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

Comparing AnyEvent/Changes (file contents):
Revision 1.61 by elmex, Sun Apr 27 19:15:43 2008 UTC vs.
Revision 1.194 by root, Thu Oct 2 06:42:39 2008 UTC

1Revision history for Perl extension AnyEvent. 1Revision history for Perl extension AnyEvent.
2 2
34.3
4 - fixed a great number of bugs and corner cases in AnyEvent::Handle:
5 - fix a bug in where in SSL connect mode, the
6 client would first wait for some data by the server and
7 otherwise hang. (Reported and analyzed in an absolutely
8 exemplary manner by Adam Rosenstein).
9 - fix a bug in where SSL EOF would not be treated
10 as stream EOF, putting the connection into a hung state.
11 (Reported and analyzed in an absolutely exemplary manner by Adam
12 Rosenstein).
13 - fix a potential segfault when the TLS context
14 would go missing in a read callback (e.g. due to stoptls) -
15 Net::SSLeay of course makes no type checking whatsoever.
16 - AnyEvent::Handle will not stop the read watcher in TLS mode.
17 - AnyEvent::Handle->stoptls will send a TLS close notify
18 instead of simply destroying the stream now.
19 - on_drain will now take the tls write buffer into account.
20 - SSL operations have been streamlined a bit and should be faster
21 now, more to come, though.
22 - added a FAQ section, will document nontrivial issues in there.
23
244.234 Mon Sep 29 04:08:13 CEST 2008
25 - fix child watcher documentation: only child exits will be reported,
26 no trace events.
27 - mention SIGPIPE in AnyEvent::Handle.
28 - perl backend will now use POSIX::times as monotonic clock
29 when available and the monotonic clock option is not.
30 - run even without Time::HiRes in some cases now, but do not
31 provide sub-second accuracy (all *supported* perl versions have
32 Time::HiRes).
33
344.233 Fri Aug 22 01:48:05 CEST 2008
35 - fix a bug in the testsuite, causing a hang.
36 - clarified lots of AnyEvent::Handle constructor arguments.
37
384.232 Thu Aug 21 20:44:25 CEST 2008
39 - fix a bug in AnyEvent::Handle that could cause two fatal error
40 callbacks on EOF, which caused AnyEvent::HTTP to signal success twice
41 (testcase by Leon Brocard).
42 - clarified on_eof/on_error documentation a bit.
43
444.231 Tue Jul 29 13:12:15 CEST 2008
45 - remove some debugging code left in AnyEvent::Util::fork_call
46 (and no, it's impossible to implement with the broken windows
47 perls without resource leaks or worse).
48
494.23 Tue Jul 29 12:19:59 CEST 2008
50 - document the first parameter passed to condvar callbacks to be
51 the callback.
52 - add AnyEvent::Socket::{ntoa,aton} aliases.
53 - optimize the AE::Handle->push_read (line) for the default
54 eol marker.
55 - optimize push_read (packstring|storable) for small packets.
56 - invoke on_error callback when no on_eof callback is set.
57 - fix a bug in push_read (storable) of unknown impact.
58
594.22 Sun Jul 20 16:34:13 CEST 2008
60 - new function AnyEvent::Socket::parse_hostport.
61 - as the bulkheads at microsoft can't even get getprotobyname reliably
62 working on their shitty fucking broken os we need to hardcode
63 some common protocol numbers in AnyEvent::Socket. How can
64 people even bother with such a pile of shit as windows.
65
664.21 Thu Jul 17 14:40:05 CEST 2008
67 - INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype
68 which is incompatible to previous invocation syntax.
69 - work around more windows perl bugs in fork_call. I cannot imagine
70 why anybody would bother implementing fork in such an obviously
71 unusable way.
72 - avoid calling $do_retry if the request has already finished in
73 AnyEvent::DNS, thus avoiding an exception (reported by Anatoly
74 K. Sharifulin).
75
764.2 Sat Jul 12 22:42:11 CEST 2008
77 - support an "interval" argument for repeating timers.
78 - fix a bug in the pure perl event loop that caused timers to
79 be delayed under some circumstances.
80 - implement optional argument checking via PERL_ANYEVENT_STRICT.
81 - not importing anything from Event causes it to malfunction
82 and not use Time::HiRes, work around this bug.
83 - more bug workarounds against the endlessly broken Tk module,
84 seems to work now on 32 bit machines, but less so on 64 bit.
85 - minor optimisations applied to most backends.
86
874.161 Fri Jul 4 14:24:48 CEST 2008
88 - fixed the t/handle/02_write.t, it read after write, with the
89 assumption that all data could be written before read
90 (reported by Jim Razmus).
91
924.160 Thu Jul 3 04:02:21 CEST 2008
93 - re-registering signal watchers after unregistering one did
94 not work if a backend falls back on AnyEvent's default
95 implementation (also affected child watchers).
96 - new AnyEvent::Handle options: autocork and no_delay.
97
984.152 Sun Jun 22 14:15:44 CEST 2008
99 - allow for 32-bit perls that implement shifts differently
100 on different architectures in parse_ipv4 (reported and
101 analysed by Keiichi DAIBA).
102
1034.151 Fri Jun 6 17:34:24 CEST 2008
104 - make sure specifying _only_ on_read and never pushing reads
105 works.
106
1074.15 Fri Jun 6 13:00:46 CEST 2008
108 - the pure perl backend would keep some watchers alive when more than
109 one watcher was registered for the same fd.
110 - new "packstring" and "storable" read and write types
111 for AnyEvent::Handle.
112 - allow on_eof handler to be called after on_error with EPIPE returns.
113 - do not immediately call on_read callback in handle constructor.
114
1154.14 Thu Jun 5 20:29:31 CEST 2008
116 - Fixed a bug in DNS SRV priority sorting.
117 - AnyEvent::Util::guard now reports runtime errors while
118 executing the guard block as warnings.
119 - handle 0-byte-reads just before EOF correctly in AnyEvent::Handle.
120
1214.13 Thu Jun 5 00:47:59 CEST 2008
122 - AnyEvent::DNS only followed cname chains with length 2,
123 contrary to documentation. bumped it up to 10, thanks to
124 microsoft, the current broken-dns-config-king.
125 - AnyEvent::DNS didn't check for socket return status
126 but instead relied on perl not creating filehandles in that
127 case - too bad it gives you a bogus file handle (reported
128 and analysed by Vladimir Timofeev).
129 - fix queue management logic in AnyEvent::Handle:
130 when on_read was registered, the queue was empty
131 and no progress could be made AnyEvent::Handle would
132 enter an endless loop.
133 - correctly start reading again when the handle became
134 busy again after idling.
135 - correctly treat tls shut-downs as EOF condition
136 (for the time being, would be nice to have a callback for that).
137 - correctly call eof callback under all conditions
138 (hopefully).
139 - the timeout callback did not expect that $self can go
140 away any time.
141
1424.12 Tue Jun 3 10:58:04 CEST 2008
143 - include AnyEvent::Intro, a tutorial for anyevent,
144 anyevent::socket and anyevent::handle.
145 - allow more options in on_error.
146
1474.11 Fri May 30 23:42:25 CEST 2008
148 - INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide
149 reverse_lookup and reverse_verify to replace it, support
150 v4mapped and v4compat addresses.
151 - provide more documentation for the resolver class.
152 - really replace longest run of :0: by :: in format_address,
153 also properly convert :: and ::1 again.
154 - support NAPTR record name and decode it.
155 - implement random weight sampling for SRV records, as per
156 rfc 2782.
157 - correctly abort on srv-record targets of ".".
158 - added AnyEvent::DNS::wait_for_slot.
159 - in the unlikely event of a virtual circuit connection
160 being invalidated by a delayed udp reply, AnyEvent::DNS
161 could die. this has been fixed.
162 - plug a probable memleak in the DNS vc code.
163 - use configured timeout also for VC request phase.
164 - implement timeout and max_outstanding methods
165 for resolver class.
166 - update version numbers in all modules.
167
1684.1 Thu May 29 05:45:40 CEST 2008
169 - INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address
170 which mirrors their purpose better (old names still available
171 for a while). Also moved AnyEvent::DNS::addr to
172 AnyEvent::Socket::resolve_sockaddr.
173 - implement AnyEvent->time and AnyEvent->now.
174 - fix IPv6 support in pack_sockaddr.
175 - officially un-experimentalise all newly introduced code.
176 - support unix domain sockets everywhere by specifying
177 a host of "unix/" and the pathname as service.
178 - implement an activity timeout in AnyEvent::Handle.
179 - added a regex read type to AnyEvent::Handle.
180 - added a json read/write type to AnyEvent::Handle.
181 - always croak in on_error in AnyEvent::Handle.
182 - document how to subclass AnyEvent::Handle.
183 - implement AnyEvent::Util::fork_call.
184 - add support for IPv6 nameservers and nameserver statements.
185 - work around _yet_ _another_ windows perl bug in where empty
186 select masks cause errors in the pure perl backend (this
187 workaround cannot be implemented for the other event loops,
188 but EV also works around this).
189 - supply AnyEvent::Socket::address_type and make good use of it.
190 - clarify and fix the Handle documentation w.r.t $self vs. $handle
191 vs. callback arguments.
192 - add some recommends to the META.yml.
193
1944.05 Mon May 26 19:44:06 CEST 2008
195 - some platforms ignore the rfc and prepend an extra sa_len member to
196 the sockaddr structure, cater for those. also use sockaddr_family,
197 didn't know it was there, but comes in handy for the workaround.
198 - undo the import WIN32 hack.
199
2004.04 Mon May 26 08:03:31 CEST 2008
201 - try to work around yet another windows bug: failed connects are reported
202 as if it were out-of-band data. windows users: you suck. Only
203 EV, Glib, Event and the pure perl backend can handle this condition.
204 - optimize write algorithm in AnyEvent::Handle.
205 - properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly.
206 - tcp_connect now properly iterates through all targets.
207 - check wether IPv6 sockets can actually be created, otherwise, disable ipv6
208 support entirely. also hardcode some AF_INET6 constants for the
209 sake of perl 5.8.
210 - vastly improved nameserver/suffix detection algorithm for windows
211 (still a hack...)
212 - try to use a fallback dns server if no dns servers could be found.
213 - splatter around some local $SIG{__DIE__}'s to work around bad
214 die handlers.
215
2164.03 Sun May 25 05:05:57 CEST 2008
217 - work around perl bugs on windows, where perl returns undocumented
218 error codes for sysread, syswrite etc.
219 - fix AnyEvent::Handle::unshift_read to really unshift and not push.
220 - AnyEvent::Handle could sometimes stumble over 0-byte writes.
221 - fix unpack_sockaddr, now works for AF_INET6 sockaddr's.
222 - updated documentation.
223 - lots of workarounds for perl 5.8 quirks.
224 - implement a more flexible type system for both reads and writes
225 in AnyEvent::Handle.
226 - support IPv6 in tcp_server.
227 - pass local hostname and port to the prepare callback in tcp_server.
228 - make types extendable in AnyEvent::Handle.
229 - croak, not die, in AnyEvent::Handle;
230
2314.0 Sat May 24 19:58:08 CEST 2008
232 - added (experimental) AnyEvent::DNS module.
233 - added (experimental) AnyEvent::Socket module.
234 - added (experiemntal) support for TLS/SSL in AnyEvent::Handle.
235 - make condvar->end call ->send as documented.
236 - make condvar constructor support the documented cb => argument.
237 - overload condvars so you can use them instead of code references.
238 - support more of the POSIX forms instead of a pure dotted quad for
239 verbatim ip addresses.
240 - AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore.
241 - try very hard not to freeze on broken windows perls.
242 - make on_eof optional in AnyEvent::Handle.
243 - support service names in tcp_connect and tcp_server.
244 - work around netbsd/openbsd bug causing unix sockets not to
245 become writable until empty.
246 - only make a dummy read for error detection when getpeername
247 returns ENOTCONN.
248 - defend AnyEvent::Util::inet_aton against Coro::LWP's brutal
249 override.
250 - new AnyEvent::Guard::cancel.
251
2523.5 Sat May 17 23:17:14 CEST 2008
253 - gracefully deal with systems not defining CLOCK_MONOTONIC
254 (as opposed to not supporting the clock itself, which worked in 3.41).
255 - added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server
256 helpers, to replace the ill-designed AnyEvent::Socket class.
257 - added some examples (eg/connect, eg/handle).
258 - added AnyEvent::Util::guard.
259 - AnyEvent::Util now exports most of its functions.
260 - fixed a bug in *_read_chunk, where the callback was called with $self as
261 first and second argument.
262 - simplified and fixed regex handling in AnyEvent::Handle::*_read_line.
263 - fix manpage condvars examples to use new syntax.
264 - considerably improve inet_aton with EV::ADNS backend to support
265 cname chains (common dns configuration bug, e.g. with www.google.de)
266 and aaaa records if no a record could be found.
267
2683.41 Sun May 11 19:53:13 CEST 2008
269 - work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao.
270
2713.4 Sun May 11 00:29:25 CEST 2008
272 - complete re-implementation of condvars with a hopefully much more useful
273 API (backward compatible functions exist but are undocumented).
274 - AnyEvent::Handle fully rewritten with a hopefully more powerful API
275 (still subject to change).
276 - take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl.
277 - provide AnyEvent::post_detect and @AnyEvent::post_detect, which
278 allows module authors to avoid forcing event loop detection.
279 (used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example).
280 - remove coro backends: Coro now provides generic support for AnyEvent
281 via Coro::AnyEvent.
282
2833.3 Mon Apr 28 09:51:06 CEST 2008
3 - added AnyEvent::Handle and AnyEvent::Socket modules 284 - added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules.
4 - fix a bug in the pure perl backend that kept watchers alive 285 - fix a bug in the pure perl backend that kept watchers alive
5 when multiple watchers were registered for the same fh/poll combo. 286 when multiple watchers were registered for the same fh/poll combo.
6 - add a benchmark section showing AnyEvent overhead and comparing 287 - add a benchmark section showing AnyEvent overhead and comparing
7 the different event loops with each other. 288 the different event loops with each other.
8 - prefer pure perl over tk when autoprobing, as it's about as fast, 289 - prefer pure perl over tk when autoprobing, as it's about as fast,
9 but doesn't crash with many watchers. 290 but doesn't crash with many watchers.
10 - make Qt support non-experimental. 291 - declare Qt support non-experimental.
11 - clarify the confusing section about the file descriptor being 292 - clarify the confusing section about the file descriptor being
12 kept alive. 293 kept alive.
13 - document the race between loading of an event module and 294 - document the race between loading of an event module and
14 child processes exiting. 295 child processes exiting.
15 - support POE as "backend" (with some caveats, POE is not generic 296 - support POE as "backend" (with some caveats, POE is not generic
16 enough, and darn slow). 297 enough, and darn slow).
17 - support Wx and Prima through POE. 298 - support Wx and Prima through POE.
18 - optimise perl backend to use 20% less memory and take advantage 299 - optimise perl backend to use 20% less memory and take advantage
19 of typical timeout behaviour. It can now compete with 300 of typical timeout behaviour. It can now compete with
20 select/poll-based C event loops in most cases (it is usually 301 select/poll-based C event loops in most cases (it is usually
21 faster than Event and Glib :). 302 faster than Event and Glib :).
22 - roughly cut EV memory use in half and increase its speed by 30%, 303 - roughly cut EV memory use in half and increase its speed by 30%,
23 by removing undocumented functionality. 304 by removing undocumented functionality. Did something similar
305 to other event loops where possible.
24 306
253.2 Thu Apr 24 10:10:40 CEST 2008 3073.2 Thu Apr 24 10:10:40 CEST 2008
26 - do not die when anyevent watchers are destroyed while running callbacks 308 - do not die when anyevent watchers are destroyed while running callbacks
27 in the pureperl backend (could only happen when two watchers are registered 309 in the pureperl backend (could only happen when two watchers are registered
28 for the same fh/poll combo). 310 for the same fh/poll combo).

Diff Legend

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