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

Comparing AnyEvent/Changes (file contents):
Revision 1.222 by root, Thu Mar 26 20:17:44 2009 UTC vs.
Revision 1.329 by root, Sun Aug 9 16:05:11 2009 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3TODO: leakthing pasth 35.0 Sun Aug 9 17:32:09 CEST 2009
4TODO: strict checks for file, and valid fd 4 - officially introduce and document the AE API.
54.341 5 - lowering the timeout at runtime did not have immediate effect
6 in AnyEvent::Handle.
7 - AE::Handle now has separate and independent read and write timeout
8 settings.
9 - probe for Guard module when AnyEvent::Util is loaded, not at
10 runtime (Event doesn't like eval "use" from callbacks, and it avoids
11 exporting the probe function).
12 - Impl::IOAsync's io watchers did not use the filehandle emulation
13 code to work around it's watcher limits.
14 - work around IO::Async's problems with overloaded objects.
15 - try to disable wearnings inside IO::Async, it's just too buggy.
16 - do no longer use Event by default, only EV or Perl, Event is too buggy.
17 - hackishly provide fast AE:: interface when EV or Perl is the backend.
18 - inherit VERSION from AnyEvent::VERSION in most modules.
19
204.91 Thu Aug 6 15:42:45 CEST 2009
21 - AE::Handle::starttls could get out of sync when the read buffer
22 already contains some TLS handshake.
23 - AE::Handle did not properly free the TLS session, causing bigger
24 memory leaks in Net::SSLeay than expected (reported by toaster).
25 - AE::Socket::tcp_connect will now artificially delay invoking
26 the callback to avoid returning after invoking the callback.
27 - convert many internal watcher uses to AE API.
28
294.9 Sat Aug 1 11:07:01 CEST 2009
30 - Glib silently fails when registering a timeout with
31 a negative value, so avoid that.
32 - call condvar callback immediately when it is set
33 after the condition is already signalled.
34 - check rbuf_max condition only after trying to consume
35 data.
36 - normalise signal numbers to names when using the ae's
37 signal handling, but do not document this (yet).
38 - pure perl signal emulation did not properly set nonblocking
39 mode on the signal pipe on win32 (this is unlikely to have caused
40 any issues).
41 - new module: AnyEvent::Debug.
42 - AnyEvent::Strict now checks that a signal was specified
43 by name (not name or number).
44 - reduce memory footprint in typical cases by ~50kb by
45 only compiling signal code when necessary.
46 - add AnyEvent::Handle->rbuf_max.
47 - grab the AE namespace for future new API, implement stubs for
48 said future/alternative API.
49 - new function: AnyEvent::Socket::format_hostport.
50
514.881 Tue Jul 28 12:51:53 CEST 2009
52 - work around a bug in local in pre-5.10 perls, causing
53 AnyEvent::Handle to recurse when it should not
54 (analyzed by elmex).
55
564.88 Tue Jul 28 04:04:37 CEST 2009
57 - re-bless the handle into a dummy package after calling
58 AnyEvent::Handle::destroy, so the user does not need to check
59 for errors after every push_write etc.
60 - do not attempt to run t/02_signals.t on obviously broken
61 platforms.
62
634.87 Sun Jul 26 02:06:16 CEST 2009
64 - do not attempt to linger when there is no longer a valid fh.
65 - enforce tls mode to be either accept or connect, do not simply
66 segfault in Net::SSLeay.
67 - AnyEvent::Handle can now call tcp_connect itself (new parameters
68 connect, on_prepare, on_connect and on_connect_error). Updated
69 tutorial accordingly.
70 - add AnyEvent::Impl::Irssi backend.
71
724.86 Mon Jul 20 23:52:29 CEST 2009
73 - since the verbose warning is not enough, explicitly document
74 that versions before 1.33 of Net::SSLeay are not secure.
75 - work around signal handling races in Event and (...) Event::Lib.
76 - try to align signal-race timer to full-second boundaries.
77 - work around Tk not liking negative timeouts.
78 - don't complain of different grades of environmental unfriendlyness
79 in IO::Async.
80
814.85 Sat Jul 18 06:16:14 CEST 2009
82 - nail the signal race problem in perl once and for all
83 (see $AnyEvent::MAX_SIGNAL_LATENCY).
84 - take advantage of Async::Interrupt if it is available.
85 - load Time::HiRes and Guard modules on demand only.
86 - add optional/recommended modules section to AnyEvent
87 documentation.
88 - reduce memory usage considerably (and reduce startup penalty)
89 by not using "strict", "warnings" and "overload" modules.
90 - work around buggy windows/openbsd perls and provide EBADMSG
91 and EPROTO ourselves when missing.
92 - improve perl 5.6 compatibility of the core event loop.
93 - made Net::SSLeay version 1.33 a soft requirement.
94
954.83 Fri Jul 17 16:56:26 CEST 2009
96 - implement AnyEvent::Socket::getprotobyname.
97 - AnyEvent::CondVar's will now detect recursive blocking
98 waits and will croak, as too many people fall into
99 this trap.
100 - AnyEvent::Handle will now call ->destroy on itself after
101 executing the on_error callback, instead of doing some
102 half-baked internal shutdown, for fatal errors.
103 - clarify on_eof behaviour w.r.t. the read queue and
104 on_read callbacks.
105 - ignore some possible spurious wake-ups in tcp_connect.
106
1074.82 Sat Jul 11 00:34:55 CEST 2009
108 - POE and Event backends didn't accept some callable objects as
109 callbacks.
110 - use Config module instead of POSIX module to detect signal names
111 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
112 module doesn't even have all POSIX signals :/.
113 - use more workarounds around the many refcnt/corruption bugs in
114 Event::Lib.
115 - work around a race condition in perl's select, causing t/03_child.t
116 to rarely fail.
117
1184.81 Thu Jul 9 10:30:30 CEST 2009
119 - AnyEvent::Handle didn't properly diagnose write errors
120 (it expected -1 from syswrite, how lame... :).
121 - support file descriptors in addition to file handles
122 in AnyEvent->io.
123 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
124 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
125 and PERL_ANYEVENT_CA_PATH.
126 - provide a sensible synopsis section for AnyEvent::TLS.
127 - add a "supported backends" section to the manpage.
128 - added simple io watcher test to testsuite, using a
129 portable_socketpair.
130 - tried to improve the stability of the Event::Lib backend,
131 YMMV.
132
1334.8 Mon Jul 6 23:45:16 CEST 2009
134 - AnyEvent::DNS did not properly follow CNAME records with
135 uppercase targets.
136 - AnyEvent::DNS would errornously return AAAA records
137 with v4 mapped addresses (a faulty record) as ipv4 addresses,
138 causing AnyEvent::Socket to throw an exception.
139 - added new module AnyEvent::TLS for easier SSL/TLS context
140 creation, with many options including hostname verification,
141 secure default configuration, lots of documentation and,
142 predefined diffie-hellman keys for perfect forward security
143 and much more. get it while it's still fresh!
144 - use AnyEvent::TLS in AnyEvent::Handle for context management.
145 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
146 so AnyEvent::Socket users have smaller memory footprint
147 in the common case.
148 - add AnyEvent::Handle->push_shutdown method.
149 - add an additional error message parameter to AnyEvent::Handle's
150 on_error callback (for TLS, $! is still available).
151 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
152 - make AnyEvent::Handle more robust against early conenction
153 failures (during new), and return C<undef> in that case
154 from the constructor.
155 - AnyEvent::Handle will now try to load only JSON::XS first,
156 then fall back to JSON.
157 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
158 and because it was documented that way.
159
1604.451 Fri Jul 3 00:28:58 CEST 2009
161 - do not clear rbuf when shutting down an AnyEvent::Handle
162 object - doing so breaks AnyEvent::HTTP.
163
1644.45 Mon Jun 29 22:59:26 CEST 2009
165 - a write error could cause AnyEvent::Handle to create
166 an I/O watcher with an undefined $fh.
167 - special-case mapped ipv4 addresses in both
168 AnyEvent::Socket::format_address and parse_address,
169 to treat them just like ipv4 addresses.
170 - updated and overhauled the AnyEvent::Intro doc.
171 - implement AnyEvent::Socket::format_ipv4/ipv6.
172 - slightly speed up portable_pipe/socketpair.
173 - expand condvar begin/end documentation.
174
1754.42 Fri Jun 26 08:32:18 CEST 2009
176 - preliminary and neccesarily incomplete support for IO::Async.
177 - reset SIGCHLD to DEFAULT when AnyEvent is loaded, in
178 case it was set to IGNORE, to ensure we can catch child statuses
179 even when the calling env acts stupidly.
180 - updated benchmarks with IO::Async, which performs very well.
181
1824.412 Wed Jun 24 01:35:57 CEST 2009
183 - support an "untaint" attribute for AnyEvent::DNS
184 and set it on the default resolver.
185 - implement and document AnyEvent::Impl::Perl::loop.
186 - remove all anyevent-env variables from %ENV
187 when running in tainted mode.
188 - mention and extend the IO::Lambda benchmark.
189
1904.411 Sun Jun 7 18:48:02 CEST 2009
191 - do not try to use F_SETFD on windows.
192
1934.41 Thu May 14 06:40:11 CEST 2009
194 - work around issues in older perls (5.8.0?) when
195 a signal handler is deleted from the %SIG hash.
196 - use POSIX::_exit in child test, to avoid running
197 destructors.
198 - speed up CHLD handling by relying on SIGCHLD being
199 synchronously delivered, even when we roll our own
200 implementation.
201 - AnyEvent::DNS: add the "dname" resource record name
202 for cosmetic reasons.
203
2044.4 Sun Apr 26 20:12:33 CEST 2009
205 - implemented idle watchers, where applicable.
206 - AnyEvent->time died when Event backend was in use.
207 - fix a memleak in the Tk backend.
208 - sped up Tk timer handling.
209 - clip negative "after" values to 0 in AnyEvent::Impl::Event
210 to avoid spamming out warning messages.
211 - fix Qt timers without interval.
212 - avoid Qt zero-timeout special case, in old libqt's.
213
2144.352 Mon Apr 20 16:31:11 CEST 2009
215 - fix AnyEvent::Strict error messages for child watchers.
216 - fix/update Makefile.PL "recommends" versions.
217 - add AnyEvent->now_update.
218
2194.351 Sat Apr 11 07:56:14 CEST 2009
220 - actually make the signal pipe work under win32.
221 - localise $! in signal handler to avoid changing $!, although
222 perl itself does not seem to save/restore errno either.
223 - set the cloexec flag on the signal pipe (normally set by perl too).
224
2254.35 Fri Mar 27 11:48:20 CET 2009
6 - event models relying on AnyEvent's signal watcher emulation did 226 - event models relying on AnyEvent's signal watcher emulation did
7 invoke the callback asynchronously, contradicting documentation 227 invoke the callback asynchronously, contradicting documentation
8 and causing signals to get lost (this includes AnyEvent's own 228 and causing signals to get lost (this includes AnyEvent's own
9 event loop). AnyEvent now uses the standard pipe trick to make 229 event loop). AnyEvent now uses the standard pipe trick to make
10 callback execution synchronous to the event loop. 230 callback execution synchronous to the event loop.
231 - AnyEvent::Handle didn't free TLS context data on DESTROY
232 (patch by Pavel Shaydo).
11 - work around the ever-incompatibly-changing API of MakeMaker. 233 - work around the ever-incompatibly-changing API of MakeMaker.
12 - document that changing global variables without restoring them 234 - document that changing global variables without restoring them
13 is a bad idea in Perl (noted by Adam Rosenstein). 235 is a bad idea in Perl (noted by Adam Rosenstein).
236 - AnyEvent::Strict now barfs if ->io is passed a file.
14 237
154.34 Thu Feb 12 18:32:45 CET 2009 2384.34 Thu Feb 12 18:32:45 CET 2009
16 - separately buffer TLS read data, as otherwise the read queue could 239 - separately buffer TLS read data, as otherwise the read queue could
17 deadlock as receiving data is not expected while draining the 240 deadlock as receiving data is not expected while draining the
18 read queue (which cna only happen with TLS). 241 read queue (which cna only happen with TLS).

Diff Legend

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