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

Comparing AnyEvent/Changes (file contents):
Revision 1.221 by root, Thu Mar 26 07:47:42 2009 UTC vs.
Revision 1.353 by root, Thu Oct 1 17:23:14 2009 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
34.341 3 - AnyEvent::Handle::run_cmd can now pass the PID of the
4 newly-created process, which is much less useful than it might
5 sound (based on patch by Yann Kerherve).
6
75.201 Tue Sep 29 12:09:25 CEST 2009
8 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
9 (reported by Torsten Foertsch).
10 - common::sense 2.0 could cause tcp_server to throw an exception
11 (analysed by elmex).
12
135.2 Mon Sep 14 07:04:49 CEST 2009
14 - INCOMPATIBLE CHANGE: do no longer support register_read_type
15 and register_write_type in AnyEvent::Handle, instead support
16 package names (the facility was mostly abused).
17 - implement "packagename-as-read/write type" support in
18 AnyEvent::Handle.
19 - AnyEvent::Handle: new options "keepalive" and "oobinline".
20 - oobinline set by default to avoid security issues.
21 - the pure-perl event loop backend wrongly detected times() fallback
22 support (spotted by Pavel Boldin).
23
245.12 Tue Sep 1 20:26:50 CEST 2009
25 - be more lenient when parsing resolv.conf files, as some people
26 use hashmarks as comment indicator inside directives (reported
27 by Michael S. Fischer).
28 - use same set of warnings as common::sense 2.0.
29 - fix a potential 32 bit overflow issue due to perl having problems
30 with large hex constants in the Perl backend.
31 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
32 by default.
33 - fix some prototypes.
34 - (optionally) require Async::Interrupt 1.02.
35 - replace opendns fallbacks by verizon/level3 public dns servers.
36
375.112 Fri Aug 21 13:59:04 CEST 2009
38 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
39 using starttls with an illegal/uncreatable tls context.
40 - port to broken win32 perls.
41
425.111 Sun Aug 16 18:44:36 CEST 2009
43 - perl errornously requires sockaddr_un structures returned by the
44 kernel to have a specific length. work around this bug by padding
45 sockaddr_un structures (found to fail on OS X, as perl only uses
46 the correct code for GNU/Linux, but likely to fail on other
47 OSes as well).
48
495.11 Wed Aug 12 17:49:37 CEST 2009
50 - fixed a bug in run_cmd with input from filename.
51 - create a json coder object when none is given in write direction
52 to reduce dependency on JSON.
53
545.1 Tue Aug 11 03:17:32 CEST 2009
55 - tcp_connect now reports the correct error in $!, not an unrelated
56 one (was broken since 4.91).
57 - AE::Handle did not properly clear rtimeout/wtimeout watchers
58 when retrying a tcp_connect.
59 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
60 - fixed wrong documentation in AE::Socket.
61
625.01 Mon Aug 10 03:16:32 CEST 2009
63 - last release broke EV child watchers :/.
64
655.0 Sun Aug 9 17:32:09 CEST 2009
66 - officially introduce and document the AE API.
67 - lowering the timeout at runtime did not have immediate effect
68 in AnyEvent::Handle.
69 - AE::Handle now has separate and independent read and write timeout
70 settings.
71 - probe for Guard module when AnyEvent::Util is loaded, not at
72 runtime (Event doesn't like eval "use" from callbacks, and it avoids
73 exporting the probe function).
74 - Impl::IOAsync's io watchers did not use the filehandle emulation
75 code to work around it's watcher limits.
76 - work around IO::Async's problems with overloaded objects.
77 - try to disable wearnings inside IO::Async, it's just too buggy.
78 - do no longer use Event by default, only EV or Perl, Event is too buggy.
79 - hackishly provide fast AE:: interface when EV or Perl is the backend.
80 - inherit VERSION from AnyEvent::VERSION in most modules.
81
824.91 Thu Aug 6 15:42:45 CEST 2009
83 - AE::Handle::starttls could get out of sync when the read buffer
84 already contains some TLS handshake.
85 - AE::Handle did not properly free the TLS session, causing bigger
86 memory leaks in Net::SSLeay than expected (reported by toaster).
87 - AE::Socket::tcp_connect will now artificially delay invoking
88 the callback to avoid returning after invoking the callback.
89 - convert many internal watcher uses to AE API.
90
914.9 Sat Aug 1 11:07:01 CEST 2009
92 - Glib silently fails when registering a timeout with
93 a negative value, so avoid that.
94 - call condvar callback immediately when it is set
95 after the condition is already signalled.
96 - check rbuf_max condition only after trying to consume
97 data.
98 - normalise signal numbers to names when using the ae's
99 signal handling, but do not document this (yet).
100 - pure perl signal emulation did not properly set nonblocking
101 mode on the signal pipe on win32 (this is unlikely to have caused
102 any issues).
103 - new module: AnyEvent::Debug.
104 - AnyEvent::Strict now checks that a signal was specified
105 by name (not name or number).
106 - reduce memory footprint in typical cases by ~50kb by
107 only compiling signal code when necessary.
108 - add AnyEvent::Handle->rbuf_max.
109 - grab the AE namespace for future new API, implement stubs for
110 said future/alternative API.
111 - new function: AnyEvent::Socket::format_hostport.
112
1134.881 Tue Jul 28 12:51:53 CEST 2009
114 - work around a bug in local in pre-5.10 perls, causing
115 AnyEvent::Handle to recurse when it should not
116 (analysed by elmex).
117
1184.88 Tue Jul 28 04:04:37 CEST 2009
119 - re-bless the handle into a dummy package after calling
120 AnyEvent::Handle::destroy, so the user does not need to check
121 for errors after every push_write etc.
122 - do not attempt to run t/02_signals.t on obviously broken
123 platforms.
124
1254.87 Sun Jul 26 02:06:16 CEST 2009
126 - do not attempt to linger when there is no longer a valid fh.
127 - enforce tls mode to be either accept or connect, do not simply
128 segfault in Net::SSLeay.
129 - AnyEvent::Handle can now call tcp_connect itself (new parameters
130 connect, on_prepare, on_connect and on_connect_error). Updated
131 tutorial accordingly.
132 - add AnyEvent::Impl::Irssi backend.
133
1344.86 Mon Jul 20 23:52:29 CEST 2009
135 - since the verbose warning is not enough, explicitly document
136 that versions before 1.33 of Net::SSLeay are not secure.
137 - work around signal handling races in Event and (...) Event::Lib.
138 - try to align signal-race timer to full-second boundaries.
139 - work around Tk not liking negative timeouts.
140 - don't complain of different grades of environmental unfriendlyness
141 in IO::Async.
142
1434.85 Sat Jul 18 06:16:14 CEST 2009
144 - nail the signal race problem in perl once and for all
145 (see $AnyEvent::MAX_SIGNAL_LATENCY).
146 - take advantage of Async::Interrupt if it is available.
147 - load Time::HiRes and Guard modules on demand only.
148 - add optional/recommended modules section to AnyEvent
149 documentation.
150 - reduce memory usage considerably (and reduce startup penalty)
151 by not using "strict", "warnings" and "overload" modules.
152 - work around buggy windows/openbsd perls and provide EBADMSG
153 and EPROTO ourselves when missing.
154 - improve perl 5.6 compatibility of the core event loop.
155 - made Net::SSLeay version 1.33 a soft requirement.
156
1574.83 Fri Jul 17 16:56:26 CEST 2009
158 - implement AnyEvent::Socket::getprotobyname.
159 - AnyEvent::CondVar's will now detect recursive blocking
160 waits and will croak, as too many people fall into
161 this trap.
162 - AnyEvent::Handle will now call ->destroy on itself after
163 executing the on_error callback, instead of doing some
164 half-baked internal shutdown, for fatal errors.
165 - clarify on_eof behaviour w.r.t. the read queue and
166 on_read callbacks.
167 - ignore some possible spurious wake-ups in tcp_connect.
168
1694.82 Sat Jul 11 00:34:55 CEST 2009
170 - POE and Event backends didn't accept some callable objects as
171 callbacks.
172 - use Config module instead of POSIX module to detect signal names
173 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
174 module doesn't even have all POSIX signals :/.
175 - use more workarounds around the many refcnt/corruption bugs in
176 Event::Lib.
177 - work around a race condition in perl's select, causing t/03_child.t
178 to rarely fail.
179
1804.81 Thu Jul 9 10:30:30 CEST 2009
181 - AnyEvent::Handle didn't properly diagnose write errors
182 (it expected -1 from syswrite, how lame... :).
183 - support file descriptors in addition to file handles
184 in AnyEvent->io.
185 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
186 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
187 and PERL_ANYEVENT_CA_PATH.
188 - provide a sensible synopsis section for AnyEvent::TLS.
189 - add a "supported backends" section to the manpage.
190 - added simple io watcher test to testsuite, using a
191 portable_socketpair.
192 - tried to improve the stability of the Event::Lib backend,
193 YMMV.
194
1954.8 Mon Jul 6 23:45:16 CEST 2009
196 - AnyEvent::DNS did not properly follow CNAME records with
197 uppercase targets.
198 - AnyEvent::DNS would errornously return AAAA records
199 with v4 mapped addresses (a faulty record) as ipv4 addresses,
200 causing AnyEvent::Socket to throw an exception.
201 - added new module AnyEvent::TLS for easier SSL/TLS context
202 creation, with many options including hostname verification,
203 secure default configuration, lots of documentation and,
204 predefined diffie-hellman keys for perfect forward security
205 and much more. get it while it's still fresh!
206 - use AnyEvent::TLS in AnyEvent::Handle for context management.
207 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
208 so AnyEvent::Socket users have smaller memory footprint
209 in the common case.
210 - add AnyEvent::Handle->push_shutdown method.
211 - add an additional error message parameter to AnyEvent::Handle's
212 on_error callback (for TLS, $! is still available).
213 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
214 - make AnyEvent::Handle more robust against early conenction
215 failures (during new), and return C<undef> in that case
216 from the constructor.
217 - AnyEvent::Handle will now try to load only JSON::XS first,
218 then fall back to JSON.
219 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
220 and because it was documented that way.
221
2224.451 Fri Jul 3 00:28:58 CEST 2009
223 - do not clear rbuf when shutting down an AnyEvent::Handle
224 object - doing so breaks AnyEvent::HTTP.
225
2264.45 Mon Jun 29 22:59:26 CEST 2009
227 - a write error could cause AnyEvent::Handle to create
228 an I/O watcher with an undefined $fh.
229 - special-case mapped ipv4 addresses in both
230 AnyEvent::Socket::format_address and parse_address,
231 to treat them just like ipv4 addresses.
232 - updated and overhauled the AnyEvent::Intro doc.
233 - implement AnyEvent::Socket::format_ipv4/ipv6.
234 - slightly speed up portable_pipe/socketpair.
235 - expand condvar begin/end documentation.
236
2374.42 Fri Jun 26 08:32:18 CEST 2009
238 - preliminary and neccesarily incomplete support for IO::Async.
239 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
240 case it was set to IGNORE, to ensure we can catch child statuses
241 even when the calling env acts stupidly.
242 - updated benchmarks with IO::Async, which performs very well.
243
2444.412 Wed Jun 24 01:35:57 CEST 2009
245 - support an "untaint" attribute for AnyEvent::DNS
246 and set it on the default resolver.
247 - implement and document AnyEvent::Impl::Perl::loop.
248 - remove all anyevent-env variables from %ENV
249 when running in tainted mode.
250 - mention and extend the IO::Lambda benchmark.
251
2524.411 Sun Jun 7 18:48:02 CEST 2009
253 - do not try to use F_SETFD on windows.
254
2554.41 Thu May 14 06:40:11 CEST 2009
256 - work around issues in older perls (5.8.0?) when
257 a signal handler is deleted from the %SIG hash.
258 - use POSIX::_exit in child test, to avoid running
259 destructors.
260 - speed up CHLD handling by relying on SIGCHLD being
261 synchronously delivered, even when we roll our own
262 implementation.
263 - AnyEvent::DNS: add the "dname" resource record name
264 for cosmetic reasons.
265
2664.4 Sun Apr 26 20:12:33 CEST 2009
267 - implemented idle watchers, where applicable.
268 - AnyEvent->time died when Event backend was in use.
269 - fix a memleak in the Tk backend.
270 - sped up Tk timer handling.
271 - clip negative "after" values to 0 in AnyEvent::Impl::Event
272 to avoid spamming out warning messages.
273 - fix Qt timers without interval.
274 - avoid Qt zero-timeout special case, in old libqt's.
275
2764.352 Mon Apr 20 16:31:11 CEST 2009
277 - fix AnyEvent::Strict error messages for child watchers.
278 - fix/update Makefile.PL "recommends" versions.
279 - add AnyEvent->now_update.
280
2814.351 Sat Apr 11 07:56:14 CEST 2009
282 - actually make the signal pipe work under win32.
283 - localise $! in signal handler to avoid changing $!, although
284 perl itself does not seem to save/restore errno either.
285 - set the cloexec flag on the signal pipe (normally set by perl too).
286
2874.35 Fri Mar 27 11:48:20 CET 2009
4 - event models relying on AnyEvent's signal watcher emulation did 288 - event models relying on AnyEvent's signal watcher emulation did
5 invoke the callback asynchronously, contradicting documentation 289 invoke the callback asynchronously, contradicting documentation
6 and causing signals to get lost (this includes AnyEvent's own 290 and causing signals to get lost (this includes AnyEvent's own
7 event loop). AnyEvent now uses the standard pipe trick to make 291 event loop). AnyEvent now uses the standard pipe trick to make
8 callback execution synchronous to the event loop. 292 callback execution synchronous to the event loop.
293 - AnyEvent::Handle didn't free TLS context data on DESTROY
294 (patch by Pavel Shaydo).
9 - work around the ever-incompatibly-changing API of MakeMaker. 295 - work around the ever-incompatibly-changing API of MakeMaker.
10 - document that changing global variables without restoring them 296 - document that changing global variables without restoring them
11 is a bad idea in Perl (noted by Adam Rosenstein). 297 is a bad idea in Perl (noted by Adam Rosenstein).
298 - AnyEvent::Strict now barfs if ->io is passed a file.
12 299
134.34 Thu Feb 12 18:32:45 CET 2009 3004.34 Thu Feb 12 18:32:45 CET 2009
14 - separately buffer TLS read data, as otherwise the read queue could 301 - separately buffer TLS read data, as otherwise the read queue could
15 deadlock as receiving data is not expected while draining the 302 deadlock as receiving data is not expected while draining the
16 read queue (which cna only happen with TLS). 303 read queue (which cna only happen with TLS).
… …
59 - warn about not loaidng AnyEvent::Impl::POE early enough 346 - warn about not loaidng AnyEvent::Impl::POE early enough
60 (patch by Adam Rosenstein). 347 (patch by Adam Rosenstein).
61 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 348 - fixed a great number of bugs and corner cases in AnyEvent::Handle:
62 - fix a bug in where in SSL connect mode, the 349 - fix a bug in where in SSL connect mode, the
63 client would first wait for some data by the server and 350 client would first wait for some data by the server and
64 otherwise hang. (reported and analyzed in an absolutely 351 otherwise hang. (reported and analysed in an absolutely
65 exemplary manner by Adam Rosenstein). 352 exemplary manner by Adam Rosenstein).
66 - fix a bug in where SSL EOF would not be treated 353 - fix a bug in where SSL EOF would not be treated
67 as stream EOF, putting the connection into a hung state. 354 as stream EOF, putting the connection into a hung state.
68 (reported and analyzed in an absolutely exemplary manner by Adam 355 (reported and analysed in an absolutely exemplary manner by Adam
69 Rosenstein). 356 Rosenstein).
70 - fix a potential segfault when the TLS context 357 - fix a potential segfault when the TLS context
71 would go missing in a read callback (e.g. due to stoptls) - 358 would go missing in a read callback (e.g. due to stoptls) -
72 Net::SSLeay of course makes no type checking whatsoever. 359 Net::SSLeay of course makes no type checking whatsoever.
73 - AnyEvent::Handle will not stop the read watcher in TLS mode. 360 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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