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

Comparing AnyEvent/Changes (file contents):
Revision 1.218 by root, Tue Feb 10 14:07:43 2009 UTC vs.
Revision 1.324 by root, Sun Aug 9 00:20:35 2009 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
34.92
4 - lowering the timeout at runtime did not have immediate effect
5 in AnyEvent::Handle.
6 - AE::Handle now has separate and independent read and write timeout
7 settings.
8 - hackishly provide fast AE:: interface when EV is the backend.
9
104.91 Thu Aug 6 15:42:45 CEST 2009
11 - AE::Handle::starttls could get out of sync when the read buffer
12 already contains some TLS handshake.
13 - AE::Handle did not properly free the TLS session, causing bigger
14 memory leaks in Net::SSLeay than expected (reported by toaster).
15 - AE::Socket::tcp_connect will now artificially delay invoking
16 the callback to avoid returning after invoking the callback.
17 - convert many internal watcher uses to AE API.
18
194.9 Sat Aug 1 11:07:01 CEST 2009
20 - Glib silently fails when registering a timeout with
21 a negative value, so avoid that.
22 - call condvar callback immediately when it is set
23 after the condition is already signalled.
24 - check rbuf_max condition only after trying to consume
25 data.
26 - normalise signal numbers to names when using the ae's
27 signal handling, but do not document this (yet).
28 - pure perl signal emulation did not properly set nonblocking
29 mode on the signal pipe on win32 (this is unlikely to have caused
30 any issues).
31 - new module: AnyEvent::Debug.
32 - AnyEvent::Strict now checks that a signal was specified
33 by name (not name or number).
34 - reduce memory footprint in typical cases by ~50kb by
35 only compiling signal code when necessary.
36 - add AnyEvent::Handle->rbuf_max.
37 - grab the AE namespace for future new API, implement stubs for
38 said future/alternative API.
39 - new function: AnyEvent::Socket::format_hostport.
40
414.881 Tue Jul 28 12:51:53 CEST 2009
42 - work around a bug in local in pre-5.10 perls, causing
43 AnyEvent::Handle to recurse when it should not
44 (analyzed by elmex).
45
464.88 Tue Jul 28 04:04:37 CEST 2009
47 - re-bless the handle into a dummy package after calling
48 AnyEvent::Handle::destroy, so the user does not need to check
49 for errors after every push_write etc.
50 - do not attempt to run t/02_signals.t on obviously broken
51 platforms.
52
534.87 Sun Jul 26 02:06:16 CEST 2009
54 - do not attempt to linger when there is no longer a valid fh.
55 - enforce tls mode to be either accept or connect, do not simply
56 segfault in Net::SSLeay.
57 - AnyEvent::Handle can now call tcp_connect itself (new parameters
58 connect, on_prepare, on_connect and on_connect_error). Updated
59 tutorial accordingly.
60 - add AnyEvent::Impl::Irssi backend.
61
624.86 Mon Jul 20 23:52:29 CEST 2009
63 - since the verbose warning is not enough, explicitly document
64 that versions before 1.33 of Net::SSLeay are not secure.
65 - work around signal handling races in Event and (...) Event::Lib.
66 - try to align signal-race timer to full-second boundaries.
67 - work around Tk not liking negative timeouts.
68 - don't complain of different grades of environmental unfriendlyness
69 in IO::Async.
70
714.85 Sat Jul 18 06:16:14 CEST 2009
72 - nail the signal race problem in perl once and for all
73 (see $AnyEvent::MAX_SIGNAL_LATENCY).
74 - take advantage of Async::Interrupt if it is available.
75 - load Time::HiRes and Guard modules on demand only.
76 - add optional/recommended modules section to AnyEvent
77 documentation.
78 - reduce memory usage considerably (and reduce startup penalty)
79 by not using "strict", "warnings" and "overload" modules.
80 - work around buggy windows/openbsd perls and provide EBADMSG
81 and EPROTO ourselves when missing.
82 - improve perl 5.6 compatibility of the core event loop.
83 - made Net::SSLeay version 1.33 a soft requirement.
84
854.83 Fri Jul 17 16:56:26 CEST 2009
86 - implement AnyEvent::Socket::getprotobyname.
87 - AnyEvent::CondVar's will now detect recursive blocking
88 waits and will croak, as too many people fall into
89 this trap.
90 - AnyEvent::Handle will now call ->destroy on itself after
91 executing the on_error callback, instead of doing some
92 half-baked internal shutdown, for fatal errors.
93 - clarify on_eof behaviour w.r.t. the read queue and
94 on_read callbacks.
95 - ignore some possible spurious wake-ups in tcp_connect.
96
974.82 Sat Jul 11 00:34:55 CEST 2009
98 - POE and Event backends didn't accept some callable objects as
99 callbacks.
100 - use Config module instead of POSIX module to detect signal names
101 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
102 module doesn't even have all POSIX signals :/.
103 - use more workarounds around the many refcnt/corruption bugs in
104 Event::Lib.
105 - work around a race condition in perl's select, causing t/03_child.t
106 to rarely fail.
107
1084.81 Thu Jul 9 10:30:30 CEST 2009
109 - AnyEvent::Handle didn't properly diagnose write errors
110 (it expected -1 from syswrite, how lame... :).
111 - support file descriptors in addition to file handles
112 in AnyEvent->io.
113 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
114 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
115 and PERL_ANYEVENT_CA_PATH.
116 - provide a sensible synopsis section for AnyEvent::TLS.
117 - add a "supported backends" section to the manpage.
118 - added simple io watcher test to testsuite, using a
119 portable_socketpair.
120 - tried to improve the stability of the Event::Lib backend,
121 YMMV.
122
1234.8 Mon Jul 6 23:45:16 CEST 2009
124 - AnyEvent::DNS did not properly follow CNAME records with
125 uppercase targets.
126 - AnyEvent::DNS would errornously return AAAA records
127 with v4 mapped addresses (a faulty record) as ipv4 addresses,
128 causing AnyEvent::Socket to throw an exception.
129 - added new module AnyEvent::TLS for easier SSL/TLS context
130 creation, with many options including hostname verification,
131 secure default configuration, lots of documentation and,
132 predefined diffie-hellman keys for perfect forward security
133 and much more. get it while it's still fresh!
134 - use AnyEvent::TLS in AnyEvent::Handle for context management.
135 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
136 so AnyEvent::Socket users have smaller memory footprint
137 in the common case.
138 - add AnyEvent::Handle->push_shutdown method.
139 - add an additional error message parameter to AnyEvent::Handle's
140 on_error callback (for TLS, $! is still available).
141 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
142 - make AnyEvent::Handle more robust against early conenction
143 failures (during new), and return C<undef> in that case
144 from the constructor.
145 - AnyEvent::Handle will now try to load only JSON::XS first,
146 then fall back to JSON.
147 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
148 and because it was documented that way.
149
1504.451 Fri Jul 3 00:28:58 CEST 2009
151 - do not clear rbuf when shutting down an AnyEvent::Handle
152 object - doing so breaks AnyEvent::HTTP.
153
1544.45 Mon Jun 29 22:59:26 CEST 2009
155 - a write error could cause AnyEvent::Handle to create
156 an I/O watcher with an undefined $fh.
157 - special-case mapped ipv4 addresses in both
158 AnyEvent::Socket::format_address and parse_address,
159 to treat them just like ipv4 addresses.
160 - updated and overhauled the AnyEvent::Intro doc.
161 - implement AnyEvent::Socket::format_ipv4/ipv6.
162 - slightly speed up portable_pipe/socketpair.
163 - expand condvar begin/end documentation.
164
1654.42 Fri Jun 26 08:32:18 CEST 2009
166 - preliminary and neccesarily incomplete support for IO::Async.
167 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
168 case it was set to IGNORE, to ensure we can catch child statuses
169 even when the calling env acts stupidly.
170 - updated benchmarks with IO::Async, which performs very well.
171
1724.412 Wed Jun 24 01:35:57 CEST 2009
173 - support an "untaint" attribute for AnyEvent::DNS
174 and set it on the default resolver.
175 - implement and document AnyEvent::Impl::Perl::loop.
176 - remove all anyevent-env variables from %ENV
177 when running in tainted mode.
178 - mention and extend the IO::Lambda benchmark.
179
1804.411 Sun Jun 7 18:48:02 CEST 2009
181 - do not try to use F_SETFD on windows.
182
1834.41 Thu May 14 06:40:11 CEST 2009
184 - work around issues in older perls (5.8.0?) when
185 a signal handler is deleted from the %SIG hash.
186 - use POSIX::_exit in child test, to avoid running
187 destructors.
188 - speed up CHLD handling by relying on SIGCHLD being
189 synchronously delivered, even when we roll our own
190 implementation.
191 - AnyEvent::DNS: add the "dname" resource record name
192 for cosmetic reasons.
193
1944.4 Sun Apr 26 20:12:33 CEST 2009
195 - implemented idle watchers, where applicable.
196 - AnyEvent->time died when Event backend was in use.
197 - fix a memleak in the Tk backend.
198 - sped up Tk timer handling.
199 - clip negative "after" values to 0 in AnyEvent::Impl::Event
200 to avoid spamming out warning messages.
201 - fix Qt timers without interval.
202 - avoid Qt zero-timeout special case, in old libqt's.
203
2044.352 Mon Apr 20 16:31:11 CEST 2009
205 - fix AnyEvent::Strict error messages for child watchers.
206 - fix/update Makefile.PL "recommends" versions.
207 - add AnyEvent->now_update.
208
2094.351 Sat Apr 11 07:56:14 CEST 2009
210 - actually make the signal pipe work under win32.
211 - localise $! in signal handler to avoid changing $!, although
212 perl itself does not seem to save/restore errno either.
213 - set the cloexec flag on the signal pipe (normally set by perl too).
214
2154.35 Fri Mar 27 11:48:20 CET 2009
216 - event models relying on AnyEvent's signal watcher emulation did
217 invoke the callback asynchronously, contradicting documentation
218 and causing signals to get lost (this includes AnyEvent's own
219 event loop). AnyEvent now uses the standard pipe trick to make
220 callback execution synchronous to the event loop.
221 - AnyEvent::Handle didn't free TLS context data on DESTROY
222 (patch by Pavel Shaydo).
223 - work around the ever-incompatibly-changing API of MakeMaker.
224 - document that changing global variables without restoring them
225 is a bad idea in Perl (noted by Adam Rosenstein).
226 - AnyEvent::Strict now barfs if ->io is passed a file.
227
2284.34 Thu Feb 12 18:32:45 CET 2009
3 - separately buffer TLS read data, as otherwise the read queue could 229 - separately buffer TLS read data, as otherwise the read queue could
4 deadlock as receiving data is not expected while draining the 230 deadlock as receiving data is not expected while draining the
5 read queue (which cna only happen with TLS). 231 read queue (which cna only happen with TLS).
6 - raise EBADMSG error on JSON decoding errors. 232 - raise EBADMSG error on JSON decoding errors.
7 - fix some minor manpage bugs (reported by Maximilian Gaß). 233 - fix some minor manpage bugs (reported by Maximilian Gaß).

Diff Legend

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