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

Comparing AnyEvent/Changes (file contents):
Revision 1.319 by root, Wed Aug 5 20:50:27 2009 UTC vs.
Revision 1.474 by root, Mon Aug 15 23:21:09 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: logging? / trace_level AnyEvent::Log
64.901 66.01
7
86.0 (not officially released)
9 - INCOMPATIBLE CHANGE: AnyEvent::DNS resource records now include
10 the TTL as fourth element - this affects ->resolve directly,
11 and ->request and ->dns_unpack indirectly. Convenience functions
12 such as AnyEvent::DNS::{a,txt,mx,srv} are not affected, neither
13 is AnyEvent::Socket.
14 - INCOMPATIBLE CHANGE: backend authors now should not implement
15 one_event or loop, but instead the AnyEvent::CondVar::_wait and _poll
16 methods.
17 - DEPRECIATION: please change
18 "use AnyEvent::Impl::Perl" to "use AnyEvent::Loop" and
19 "AnyEvent::Impl::Perl::loop" to "AnyEvent::Loop::run"
20 in your programs, for future compatibility.
21 - *blush* idle watcher emulation was borked.
22 - AnyEvent::Strict now checks the AE::xxx functions as well.
23 - querying too long domain names would barf deep inside AnyEvent::DNS - since
24 this is hard to avoid for applications, AnyEvent:.DNS::request now checks for
25 for too long domain names explicitly and passes an undef to the callback
26 instead (reported by James Bromberger).
27 - make AnyEvent::Util more robust against EINTR (analyzed by Andrew Suffield).
28 - new function: AnyEvent::postpone.
29 - new env variable, PERL_ANYEVENT_DEBUG_SHELL, to automatically start
30 a debug shell.
31 - new env variable, PERL_ANYEVENT_DEBUG_WRAP, to automatically enable
32 watcher instrumentation/wrapping (see AnyEvent::Debug::wrap).
33 - lots of new AnyEvent::Debug functionality.
34 - AnyEvent::Strict now checks for modifications $_ by the callback.
35 - IO::Async sometimes cancels the wrong timer when you call cancel_timer.
36 Invest extra overhead to work around this peculiar behaviour.
37 - split the pure perl backend into a loop module and an impl
38 module (AnyEvent::Loop and AnyEvent::Impl::Perl).
39 - some read types (regex, netstring, json, storable) could
40 errornously keep an AE::Handle object alive on their own
41 (reported by Mohammad Toossi).
42 - add more workarounds around child watcher bugs in IO::Async
43 and (newly introduced) in POE.
44 - use glib child watchers and try to work around its limitations,
45 if possible (based on a patch by Kevin Ryde).
46 - add support for FLTK (fltk2), via AnyEvent::Impl::FLTK2.
47 - many read types in AnyEvent::Handle would malfunction during
48 transitions to TLS because of a "delete $self->{rbuf}".
49 (reported by Gerald Galster).
50 - improve condvar blocking wait performance (probably only noticable
51 with faster backends), also make it easier for event loops that
52 do not support blocking waits.
53 - PERL_ANYEVENT_MODEL now also allows full module names, although
54 the usefulness of this feature is not yet clear.
55 - add tests for all supported event loops, conditional on PERL_ANYEVENT_LOOP_TESTS.
56 - add t/80_ssltext.t.
57 - no longer use AUTOLOAD for the initial loading (for no very good reason).
58 - tcp_server now tries to unlink unix domain sockets when it gets destroyed.
59 - do not run event loop detection multiple times when programs call their cached
60 copies of AnyEvent::detect.
61 - parse_hostport now accepts absolute paths als unix domain sockets.
62
635.34 Thu May 12 09:59:41 CEST 2011
64 - AE::TLS didn't support both cert_file and key at the same time,
65 due to some copy&paste bug (reported and analyzed by vti).
66 - improve AE::Handle timeout documentation slightly.
67
685.33 Mon Feb 7 21:51:41 CET 2011
69 - never appeared on CPAN, for unknown reasons.
70 - AnyEvent::Handle on_connect_error - do not try to destroy a handle
71 object when it's already gone (analyzed by Ryan Bullock).
72 - update idna algorithm(s) to UTS#46 version 6.0.0.
73 - preserve trailing dot in idn_to_ascii - testsuite says to preserve,
74 uts46 says it's an error, and testsuite makes more sense.
75 - remove leading dots in idn_to_ascii - testsuite says to remove,
76 uts46 says it's an error, and testsuite makes more sense.
77 - fixed a bug in idn ignored character processing.
78 - passes the uts46 6.0.0 testsuite except for three cases which
79 apparently are bugs in the testsuite itself.
80 - passes the uts46 6.0.1 testsuite except for eleven cases which
81 apparently are bugs in the testsuite itself.
82
835.31 Sun Jan 23 11:49:19 CET 2011
84 - work around a POE bug where POE would first poll for new events and
85 invoke handlers BEFORE handling the message queue.
86 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
87 - stop_read now does just that in TLS mode. see its
88 documentation for caveats.
89 - update to common::sense 3.4 warning set (some warnings in perl impose
90 an arbitrary 32kb limit on regex matches, which affects
91 AnyEvent::Handle).
92 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
93 - new experimental AE::Handle::resettls method.
94
955.3 Fri Dec 31 05:48:13 CET 2010
96 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
97 issues with IO::Async have been solved in that version, including
98 the availabilty of a default loop!
99 - implement new wbuf_max parameter for AnyEvent::Handle.
100 - added Cocoa::EventLoop backend by Daisuke Murase.
101 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
102 - clarify handle on_prepare documentation.
103 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
104 to avoid leaking them to any child processes.
105 - try to force files into the arch-dependend lib directory, for the
106 half-"turly obsessive compulsive person"(s) that depend on it.
107
1085.29 Sun Dec 5 10:49:21 CET 2010
109 - convert EV backend to EV 4.00 API (so better upgrade EV too).
110 - AE::Handle: implement dynamic read_size adjustment, add
111 max_read_size parameter, reduce default read_size to 2048.
112 - add some FAQ entries for servers and/or on_eof vs. on_error.
113 - work around OS bugs (cygwin again) when lingering in AE::Handle.
114 - reduce memory usage slightly when connect was used in AE::Handle.
115 - wrok around more segfaults in Tk.
116 - document tls_ctx => undef as valid.
117 - detect WNOHANG value at compiletime.
118 - include some probably linux-only support for building AnyEvent
119 as part of the perl core.
120 - improve compatibility of t/08_idna.t to perl 5.8.
121 - make AnyEvent::Strict truly optional (in case somebody builds a
122 minimal perl...).
123
1245.28 Wed Oct 13 04:14:23 CEST 2010
125 - due to a glitch, AnyEvent's internal getprotobyname wasn't
126 actually used in tcp_server.
127 - implement AnyEvent::CondVar->new method, which was documented but
128 missing.
129 - support autoloading of read types in unshift_read.
130 - AnyEvent::Strict: do the test for invalid fh arguments differently,
131 resulting in more informative error messages.
132 - new AnyEvent::FAQ manpage.
133 - clarify that you can actually call push_read etc. in on_read.
134 - put keys and certs in /tmp, not into the current directory,
135 and document this in the SECURITY CONSIDERATIONS sections
136 of AnyEvent::TLS.
137 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
138 requested by Ron Isaacson.
139 - document $AnyEvent::DNS::RESOLVER variable.
140 - incorporated major typo patches by Abhijit Menon-Sen.
141 - AnyEvent::Handle now croaks when negative timeouts are passed.
142 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
143 undocumented, but who knows when they might come in handy.
144
1455.271 Tue Jun 8 12:05:46 CEST 2010
146 - backport to perl 5.8.x.
147
1485.27 Sun Jun 6 12:12:05 CEST 2010
149 - postpone differently in AnyEvent::Socket now, as
150 when not, canceling the connection attempt might fail
151 (found by Felix Antonius Wilhelm Ostmann).
152 - explicitly check for non-stream sockets in AE::Handle, too many
153 clueless people fell into the trap of this somehow working.
154 - simplified and reworked the "OTHER MODULES" section.
155 - better/more condvar examples.
156
1575.261 Wed Apr 28 16:13:36 CEST 2010
158 - AF_INET6 was not properly used from Socket6 during configuration
159 time (found by Andy Grundman).
160
1615.26 Mon Apr 12 04:49:35 CEST 2010
162 - don't generate spurious readyness notifications when select
163 returns EINTR in the pure perl backend (analysed and testcase
164 by Bas Denissen).
165 - give same examples in SYNOPSIS sections of AnyEvent and AE.
166 - provider faster implementations for AE::time/now/now_update for
167 EV and Perl backends.
168 - provide faster AE::cv and AE::time implementations for all
169 backends.
170 - no longer support lower/mixed-case signal names (this was
171 never documented nor universally supported).
172 - some more memory ยต-optimisations, and somewhat less messy
173 code to cope with both AnyEvent and AE APIs.
174
1755.251 Sat Mar 13 00:58:21 CET 2010
176 - make test failed if it was run before make install on perl <=
177 5.8.8, and older versions of perl make our live very difficult,
178 so write constants.pl during Makefile.PL time.
179
1805.25 Sat Mar 13 00:23:14 CET 2010
181 - fix a race condition in AnyEvent::Handle that would cause
182 a "bio_read: unsupported method" error (in _another_
183 ssl connection) after user code threw an exception.
184 - added AnyEvent::Handle->destroyed method.
185 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
186 - the AnyEvent::Util::fh_nonblocking function worked only
187 by lucky accident on win32.
188 - smaller and faster AnyEvent::Util::fh_nonblocking.
189 - when the (required!) Time::HiRes module is missing, AnyEvent
190 did not fall back to built-in time properly.
191 - do not load Fcntl at runtime, saving memory and loading time.
192 - precompile a number of constants and use them instead of runtime
193 detection and eval.
194 - free detection code after detection and similar memory optimisations.
195 - Perl backend timer interval best effort drift has been improved
196 (same algorithm as EV).
197 - update unicode idna mapping table.
198
1995.24 Tue Jan 5 11:39:43 CET 2010
200 - cygwin never reports errors from failed connects. 1.5 just gives
201 you continous readyness and EAGAIN, 1.7 is even more broken and
202 just hangs. work around both issues in a cygwin-specific hack.
203 - improve idle watcher documentation slightly.
204
2055.23 Sun Dec 20 23:48:00 CET 2009
206 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
207 - implement punycode_encode/decode, idn_nameprep,
208 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
209 - provide $AE::VERSION.
210 - removed traces of "no strict 'refs'".
211
2125.22 Sat Dec 5 03:51:13 CET 2009
213 - downgrade-or-fail in AnyEvent::Handle::push_write, to
214 diagnose encoding failures earlier and more succinctly.
215 (this works around bugs in perl, throwing away encoding info
216 when passing scalar data to extensions).
217 - add more examples to AnyEvent::Socket manpage.
218 - upgrade internal warning set to the same as common::sense 2.03.
219 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
220 (John Beppu).
221 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
222 - add google public dns servers to fallback server set and make sure
223 we load-balance properly between the three sets. also add all
224 fallback dns servers, not just a random one, to each dns config.
225
2265.21 Thu Nov 19 02:48:47 CET 2009
227 - fix a problem where socket constants were called with parameters
228 (spotted by David Friedland).
229 - fork_call never use'd POSIX (reported by Daisuke Maki).
230 - improve perl 5.6 compatibility further (but it still won't work
231 unless you rip out everything but the core).
232 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
233 uses 10MB of RAM, but doesn't open a console window. *sigh*.
234
2355.202 Wed Oct 14 22:35:44 CEST 2009
236 - AnyEvent::DNS would unexpectedly clobber $_ under windows
237 (analysed by Matthias Waldorf).
238 - AnyEvent::Handle::run_cmd can now pass the PID of the
239 newly-created process, which is much less useful than it might
240 sound (based on patch by Yann Kerherve).
241
2425.201 Tue Sep 29 12:09:25 CEST 2009
243 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
244 (reported by Torsten Foertsch).
245 - common::sense 2.0 could cause tcp_server to throw an exception
246 (analysed by elmex).
247
2485.2 Mon Sep 14 07:04:49 CEST 2009
249 - INCOMPATIBLE CHANGE: do no longer support register_read_type
250 and register_write_type in AnyEvent::Handle, instead support
251 package names (the facility was mostly abused).
252 - implement "packagename-as-read/write type" support in
253 AnyEvent::Handle.
254 - AnyEvent::Handle: new options "keepalive" and "oobinline".
255 - oobinline set by default to avoid security issues.
256 - the pure-perl event loop backend wrongly detected times() fallback
257 support (spotted by Pavel Boldin).
258
2595.12 Tue Sep 1 20:26:50 CEST 2009
260 - be more lenient when parsing resolv.conf files, as some people
261 use hashmarks as comment indicator inside directives (reported
262 by Michael S. Fischer).
263 - use same set of warnings as common::sense 2.0.
264 - fix a potential 32 bit overflow issue due to perl having problems
265 with large hex constants in the Perl backend.
266 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
267 by default.
268 - fix some prototypes.
269 - (optionally) require Async::Interrupt 1.02.
270 - replace opendns fallbacks by verizon/level3 public dns servers.
271
2725.112 Fri Aug 21 13:59:04 CEST 2009
273 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
274 using starttls with an illegal/uncreatable tls context.
275 - port to broken win32 perls.
276
2775.111 Sun Aug 16 18:44:36 CEST 2009
278 - perl errornously requires sockaddr_un structures returned by the
279 kernel to have a specific length. work around this bug by padding
280 sockaddr_un structures (found to fail on OS X, as perl only uses
281 the correct code for GNU/Linux, but likely to fail on other
282 OSes as well).
283
2845.11 Wed Aug 12 17:49:37 CEST 2009
285 - fixed a bug in run_cmd with input from filename.
286 - create a json coder object when none is given in write direction
287 to reduce dependency on JSON.
288
2895.1 Tue Aug 11 03:17:32 CEST 2009
290 - tcp_connect now reports the correct error in $!, not an unrelated
291 one (was broken since 4.91).
292 - AE::Handle did not properly clear rtimeout/wtimeout watchers
293 when retrying a tcp_connect.
294 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
295 - fixed wrong documentation in AE::Socket.
296
2975.01 Mon Aug 10 03:16:32 CEST 2009
298 - last release broke EV child watchers :/.
299
3005.0 Sun Aug 9 17:32:09 CEST 2009
301 - officially introduce and document the AE API.
302 - lowering the timeout at runtime did not have immediate effect
303 in AnyEvent::Handle.
304 - AE::Handle now has separate and independent read and write timeout
305 settings.
306 - probe for Guard module when AnyEvent::Util is loaded, not at
307 runtime (Event doesn't like eval "use" from callbacks, and it avoids
308 exporting the probe function).
309 - Impl::IOAsync's io watchers did not use the filehandle emulation
310 code to work around it's watcher limits.
311 - work around IO::Async's problems with overloaded objects.
312 - try to disable wearnings inside IO::Async, it's just too buggy.
313 - do no longer use Event by default, only EV or Perl, Event is too buggy.
314 - hackishly provide fast AE:: interface when EV or Perl is the backend.
315 - inherit VERSION from AnyEvent::VERSION in most modules.
316
3174.91 Thu Aug 6 15:42:45 CEST 2009
7 - AE::Handle::starttls could get out of sync when the read buffer 318 - AE::Handle::starttls could get out of sync when the read buffer
8 already contains some TLS handshake. 319 already contains some TLS handshake.
9 - AE::Handle did not properly free the TLS session, causing bigger 320 - AE::Handle did not properly free the TLS session, causing bigger
10 memory leaks in Net::SSLeay than expected (reported by toaster). 321 memory leaks in Net::SSLeay than expected (reported by toaster).
322 - AE::Socket::tcp_connect will now artificially delay invoking
323 the callback to avoid returning after invoking the callback.
324 - convert many internal watcher uses to AE API.
11 325
124.9 Sat Aug 1 11:07:01 CEST 2009 3264.9 Sat Aug 1 11:07:01 CEST 2009
13 - Glib silently fails when registering a timeout with 327 - Glib silently fails when registering a timeout with
14 a negative value, so avoid that. 328 a negative value, so avoid that.
15 - call condvar callback immediately when it is set 329 - call condvar callback immediately when it is set
32 - new function: AnyEvent::Socket::format_hostport. 346 - new function: AnyEvent::Socket::format_hostport.
33 347
344.881 Tue Jul 28 12:51:53 CEST 2009 3484.881 Tue Jul 28 12:51:53 CEST 2009
35 - work around a bug in local in pre-5.10 perls, causing 349 - work around a bug in local in pre-5.10 perls, causing
36 AnyEvent::Handle to recurse when it should not 350 AnyEvent::Handle to recurse when it should not
37 (analyzed by elmex). 351 (analysed by elmex).
38 352
394.88 Tue Jul 28 04:04:37 CEST 2009 3534.88 Tue Jul 28 04:04:37 CEST 2009
40 - re-bless the handle into a dummy package after calling 354 - re-bless the handle into a dummy package after calling
41 AnyEvent::Handle::destroy, so the user does not need to check 355 AnyEvent::Handle::destroy, so the user does not need to check
42 for errors after every push_write etc. 356 for errors after every push_write etc.
2644.3 Fri Oct 3 09:18:43 CEST 2008 5784.3 Fri Oct 3 09:18:43 CEST 2008
265 - AnyEvent will now install a no-op signal handler for SIGPIPE, 579 - AnyEvent will now install a no-op signal handler for SIGPIPE,
266 unless one has been installed already. 580 unless one has been installed already.
267 - warn about not loaidng AnyEvent::Impl::POE early enough 581 - warn about not loaidng AnyEvent::Impl::POE early enough
268 (patch by Adam Rosenstein). 582 (patch by Adam Rosenstein).
269 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 583 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
270 - fix a bug in where in SSL connect mode, the 584 - fix a bug in where in SSL connect mode, the
271 client would first wait for some data by the server and 585 client would first wait for some data by the server and
272 otherwise hang. (reported and analyzed in an absolutely 586 otherwise hang. (reported and analysed in an absolutely
273 exemplary manner by Adam Rosenstein). 587 exemplary manner by Adam Rosenstein).
274 - fix a bug in where SSL EOF would not be treated 588 - fix a bug in where SSL EOF would not be treated
275 as stream EOF, putting the connection into a hung state. 589 as stream EOF, putting the connection into a hung state.
276 (reported and analyzed in an absolutely exemplary manner by Adam 590 (reported and analysed in an absolutely exemplary manner by Adam
277 Rosenstein). 591 Rosenstein).
278 - fix a potential segfault when the TLS context 592 - fix a potential segfault when the TLS context
279 would go missing in a read callback (e.g. due to stoptls) - 593 would go missing in a read callback (e.g. due to stoptls) -
280 Net::SSLeay of course makes no type checking whatsoever. 594 Net::SSLeay of course makes no type checking whatsoever.
281 - AnyEvent::Handle will not stop the read watcher in TLS mode. 595 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines