| 1 |
Revision history for Perl extension AnyEvent. |
| 2 |
|
| 3 |
4.88 Tue Jul 28 04:04:37 CEST 2009 |
| 4 |
- re-bless the handle into a dummy package after calling |
| 5 |
AnyEvent::Handle::destroy, so the user does not need to check |
| 6 |
for errors after every push_write etc. |
| 7 |
- do not attempt to run t/02_signals.t on obviously broken |
| 8 |
platforms. |
| 9 |
|
| 10 |
4.87 Sun Jul 26 02:06:16 CEST 2009 |
| 11 |
- do not attempt to linger when there is no longer a valid fh. |
| 12 |
- enforce tls mode to be either accept or connect, do not simply |
| 13 |
segfault in Net::SSLeay. |
| 14 |
- AnyEvent::Handle can now call tcp_connect itself (new parameters |
| 15 |
connect, on_prepare, on_connect and on_connect_error). Updated |
| 16 |
tutorial accordingly. |
| 17 |
- add AnyEvent::Impl::Irssi backend. |
| 18 |
|
| 19 |
4.86 Mon Jul 20 23:52:29 CEST 2009 |
| 20 |
- since the verbose warning is not enough, explicitly document |
| 21 |
that versions before 1.33 of Net::SSLeay are not secure. |
| 22 |
- work around signal handling races in Event and (...) Event::Lib. |
| 23 |
- try to align signal-race timer to full-second boundaries. |
| 24 |
- work around Tk not liking negative timeouts. |
| 25 |
- don't complain of different grades of environmental unfriendlyness |
| 26 |
in IO::Async. |
| 27 |
|
| 28 |
4.85 Sat Jul 18 06:16:14 CEST 2009 |
| 29 |
- nail the signal race problem in perl once and for all |
| 30 |
(see $AnyEvent::MAX_SIGNAL_LATENCY). |
| 31 |
- take advantage of Async::Interrupt if it is available. |
| 32 |
- load Time::HiRes and Guard modules on demand only. |
| 33 |
- add optional/recommended modules section to AnyEvent |
| 34 |
documentation. |
| 35 |
- reduce memory usage considerably (and reduce startup penalty) |
| 36 |
by not using "strict", "warnings" and "overload" modules. |
| 37 |
- work around buggy windows/openbsd perls and provide EBADMSG |
| 38 |
and EPROTO ourselves when missing. |
| 39 |
- improve perl 5.6 compatibility of the core event loop. |
| 40 |
- made Net::SSLeay version 1.33 a soft requirement. |
| 41 |
|
| 42 |
4.83 Fri Jul 17 16:56:26 CEST 2009 |
| 43 |
- implement AnyEvent::Socket::getprotobyname. |
| 44 |
- AnyEvent::CondVar's will now detect recursive blocking |
| 45 |
waits and will croak, as too many people fall into |
| 46 |
this trap. |
| 47 |
- AnyEvent::Handle will now call ->destroy on itself after |
| 48 |
executing the on_error callback, instead of doing some |
| 49 |
half-baked internal shutdown, for fatal errors. |
| 50 |
- clarify on_eof behaviour w.r.t. the read queue and |
| 51 |
on_read callbacks. |
| 52 |
- ignore some possible spurious wake-ups in tcp_connect. |
| 53 |
|
| 54 |
4.82 Sat Jul 11 00:34:55 CEST 2009 |
| 55 |
- POE and Event backends didn't accept some callable objects as |
| 56 |
callbacks. |
| 57 |
- use Config module instead of POSIX module to detect signal names |
| 58 |
in AnyEvent::Strict and AnyEvent::Impl::EventLib, as the POSIX |
| 59 |
module doesn't even have all POSIX signals :/. |
| 60 |
- use more workarounds around the many refcnt/corruption bugs in |
| 61 |
Event::Lib. |
| 62 |
- work around a race condition in perl's select, causing t/03_child.t |
| 63 |
to rarely fail. |
| 64 |
|
| 65 |
4.81 Thu Jul 9 10:30:30 CEST 2009 |
| 66 |
- AnyEvent::Handle didn't properly diagnose write errors |
| 67 |
(it expected -1 from syswrite, how lame... :). |
| 68 |
- support file descriptors in addition to file handles |
| 69 |
in AnyEvent->io. |
| 70 |
- new env variables: PERL_ANYEVENT_RESOLV_CONF, |
| 71 |
PERL_ANYEVENT_MAX_OUTSTANDING_DNS, PERL_ANYEVENT_CA_FILE |
| 72 |
and PERL_ANYEVENT_CA_PATH. |
| 73 |
- provide a sensible synopsis section for AnyEvent::TLS. |
| 74 |
- add a "supported backends" section to the manpage. |
| 75 |
- added simple io watcher test to testsuite, using a |
| 76 |
portable_socketpair. |
| 77 |
- tried to improve the stability of the Event::Lib backend, |
| 78 |
YMMV. |
| 79 |
|
| 80 |
4.8 Mon Jul 6 23:45:16 CEST 2009 |
| 81 |
- AnyEvent::DNS did not properly follow CNAME records with |
| 82 |
uppercase targets. |
| 83 |
- AnyEvent::DNS would errornously return AAAA records |
| 84 |
with v4 mapped addresses (a faulty record) as ipv4 addresses, |
| 85 |
causing AnyEvent::Socket to throw an exception. |
| 86 |
- added new module AnyEvent::TLS for easier SSL/TLS context |
| 87 |
creation, with many options including hostname verification, |
| 88 |
secure default configuration, lots of documentation and, |
| 89 |
predefined diffie-hellman keys for perfect forward security |
| 90 |
and much more. get it while it's still fresh! |
| 91 |
- use AnyEvent::TLS in AnyEvent::Handle for context management. |
| 92 |
- load AnyEvent::Handle only on demand in AnyEvent::DNS, |
| 93 |
so AnyEvent::Socket users have smaller memory footprint |
| 94 |
in the common case. |
| 95 |
- add AnyEvent::Handle->push_shutdown method. |
| 96 |
- add an additional error message parameter to AnyEvent::Handle's |
| 97 |
on_error callback (for TLS, $! is still available). |
| 98 |
- add AnyEvent::Handle on_starttls/on_stoptls callbacks. |
| 99 |
- make AnyEvent::Handle more robust against early conenction |
| 100 |
failures (during new), and return C<undef> in that case |
| 101 |
from the constructor. |
| 102 |
- AnyEvent::Handle will now try to load only JSON::XS first, |
| 103 |
then fall back to JSON. |
| 104 |
- format_ipv4/format_ipv6 are now exported by default, for symmetry, |
| 105 |
and because it was documented that way. |
| 106 |
|
| 107 |
4.451 Fri Jul 3 00:28:58 CEST 2009 |
| 108 |
- do not clear rbuf when shutting down an AnyEvent::Handle |
| 109 |
object - doing so breaks AnyEvent::HTTP. |
| 110 |
|
| 111 |
4.45 Mon Jun 29 22:59:26 CEST 2009 |
| 112 |
- a write error could cause AnyEvent::Handle to create |
| 113 |
an I/O watcher with an undefined $fh. |
| 114 |
- special-case mapped ipv4 addresses in both |
| 115 |
AnyEvent::Socket::format_address and parse_address, |
| 116 |
to treat them just like ipv4 addresses. |
| 117 |
- updated and overhauled the AnyEvent::Intro doc. |
| 118 |
- implement AnyEvent::Socket::format_ipv4/ipv6. |
| 119 |
- slightly speed up portable_pipe/socketpair. |
| 120 |
- expand condvar begin/end documentation. |
| 121 |
|
| 122 |
4.42 Fri Jun 26 08:32:18 CEST 2009 |
| 123 |
- preliminary and neccesarily incomplete support for IO::Async. |
| 124 |
- reset SIGCHLD to DEFAULT when AnyEvent is loaded, in |
| 125 |
case it was set to IGNORE, to ensure we can catch child statuses |
| 126 |
even when the calling env acts stupidly. |
| 127 |
- updated benchmarks with IO::Async, which performs very well. |
| 128 |
|
| 129 |
4.412 Wed Jun 24 01:35:57 CEST 2009 |
| 130 |
- support an "untaint" attribute for AnyEvent::DNS |
| 131 |
and set it on the default resolver. |
| 132 |
- implement and document AnyEvent::Impl::Perl::loop. |
| 133 |
- remove all anyevent-env variables from %ENV |
| 134 |
when running in tainted mode. |
| 135 |
- mention and extend the IO::Lambda benchmark. |
| 136 |
|
| 137 |
4.411 Sun Jun 7 18:48:02 CEST 2009 |
| 138 |
- do not try to use F_SETFD on windows. |
| 139 |
|
| 140 |
4.41 Thu May 14 06:40:11 CEST 2009 |
| 141 |
- work around issues in older perls (5.8.0?) when |
| 142 |
a signal handler is deleted from the %SIG hash. |
| 143 |
- use POSIX::_exit in child test, to avoid running |
| 144 |
destructors. |
| 145 |
- speed up CHLD handling by relying on SIGCHLD being |
| 146 |
synchronously delivered, even when we roll our own |
| 147 |
implementation. |
| 148 |
- AnyEvent::DNS: add the "dname" resource record name |
| 149 |
for cosmetic reasons. |
| 150 |
|
| 151 |
4.4 Sun Apr 26 20:12:33 CEST 2009 |
| 152 |
- implemented idle watchers, where applicable. |
| 153 |
- AnyEvent->time died when Event backend was in use. |
| 154 |
- fix a memleak in the Tk backend. |
| 155 |
- sped up Tk timer handling. |
| 156 |
- clip negative "after" values to 0 in AnyEvent::Impl::Event |
| 157 |
to avoid spamming out warning messages. |
| 158 |
- fix Qt timers without interval. |
| 159 |
- avoid Qt zero-timeout special case, in old libqt's. |
| 160 |
|
| 161 |
4.352 Mon Apr 20 16:31:11 CEST 2009 |
| 162 |
- fix AnyEvent::Strict error messages for child watchers. |
| 163 |
- fix/update Makefile.PL "recommends" versions. |
| 164 |
- add AnyEvent->now_update. |
| 165 |
|
| 166 |
4.351 Sat Apr 11 07:56:14 CEST 2009 |
| 167 |
- actually make the signal pipe work under win32. |
| 168 |
- localise $! in signal handler to avoid changing $!, although |
| 169 |
perl itself does not seem to save/restore errno either. |
| 170 |
- set the cloexec flag on the signal pipe (normally set by perl too). |
| 171 |
|
| 172 |
4.35 Fri Mar 27 11:48:20 CET 2009 |
| 173 |
- event models relying on AnyEvent's signal watcher emulation did |
| 174 |
invoke the callback asynchronously, contradicting documentation |
| 175 |
and causing signals to get lost (this includes AnyEvent's own |
| 176 |
event loop). AnyEvent now uses the standard pipe trick to make |
| 177 |
callback execution synchronous to the event loop. |
| 178 |
- AnyEvent::Handle didn't free TLS context data on DESTROY |
| 179 |
(patch by Pavel Shaydo). |
| 180 |
- work around the ever-incompatibly-changing API of MakeMaker. |
| 181 |
- document that changing global variables without restoring them |
| 182 |
is a bad idea in Perl (noted by Adam Rosenstein). |
| 183 |
- AnyEvent::Strict now barfs if ->io is passed a file. |
| 184 |
|
| 185 |
4.34 Thu Feb 12 18:32:45 CET 2009 |
| 186 |
- separately buffer TLS read data, as otherwise the read queue could |
| 187 |
deadlock as receiving data is not expected while draining the |
| 188 |
read queue (which cna only happen with TLS). |
| 189 |
- raise EBADMSG error on JSON decoding errors. |
| 190 |
- fix some minor manpage bugs (reported by Maximilian Gaß). |
| 191 |
- speed up select bitmask parsing quite a bit in the pure perl |
| 192 |
backend. |
| 193 |
- use CORE::select instead of just select in AnyEvent::Impl::Perl. |
| 194 |
|
| 195 |
4.331 Tue Jan 6 21:07:25 CET 2009 |
| 196 |
- socketpair fails on many vista machines because vista has |
| 197 |
completely broken accept/getpeername and getsockname functions, |
| 198 |
so we provide our own socketpair emulation that kind of works |
| 199 |
(AnyEvent::Util::portable_pipe). |
| 200 |
- new function: AnyEvent::Util::portable_socketpair. |
| 201 |
- take advantage of the Guard module if it exists. |
| 202 |
|
| 203 |
4.33 Fri Nov 21 02:35:40 CET 2008 |
| 204 |
- AnyEvent::Strict did errornously flag a fileno of 0 as |
| 205 |
illegal. |
| 206 |
- reduce memory usage and slightly speed up the pure perl backend |
| 207 |
by only storing the file descriptor, not the file handle. |
| 208 |
- add missing autocork method to AnyEvent::Handle (reported |
| 209 |
by Adam Rosenstein). |
| 210 |
- AnyEvent::DNS->resolve errornously documented an $rcode result |
| 211 |
argument, but there isn't (spotted by Henrik Krohns). |
| 212 |
- the naked truth about Tk - it's basically unmaintained. |
| 213 |
|
| 214 |
4.32 Mon Nov 3 22:46:32 CET 2008 |
| 215 |
- fix AnyEvent::Socket::resolve_sockaddr to properly support |
| 216 |
unix sockets again and choose a proper default for |
| 217 |
the socket type. |
| 218 |
- fix call to resolve_sockaddr in tcp_connect. |
| 219 |
|
| 220 |
4.31 Thu Oct 30 04:41:48 CET 2008 |
| 221 |
- implemented AnyEvent::Handle->destroy method. |
| 222 |
- hint about unexpected effects in TLS mode. |
| 223 |
- speed up AnyEvent::Impl::Perl by using more arrays |
| 224 |
instead of hashes without the slightest loss of readability :) |
| 225 |
- work around a perl argument refcounting bug. |
| 226 |
- with some perl patches applied, AnyEvent now seems leak-free. |
| 227 |
|
| 228 |
4.3 Fri Oct 3 09:18:43 CEST 2008 |
| 229 |
- AnyEvent will now install a no-op signal handler for SIGPIPE, |
| 230 |
unless one has been installed already. |
| 231 |
- warn about not loaidng AnyEvent::Impl::POE early enough |
| 232 |
(patch by Adam Rosenstein). |
| 233 |
- fixed a great number of bugs and corner cases in AnyEvent::Handle: |
| 234 |
- fix a bug in where in SSL connect mode, the |
| 235 |
client would first wait for some data by the server and |
| 236 |
otherwise hang. (reported and analyzed in an absolutely |
| 237 |
exemplary manner by Adam Rosenstein). |
| 238 |
- fix a bug in where SSL EOF would not be treated |
| 239 |
as stream EOF, putting the connection into a hung state. |
| 240 |
(reported and analyzed in an absolutely exemplary manner by Adam |
| 241 |
Rosenstein). |
| 242 |
- fix a potential segfault when the TLS context |
| 243 |
would go missing in a read callback (e.g. due to stoptls) - |
| 244 |
Net::SSLeay of course makes no type checking whatsoever. |
| 245 |
- AnyEvent::Handle will not stop the read watcher in TLS mode. |
| 246 |
- AnyEvent::Handle->stoptls will send a TLS close notify |
| 247 |
instead of simply destroying the stream now. |
| 248 |
- fix a bug where large blocks of data written in TLS mode |
| 249 |
would not be sent unless triggered by receives. |
| 250 |
- on_drain will now take the tls write buffer into account. |
| 251 |
- SSL operations have been streamlined a bit and should be faster |
| 252 |
now, more to come, though. |
| 253 |
- added a FAQ section, will document nontrivial issues in there. |
| 254 |
|
| 255 |
4.234 Mon Sep 29 04:08:13 CEST 2008 |
| 256 |
- fix child watcher documentation: only child exits will be reported, |
| 257 |
no trace events. |
| 258 |
- mention SIGPIPE in AnyEvent::Handle. |
| 259 |
- perl backend will now use POSIX::times as monotonic clock |
| 260 |
when available and the monotonic clock option is not. |
| 261 |
- run even without Time::HiRes in some cases now, but do not |
| 262 |
provide sub-second accuracy (all *supported* perl versions have |
| 263 |
Time::HiRes). |
| 264 |
|
| 265 |
4.233 Fri Aug 22 01:48:05 CEST 2008 |
| 266 |
- fix a bug in the testsuite, causing a hang. |
| 267 |
- clarified lots of AnyEvent::Handle constructor arguments. |
| 268 |
|
| 269 |
4.232 Thu Aug 21 20:44:25 CEST 2008 |
| 270 |
- fix a bug in AnyEvent::Handle that could cause two fatal error |
| 271 |
callbacks on EOF, which caused AnyEvent::HTTP to signal success twice |
| 272 |
(testcase by Leon Brocard). |
| 273 |
- clarified on_eof/on_error documentation a bit. |
| 274 |
|
| 275 |
4.231 Tue Jul 29 13:12:15 CEST 2008 |
| 276 |
- remove some debugging code left in AnyEvent::Util::fork_call |
| 277 |
(and no, it's impossible to implement with the broken windows |
| 278 |
perls without resource leaks or worse). |
| 279 |
|
| 280 |
4.23 Tue Jul 29 12:19:59 CEST 2008 |
| 281 |
- document the first parameter passed to condvar callbacks to be |
| 282 |
the callback. |
| 283 |
- add AnyEvent::Socket::{ntoa,aton} aliases. |
| 284 |
- optimize the AE::Handle->push_read (line) for the default |
| 285 |
eol marker. |
| 286 |
- optimize push_read (packstring|storable) for small packets. |
| 287 |
- invoke on_error callback when no on_eof callback is set. |
| 288 |
- fix a bug in push_read (storable) of unknown impact. |
| 289 |
|
| 290 |
4.22 Sun Jul 20 16:34:13 CEST 2008 |
| 291 |
- new function AnyEvent::Socket::parse_hostport. |
| 292 |
- as the bulkheads at microsoft can't even get getprotobyname reliably |
| 293 |
working on their shitty fucking broken os we need to hardcode |
| 294 |
some common protocol numbers in AnyEvent::Socket. How can |
| 295 |
people even bother with such a pile of shit as windows. |
| 296 |
|
| 297 |
4.21 Thu Jul 17 14:40:05 CEST 2008 |
| 298 |
- INCOMPATIBLE CHANGE: AnyEvent::Util::fork_call now has a prototype |
| 299 |
which is incompatible to previous invocation syntax. |
| 300 |
- work around more windows perl bugs in fork_call. I cannot imagine |
| 301 |
why anybody would bother implementing fork in such an obviously |
| 302 |
unusable way. |
| 303 |
- avoid calling $do_retry if the request has already finished in |
| 304 |
AnyEvent::DNS, thus avoiding an exception (reported by Anatoly |
| 305 |
K. Sharifulin). |
| 306 |
|
| 307 |
4.2 Sat Jul 12 22:42:11 CEST 2008 |
| 308 |
- support an "interval" argument for repeating timers. |
| 309 |
- fix a bug in the pure perl event loop that caused timers to |
| 310 |
be delayed under some circumstances. |
| 311 |
- implement optional argument checking via PERL_ANYEVENT_STRICT. |
| 312 |
- not importing anything from Event causes it to malfunction |
| 313 |
and not use Time::HiRes, work around this bug. |
| 314 |
- more bug workarounds against the endlessly broken Tk module, |
| 315 |
seems to work now on 32 bit machines, but less so on 64 bit. |
| 316 |
- minor optimisations applied to most backends. |
| 317 |
|
| 318 |
4.161 Fri Jul 4 14:24:48 CEST 2008 |
| 319 |
- fixed the t/handle/02_write.t, it read after write, with the |
| 320 |
assumption that all data could be written before read |
| 321 |
(reported by Jim Razmus). |
| 322 |
|
| 323 |
4.160 Thu Jul 3 04:02:21 CEST 2008 |
| 324 |
- re-registering signal watchers after unregistering one did |
| 325 |
not work if a backend falls back on AnyEvent's default |
| 326 |
implementation (also affected child watchers). |
| 327 |
- new AnyEvent::Handle options: autocork and no_delay. |
| 328 |
|
| 329 |
4.152 Sun Jun 22 14:15:44 CEST 2008 |
| 330 |
- allow for 32-bit perls that implement shifts differently |
| 331 |
on different architectures in parse_ipv4 (reported and |
| 332 |
analysed by Keiichi DAIBA). |
| 333 |
|
| 334 |
4.151 Fri Jun 6 17:34:24 CEST 2008 |
| 335 |
- make sure specifying _only_ on_read and never pushing reads |
| 336 |
works. |
| 337 |
|
| 338 |
4.15 Fri Jun 6 13:00:46 CEST 2008 |
| 339 |
- the pure perl backend would keep some watchers alive when more than |
| 340 |
one watcher was registered for the same fd. |
| 341 |
- new "packstring" and "storable" read and write types |
| 342 |
for AnyEvent::Handle. |
| 343 |
- allow on_eof handler to be called after on_error with EPIPE returns. |
| 344 |
- do not immediately call on_read callback in handle constructor. |
| 345 |
|
| 346 |
4.14 Thu Jun 5 20:29:31 CEST 2008 |
| 347 |
- Fixed a bug in DNS SRV priority sorting. |
| 348 |
- AnyEvent::Util::guard now reports runtime errors while |
| 349 |
executing the guard block as warnings. |
| 350 |
- handle 0-byte-reads just before EOF correctly in AnyEvent::Handle. |
| 351 |
|
| 352 |
4.13 Thu Jun 5 00:47:59 CEST 2008 |
| 353 |
- AnyEvent::DNS only followed cname chains with length 2, |
| 354 |
contrary to documentation. bumped it up to 10, thanks to |
| 355 |
microsoft, the current broken-dns-config-king. |
| 356 |
- AnyEvent::DNS didn't check for socket return status |
| 357 |
but instead relied on perl not creating filehandles in that |
| 358 |
case - too bad it gives you a bogus file handle (reported |
| 359 |
and analysed by Vladimir Timofeev). |
| 360 |
- fix queue management logic in AnyEvent::Handle: |
| 361 |
when on_read was registered, the queue was empty |
| 362 |
and no progress could be made AnyEvent::Handle would |
| 363 |
enter an endless loop. |
| 364 |
- correctly start reading again when the handle became |
| 365 |
busy again after idling. |
| 366 |
- correctly treat tls shut-downs as EOF condition |
| 367 |
(for the time being, would be nice to have a callback for that). |
| 368 |
- correctly call eof callback under all conditions |
| 369 |
(hopefully). |
| 370 |
- the timeout callback did not expect that $self can go |
| 371 |
away any time. |
| 372 |
|
| 373 |
4.12 Tue Jun 3 10:58:04 CEST 2008 |
| 374 |
- include AnyEvent::Intro, a tutorial for anyevent, |
| 375 |
anyevent::socket and anyevent::handle. |
| 376 |
- allow more options in on_error. |
| 377 |
|
| 378 |
4.11 Fri May 30 23:42:25 CEST 2008 |
| 379 |
- INCOMPATIBLE CHANGE: replace ptr by real PTR lookup, provide |
| 380 |
reverse_lookup and reverse_verify to replace it, support |
| 381 |
v4mapped and v4compat addresses. |
| 382 |
- provide more documentation for the resolver class. |
| 383 |
- really replace longest run of :0: by :: in format_address, |
| 384 |
also properly convert :: and ::1 again. |
| 385 |
- support NAPTR record name and decode it. |
| 386 |
- implement random weight sampling for SRV records, as per |
| 387 |
rfc 2782. |
| 388 |
- correctly abort on srv-record targets of ".". |
| 389 |
- added AnyEvent::DNS::wait_for_slot. |
| 390 |
- in the unlikely event of a virtual circuit connection |
| 391 |
being invalidated by a delayed udp reply, AnyEvent::DNS |
| 392 |
could die. this has been fixed. |
| 393 |
- plug a probable memleak in the DNS vc code. |
| 394 |
- use configured timeout also for VC request phase. |
| 395 |
- implement timeout and max_outstanding methods |
| 396 |
for resolver class. |
| 397 |
- update version numbers in all modules. |
| 398 |
|
| 399 |
4.1 Thu May 29 05:45:40 CEST 2008 |
| 400 |
- INCOMPATIBLE CHANGE: renamed xxx_ip functions to xxx_address |
| 401 |
which mirrors their purpose better (old names still available |
| 402 |
for a while). Also moved AnyEvent::DNS::addr to |
| 403 |
AnyEvent::Socket::resolve_sockaddr. |
| 404 |
- implement AnyEvent->time and AnyEvent->now. |
| 405 |
- fix IPv6 support in pack_sockaddr. |
| 406 |
- officially un-experimentalise all newly introduced code. |
| 407 |
- support unix domain sockets everywhere by specifying |
| 408 |
a host of "unix/" and the pathname as service. |
| 409 |
- implement an activity timeout in AnyEvent::Handle. |
| 410 |
- added a regex read type to AnyEvent::Handle. |
| 411 |
- added a json read/write type to AnyEvent::Handle. |
| 412 |
- always croak in on_error in AnyEvent::Handle. |
| 413 |
- document how to subclass AnyEvent::Handle. |
| 414 |
- implement AnyEvent::Util::fork_call. |
| 415 |
- add support for IPv6 nameservers and nameserver statements. |
| 416 |
- work around _yet_ _another_ windows perl bug in where empty |
| 417 |
select masks cause errors in the pure perl backend (this |
| 418 |
workaround cannot be implemented for the other event loops, |
| 419 |
but EV also works around this). |
| 420 |
- supply AnyEvent::Socket::address_type and make good use of it. |
| 421 |
- clarify and fix the Handle documentation w.r.t $self vs. $handle |
| 422 |
vs. callback arguments. |
| 423 |
- add some recommends to the META.yml. |
| 424 |
|
| 425 |
4.05 Mon May 26 19:44:06 CEST 2008 |
| 426 |
- some platforms ignore the rfc and prepend an extra sa_len member to |
| 427 |
the sockaddr structure, cater for those. also use sockaddr_family, |
| 428 |
didn't know it was there, but comes in handy for the workaround. |
| 429 |
- undo the import WIN32 hack. |
| 430 |
|
| 431 |
4.04 Mon May 26 08:03:31 CEST 2008 |
| 432 |
- try to work around yet another windows bug: failed connects are reported |
| 433 |
as if it were out-of-band data. windows users: you suck. Only |
| 434 |
EV, Glib, Event and the pure perl backend can handle this condition. |
| 435 |
- optimize write algorithm in AnyEvent::Handle. |
| 436 |
- properly parse PERL_ANYEVENT_PROTOCOLS and use the info correctly. |
| 437 |
- tcp_connect now properly iterates through all targets. |
| 438 |
- check wether IPv6 sockets can actually be created, otherwise, disable ipv6 |
| 439 |
support entirely. also hardcode some AF_INET6 constants for the |
| 440 |
sake of perl 5.8. |
| 441 |
- vastly improved nameserver/suffix detection algorithm for windows |
| 442 |
(still a hack...) |
| 443 |
- try to use a fallback dns server if no dns servers could be found. |
| 444 |
- splatter around some local $SIG{__DIE__}'s to work around bad |
| 445 |
die handlers. |
| 446 |
|
| 447 |
4.03 Sun May 25 05:05:57 CEST 2008 |
| 448 |
- work around perl bugs on windows, where perl returns undocumented |
| 449 |
error codes for sysread, syswrite etc. |
| 450 |
- fix AnyEvent::Handle::unshift_read to really unshift and not push. |
| 451 |
- AnyEvent::Handle could sometimes stumble over 0-byte writes. |
| 452 |
- fix unpack_sockaddr, now works for AF_INET6 sockaddr's. |
| 453 |
- updated documentation. |
| 454 |
- lots of workarounds for perl 5.8 quirks. |
| 455 |
- implement a more flexible type system for both reads and writes |
| 456 |
in AnyEvent::Handle. |
| 457 |
- support IPv6 in tcp_server. |
| 458 |
- pass local hostname and port to the prepare callback in tcp_server. |
| 459 |
- make types extendable in AnyEvent::Handle. |
| 460 |
- croak, not die, in AnyEvent::Handle; |
| 461 |
|
| 462 |
4.0 Sat May 24 19:58:08 CEST 2008 |
| 463 |
- added (experimental) AnyEvent::DNS module. |
| 464 |
- added (experimental) AnyEvent::Socket module. |
| 465 |
- added (experiemntal) support for TLS/SSL in AnyEvent::Handle. |
| 466 |
- make condvar->end call ->send as documented. |
| 467 |
- make condvar constructor support the documented cb => argument. |
| 468 |
- overload condvars so you can use them instead of code references. |
| 469 |
- support more of the POSIX forms instead of a pure dotted quad for |
| 470 |
verbatim ip addresses. |
| 471 |
- AnyEvent::Socket::inet_aton will use AnyEvent::DNS now and not fork anymore. |
| 472 |
- try very hard not to freeze on broken windows perls. |
| 473 |
- make on_eof optional in AnyEvent::Handle. |
| 474 |
- support service names in tcp_connect and tcp_server. |
| 475 |
- work around netbsd/openbsd bug causing unix sockets not to |
| 476 |
become writable until empty. |
| 477 |
- only make a dummy read for error detection when getpeername |
| 478 |
returns ENOTCONN. |
| 479 |
- defend AnyEvent::Util::inet_aton against Coro::LWP's brutal |
| 480 |
override. |
| 481 |
- new AnyEvent::Guard::cancel. |
| 482 |
|
| 483 |
3.5 Sat May 17 23:17:14 CEST 2008 |
| 484 |
- gracefully deal with systems not defining CLOCK_MONOTONIC |
| 485 |
(as opposed to not supporting the clock itself, which worked in 3.41). |
| 486 |
- added AnyEvent::Util::tcp_connect and AnyEvent::Util::tcp_server |
| 487 |
helpers, to replace the ill-designed AnyEvent::Socket class. |
| 488 |
- added some examples (eg/connect, eg/handle). |
| 489 |
- added AnyEvent::Util::guard. |
| 490 |
- AnyEvent::Util now exports most of its functions. |
| 491 |
- fixed a bug in *_read_chunk, where the callback was called with $self as |
| 492 |
first and second argument. |
| 493 |
- simplified and fixed regex handling in AnyEvent::Handle::*_read_line. |
| 494 |
- fix manpage condvars examples to use new syntax. |
| 495 |
- considerably improve inet_aton with EV::ADNS backend to support |
| 496 |
cname chains (common dns configuration bug, e.g. with www.google.de) |
| 497 |
and aaaa records if no a record could be found. |
| 498 |
|
| 499 |
3.41 Sun May 11 19:53:13 CEST 2008 |
| 500 |
- work around a bug in perl 5.8.8's local. Reported by Yi Ma Mao. |
| 501 |
|
| 502 |
3.4 Sun May 11 00:29:25 CEST 2008 |
| 503 |
- complete re-implementation of condvars with a hopefully much more useful |
| 504 |
API (backward compatible functions exist but are undocumented). |
| 505 |
- AnyEvent::Handle fully rewritten with a hopefully more powerful API |
| 506 |
(still subject to change). |
| 507 |
- take advantage of CLOCK_MONOTONIC, if available, in AnyEvent::Impl::Perl. |
| 508 |
- provide AnyEvent::post_detect and @AnyEvent::post_detect, which |
| 509 |
allows module authors to avoid forcing event loop detection. |
| 510 |
(used by Coro::AIO, Coro::BDB, Coro::AnyEvent for example). |
| 511 |
- remove coro backends: Coro now provides generic support for AnyEvent |
| 512 |
via Coro::AnyEvent. |
| 513 |
|
| 514 |
3.3 Mon Apr 28 09:51:06 CEST 2008 |
| 515 |
- added AnyEvent::Handle, AnyEvent::Socket and AnyEvent::Util modules. |
| 516 |
- fix a bug in the pure perl backend that kept watchers alive |
| 517 |
when multiple watchers were registered for the same fh/poll combo. |
| 518 |
- add a benchmark section showing AnyEvent overhead and comparing |
| 519 |
the different event loops with each other. |
| 520 |
- prefer pure perl over tk when autoprobing, as it's about as fast, |
| 521 |
but doesn't crash with many watchers. |
| 522 |
- declare Qt support non-experimental. |
| 523 |
- clarify the confusing section about the file descriptor being |
| 524 |
kept alive. |
| 525 |
- document the race between loading of an event module and |
| 526 |
child processes exiting. |
| 527 |
- support POE as "backend" (with some caveats, POE is not generic |
| 528 |
enough, and darn slow). |
| 529 |
- support Wx and Prima through POE. |
| 530 |
- optimise perl backend to use 20% less memory and take advantage |
| 531 |
of typical timeout behaviour. It can now compete with |
| 532 |
select/poll-based C event loops in most cases (it is usually |
| 533 |
faster than Event and Glib :). |
| 534 |
- roughly cut EV memory use in half and increase its speed by 30%, |
| 535 |
by removing undocumented functionality. Did something similar |
| 536 |
to other event loops where possible. |
| 537 |
|
| 538 |
3.2 Thu Apr 24 10:10:40 CEST 2008 |
| 539 |
- do not die when anyevent watchers are destroyed while running callbacks |
| 540 |
in the pureperl backend (could only happen when two watchers are registered |
| 541 |
for the same fh/poll combo). |
| 542 |
- support autoloading for child watchers, was broken in all versions. |
| 543 |
- implement PERL_ANYEVENT_MODEL env variable. |
| 544 |
- (experimental) implement interface to the (very crashy/buggy) Event::Lib module. |
| 545 |
- (experimental) implement interface to the Qt module (cannot be autoprobed). |
| 546 |
- this release is IO:AnyEvent-proof. |
| 547 |
|
| 548 |
3.12 Tue Apr 22 07:11:46 CEST 2008 |
| 549 |
- reinstate AUTHOR section that got lost somehow. |
| 550 |
- do not hang in the testsuite with badly broken perls (activestate, |
| 551 |
strawberry...), but instead diagnose the problem and continue. |
| 552 |
- use INT instead of CHLD in an attempt to support broken windows perls |
| 553 |
better (this decreases the test quality, unfortunately). |
| 554 |
- do not send the signal to the process group |
| 555 |
(no problem for CHLD, bad for INT :). |
| 556 |
|
| 557 |
3.11 Sat Apr 19 06:57:31 CEST 2008 |
| 558 |
- major documentation rework. |
| 559 |
- document the fact that child watchers only watch for zombies. |
| 560 |
- fix the child watcher example. |
| 561 |
|
| 562 |
3.1 Wed Apr 16 17:09:01 CEST 2008 |
| 563 |
- work around recurring bugs in Tk by dup'ing filehandles, the |
| 564 |
only method with good success chances on Tk (the bugs apparently |
| 565 |
don't get fixed anytime soon). |
| 566 |
- lift the restriction of only one watcher per fh direction |
| 567 |
(as the Tk bug workaround also lifts it and only Tk imposed |
| 568 |
such strong limits). |
| 569 |
- changed probe order to prefer coro adaptors. |
| 570 |
- explain why recursion into the event loop is not supported |
| 571 |
unless the backend supports it (only Coro::EV does without |
| 572 |
any restrictions...). |
| 573 |
- add simple manpages for all backend modules. |
| 574 |
|
| 575 |
3.0 Mon Apr 7 21:30:23 CEST 2008 |
| 576 |
- Coro::Signal changed semantics, roll our own, also cleaning |
| 577 |
up the Coro implementation in general. |
| 578 |
- rename Coro backend to CoroEvent. |
| 579 |
- add some decision helping paragraph to the manpage that should |
| 580 |
help people to decide whether AnyEvent is the right thing for them. |
| 581 |
|
| 582 |
2.9 Mon Jan 28 13:31:54 CET 2008 |
| 583 |
- update for EV 3.0 API changes. |
| 584 |
|
| 585 |
2.8 Sun Nov 25 15:06:03 CET 2007 |
| 586 |
- waitpid can validly return 0. accept this fact of life |
| 587 |
instead of reporting it to any watchers. |
| 588 |
|
| 589 |
2.7 Fri Nov 23 11:41:14 CET 2007 |
| 590 |
- force use of AnyEvent::Impl::Perl in testsuite, there is |
| 591 |
too much breakage outside AnyEvent. |
| 592 |
- deliver signals synchronously in AnyEvent::Impl::Perl |
| 593 |
or any other event loop that relies on AnyEvents child |
| 594 |
watcher emulation. *Could* help with hanging testsuite |
| 595 |
(except when Event or EV are installed). |
| 596 |
|
| 597 |
2.6 Fri Nov 9 20:36:35 CET 2007 |
| 598 |
- fix bug in testsuite. |
| 599 |
- move EV adaptor modules to AnyEvent. |
| 600 |
- add Coro+EV adaptor module. |
| 601 |
|
| 602 |
2.55 Tue Nov 6 17:41:32 CET 2007 |
| 603 |
- add EV to the list of supported event models. |
| 604 |
- do not auto-reset pid watchers, pass pid and status to them. |
| 605 |
- allow a pid of zero to watch for all children in child watchers. |
| 606 |
|
| 607 |
2.54 Wed Jul 18 17:36:23 CEST 2007 |
| 608 |
- work around a perl bug that results in |
| 609 |
BEGIN not safe after errors--compilation aborted without |
| 610 |
any discernible reason or error message by once |
| 611 |
again not use'ing strict. |
| 612 |
|
| 613 |
2.53 Sun Jul 8 10:51:53 CEST 2007 |
| 614 |
- make Glib interface work again (spotted by elmex). |
| 615 |
|
| 616 |
2.52 Wed Mar 7 18:36:16 CET 2007 |
| 617 |
- child watchers stopped working permanently when |
| 618 |
all current child watchers were destroyed. |
| 619 |
|
| 620 |
2.51 Mon Dec 11 21:33:24 CET 2006 |
| 621 |
- work around bugs in perl where eval "require Module" returns |
| 622 |
true even if the module couldn't be loaded :/. |
| 623 |
|
| 624 |
2.5 Mon Dec 11 02:15:28 CET 2006 |
| 625 |
- avoid $AUTOLOAD because many perls corrupt it. |
| 626 |
- AnyEvent::detect forces autodetection. |
| 627 |
- implement signal watchers (experimental) |
| 628 |
(Tk does not support async signals (<= 804.027 at least), |
| 629 |
so they might get delayed indefinitely on Tk). |
| 630 |
- implement child watchers (experimental). |
| 631 |
- moved default condvar implementation into base module, simplifying |
| 632 |
most implementation modules. |
| 633 |
|
| 634 |
2.1 Fri Nov 24 15:50:48 CET 2006 |
| 635 |
- better docs. |
| 636 |
- simple testsuite. |
| 637 |
- added AnyEvent->one_event method for special purposes. |
| 638 |
|
| 639 |
2.0 Wed Nov 1 02:21:30 CET 2006 |
| 640 |
- INCOMPATIBLE CHANGE: poll can now either be "r" or "w" but not both. |
| 641 |
- INCOMPATIBLE CHANGE: io watcher callbacks have no arguments anymore. |
| 642 |
- EXPERIMENTAL: |
| 643 |
new pure-perl model added, AnyEvent now always finds a suitable |
| 644 |
event model. |
| 645 |
- improved documentation. |
| 646 |
- reduced cpu overhead. |
| 647 |
|
| 648 |
1.02 Fri Jan 13 14:15:40 CET 2006 |
| 649 |
- add COPYING to clarify license. |
| 650 |
|
| 651 |
1.01 Sun Jan 8 05:48:09 CET 2006 |
| 652 |
- correctly cancel Event timers, avoid memory leaks. |
| 653 |
|
| 654 |
1.0 Sun Jan 8 05:40:58 CET 2006 |
| 655 |
- allow third-party interfaces to be added at runtime. |
| 656 |
- fix version number. |
| 657 |
|
| 658 |
0.9 Fri Dec 30 02:25:41 CET 2005 |
| 659 |
- fixed errors in the documentation. |
| 660 |
- going back to using $VERSION as "load check", |
| 661 |
as suggested by Jörn Reder. Let's see why I stopped using |
| 662 |
this method. |
| 663 |
- add 'hup' to the list of events to receive for glib |
| 664 |
and generate 'rw' events for it, as some versions of |
| 665 |
glib seem to require this, while this breaks other |
| 666 |
(older) versions of glib. |
| 667 |
- implemented PERL_ANYEVENT_VERBOSE env variable. |
| 668 |
|
| 669 |
0.3 Sun Dec 4 10:43:33 CET 2005 |
| 670 |
- no changes to the code. |
| 671 |
- fixed documentation. |
| 672 |
- clarified some corner cases. |
| 673 |
- vastly improved documentation with a long example. |
| 674 |
|
| 675 |
0.2 Thu Dec 1 22:18:49 CET 2005 |
| 676 |
- check whether a valid method was called, to avoid endless loops. |
| 677 |
- move condvars into their own class (cleaner, safer). |
| 678 |
- don't die when some modules are not available. |
| 679 |
- Tk is too broken to allow for words. Will work once Tk works. |
| 680 |
|
| 681 |
0.1 Thu Dec 1 19:44:57 CET 2005 |
| 682 |
- spiced up and cleaned up for release. |
| 683 |
|
| 684 |
0.01 Tue Sep 21 18:12:09 CEST 2004 |
| 685 |
- original version; copied from Convert::Scalar. |
| 686 |
|
