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

Comparing AnyEvent/Changes (file contents):
Revision 1.269 by root, Wed Jul 8 04:49:18 2009 UTC vs.
Revision 1.438 by root, Sun Jan 23 10:44:48 2011 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
34.801 3TODO: document TCP_*** constants
4TODO: catch AE::xxx for anyevent strict and anyevent debug
5TODO: anyevent::trace? or debug tracing all watchers?
6TODO: anyevent strict checking for $_ changes?
7
8TODO: resetttls?
9 - work around a POE bug where POE would first poll for new events and
10 invoke handlers BEFORE handling the message queue.
11 - fix parsing of "domain" statement in resolv.conf (Johannes Plunien).
12 - stop_read now does just that in TLS mode. see its
13 documentation for caveats.
14 - update to common::sense 3.4 warning set.
15 - AE::Handle had a broken rbuf_max method and missing wbuf_max method.
16
175.3 Fri Dec 31 05:48:13 CET 2010
18 - major IOAsync patch by Paul Evans - 0.33 is required, as the major
19 issues with IO::Async have been solved in that version, including
20 the availabilty of a default loop!
21 - implement new wbuf_max parameter for AnyEvent::Handle.
22 - added Cocoa::EventLoop backend by Daisuke Murase.
23 - the default testsuite now honors PERL_ANYEVENT_MODEL (Paul Evans).
24 - clarify handle on_prepare documentation.
25 - set CLOEXEC flag on parent-side fd's in AnyEvent::Util::run_cmd,
26 to avoid leaking them to any child processes.
27 - try to force files into the arch-dependend lib directory, for the
28 half-"turly obsessive compulsive person"(s) that depend on it.
29
305.29 Sun Dec 5 10:49:21 CET 2010
31 - convert EV backend to EV 4.00 API (so better upgrade EV too).
32 - AE::Handle: implement dynamic read_size adjustment, add
33 max_read_size parameter, reduce default read_size to 2048.
34 - add some FAQ entries for servers and/or on_eof vs. on_error.
35 - work around OS bugs (cygwin again) when lingering in AE::Handle.
36 - reduce memory usage slightly when connect was used in AE::Handle.
37 - wrok around more segfaults in Tk.
38 - document tls_ctx => undef as valid.
39 - detect WNOHANG value at compiletime.
40 - include some probably linux-only support for building AnyEvent
41 as part of the perl core.
42 - improve compatibility of t/08_idna.t to perl 5.8.
43 - make AnyEvent::Strict truly optional (in case somebody builds a
44 minimal perl...).
45
465.28 Wed Oct 13 04:14:23 CEST 2010
47 - due to a glitch, AnyEvent's internal getprotobyname wasn't
48 actually used in tcp_server.
49 - implement AnyEvent::CondVar->new method, which was documented but
50 missing.
51 - support autoloading of read types in unshift_read.
52 - AnyEvent::Strict: do the test for invalid fh arguments differently,
53 resulting in more informative error messages.
54 - new AnyEvent::FAQ manpage.
55 - clarify that you can actually call push_read etc. in on_read.
56 - put keys and certs in /tmp, not into the current directory,
57 and document this in the SECURITY CONSIDERATIONS sections
58 of AnyEvent::TLS.
59 - support ";" as resolv.conf comment character in AnyEvent::DNS, as
60 requested by Ron Isaacson.
61 - document $AnyEvent::DNS::RESOLVER variable.
62 - incorporated major typo patches by Abhijit Menon-Sen.
63 - AnyEvent::Handle now croaks when negative timeouts are passed.
64 - add a shitload of TCP_xxx sockopt name constants to AnyEvent::Util,
65 undocumented, but who knows when they might come in handy.
66
675.271 Tue Jun 8 12:05:46 CEST 2010
68 - backport to perl 5.8.x.
69
705.27 Sun Jun 6 12:12:05 CEST 2010
71 - postpone differently in AnyEvent::Socket now, as
72 when not, canceling the connection attempt might fail
73 (found by Felix Antonius Wilhelm Ostmann).
74 - explicitly check for non-stream sockets in AE::Handle, too many
75 clueless people fell into the trap of this somehow working.
76 - simplified and reworked the "OTHER MODULES" section.
77 - better/more condvar examples.
78
795.261 Wed Apr 28 16:13:36 CEST 2010
80 - AF_INET6 was not properly used from Socket6 during configuration
81 time (found by Andy Grundman).
82
835.26 Mon Apr 12 04:49:35 CEST 2010
84 - don't generate spurious readyness notifications when select
85 returns EINTR in the pure perl backend (analysed and testcase
86 by Bas Denissen).
87 - give same examples in SYNOPSIS sections of AnyEvent and AE.
88 - provider faster implementations for AE::time/now/now_update for
89 EV and Perl backends.
90 - provide faster AE::cv and AE::time implementations for all
91 backends.
92 - no longer support lower/mixed-case signal names (this was
93 never documented nor universally supported).
94 - some more memory ยต-optimisations, and somewhat less messy
95 code to cope with both AnyEvent and AE APIs.
96
975.251 Sat Mar 13 00:58:21 CET 2010
98 - make test failed if it was run before make install on perl <=
99 5.8.8, and older versions of perl make our live very difficult,
100 so write constants.pl during Makefile.PL time.
101
1025.25 Sat Mar 13 00:23:14 CET 2010
103 - fix a race condition in AnyEvent::Handle that would cause
104 a "bio_read: unsupported method" error (in _another_
105 ssl connection) after user code threw an exception.
106 - added AnyEvent::Handle->destroyed method.
107 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
108 - the AnyEvent::Util::fh_nonblocking function worked only
109 by lucky accident on win32.
110 - smaller and faster AnyEvent::Util::fh_nonblocking.
111 - when the (required!) Time::HiRes module is missing, AnyEvent
112 did not fall back to built-in time properly.
113 - do not load Fcntl at runtime, saving memory and loading time.
114 - precompile a number of constants and use them instead of runtime
115 detection and eval.
116 - free detection code after detection and similar memory optimisations.
117 - Perl backend timer interval best effort drift has been improved
118 (same algorithm as EV).
119 - update unicode idna mapping table.
120
1215.24 Tue Jan 5 11:39:43 CET 2010
122 - cygwin never reports errors from failed connects. 1.5 just gives
123 you continous readyness and EAGAIN, 1.7 is even more broken and
124 just hangs. work around both issues in a cygwin-specific hack.
125 - improve idle watcher documentation slightly.
126
1275.23 Sun Dec 20 23:48:00 CET 2009
128 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
129 - implement punycode_encode/decode, idn_nameprep,
130 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
131 - provide $AE::VERSION.
132 - removed traces of "no strict 'refs'".
133
1345.22 Sat Dec 5 03:51:13 CET 2009
135 - downgrade-or-fail in AnyEvent::Handle::push_write, to
136 diagnose encoding failures earlier and more succinctly.
137 (this works around bugs in perl, throwing away encoding info
138 when passing scalar data to extensions).
139 - add more examples to AnyEvent::Socket manpage.
140 - upgrade internal warning set to the same as common::sense 2.03.
141 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
142 (John Beppu).
143 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
144 - add google public dns servers to fallback server set and make sure
145 we load-balance properly between the three sets. also add all
146 fallback dns servers, not just a random one, to each dns config.
147
1485.21 Thu Nov 19 02:48:47 CET 2009
149 - fix a problem where socket constants were called with parameters
150 (spotted by David Friedland).
151 - fork_call never use'd POSIX (reported by Daisuke Maki).
152 - improve perl 5.6 compatibility further (but it still won't work
153 unless you rip out everything but the core).
154 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
155 uses 10MB of RAM, but doesn't open a console window. *sigh*.
156
1575.202 Wed Oct 14 22:35:44 CEST 2009
158 - AnyEvent::DNS would unexpectedly clobber $_ under windows
159 (analysed by Matthias Waldorf).
160 - AnyEvent::Handle::run_cmd can now pass the PID of the
161 newly-created process, which is much less useful than it might
162 sound (based on patch by Yann Kerherve).
163
1645.201 Tue Sep 29 12:09:25 CEST 2009
165 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
166 (reported by Torsten Foertsch).
167 - common::sense 2.0 could cause tcp_server to throw an exception
168 (analysed by elmex).
169
1705.2 Mon Sep 14 07:04:49 CEST 2009
171 - INCOMPATIBLE CHANGE: do no longer support register_read_type
172 and register_write_type in AnyEvent::Handle, instead support
173 package names (the facility was mostly abused).
174 - implement "packagename-as-read/write type" support in
175 AnyEvent::Handle.
176 - AnyEvent::Handle: new options "keepalive" and "oobinline".
177 - oobinline set by default to avoid security issues.
178 - the pure-perl event loop backend wrongly detected times() fallback
179 support (spotted by Pavel Boldin).
180
1815.12 Tue Sep 1 20:26:50 CEST 2009
182 - be more lenient when parsing resolv.conf files, as some people
183 use hashmarks as comment indicator inside directives (reported
184 by Michael S. Fischer).
185 - use same set of warnings as common::sense 2.0.
186 - fix a potential 32 bit overflow issue due to perl having problems
187 with large hex constants in the Perl backend.
188 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
189 by default.
190 - fix some prototypes.
191 - (optionally) require Async::Interrupt 1.02.
192 - replace opendns fallbacks by verizon/level3 public dns servers.
193
1945.112 Fri Aug 21 13:59:04 CEST 2009
195 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
196 using starttls with an illegal/uncreatable tls context.
197 - port to broken win32 perls.
198
1995.111 Sun Aug 16 18:44:36 CEST 2009
200 - perl errornously requires sockaddr_un structures returned by the
201 kernel to have a specific length. work around this bug by padding
202 sockaddr_un structures (found to fail on OS X, as perl only uses
203 the correct code for GNU/Linux, but likely to fail on other
204 OSes as well).
205
2065.11 Wed Aug 12 17:49:37 CEST 2009
207 - fixed a bug in run_cmd with input from filename.
208 - create a json coder object when none is given in write direction
209 to reduce dependency on JSON.
210
2115.1 Tue Aug 11 03:17:32 CEST 2009
212 - tcp_connect now reports the correct error in $!, not an unrelated
213 one (was broken since 4.91).
214 - AE::Handle did not properly clear rtimeout/wtimeout watchers
215 when retrying a tcp_connect.
216 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
217 - fixed wrong documentation in AE::Socket.
218
2195.01 Mon Aug 10 03:16:32 CEST 2009
220 - last release broke EV child watchers :/.
221
2225.0 Sun Aug 9 17:32:09 CEST 2009
223 - officially introduce and document the AE API.
224 - lowering the timeout at runtime did not have immediate effect
225 in AnyEvent::Handle.
226 - AE::Handle now has separate and independent read and write timeout
227 settings.
228 - probe for Guard module when AnyEvent::Util is loaded, not at
229 runtime (Event doesn't like eval "use" from callbacks, and it avoids
230 exporting the probe function).
231 - Impl::IOAsync's io watchers did not use the filehandle emulation
232 code to work around it's watcher limits.
233 - work around IO::Async's problems with overloaded objects.
234 - try to disable wearnings inside IO::Async, it's just too buggy.
235 - do no longer use Event by default, only EV or Perl, Event is too buggy.
236 - hackishly provide fast AE:: interface when EV or Perl is the backend.
237 - inherit VERSION from AnyEvent::VERSION in most modules.
238
2394.91 Thu Aug 6 15:42:45 CEST 2009
240 - AE::Handle::starttls could get out of sync when the read buffer
241 already contains some TLS handshake.
242 - AE::Handle did not properly free the TLS session, causing bigger
243 memory leaks in Net::SSLeay than expected (reported by toaster).
244 - AE::Socket::tcp_connect will now artificially delay invoking
245 the callback to avoid returning after invoking the callback.
246 - convert many internal watcher uses to AE API.
247
2484.9 Sat Aug 1 11:07:01 CEST 2009
249 - Glib silently fails when registering a timeout with
250 a negative value, so avoid that.
251 - call condvar callback immediately when it is set
252 after the condition is already signalled.
253 - check rbuf_max condition only after trying to consume
254 data.
255 - normalise signal numbers to names when using the ae's
256 signal handling, but do not document this (yet).
257 - pure perl signal emulation did not properly set nonblocking
258 mode on the signal pipe on win32 (this is unlikely to have caused
259 any issues).
260 - new module: AnyEvent::Debug.
261 - AnyEvent::Strict now checks that a signal was specified
262 by name (not name or number).
263 - reduce memory footprint in typical cases by ~50kb by
264 only compiling signal code when necessary.
265 - add AnyEvent::Handle->rbuf_max.
266 - grab the AE namespace for future new API, implement stubs for
267 said future/alternative API.
268 - new function: AnyEvent::Socket::format_hostport.
269
2704.881 Tue Jul 28 12:51:53 CEST 2009
271 - work around a bug in local in pre-5.10 perls, causing
272 AnyEvent::Handle to recurse when it should not
273 (analysed by elmex).
274
2754.88 Tue Jul 28 04:04:37 CEST 2009
276 - re-bless the handle into a dummy package after calling
277 AnyEvent::Handle::destroy, so the user does not need to check
278 for errors after every push_write etc.
279 - do not attempt to run t/02_signals.t on obviously broken
280 platforms.
281
2824.87 Sun Jul 26 02:06:16 CEST 2009
283 - do not attempt to linger when there is no longer a valid fh.
284 - enforce tls mode to be either accept or connect, do not simply
285 segfault in Net::SSLeay.
286 - AnyEvent::Handle can now call tcp_connect itself (new parameters
287 connect, on_prepare, on_connect and on_connect_error). Updated
288 tutorial accordingly.
289 - add AnyEvent::Impl::Irssi backend.
290
2914.86 Mon Jul 20 23:52:29 CEST 2009
292 - since the verbose warning is not enough, explicitly document
293 that versions before 1.33 of Net::SSLeay are not secure.
294 - work around signal handling races in Event and (...) Event::Lib.
295 - try to align signal-race timer to full-second boundaries.
296 - work around Tk not liking negative timeouts.
297 - don't complain of different grades of environmental unfriendlyness
298 in IO::Async.
299
3004.85 Sat Jul 18 06:16:14 CEST 2009
301 - nail the signal race problem in perl once and for all
302 (see $AnyEvent::MAX_SIGNAL_LATENCY).
303 - take advantage of Async::Interrupt if it is available.
304 - load Time::HiRes and Guard modules on demand only.
305 - add optional/recommended modules section to AnyEvent
306 documentation.
307 - reduce memory usage considerably (and reduce startup penalty)
308 by not using "strict", "warnings" and "overload" modules.
309 - work around buggy windows/openbsd perls and provide EBADMSG
310 and EPROTO ourselves when missing.
311 - improve perl 5.6 compatibility of the core event loop.
312 - made Net::SSLeay version 1.33 a soft requirement.
313
3144.83 Fri Jul 17 16:56:26 CEST 2009
315 - implement AnyEvent::Socket::getprotobyname.
316 - AnyEvent::CondVar's will now detect recursive blocking
317 waits and will croak, as too many people fall into
318 this trap.
319 - AnyEvent::Handle will now call ->destroy on itself after
320 executing the on_error callback, instead of doing some
321 half-baked internal shutdown, for fatal errors.
322 - clarify on_eof behaviour w.r.t. the read queue and
323 on_read callbacks.
324 - ignore some possible spurious wake-ups in tcp_connect.
325
3264.82 Sat Jul 11 00:34:55 CEST 2009
327 - POE and Event backends didn't accept some callable objects as
328 callbacks.
329 - use Config module instead of POSIX module to detect signal names
330 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
331 module doesn't even have all POSIX signals :/.
332 - use more workarounds around the many refcnt/corruption bugs in
333 Event::Lib.
334 - work around a race condition in perl's select, causing t/03_child.t
335 to rarely fail.
336
3374.81 Thu Jul 9 10:30:30 CEST 2009
338 - AnyEvent::Handle didn't properly diagnose write errors
339 (it expected -1 from syswrite, how lame... :).
4 - support file descriptors in addition to file handles 340 - support file descriptors in addition to file handles
5 in AnyEvent->io. 341 in AnyEvent->io.
6 - new env variables: PERL_ANYEVENT_RESOLV_CONF, 342 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
7 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE 343 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
8 and PERL_ANYEVENT_CA_PATH. 344 and PERL_ANYEVENT_CA_PATH.
9 - provide a sensible synopsis section for AnyEvent::TLS. 345 - provide a sensible synopsis section for AnyEvent::TLS.
346 - add a "supported backends" section to the manpage.
347 - added simple io watcher test to testsuite, using a
348 portable_socketpair.
349 - tried to improve the stability of the Event::Lib backend,
350 YMMV.
10 351
114.8 Mon Jul 6 23:45:16 CEST 2009 3524.8 Mon Jul 6 23:45:16 CEST 2009
12 - AnyEvent::DNS did not properly follow CNAME records with 353 - AnyEvent::DNS did not properly follow CNAME records with
13 uppercase targets. 354 uppercase targets.
14 - AnyEvent::DNS would errornously return AAAA records 355 - AnyEvent::DNS would errornously return AAAA records
1594.3 Fri Oct 3 09:18:43 CEST 2008 5004.3 Fri Oct 3 09:18:43 CEST 2008
160 - AnyEvent will now install a no-op signal handler for SIGPIPE, 501 - AnyEvent will now install a no-op signal handler for SIGPIPE,
161 unless one has been installed already. 502 unless one has been installed already.
162 - warn about not loaidng AnyEvent::Impl::POE early enough 503 - warn about not loaidng AnyEvent::Impl::POE early enough
163 (patch by Adam Rosenstein). 504 (patch by Adam Rosenstein).
164 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 505 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
165 - fix a bug in where in SSL connect mode, the 506 - fix a bug in where in SSL connect mode, the
166 client would first wait for some data by the server and 507 client would first wait for some data by the server and
167 otherwise hang. (reported and analyzed in an absolutely 508 otherwise hang. (reported and analysed in an absolutely
168 exemplary manner by Adam Rosenstein). 509 exemplary manner by Adam Rosenstein).
169 - fix a bug in where SSL EOF would not be treated 510 - fix a bug in where SSL EOF would not be treated
170 as stream EOF, putting the connection into a hung state. 511 as stream EOF, putting the connection into a hung state.
171 (reported and analyzed in an absolutely exemplary manner by Adam 512 (reported and analysed in an absolutely exemplary manner by Adam
172 Rosenstein). 513 Rosenstein).
173 - fix a potential segfault when the TLS context 514 - fix a potential segfault when the TLS context
174 would go missing in a read callback (e.g. due to stoptls) - 515 would go missing in a read callback (e.g. due to stoptls) -
175 Net::SSLeay of course makes no type checking whatsoever. 516 Net::SSLeay of course makes no type checking whatsoever.
176 - AnyEvent::Handle will not stop the read watcher in TLS mode. 517 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines