors now should not implement
19 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
20 methods.
21 - DEPRECIATION: please change
22 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
23 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
24 in your programs, for future compatibility.
25 - *blush* idle watcher emulation was borked.
26 - AnyEvent::Strict now checks the AE::xxx functions as well.
27 - querying too long domain names would barf deep inside AnyEvent::DNS - since
28 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
29 for too long domain names explicitly and passes an undef to the callback
30 instead (reported by James Bromberger).
31 - new function: AnyEvent::postpone.
32 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
33 Invest extra overhead to work around this peculiar behaviour.
34 - split the pure perl backend into a loop module and an impl
35 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
36 - some read types (regex, netstring, json, storable) could
37 errornously keep an AE::Handle object alive on their own
38 (reported by Mohammad Toossi).
39 - add more workarounds around child watcher bugs in IO::Async
40 and (newly introduced) in POE.
41 - use glib child watchers and try to work around its limitations,
42 if possible (based on a patch by Kevin Ryde).
43 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
44 - many read types in AnyEvent::Handle would malfunction during
45 transitions to TLS because of a "delete $self->{rbuf}".
46 (reported by Gerald Galster).
47 - resolve AnyEvent::Util::guard at first-use-time, not at load time,
48 for further speed and memory savings when not used.
49 - improve condvar blocking wait performance (probably only noticable
50 with faster backends), also make it easier for event loops that
51 do not support blocking waits.
52 - PERL_ANYEVENT_MODEL now also allows full module names, although
53 the usefulness of this feature is not yet clear.
54 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
55 - add t/80_ssltext.t.
56 - no longer use AUTOLOAD for the initial loading (for no very good reason).
57
585.34 Thu May 12 09:59:41 CEST 2011
59 - AE::TLS didn't support both cert_file and key at the same time,
60 due to some copy&paste bug (reported and analyzed by vti).
61 - improve AE::Handle timeout documentation slightly.
62
635.33 Mon Feb 7 21:51:41 CET 2011
64 - never appeared on CPAN, for unknown reasons.
65 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
66 object when it's already gone (analyzed by Ryan Bullock).
67 - update idna algorithm(s) to UTS#46 version 6.0.0.
68 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
69 uts46 says it's an error, and testsuite makes more sense.
70 - remove leading dots in idn_to_ascii - testsuite says to remove,
71 uts46 says it's an error, and testsuite makes more sense.
72 - fixed a bug in idn ignored character processing.
73 - passes the uts46 6.0.0 testsuite except for three cases which
74 apparently are bugs in the testsuite itself.
75 - passes the uts46 6.0.1 testsuite except for eleven cases which
76 apparently are bugs in the testsuite itself.
77
785.31 Sun Jan 23 11:49:19 CET 2011
79 - work around a POE bug where POE would first poll for new events and
80 invoke handlers BEFORE handling the message queue.
81 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
82 - stop_read now does just that in TLS mode. see its
83 documentation for caveats.
84 - update to common::sense 3.4 warning set (some warnings in perl impose
85 an arbitrary 32kb limit on regex matches, which affects
86 AnyEvent::Handle).
87 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
88 - new experimental AE::Handle::resettls method.
89
905.3 Fri Dec 31 05:48:13 CET 2010
91 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
92 issues with IO::Async have been solved in that version, including
93 the availabilty of a default loop!
94 - implement new wbuf_max parameter for AnyEvent::Handle.
95 - added Cocoa::EventLoop backend by Daisuke Murase.
96 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
97 - clarify handle on_prepare documentation.
98 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
99 to avoid leaking them to any child processes.
100 - try to force files into the arch-dependend lib directory, for the
101 half-"turly obsessive compulsive person"(s) that depend on it.
102
1035.29 Sun Dec 5 10:49:21 CET 2010
104 - convert EV backend to EV 4.00 API (so better upgrade EV too).
105 - AE::Handle: implement dynamic read_size adjustment, add
106 max_read_size parameter, reduce default read_size to 2048.
107 - add some FAQ entries for servers and/or on_eof vs. on_error.
108 - work around OS bugs (cygwin again) when lingering in AE::Handle.
109 - reduce memory usage slightly when connect was used in AE::Handle.
110 - wrok around more segfaults in Tk.
111 - document tls_ctx => undef as valid.
112 - detect WNOHANG value at compiletime.
113 - include some probably linux-only support for building AnyEvent
114 as part of the perl core.
115 - improve compatibility of t/08_idna.t to perl 5.8.
116 - make AnyEvent::Strict truly optional (in case somebody builds a
117 minimal perl...).
118
1195.28 Wed Oct 13 04:14:23 CEST 2010
120 - due to a glitch, AnyEvent's internal getprotobyname wasn't
121 actually used in tcp_server.
122 - implement AnyEvent::CondVar->new method, which was documented but
123 missing.
124 - support autoloading of read types in unshift_read.
125 - AnyEvent::Strict: do the test for invalid fh arguments differently,
126 resulting in more informative error messages.
127 - new AnyEvent::FAQ manpage.
128 - clarify that you can actually call push_read etc. in on_read.
129 - put keys and certs in /tmp, not into the current directory,
130 and document this in the SECURITY CONSIDERATIONS sections
131 of AnyEvent::TLS.
132 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
133 requested by Ron Isaacson.
134 - document $AnyEvent::DNS::RESOLVER variable.
135 - incorporated major typo patches by Abhijit Menon-Sen.
136 - AnyEvent::Handle now croaks when negative timeouts are passed.
137 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
138 undocumented, but who knows when they might come in handy.
139
1405.271 Tue Jun 8 12:05:46 CEST 2010
141 - backport to perl 5.8.x.
142
1435.27 Sun Jun 6 12:12:05 CEST 2010
144 - postpone differently in AnyEvent::Socket now, as
145 when not, canceling the connection attempt might fail
146 (found by Felix Antonius Wilhelm Ostmann).
147 - explicitly check for non-stream sockets in AE::Handle, too many
148 clueless people fell into the trap of this somehow working.
149 - simplified and reworked the "OTHER MODULES" section.
150 - better/more condvar examples.
151
1525.261 Wed Apr 28 16:13:36 CEST 2010
153 - AF_INET6 was not properly used from Socket6 during configuration
154 time (found by Andy Grundman).
155
1565.26 Mon Apr 12 04:49:35 CEST 2010
157 - don't generate spurious readyness notifications when select
158 returns EINTR in the pure perl backend (analysed and testcase
159 by Bas Denissen).
160 - give same examples in SYNOPSIS sections of AnyEvent and AE.
161 - provider faster implementations for AE::time/now/now_update for
162 EV and Perl backends.
163 - provide faster AE::cv and AE::time implementations for all
164 backends.
165 - no longer support lower/mixed-case signal names (this was
166 never documented nor universally supported).
167 - some more memory ยต-optimisations, and somewhat less messy
168 code to cope with both AnyEvent and AE APIs.
169
1705.251 Sat Mar 13 00:58:21 CET 2010
171 - make test failed if it was run before make install on perl <=
172 5.8.8, and older versions of perl make our live very difficult,
173 so write constants.pl during Makefile.PL time.
174
1755.25 Sat Mar 13 00:23:14 CET 2010
176 - fix a race condition in AnyEvent::Handle that would cause
177 a "bio_read: unsupported method" error (in _another_
178 ssl connection) after user code threw an exception.
179 - added AnyEvent::Handle->destroyed method.
180 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
181 - the AnyEvent::Util::fh_nonblocking function worked only
182 by lucky accident on win32.
183 - smaller and faster AnyEvent::Util::fh_nonblocking.
184 - when the (required!) Time::HiRes module is missing, AnyEvent
185 did not fall back to built-in time properly.
186 - do not load Fcntl at runtime, saving memory and loading time.
187 - precompile a number of constants and use them instead of runtime
188 detection and eval.
189 - free detection code after detection and similar memory optimisations.
190 - Perl backend timer interval best effort drift has been improved
191 (same algorithm as EV).
192 - update unicode idna mapping table.
193
1945.24 Tue Jan 5 11:39:43 CET 2010
195 - cygwin never reports errors from failed connects. 1.5 just gives
196 you continous readyness and EAGAIN, 1.7 is even more broken and
197 just hangs. work around both issues in a cygwin-specific hack.
198 - improve idle watcher documentation slightly.
199
2005.23 Sun Dec 20 23:48:00 CET 2009
201 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
202 - implement punycode_encode/decode, idn_nameprep,
203 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
204 - provide $AE::VERSION.
205 - removed traces of "no strict 'refs'".
206
2075.22 Sat Dec 5 03:51:13 CET 2009
208 - downgrade-or-fail in AnyEvent::Handle::push_write, to
209 diagnose encoding failures earlier and more succinctly.
210 (this works around bugs in perl, throwing away encoding info
211 when passing scalar data to extensions).
212 - add more examples to AnyEvent::Socket manpage.
213 - upgrade internal warning set to the same as common::sense 2.03.
214 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
215 (John Beppu).
216 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
217 - add google public dns servers to fallback server set and make sure
218 we load-balance properly between the three sets. also add all
219 fallback dns servers, not just a random one, to each dns config.
220
2215.21 Thu Nov 19 02:48:47 CET 2009
222 - fix a problem where socket constants were called with parameters
223 (spotted by David Friedland).
224 - fork_call never use'd POSIX (reported by Daisuke Maki).
225 - improve perl 5.6 compatibility further (but it still won't work
226 unless you rip out everything but the core).
227 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
228 uses 10MB of RAM, but doesn't open a console window. *sigh*.
229
2305.202 Wed Oct 14 22:35:44 CEST 2009
231 - AnyEvent::DNS would unexpectedly clobber $_ under windows
232 (analysed by Matthias Waldorf).
233 - AnyEvent::Handle::run_cmd can now pass the PID of the
234 newly-created process, which is much less useful than it might
235 sound (based on patch by Yann Kerherve).
236
2375.201 Tue Sep 29 12:09:25 CEST 2009
238 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
239 (reported by Torsten Foertsch).
240 - common::sense 2.0 could cause tcp_server to throw an exception
241 (analysed by elmex).
242
2435.2 Mon Sep 14 07:04:49 CEST 2009
244 - INCOMPATIBLE CHANGE: do no longer support register_read_type
245 and register_write_type in AnyEvent::Handle, instead support
246 package names (the facility was mostly abused).
247 - implement "packagename-as-read/write type" support in
248 AnyEvent::Handle.
249 - AnyEvent::Handle: new options "keepalive" and "oobinline".
250 - oobinline set by default to avoid security issues.
251 - the pure-perl event loop backend wrongly detected times() fallback
252 support (spotted by Pavel Boldin).
253
2545.12 Tue Sep 1 20:26:50 CEST 2009
255 - be more lenient when parsing resolv.conf files, as some people
256 use hashmarks as comment indicator inside directives (reported
257 by Michael S. Fischer).
258 - use same set of warnings as common::sense 2.0.
259 - fix a potential 32 bit overflow issue due to perl having problems
260 with large hex constants in the Perl backend.
261 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
262 by default.
263 - fix some prototypes.
264 - (optionally) require Async::Interrupt 1.02.
265 - replace opendns fallbacks by verizon/level3 public dns servers.
266
2675.112 Fri Aug 21 13:59:04 CEST 2009
268 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
269 using starttls with an illegal/uncreatable tls context.
270 - port to broken win32 perls.
271
2725.111 Sun Aug 16 18:44:36 CEST 2009
273 - perl errornously requires sockaddr_un structures returned by the
274 kernel to have a specific length. work around this bug by padding
275 sockaddr_un structures (found to fail on OS X, as perl only uses
276 the correct code for GNU/Linux, but likely to fail on other
277 OSes as well).
278
2795.11 Wed Aug 12 17:49:37 CEST 2009
280 - fixed a bug in run_cmd with input from filename.
281 - create a json coder object when none is given in write direction
282 to reduce dependency on JSON.
283
2845.1 Tue Aug 11 03:17:32 CEST 2009
285 - tcp_connect now reports the correct error in $!, not an unrelated
286 one (was broken since 4.91).
287 - AE::Handle did not properly clear rtimeout/wtimeout watchers
288 when retrying a tcp_connect.
289 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
290 - fixed wrong documentation in AE::Socket.
291
2925.01 Mon Aug 10 03:16:32 CEST 2009
293 - last release broke EV child watchers :/.
294
2955.0 Sun Aug 9 17:32:09 CEST 2009
296 - officially introduce and document the AE API.
297 - lowering the timeout at runtime did not have immediate effect
298 in AnyEvent::Handle.
299 - AE::Handle now has separate and independent read and write timeout
300 settings.
301 - probe for Guard module when AnyEvent::Util is loaded, not at
302 runtime (Event doesn't like eval "use" from callbacks, and it avoids
303 exporting the probe function).
304 - Impl::IOAsync's io watchers did not use the filehandle emulation
305 code to work around it's watcher limits.
306 - work around IO::Async's problems with overloaded objects.
307 - try to disable wearnings inside IO::Async, it's just too buggy.
308 - do no longer use Event by default, only EV or Perl, Event is too buggy.
309 - hackishly provide fast AE:: interface when EV or Perl is the backend.
310 - inherit VERSION from AnyEvent::VERSION in most modules.
311
3124.91 Thu Aug 6 15:42:45 CEST 2009
313 - AE::Handle::starttls could get out of sync when the read buffer
314 already contains some TLS handshake.
315 - AE::Handle did not properly free the TLS session, causing bigger
316 memory leaks in Net::SSLeay than expected (reported by toaster).
317 - AE::Socket::tcp_connect will now artificially delay invoking
318 the callback to avoid returning after invoking the callback.
319 - convert many internal watcher uses to AE API.
320
3214.9 Sat Aug 1 11:07:01 CEST 2009
322 - Glib silently fails when registering a timeout with
323 a negative value, so avoid that.
324 - call condvar callback immediately when it is set
325 after the condition is already signalled.
326 - check rbuf_max condition only after trying to consume
327 data.
328 - normalise signal numbers to names when using the ae's
329 signal handling, but do not document this (yet).
330 - pure perl signal emulation did not properly set nonblocking
331 mode on the signal pipe on win32 (this is unlikely to have caused
332 any issues).
333 - new module: AnyEvent::Debug.
334 - AnyEvent::Strict now checks that a signal was specified
335 by name (not name or number).
336 - reduce memory footprint in typical cases by ~50kb by
337 only compiling signal code when necessary.
338 - add AnyEvent::Handle->rbuf_max.
339 - grab the AE namespace for future new API, implement stubs for
340 said future/alternative API.
341 - new function: AnyEvent::Socket::format_hostport.
342
3434.881 Tue Jul 28 12:51:53 CEST 2009
344 - work around a bug in local in pre-5.10 perls, causing
345 AnyEvent::Handle to recurse when it should not
346 (analysed by elmex).
347
3484.88 Tue Jul 28 04:04:37 CEST 2009
349 - re-bless the handle into a dummy package after calling
350 AnyEvent::Handle::destroy, so the user does not need to check
351 for errors after every push_write etc.
352 - do not attempt to run t/02_signals.t on obviously broken
353 platforms.
354
3554.87 Sun Jul 26 02:06:16 CEST 2009
6 - do not attempt to linger when there is no longer a valid fh. 356 - do not attempt to linger when there is no longer a valid fh.
7 - enforce tls mode to be either accept or connect, do not simply 357 - enforce tls mode to be either accept or connect, do not simply
8 segfault in Net::SSLeay. 358 segfault in Net::SSLeay.
359 - AnyEvent::Handle can now call tcp_connect itself (new parameters
360 connect, on_prepare, on_connect and on_connect_error). Updated
361 tutorial accordingly.
362 - add AnyEvent::Impl::Irssi backend.
9 363
104.86 Mon Jul 20 23:52:29 CEST 2009 3644.86 Mon Jul 20 23:52:29 CEST 2009
11 - since the verbose warning is not enough, explicitly document 365 - since the verbose warning is not enough, explicitly document
12 that versions before 1.33 of Net::SSLeay are not secure. 366 that versions before 1.33 of Net::SSLeay are not secure.
13 - work around signal handling races in Event and (...) Event::Lib. 367 - work around signal handling races in Event and (...) Event::Lib.
14 - try to align signal-race timer to full-second boundaries. 368 - try to align signal-race timer to full-second boundaries.
15 - work around Tk not liking negative timeouts. 369 - work around Tk not liking negative timeouts.
16 - don't complain of different grades of environmental unfriendlyness 370 - don't complain of different grades of environmental unfriendlyness
17 in IO::ASync. 371 in IO::Async.
18 372
194.85 Sat Jul 18 06:16:14 CEST 2009 3734.85 Sat Jul 18 06:16:14 CEST 2009
20 - nail the signal race problem in perl once and for all 374 - nail the signal race problem in perl once and for all
21 (see $AnyEvent::MAX_SIGNAL_LATENCY). 375 (see $AnyEvent::MAX_SIGNAL_LATENCY).
22 - take advantage of Async::Interrupt if it is available. 376 - take advantage of Async::Interrupt if it is available.
… …
2194.3 Fri Oct 3 09:18:43 CEST 2008 5734.3 Fri Oct 3 09:18:43 CEST 2008
220 - AnyEvent will now install a no-op signal handler for SIGPIPE, 574 - AnyEvent will now install a no-op signal handler for SIGPIPE,
221 unless one has been installed already. 575 unless one has been installed already.
222 - warn about not loaidng AnyEvent::Impl::POE early enough 576 - warn about not loaidng AnyEvent::Impl::POE early enough
223 (patch by Adam Rosenstein). 577 (patch by Adam Rosenstein).
224 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 578 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
225 - fix a bug in where in SSL connect mode, the 579 - fix a bug in where in SSL connect mode, the
226 client would first wait for some data by the server and 580 client would first wait for some data by the server and
227 otherwise hang. (reported and analyzed in an absolutely 581 otherwise hang. (reported and analysed in an absolutely
228 exemplary manner by Adam Rosenstein). 582 exemplary manner by Adam Rosenstein).
229 - fix a bug in where SSL EOF would not be treated 583 - fix a bug in where SSL EOF would not be treated
230 as stream EOF, putting the connection into a hung state. 584 as stream EOF, putting the connection into a hung state.
231 (reported and analyzed in an absolutely exemplary manner by Adam 585 (reported and analysed in an absolutely exemplary manner by Adam
232 Rosenstein). 586 Rosenstein).
233 - fix a potential segfault when the TLS context 587 - fix a potential segfault when the TLS context
234 would go missing in a read callback (e.g. due to stoptls) - 588 would go missing in a read callback (e.g. due to stoptls) -
235 Net::SSLeay of course makes no type checking whatsoever. 589 Net::SSLeay of course makes no type checking whatsoever.
236 - AnyEvent::Handle will not stop the read watcher in TLS mode. 590 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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