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

Comparing AnyEvent/Changes (file contents):
Revision 1.262 by root, Sun Jul 5 01:38:43 2009 UTC vs.
Revision 1.391 by root, Wed Mar 24 23:28:06 2010 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3ZERO_RETURN - stoptls? 3 - don't generate spurious readyness notifications when select
4 returns EINTR in the pure perl backend (analysed and testcase
5 by Bas Denissen).
6 - give same examples in SYNOPSIS sections of AnyEvent and AE.
7 - provider faster implementations for AE::time/now/now_update for
8 EV and Perl backends.
9 - provide faster AE::cv and AE::time implementations for all
10 backends.
11 - no longer support lower/mixed-case signal names (this was
12 never documented nor universally supported).
13 - some more memory ยต-optimisations, and somewhat less messy
14 code to cope with both AnyEvent and AE APIs.
15
165.251 Sat Mar 13 00:58:21 CET 2010
17 - make test failed if it was run before make install on perl <=
18 5.8.8, and older versions of perl make our live very difficult,
19 so write constants.pl during Makefile.PL time.
20
215.25 Sat Mar 13 00:23:14 CET 2010
22 - fix a race condition in AnyEvent::Handle that would cause
23 a "bio_read: unsupported method" error (in _another_
24 ssl connection) after user code threw an exception.
25 - added AnyEvent::Handle->destroyed method.
26 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
27 - the AnyEvent::Util::fh_nonblocking function worked only
28 by lucky accident on win32.
29 - smaller and faster AnyEvent::Util::fh_nonblocking.
30 - when the (required!) Time::HiRes module is missing, AnyEvent
31 did not fall back to built-in time properly.
32 - do not load Fcntl at runtime, saving memory and loading time.
33 - precompile a number of constants and use them instead of runtime
34 detection and eval.
35 - free detection code after detection and similar memory optimisations.
36 - Perl backend timer interval best effort drift has been improved
37 (same algorithm as EV).
38 - update unicode idna mapping table.
39
405.24 Tue Jan 5 11:39:43 CET 2010
41 - cygwin never reports errors from failed connects. 1.5 just gives
42 you continous readyness and EAGAIN, 1.7 is even more broken and
43 just hangs. work around both issues in a cygwin-specific hack.
44 - improve idle watcher documentation slightly.
45
465.23 Sun Dec 20 23:48:00 CET 2009
47 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
48 - implement punycode_encode/decode, idn_nameprep,
49 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
50 - provide $AE::VERSION.
51 - removed traces of "no strict 'refs'".
52
535.22 Sat Dec 5 03:51:13 CET 2009
54 - downgrade-or-fail in AnyEvent::Handle::push_write, to
55 diagnose encoding failures earlier and more succinctly.
56 (this works around bugs in perl, throwing away encoding info
57 when passing scalar data to extensions).
58 - add more examples to AnyEvent::Socket manpage.
59 - upgrade internal warning set to the same as common::sense 2.03.
60 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
61 (John Beppu).
62 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
63 - add google public dns servers to fallback server set and make sure
64 we load-balance properly between the three sets. also add all
65 fallback dns servers, not just a random one, to each dns config.
66
675.21 Thu Nov 19 02:48:47 CET 2009
68 - fix a problem where socket constants were called with parameters
69 (spotted by David Friedland).
70 - fork_call never use'd POSIX (reported by Daisuke Maki).
71 - improve perl 5.6 compatibility further (but it still won't work
72 unless you rip out everything but the core).
73 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
74 uses 10MB of RAM, but doesn't open a console window. *sigh*.
75
765.202 Wed Oct 14 22:35:44 CEST 2009
77 - AnyEvent::DNS would unexpectedly clobber $_ under windows
78 (analysed by Matthias Waldorf).
79 - AnyEvent::Handle::run_cmd can now pass the PID of the
80 newly-created process, which is much less useful than it might
81 sound (based on patch by Yann Kerherve).
82
835.201 Tue Sep 29 12:09:25 CEST 2009
84 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
85 (reported by Torsten Foertsch).
86 - common::sense 2.0 could cause tcp_server to throw an exception
87 (analysed by elmex).
88
895.2 Mon Sep 14 07:04:49 CEST 2009
90 - INCOMPATIBLE CHANGE: do no longer support register_read_type
91 and register_write_type in AnyEvent::Handle, instead support
92 package names (the facility was mostly abused).
93 - implement "packagename-as-read/write type" support in
94 AnyEvent::Handle.
95 - AnyEvent::Handle: new options "keepalive" and "oobinline".
96 - oobinline set by default to avoid security issues.
97 - the pure-perl event loop backend wrongly detected times() fallback
98 support (spotted by Pavel Boldin).
99
1005.12 Tue Sep 1 20:26:50 CEST 2009
101 - be more lenient when parsing resolv.conf files, as some people
102 use hashmarks as comment indicator inside directives (reported
103 by Michael S. Fischer).
104 - use same set of warnings as common::sense 2.0.
105 - fix a potential 32 bit overflow issue due to perl having problems
106 with large hex constants in the Perl backend.
107 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
108 by default.
109 - fix some prototypes.
110 - (optionally) require Async::Interrupt 1.02.
111 - replace opendns fallbacks by verizon/level3 public dns servers.
112
1135.112 Fri Aug 21 13:59:04 CEST 2009
114 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
115 using starttls with an illegal/uncreatable tls context.
116 - port to broken win32 perls.
117
1185.111 Sun Aug 16 18:44:36 CEST 2009
119 - perl errornously requires sockaddr_un structures returned by the
120 kernel to have a specific length. work around this bug by padding
121 sockaddr_un structures (found to fail on OS X, as perl only uses
122 the correct code for GNU/Linux, but likely to fail on other
123 OSes as well).
124
1255.11 Wed Aug 12 17:49:37 CEST 2009
126 - fixed a bug in run_cmd with input from filename.
127 - create a json coder object when none is given in write direction
128 to reduce dependency on JSON.
129
1305.1 Tue Aug 11 03:17:32 CEST 2009
131 - tcp_connect now reports the correct error in $!, not an unrelated
132 one (was broken since 4.91).
133 - AE::Handle did not properly clear rtimeout/wtimeout watchers
134 when retrying a tcp_connect.
135 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
136 - fixed wrong documentation in AE::Socket.
137
1385.01 Mon Aug 10 03:16:32 CEST 2009
139 - last release broke EV child watchers :/.
140
1415.0 Sun Aug 9 17:32:09 CEST 2009
142 - officially introduce and document the AE API.
143 - lowering the timeout at runtime did not have immediate effect
144 in AnyEvent::Handle.
145 - AE::Handle now has separate and independent read and write timeout
146 settings.
147 - probe for Guard module when AnyEvent::Util is loaded, not at
148 runtime (Event doesn't like eval "use" from callbacks, and it avoids
149 exporting the probe function).
150 - Impl::IOAsync's io watchers did not use the filehandle emulation
151 code to work around it's watcher limits.
152 - work around IO::Async's problems with overloaded objects.
153 - try to disable wearnings inside IO::Async, it's just too buggy.
154 - do no longer use Event by default, only EV or Perl, Event is too buggy.
155 - hackishly provide fast AE:: interface when EV or Perl is the backend.
156 - inherit VERSION from AnyEvent::VERSION in most modules.
157
1584.91 Thu Aug 6 15:42:45 CEST 2009
159 - AE::Handle::starttls could get out of sync when the read buffer
160 already contains some TLS handshake.
161 - AE::Handle did not properly free the TLS session, causing bigger
162 memory leaks in Net::SSLeay than expected (reported by toaster).
163 - AE::Socket::tcp_connect will now artificially delay invoking
164 the callback to avoid returning after invoking the callback.
165 - convert many internal watcher uses to AE API.
166
1674.9 Sat Aug 1 11:07:01 CEST 2009
168 - Glib silently fails when registering a timeout with
169 a negative value, so avoid that.
170 - call condvar callback immediately when it is set
171 after the condition is already signalled.
172 - check rbuf_max condition only after trying to consume
173 data.
174 - normalise signal numbers to names when using the ae's
175 signal handling, but do not document this (yet).
176 - pure perl signal emulation did not properly set nonblocking
177 mode on the signal pipe on win32 (this is unlikely to have caused
178 any issues).
179 - new module: AnyEvent::Debug.
180 - AnyEvent::Strict now checks that a signal was specified
181 by name (not name or number).
182 - reduce memory footprint in typical cases by ~50kb by
183 only compiling signal code when necessary.
184 - add AnyEvent::Handle->rbuf_max.
185 - grab the AE namespace for future new API, implement stubs for
186 said future/alternative API.
187 - new function: AnyEvent::Socket::format_hostport.
188
1894.881 Tue Jul 28 12:51:53 CEST 2009
190 - work around a bug in local in pre-5.10 perls, causing
191 AnyEvent::Handle to recurse when it should not
192 (analysed by elmex).
193
1944.88 Tue Jul 28 04:04:37 CEST 2009
195 - re-bless the handle into a dummy package after calling
196 AnyEvent::Handle::destroy, so the user does not need to check
197 for errors after every push_write etc.
198 - do not attempt to run t/02_signals.t on obviously broken
199 platforms.
200
2014.87 Sun Jul 26 02:06:16 CEST 2009
202 - do not attempt to linger when there is no longer a valid fh.
203 - enforce tls mode to be either accept or connect, do not simply
204 segfault in Net::SSLeay.
205 - AnyEvent::Handle can now call tcp_connect itself (new parameters
206 connect, on_prepare, on_connect and on_connect_error). Updated
207 tutorial accordingly.
208 - add AnyEvent::Impl::Irssi backend.
209
2104.86 Mon Jul 20 23:52:29 CEST 2009
211 - since the verbose warning is not enough, explicitly document
212 that versions before 1.33 of Net::SSLeay are not secure.
213 - work around signal handling races in Event and (...) Event::Lib.
214 - try to align signal-race timer to full-second boundaries.
215 - work around Tk not liking negative timeouts.
216 - don't complain of different grades of environmental unfriendlyness
217 in IO::Async.
218
2194.85 Sat Jul 18 06:16:14 CEST 2009
220 - nail the signal race problem in perl once and for all
221 (see $AnyEvent::MAX_SIGNAL_LATENCY).
222 - take advantage of Async::Interrupt if it is available.
223 - load Time::HiRes and Guard modules on demand only.
224 - add optional/recommended modules section to AnyEvent
225 documentation.
226 - reduce memory usage considerably (and reduce startup penalty)
227 by not using "strict", "warnings" and "overload" modules.
228 - work around buggy windows/openbsd perls and provide EBADMSG
229 and EPROTO ourselves when missing.
230 - improve perl 5.6 compatibility of the core event loop.
231 - made Net::SSLeay version 1.33 a soft requirement.
232
2334.83 Fri Jul 17 16:56:26 CEST 2009
234 - implement AnyEvent::Socket::getprotobyname.
235 - AnyEvent::CondVar's will now detect recursive blocking
236 waits and will croak, as too many people fall into
237 this trap.
238 - AnyEvent::Handle will now call ->destroy on itself after
239 executing the on_error callback, instead of doing some
240 half-baked internal shutdown, for fatal errors.
241 - clarify on_eof behaviour w.r.t. the read queue and
242 on_read callbacks.
243 - ignore some possible spurious wake-ups in tcp_connect.
244
2454.82 Sat Jul 11 00:34:55 CEST 2009
246 - POE and Event backends didn't accept some callable objects as
247 callbacks.
248 - use Config module instead of POSIX module to detect signal names
249 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
250 module doesn't even have all POSIX signals :/.
251 - use more workarounds around the many refcnt/corruption bugs in
252 Event::Lib.
253 - work around a race condition in perl's select, causing t/03_child.t
254 to rarely fail.
255
2564.81 Thu Jul 9 10:30:30 CEST 2009
257 - AnyEvent::Handle didn't properly diagnose write errors
258 (it expected -1 from syswrite, how lame... :).
259 - support file descriptors in addition to file handles
260 in AnyEvent->io.
261 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
262 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
263 and PERL_ANYEVENT_CA_PATH.
264 - provide a sensible synopsis section for AnyEvent::TLS.
265 - add a "supported backends" section to the manpage.
266 - added simple io watcher test to testsuite, using a
267 portable_socketpair.
268 - tried to improve the stability of the Event::Lib backend,
269 YMMV.
270
2714.8 Mon Jul 6 23:45:16 CEST 2009
4 - AnyEvent::DNS did not properly follow CNAME records with 272 - AnyEvent::DNS did not properly follow CNAME records with
5 uppercase targets. 273 uppercase targets.
6 - AnyEvent::DNS would errornously return AAAA records 274 - AnyEvent::DNS would errornously return AAAA records
7 with v4 mapped addresses (a faulty record) as ipv4 addresses, 275 with v4 mapped addresses (a faulty record) as ipv4 addresses,
8 causing AnyEvent::Socket to throw an exception. 276 causing AnyEvent::Socket to throw an exception.
… …
16 so AnyEvent::Socket users have smaller memory footprint 284 so AnyEvent::Socket users have smaller memory footprint
17 in the common case. 285 in the common case.
18 - add AnyEvent::Handle->push_shutdown method. 286 - add AnyEvent::Handle->push_shutdown method.
19 - add an additional error message parameter to AnyEvent::Handle's 287 - add an additional error message parameter to AnyEvent::Handle's
20 on_error callback (for TLS, $! is still available). 288 on_error callback (for TLS, $! is still available).
289 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
21 - make AnyEvent::Handle more robust against early conenction 290 - make AnyEvent::Handle more robust against early conenction
22 failures (during new), and return C<undef> in that case 291 failures (during new), and return C<undef> in that case
23 from the constructor. 292 from the constructor.
24 - AnyEvent::Handle will now try to load only JSON::XS first, 293 - AnyEvent::Handle will now try to load only JSON::XS first,
25 then fall back to JSON. 294 then fall back to JSON.
… …
1504.3 Fri Oct 3 09:18:43 CEST 2008 4194.3 Fri Oct 3 09:18:43 CEST 2008
151 - AnyEvent will now install a no-op signal handler for SIGPIPE, 420 - AnyEvent will now install a no-op signal handler for SIGPIPE,
152 unless one has been installed already. 421 unless one has been installed already.
153 - warn about not loaidng AnyEvent::Impl::POE early enough 422 - warn about not loaidng AnyEvent::Impl::POE early enough
154 (patch by Adam Rosenstein). 423 (patch by Adam Rosenstein).
155 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 424 - fixed a great number of bugs and corner cases in AnyEvent::Handle.
156 - fix a bug in where in SSL connect mode, the 425 - fix a bug in where in SSL connect mode, the
157 client would first wait for some data by the server and 426 client would first wait for some data by the server and
158 otherwise hang. (reported and analyzed in an absolutely 427 otherwise hang. (reported and analysed in an absolutely
159 exemplary manner by Adam Rosenstein). 428 exemplary manner by Adam Rosenstein).
160 - fix a bug in where SSL EOF would not be treated 429 - fix a bug in where SSL EOF would not be treated
161 as stream EOF, putting the connection into a hung state. 430 as stream EOF, putting the connection into a hung state.
162 (reported and analyzed in an absolutely exemplary manner by Adam 431 (reported and analysed in an absolutely exemplary manner by Adam
163 Rosenstein). 432 Rosenstein).
164 - fix a potential segfault when the TLS context 433 - fix a potential segfault when the TLS context
165 would go missing in a read callback (e.g. due to stoptls) - 434 would go missing in a read callback (e.g. due to stoptls) -
166 Net::SSLeay of course makes no type checking whatsoever. 435 Net::SSLeay of course makes no type checking whatsoever.
167 - AnyEvent::Handle will not stop the read watcher in TLS mode. 436 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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