| 1 |
Revision history for Perl extension AnyEvent. |
| 2 |
|
| 3 |
4.21 Thu Jul 17 14:40:05 CEST 2008 |
| 4 |
- INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype |
| 5 |
which is incompatible to previous invocation syntax. |
| 6 |
- work around more windows perl bugs in fork_call. I cannot imagine |
| 7 |
why anybody would bother implementing fork in such an obviously |
| 8 |
unusable way. |
| 9 |
- avoid calling $do_retry if the request has already finished in |
| 10 |
AnyEvent::DNS, thus avoiding an exception (reported by Anatoly |
| 11 |
K. Sharifulin). |
| 12 |
|
| 13 |
4.2 Sat Jul 12 22:42:11 CEST 2008 |
| 14 |
- support an "interval" argument for repeating timers. |
| 15 |
- fix a bug in the pure perl event loop that caused timers to |
| 16 |
be delayed under some circumstances. |
| 17 |
- implement optional argument checking via PERL_ANYEVENT_STRICT. |
| 18 |
- not importing anything from Event causes it to malfunction |
| 19 |
and not use Time::HiRes, work around this bug. |
| 20 |
- more bug workarounds against the endlessly broken Tk module, |
| 21 |
seems to work now on 32 bit machines, but less so on 64 bit. |
| 22 |
- minor optimisations applied to most backends. |
| 23 |
|
| 24 |
4.161 Fri Jul 4 14:24:48 CEST 2008 |
| 25 |
- fixed the t/handle/02_write.t, it read after write, with the |
| 26 |
assumption that all data could be written before read |
| 27 |
(reported by Jim Razmus). |
| 28 |
|
| 29 |
4.160 Thu Jul 3 04:02:21 CEST 2008 |
| 30 |
- re-registering signal watchers after unregistering one did |
| 31 |
not work if a backend falls back on AnyEvent's default |
| 32 |
implementation (also affected child watchers). |
| 33 |
- new AnyEvent::Handle options: autocork and no_delay. |
| 34 |
|
| 35 |
4.152 Sun Jun 22 14:15:44 CEST 2008 |
| 36 |
- allow for 32-bit perls that implement shifts differently |
| 37 |
on different architectures in parse_ipv4 (reported and |
| 38 |
analysed by Keiichi DAIBA). |
| 39 |
|
| 40 |
4.151 Fri Jun 6 17:34:24 CEST 2008 |
| 41 |
- make sure specifying _only_ on_read and never pushing reads |
| 42 |
works. |
| 43 |
|
| 44 |
4.15 Fri Jun 6 13:00:46 CEST 2008 |
| 45 |
- the pure perl backend would keep some watchers alive when more than |
| 46 |
one watcher was registered for the same fd. |
| 47 |
- new "packstring" and "storable" read and write types |
| 48 |
for AnyEvent::Handle. |
| 49 |
- allow on_eof handler to be called after on_error with EPIPE returns. |
| 50 |
- do not immediately call on_read callback in handle constructor. |
| 51 |
|
| 52 |
4.14 Thu Jun 5 20:29:31 CEST 2008 |
| 53 |
- Fixed a bug in DNS SRV priority sorting. |
| 54 |
- AnyEvent::Util::guard now reports runtime errors while |
| 55 |
executing the guard block as warnings. |
| 56 |
- handle 0-byte-reads just before EOF correctly in AnyEvent::Handle. |
| 57 |
|
| 58 |
4.13 Thu Jun 5 00:47:59 CEST 2008 |
| 59 |
- AnyEvent::DNS only followed cname chains with length 2, |
| 60 |
contrary to documentation. bumped it up to 10, thanks to |
| 61 |
microsoft, the current broken-dns-config-king. |
| 62 |
- AnyEvent::DNS didn't check for socket return status |
| 63 |
but instead relied on perl not creating filehandles in that |
| 64 |
case - too bad it gives you a bogus file handle (reported |
| 65 |
and analysed by Vladimir Timofeev). |
| 66 |
- fix queue management logic in AnyEvent::Handle: |
| 67 |
when on_read was registered, the queue was empty |
| 68 |
and no progress could be made AnyEvent::Handle would |
| 69 |
enter an endless loop. |
| 70 |
- correctly start reading again when the handle became |
| 71 |
busy again after idling. |
| 72 |
- correctly treat tls shut-downs as EOF condition |
| 73 |
(for the time being, would be nice to have a callback for that). |
| 74 |
- correctly call eof callback under all conditions |
| 75 |
(hopefully). |
| 76 |
- the timeout callback did not expect that $self can go |
| 77 |
away any time. |
| 78 |
|
| 79 |
4.12 Tue Jun 3 10:58:04 CEST 2008 |
| 80 |
- include AnyEvent::Intro, a tutorial for anyevent, |
| 81 |
anyevent::socket and anyevent::handle. |
| 82 |
- allow more options in on_error. |
| 83 |
|
| 84 |
4.11 Fri May 30 23:42:25 CEST 2008 |
| 85 |
- INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide |
| 86 |
reverse_lookup and reverse_verify to replace it, support |
| 87 |
v4mapped and v4compat addresses. |
| 88 |
- provide more documentation for the resolver class. |
| 89 |
- really replace longest run of :0: by :: in format_address, |
| 90 |
also properly convert :: and ::1 again. |
| 91 |
- support NAPTR record name and decode it. |
| 92 |
- implement random weight sampling for SRV records, as per |
| 93 |
rfc 2782. |
| 94 |
- correctly abort on srv-record targets of ".". |
| 95 |
- added AnyEvent::DNS::wait_for_slot. |
| 96 |
- in the unlikely event of a virtual circuit connection |
| 97 |
being invalidated by a delayed udp reply, AnyEvent::DNS |
| 98 |
could die. this has been fixed. |
| 99 |
- plug a probable memleak in the DNS vc code. |
| 100 |
- use configured timeout also for VC request phase. |
| 101 |
- implement timeout and max_outstanding methods |
| 102 |
for resolver class. |
| 103 |
- update version numbers in all modules. |
| 104 |
|
| 105 |
4.1 Thu May 29 05:45:40 CEST 2008 |
| 106 |
- INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address |
| 107 |
which mirrors their purpose better (old names still available |
| 108 |
for a while). Also moved AnyEvent::DNS::addr to |
| 109 |
AnyEvent::Socket::resolve_sockaddr. |
| 110 |
- implement AnyEvent->time and AnyEvent->now. |
| 111 |
- fix IPv6 support in pack_sockaddr. |
| 112 |
- officially un-experimentalise all newly introduced code. |
| 113 |
- support unix domain sockets everywhere by specifying |
| 114 |
a host of "unix/" and the pathname as service. |
| 115 |
- implement an activity timeout in AnyEvent::Handle. |
| 116 |
- added a regex read type to AnyEvent::Handle. |
| 117 |
- added a json read/write type to AnyEvent::Handle. |
| 118 |
- always croak in on_error in AnyEvent::Handle. |
| 119 |
- document how to subclass AnyEvent::Handle. |
| 120 |
- implement AnyEvent::Util::fork_call. |
| 121 |
- add support for IPv6 nameservers and nameserver statements. |
| 122 |
- work around _yet_ _another_ windows perl bug in where empty |
| 123 |
select masks cause errors in the pure perl backend (this |
| 124 |
workaround cannot be implemented for the other event loops, |
| 125 |
but EV also works around this). |
| 126 |
- supply AnyEvent::Socket::address_type and make good use of it. |
| 127 |
- clarify and fix the Handle documentation w.r.t $self vs. $handle |
| 128 |
vs. callback arguments. |
| 129 |
- add some recommends to the META.yml. |
| 130 |
|
| 131 |
4.05 Mon May 26 19:44:06 CEST 2008 |
| 132 |
- some platforms ignore the rfc and prepend an extra sa_len member to |
| 133 |
the sockaddr structure, cater for those. also use sockaddr_family, |
| 134 |
didn't know it was there, but comes in handy for the workaround. |
| 135 |
- undo the import WIN32 hack. |
| 136 |
|
| 137 |
4.04 Mon May 26 08:03:31 CEST 2008 |
| 138 |
- try to work around yet another windows bug: failed connects are reported |
| 139 |
as if it were out-of-band data. windows users: you suck. Only |
| 140 |
EV, Glib, Event and the pure perl backend can handle this condition. |
| 141 |
- optimize write algorithm in AnyEvent::Handle. |
| 142 |
- properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly. |
| 143 |
- tcp_connect now properly iterates through all targets. |
| 144 |
- check wether IPv6 sockets can actually be created, otherwise, disable ipv6 |
| 145 |
support entirely. also hardcode some AF_INET6 constants for the |
| 146 |
sake of perl 5.8. |
| 147 |
- vastly improved nameserver/suffix detection algorithm for windows |
| 148 |
(still a hack...) |
| 149 |
- try to use a fallback dns server if no dns servers could be found. |
| 150 |
- splatter around some local $SIG{__DIE__}'s to work around bad |
| 151 |
die handlers. |
| 152 |
|
| 153 |
4.03 Sun May 25 05:05:57 CEST 2008 |
| 154 |
- work around perl bugs on windows, where perl returns undocumented |
| 155 |
error codes for sysread, syswrite etc. |
| 156 |
- fix AnyEvent::Handle::unshift_read to really unshift and not push. |
| 157 |
- AnyEvent::Handle could sometimes stumble over 0-byte writes. |
| 158 |
- fix unpack_sockaddr, now works for AF_INET6 sockaddr's. |
| 159 |
- updated documentation. |
| 160 |
- lots of workarounds for perl 5.8 quirks. |
| 161 |
- implement a more flexible type system for both reads and writes |
| 162 |
in AnyEvent::Handle. |
| 163 |
- support IPv6 in tcp_server. |
| 164 |
- pass local hostname and port to the prepare callback in tcp_server. |
| 165 |
- make types extendable in AnyEvent::Handle. |
| 166 |
- croak, not die, in AnyEvent::Handle; |
| 167 |
|
| 168 |
4.0 Sat May 24 19:58:08 CEST 2008 |
| 169 |
- added (experimental) AnyEvent::DNS module. |
| 170 |
- added (experimental) AnyEvent::Socket module. |
| 171 |
- added (experiemntal) support for TLS/SSL in AnyEvent::Handle. |
| 172 |
- make condvar->end call ->send as documented. |
| 173 |
- make condvar constructor support the documented cb => argument. |
| 174 |
- overload condvars so you can use them instead of code references. |
| 175 |
- support more of the POSIX forms instead of a pure dotted quad for |
| 176 |
verbatim ip addresses. |
| 177 |
- AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. |
| 178 |
- try very hard not to freeze on broken windows perls. |
| 179 |
- make on_eof optional in AnyEvent::Handle. |
| 180 |
- support service names in tcp_connect and tcp_server. |
| 181 |
- work around netbsd/openbsd bug causing unix sockets not to |
| 182 |
become writable until empty. |
| 183 |
- only make a dummy read for error detection when getpeername |
| 184 |
returns ENOTCONN. |
| 185 |
- defend AnyEvent::Util::inet_aton against Coro::LWP brutal |
| 186 |
override. |
| 187 |
- new AnyEvent::Guard::cancel. |
| 188 |
|
| 189 |
3.5 Sat May 17 23:17:14 CEST 2008 |
| 190 |
- gracefully deal with systems not defining CLOCK_MONOTONIC |
| 191 |
(as opposed to not supporting the clock itself, which worked in 3.41). |
| 192 |
- added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server |
| 193 |
helpers, to replace the ill-designed AnyEvent::Socket class. |
| 194 |
- added some examples (eg/connect, eg/handle). |
| 195 |
- added AnyEvent::Util::guard. |
| 196 |
- AnyEvent::Util now exports most of its functions. |
| 197 |
- fixed a bug in *_read_chunk, where the callback was called with $self as |
| 198 |
first and second argument. |
| 199 |
- simplified and fixed regex handling in AnyEvent::Handle::*_read_line. |
| 200 |
- fix manpage condvars examples to use new syntax. |
| 201 |
- considerably improve inet_aton with EV::ADNS backend to support |
| 202 |
cname chains (common dns configuration bug, e.g. with www.google.de) |
| 203 |
and aaaa records if no a record could be found. |
| 204 |
|
| 205 |
3.41 Sun May 11 19:53:13 CEST 2008 |
| 206 |
- work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. |
| 207 |
|
| 208 |
3.4 Sun May 11 00:29:25 CEST 2008 |
| 209 |
- complete re-implementation of condvars with a hopefully much mroe useful |
| 210 |
API (backward compatible functions exist but are undocumented). |
| 211 |
- AnyEvent::Handle fully rewritten with a hopefully more powerful API |
| 212 |
(still subject to change). |
| 213 |
- take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
| 214 |
- provide AnyEvent::post_detect and @AnyEvent::post_detect, which |
| 215 |
allows module authors to avoid forcing event loop detection. |
| 216 |
(used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example). |
| 217 |
- remove coro backends: Coro now provides generic support for AnyEvent |
| 218 |
via Coro::AnyEvent. |
| 219 |
|
| 220 |
3.3 Mon Apr 28 09:51:06 CEST 2008 |
| 221 |
- added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. |
| 222 |
- fix a bug in the pure perl backend that kept watchers alive |
| 223 |
when multiple watchers were registered for the same fh/poll combo. |
| 224 |
- add a benchmark section showing AnyEvent overhead and comparing |
| 225 |
the different event loops with each other. |
| 226 |
- prefer pure perl over tk when autoprobing, as it's about as fast, |
| 227 |
but doesn't crash with many watchers. |
| 228 |
- declare Qt support non-experimental. |
| 229 |
- clarify the confusing section about the file descriptor being |
| 230 |
kept alive. |
| 231 |
- document the race between loading of an event module and |
| 232 |
child processes exiting. |
| 233 |
- support POE as "backend" (with some caveats, POE is not generic |
| 234 |
enough, and darn slow). |
| 235 |
- support Wx and Prima through POE. |
| 236 |
- optimise perl backend to use 20% less memory and take advantage |
| 237 |
of typical timeout behaviour. It can now compete with |
| 238 |
select/poll-based C event loops in most cases (it is usually |
| 239 |
faster than Event and Glib :). |
| 240 |
- roughly cut EV memory use in half and increase its speed by 30%, |
| 241 |
by removing undocumented functionality. Did something similar |
| 242 |
to other event loops where possible. |
| 243 |
|
| 244 |
3.2 Thu Apr 24 10:10:40 CEST 2008 |
| 245 |
- do not die when anyevent watchers are destroyed while running callbacks |
| 246 |
in the pureperl backend (could only happen when two watchers are registered |
| 247 |
for the same fh/poll combo). |
| 248 |
- support autoloading for child watchers, was broken in all versions. |
| 249 |
- implement PERL_ANYEVENT_MODEL env variable. |
| 250 |
- (experimental) implement interface to the (very crashy/buggy) Event::Lib module. |
| 251 |
- (experimental) implement interface to the Qt module (cannot be autoprobed). |
| 252 |
- this release is IO:AnyEvent-proof. |
| 253 |
|
| 254 |
3.12 Tue Apr 22 07:11:46 CEST 2008 |
| 255 |
- reinstate AUTHOR section that got lost somehow. |
| 256 |
- do not hang in the testsuite with badly broken perls (activestate, |
| 257 |
strawberry...), but instead diagnose the problem and continue. |
| 258 |
- use INT instead of CHLD in an attempt to support broken windows perls |
| 259 |
better (this decreases the test quality, unfortunately). |
| 260 |
- do not send the signal to the process group |
| 261 |
(no problem for CHLD, bad for INT :). |
| 262 |
|
| 263 |
3.11 Sat Apr 19 06:57:31 CEST 2008 |
| 264 |
- major documentation rework. |
| 265 |
- document the fact that child watchers only watch for zombies. |
| 266 |
- fix the child watcher example. |
| 267 |
|
| 268 |
3.1 Wed Apr 16 17:09:01 CEST 2008 |
| 269 |
- work around recurring bugs in Tk by dup'ing filehandles, the |
| 270 |
only method with good success chances on Tk (the bugs apparently |
| 271 |
don't get fixed anytime soon). |
| 272 |
- lift the restriction of only one watcher per fh direction |
| 273 |
(as the Tk bug workaround also lifts it and only Tk imposed |
| 274 |
such strong limits). |
| 275 |
- changed probe order to prefer coro adaptors. |
| 276 |
- explain why recursion into the event loop is not supported |
| 277 |
unless the backend supports it (only Coro::EV does without |
| 278 |
any restrictions...). |
| 279 |
- add simple manpages for all backend modules. |
| 280 |
|
| 281 |
3.0 Mon Apr 7 21:30:23 CEST 2008 |
| 282 |
- Coro::Signal changed semantics, roll our own, also cleaning |
| 283 |
up the Coro implementation in general. |
| 284 |
- rename Coro backend to CoroEvent. |
| 285 |
- add some decision helping paragraph to the manpage that should |
| 286 |
help people to decide whether AnyEvent is the right thing for them. |
| 287 |
|
| 288 |
2.9 Mon Jan 28 13:31:54 CET 2008 |
| 289 |
- update for EV 3.0 API changes. |
| 290 |
|
| 291 |
2.8 Sun Nov 25 15:06:03 CET 2007 |
| 292 |
- waitpid can validly return 0. accept this fact of life |
| 293 |
instead of reporting it to any watchers. |
| 294 |
|
| 295 |
2.7 Fri Nov 23 11:41:14 CET 2007 |
| 296 |
- force use of AnyEvent::Impl::Perl in testsuite, there is |
| 297 |
too much breakage outside AnyEvent. |
| 298 |
- deliver signals synchronously in AnyEvent::Impl::Perl |
| 299 |
or any other event loop that relies on AnyEvents child |
| 300 |
watcher emulation. *Could* help with hanging testsuite |
| 301 |
(except when Event or EV are installed). |
| 302 |
|
| 303 |
2.6 Fri Nov 9 20:36:35 CET 2007 |
| 304 |
- fix bug in testsuite. |
| 305 |
- move EV adaptor modules to AnyEvent. |
| 306 |
- add Coro+EV adaptor module. |
| 307 |
|
| 308 |
2.55 Tue Nov 6 17:41:32 CET 2007 |
| 309 |
- add EV to the list of supported event models. |
| 310 |
- do not auto-reset pid watchers, pass pid and status to them. |
| 311 |
- allow a pid of zero to watch for all children in child watchers. |
| 312 |
|
| 313 |
2.54 Wed Jul 18 17:36:23 CEST 2007 |
| 314 |
- work around a perl bug that results in |
| 315 |
BEGIN not safe after errors--compilation aborted without |
| 316 |
any discernible reason or error message by once |
| 317 |
again not use'ing strict. |
| 318 |
|
| 319 |
2.53 Sun Jul 8 10:51:53 CEST 2007 |
| 320 |
- make Glib interface work again (spotted by elmex). |
| 321 |
|
| 322 |
2.52 Wed Mar 7 18:36:16 CET 2007 |
| 323 |
- child watchers stopped working permanently when |
| 324 |
all current child watchers were destroyed. |
| 325 |
|
| 326 |
2.51 Mon Dec 11 21:33:24 CET 2006 |
| 327 |
- work around bugs in perl where eval "require Module" returns |
| 328 |
true even if the module couldn't be loaded :/. |
| 329 |
|
| 330 |
2.5 Mon Dec 11 02:15:28 CET 2006 |
| 331 |
- avoid $AUTOLOAD because many perls corrupt it. |
| 332 |
- AnyEvent::detect forces autodetection. |
| 333 |
- implement signal watchers (experimental) |
| 334 |
(Tk does not support async signals (<= 804.027 at least), |
| 335 |
so they might get delayed indefinitely on Tk). |
| 336 |
- implement child watchers (experimental). |
| 337 |
- moved default condvar implementation into base module, simplifying |
| 338 |
most implementation modules. |
| 339 |
|
| 340 |
2.1 Fri Nov 24 15:50:48 CET 2006 |
| 341 |
- better docs. |
| 342 |
- simple testsuite. |
| 343 |
- added AnyEvent->one_event method for special purposes. |
| 344 |
|
| 345 |
2.0 Wed Nov 1 02:21:30 CET 2006 |
| 346 |
- INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both. |
| 347 |
- INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore. |
| 348 |
- EXPERIMENTAL: |
| 349 |
new pure-perl model added, AnyEvent now always finds a suitable |
| 350 |
event model. |
| 351 |
- improved documentation. |
| 352 |
- reduced cpu overhead. |
| 353 |
|
| 354 |
1.02 Fri Jan 13 14:15:40 CET 2006 |
| 355 |
- add COPYING to clarify license. |
| 356 |
|
| 357 |
1.01 Sun Jan 8 05:48:09 CET 2006 |
| 358 |
- correctly cancel Event timers, avoid memory leaks. |
| 359 |
|
| 360 |
1.0 Sun Jan 8 05:40:58 CET 2006 |
| 361 |
- allow third-party interfaces to be added at runtime. |
| 362 |
- fix version number. |
| 363 |
|
| 364 |
0.9 Fri Dec 30 02:25:41 CET 2005 |
| 365 |
- fixed errors in the documentation. |
| 366 |
- going back to using $VERSION as "load check", |
| 367 |
as suggested by Jörn Reder. Let's see why I stopped using |
| 368 |
this method. |
| 369 |
- add 'hup' to the list of events to receive for glib |
| 370 |
and generate 'rw' events for it, as some versions of |
| 371 |
glib seem to require this, while this breaks other |
| 372 |
(older) versions of glib. |
| 373 |
- implemented PERL_ANYEVENT_VERBOSE env variable. |
| 374 |
|
| 375 |
0.3 Sun Dec 4 10:43:33 CET 2005 |
| 376 |
- no changes to the code. |
| 377 |
- fixed documentation. |
| 378 |
- clarified some corner cases. |
| 379 |
- vastly improved documentation with a long example. |
| 380 |
|
| 381 |
0.2 Thu Dec 1 22:18:49 CET 2005 |
| 382 |
- check whether a valid method was called, to avoid endless loops. |
| 383 |
- move condvars into their own class (cleaner, safer). |
| 384 |
- don't die when some modules are not available. |
| 385 |
- Tk is too broken to allow for words. Will work once Tk works. |
| 386 |
|
| 387 |
0.1 Thu Dec 1 19:44:57 CET 2005 |
| 388 |
- spiced up and cleaned up for release. |
| 389 |
|
| 390 |
0.01 Tue Sep 21 18:12:09 CEST 2004 |
| 391 |
- original version; copied from Convert::Scalar. |
| 392 |
|
