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

Comparing AnyEvent/Changes (file contents):
Revision 1.259 by root, Fri Jul 3 08:51:48 2009 UTC vs.
Revision 1.381 by root, Sun Feb 21 09:28:18 2010 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
3ZERO_RETURN - stoptls? 3 - speed up AnyEvent::Socket::format_address and ::format_ipv6.
4 - the AnyEvent::Util::fh_nonblocking function worked only
5 by lucky accident on win32.
6 - smaller and faster AnyEvent::Util::fh_nonblocking.
7 - when the (required!) Time::HiRes module is missing, AnyEvent
8 did not fall back to built-in time properly.
9 - do not load Fcntl at runtime, saving memory and loading time.
10 - precompile a number of constants and use them instead of runtime
11 detection and eval.
12 - free detection code after detection and similar memory optimisations.
13
145.24 Tue Jan 5 11:39:43 CET 2010
15 - cygwin never reports errors from failed connects. 1.5 just gives
16 you continous readyness and EAGAIN, 1.7 is even more broken and
17 just hangs. work around both issues in a cygwin-specific hack.
18 - improve idle watcher documentation slightly.
19
205.23 Sun Dec 20 23:48:00 CET 2009
21 - support IDNs in resolve_sockaddr, and therefore in tcp_connect.
22 - implement punycode_encode/decode, idn_nameprep,
23 idn_to_ascii and idn_to_unicode operations in AnyEvent::Util.
24 - provide $AE::VERSION.
25 - removed traces of "no strict 'refs'".
26
275.22 Sat Dec 5 03:51:13 CET 2009
28 - downgrade-or-fail in AnyEvent::Handle::push_write, to
29 diagnose encoding failures earlier and more succinctly.
30 (this works around bugs in perl, throwing away encoding info
31 when passing scalar data to extensions).
32 - add more examples to AnyEvent::Socket manpage.
33 - upgrade internal warning set to the same as common::sense 2.03.
34 - use pack "n/a*" for pre-5.8.9 perl compatibility in AnyEvent::DNS
35 (John Beppu).
36 - AnyEvent::Socket::inet_aton now properly supports ipv6, as documented.
37 - add google public dns servers to fallback server set and make sure
38 we load-balance properly between the three sets. also add all
39 fallback dns servers, not just a random one, to each dns config.
40
415.21 Thu Nov 19 02:48:47 CET 2009
42 - fix a problem where socket constants were called with parameters
43 (spotted by David Friedland).
44 - fork_call never use'd POSIX (reported by Daisuke Maki).
45 - improve perl 5.6 compatibility further (but it still won't work
46 unless you rip out everything but the core).
47 - prefer Net::DNS::Resolver over ipconfig, if installed, on win32.
48 uses 10MB of RAM, but doesn't open a console window. *sigh*.
49
505.202 Wed Oct 14 22:35:44 CEST 2009
51 - AnyEvent::DNS would unexpectedly clobber $_ under windows
52 (analysed by Matthias Waldorf).
53 - AnyEvent::Handle::run_cmd can now pass the PID of the
54 newly-created process, which is much less useful than it might
55 sound (based on patch by Yann Kerherve).
56
575.201 Tue Sep 29 12:09:25 CEST 2009
58 - AnyEvent:Handle::on_starttls/on_stoptls methods were broken
59 (reported by Torsten Foertsch).
60 - common::sense 2.0 could cause tcp_server to throw an exception
61 (analysed by elmex).
62
635.2 Mon Sep 14 07:04:49 CEST 2009
64 - INCOMPATIBLE CHANGE: do no longer support register_read_type
65 and register_write_type in AnyEvent::Handle, instead support
66 package names (the facility was mostly abused).
67 - implement "packagename-as-read/write type" support in
68 AnyEvent::Handle.
69 - AnyEvent::Handle: new options "keepalive" and "oobinline".
70 - oobinline set by default to avoid security issues.
71 - the pure-perl event loop backend wrongly detected times() fallback
72 support (spotted by Pavel Boldin).
73
745.12 Tue Sep 1 20:26:50 CEST 2009
75 - be more lenient when parsing resolv.conf files, as some people
76 use hashmarks as comment indicator inside directives (reported
77 by Michael S. Fischer).
78 - use same set of warnings as common::sense 2.0.
79 - fix a potential 32 bit overflow issue due to perl having problems
80 with large hex constants in the Perl backend.
81 - do not use the slower 5.6 version fo the enc_name code in AnyEvent::DNS
82 by default.
83 - fix some prototypes.
84 - (optionally) require Async::Interrupt 1.02.
85 - replace opendns fallbacks by verizon/level3 public dns servers.
86
875.112 Fri Aug 21 13:59:04 CEST 2009
88 - AnyEvent::Handle could cause hard crashes in Net::SSLeay when
89 using starttls with an illegal/uncreatable tls context.
90 - port to broken win32 perls.
91
925.111 Sun Aug 16 18:44:36 CEST 2009
93 - perl errornously requires sockaddr_un structures returned by the
94 kernel to have a specific length. work around this bug by padding
95 sockaddr_un structures (found to fail on OS X, as perl only uses
96 the correct code for GNU/Linux, but likely to fail on other
97 OSes as well).
98
995.11 Wed Aug 12 17:49:37 CEST 2009
100 - fixed a bug in run_cmd with input from filename.
101 - create a json coder object when none is given in write direction
102 to reduce dependency on JSON.
103
1045.1 Tue Aug 11 03:17:32 CEST 2009
105 - tcp_connect now reports the correct error in $!, not an unrelated
106 one (was broken since 4.91).
107 - AE::Handle did not properly clear rtimeout/wtimeout watchers
108 when retrying a tcp_connect.
109 - new functions: AnyEvent::Util::run_cmd and close_all_fds_except.
110 - fixed wrong documentation in AE::Socket.
111
1125.01 Mon Aug 10 03:16:32 CEST 2009
113 - last release broke EV child watchers :/.
114
1155.0 Sun Aug 9 17:32:09 CEST 2009
116 - officially introduce and document the AE API.
117 - lowering the timeout at runtime did not have immediate effect
118 in AnyEvent::Handle.
119 - AE::Handle now has separate and independent read and write timeout
120 settings.
121 - probe for Guard module when AnyEvent::Util is loaded, not at
122 runtime (Event doesn't like eval "use" from callbacks, and it avoids
123 exporting the probe function).
124 - Impl::IOAsync's io watchers did not use the filehandle emulation
125 code to work around it's watcher limits.
126 - work around IO::Async's problems with overloaded objects.
127 - try to disable wearnings inside IO::Async, it's just too buggy.
128 - do no longer use Event by default, only EV or Perl, Event is too buggy.
129 - hackishly provide fast AE:: interface when EV or Perl is the backend.
130 - inherit VERSION from AnyEvent::VERSION in most modules.
131
1324.91 Thu Aug 6 15:42:45 CEST 2009
133 - AE::Handle::starttls could get out of sync when the read buffer
134 already contains some TLS handshake.
135 - AE::Handle did not properly free the TLS session, causing bigger
136 memory leaks in Net::SSLeay than expected (reported by toaster).
137 - AE::Socket::tcp_connect will now artificially delay invoking
138 the callback to avoid returning after invoking the callback.
139 - convert many internal watcher uses to AE API.
140
1414.9 Sat Aug 1 11:07:01 CEST 2009
142 - Glib silently fails when registering a timeout with
143 a negative value, so avoid that.
144 - call condvar callback immediately when it is set
145 after the condition is already signalled.
146 - check rbuf_max condition only after trying to consume
147 data.
148 - normalise signal numbers to names when using the ae's
149 signal handling, but do not document this (yet).
150 - pure perl signal emulation did not properly set nonblocking
151 mode on the signal pipe on win32 (this is unlikely to have caused
152 any issues).
153 - new module: AnyEvent::Debug.
154 - AnyEvent::Strict now checks that a signal was specified
155 by name (not name or number).
156 - reduce memory footprint in typical cases by ~50kb by
157 only compiling signal code when necessary.
158 - add AnyEvent::Handle->rbuf_max.
159 - grab the AE namespace for future new API, implement stubs for
160 said future/alternative API.
161 - new function: AnyEvent::Socket::format_hostport.
162
1634.881 Tue Jul 28 12:51:53 CEST 2009
164 - work around a bug in local in pre-5.10 perls, causing
165 AnyEvent::Handle to recurse when it should not
166 (analysed by elmex).
167
1684.88 Tue Jul 28 04:04:37 CEST 2009
169 - re-bless the handle into a dummy package after calling
170 AnyEvent::Handle::destroy, so the user does not need to check
171 for errors after every push_write etc.
172 - do not attempt to run t/02_signals.t on obviously broken
173 platforms.
174
1754.87 Sun Jul 26 02:06:16 CEST 2009
176 - do not attempt to linger when there is no longer a valid fh.
177 - enforce tls mode to be either accept or connect, do not simply
178 segfault in Net::SSLeay.
179 - AnyEvent::Handle can now call tcp_connect itself (new parameters
180 connect, on_prepare, on_connect and on_connect_error). Updated
181 tutorial accordingly.
182 - add AnyEvent::Impl::Irssi backend.
183
1844.86 Mon Jul 20 23:52:29 CEST 2009
185 - since the verbose warning is not enough, explicitly document
186 that versions before 1.33 of Net::SSLeay are not secure.
187 - work around signal handling races in Event and (...) Event::Lib.
188 - try to align signal-race timer to full-second boundaries.
189 - work around Tk not liking negative timeouts.
190 - don't complain of different grades of environmental unfriendlyness
191 in IO::Async.
192
1934.85 Sat Jul 18 06:16:14 CEST 2009
194 - nail the signal race problem in perl once and for all
195 (see $AnyEvent::MAX_SIGNAL_LATENCY).
196 - take advantage of Async::Interrupt if it is available.
197 - load Time::HiRes and Guard modules on demand only.
198 - add optional/recommended modules section to AnyEvent
199 documentation.
200 - reduce memory usage considerably (and reduce startup penalty)
201 by not using "strict", "warnings" and "overload" modules.
202 - work around buggy windows/openbsd perls and provide EBADMSG
203 and EPROTO ourselves when missing.
204 - improve perl 5.6 compatibility of the core event loop.
205 - made Net::SSLeay version 1.33 a soft requirement.
206
2074.83 Fri Jul 17 16:56:26 CEST 2009
208 - implement AnyEvent::Socket::getprotobyname.
209 - AnyEvent::CondVar's will now detect recursive blocking
210 waits and will croak, as too many people fall into
211 this trap.
212 - AnyEvent::Handle will now call ->destroy on itself after
213 executing the on_error callback, instead of doing some
214 half-baked internal shutdown, for fatal errors.
215 - clarify on_eof behaviour w.r.t. the read queue and
216 on_read callbacks.
217 - ignore some possible spurious wake-ups in tcp_connect.
218
2194.82 Sat Jul 11 00:34:55 CEST 2009
220 - POE and Event backends didn't accept some callable objects as
221 callbacks.
222 - use Config module instead of POSIX module to detect signal names
223 in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX
224 module doesn't even have all POSIX signals :/.
225 - use more workarounds around the many refcnt/corruption bugs in
226 Event::Lib.
227 - work around a race condition in perl's select, causing t/03_child.t
228 to rarely fail.
229
2304.81 Thu Jul 9 10:30:30 CEST 2009
231 - AnyEvent::Handle didn't properly diagnose write errors
232 (it expected -1 from syswrite, how lame... :).
233 - support file descriptors in addition to file handles
234 in AnyEvent->io.
235 - new env variables: PERL_ANYEVENT_RESOLV_CONF,
236 PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE
237 and PERL_ANYEVENT_CA_PATH.
238 - provide a sensible synopsis section for AnyEvent::TLS.
239 - add a "supported backends" section to the manpage.
240 - added simple io watcher test to testsuite, using a
241 portable_socketpair.
242 - tried to improve the stability of the Event::Lib backend,
243 YMMV.
244
2454.8 Mon Jul 6 23:45:16 CEST 2009
246 - AnyEvent::DNS did not properly follow CNAME records with
247 uppercase targets.
248 - AnyEvent::DNS would errornously return AAAA records
249 with v4 mapped addresses (a faulty record) as ipv4 addresses,
250 causing AnyEvent::Socket to throw an exception.
251 - added new module AnyEvent::TLS for easier SSL/TLS context
252 creation, with many options including hostname verification,
253 secure default configuration, lots of documentation and,
254 predefined diffie-hellman keys for perfect forward security
255 and much more. get it while it's still fresh!
256 - use AnyEvent::TLS in AnyEvent::Handle for context management.
4 - load AnyEvent::Handle only on demand in AnyEvent::DNS, 257 - load AnyEvent::Handle only on demand in AnyEvent::DNS,
5 so AnyEvent::Socket users have smaller memory footprint 258 so AnyEvent::Socket users have smaller memory footprint
6 in the common case. 259 in the common case.
7 - add AnyEvent::Handle->push_shutdown method. 260 - add AnyEvent::Handle->push_shutdown method.
8 - add an additional error message parameter to AnyEvent::Handle's 261 - add an additional error message parameter to AnyEvent::Handle's
9 on_error callback (for TLS, $! is still available). 262 on_error callback (for TLS, $! is still available).
263 - add AnyEvent::Handle on_starttls/on_stoptls callbacks.
10 - make AnyEvent::Handle more robust against early conenction 264 - make AnyEvent::Handle more robust against early conenction
11 failures (during new), and return C<undef> in that case 265 failures (during new), and return C<undef> in that case
12 from the constructor. 266 from the constructor.
13 - AnyEvent::Handle will now try to load only JSON::XS first, 267 - AnyEvent::Handle will now try to load only JSON::XS first,
14 then fall back to JSON. 268 then fall back to JSON.
269 - format_ipv4/format_ipv6 are now exported by default, for symmetry,
270 and because it was documented that way.
15 271
164.451 Fri Jul 3 00:28:58 CEST 2009 2724.451 Fri Jul 3 00:28:58 CEST 2009
17 - do not clear rbuf when shutting down an AnyEvent::Handle 273 - do not clear rbuf when shutting down an AnyEvent::Handle
18 object - doing so breaks AnyEvent::HTTP. 274 object - doing so breaks AnyEvent::HTTP.
19 275
… …
140 - warn about not loaidng AnyEvent::Impl::POE early enough 396 - warn about not loaidng AnyEvent::Impl::POE early enough
141 (patch by Adam Rosenstein). 397 (patch by Adam Rosenstein).
142 - fixed a great number of bugs and corner cases in AnyEvent::Handle: 398 - fixed a great number of bugs and corner cases in AnyEvent::Handle:
143 - fix a bug in where in SSL connect mode, the 399 - fix a bug in where in SSL connect mode, the
144 client would first wait for some data by the server and 400 client would first wait for some data by the server and
145 otherwise hang. (reported and analyzed in an absolutely 401 otherwise hang. (reported and analysed in an absolutely
146 exemplary manner by Adam Rosenstein). 402 exemplary manner by Adam Rosenstein).
147 - fix a bug in where SSL EOF would not be treated 403 - fix a bug in where SSL EOF would not be treated
148 as stream EOF, putting the connection into a hung state. 404 as stream EOF, putting the connection into a hung state.
149 (reported and analyzed in an absolutely exemplary manner by Adam 405 (reported and analysed in an absolutely exemplary manner by Adam
150 Rosenstein). 406 Rosenstein).
151 - fix a potential segfault when the TLS context 407 - fix a potential segfault when the TLS context
152 would go missing in a read callback (e.g. due to stoptls) - 408 would go missing in a read callback (e.g. due to stoptls) -
153 Net::SSLeay of course makes no type checking whatsoever. 409 Net::SSLeay of course makes no type checking whatsoever.
154 - AnyEvent::Handle will not stop the read watcher in TLS mode. 410 - AnyEvent::Handle will not stop the read watcher in TLS mode.

Diff Legend

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