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

Comparing AnyEvent/lib/AnyEvent.pm (file contents):
Revision 1.193 by root, Thu Feb 12 17:33:37 2009 UTC vs.
Revision 1.308 by root, Fri Dec 25 07:39:41 2009 UTC

1=head1 NAME 1=head1 NAME
2 2
3AnyEvent - provide framework for multiple event loops 3AnyEvent - the DBI of event loop programming
4 4
5EV, Event, Glib, Tk, Perl, Event::Lib, Qt, POE - various supported event loops 5EV, Event, Glib, Tk, Perl, Event::Lib, Irssi, rxvt-unicode, IO::Async, Qt
6and POE are various supported event loops/environments.
6 7
7=head1 SYNOPSIS 8=head1 SYNOPSIS
8 9
9 use AnyEvent; 10 use AnyEvent;
10 11
12 # file descriptor readable
11 my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub { ... }); 13 my $w = AnyEvent->io (fh => $fh, poll => "r", cb => sub { ... });
12 14
15 # one-shot or repeating timers
13 my $w = AnyEvent->timer (after => $seconds, cb => sub { ... }); 16 my $w = AnyEvent->timer (after => $seconds, cb => sub { ... });
14 my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ... 17 my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ...
15 18
16 print AnyEvent->now; # prints current event loop time 19 print AnyEvent->now; # prints current event loop time
17 print AnyEvent->time; # think Time::HiRes::time or simply CORE::time. 20 print AnyEvent->time; # think Time::HiRes::time or simply CORE::time.
18 21
22 # POSIX signal
19 my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... }); 23 my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... });
20 24
25 # child process exit
21 my $w = AnyEvent->child (pid => $pid, cb => sub { 26 my $w = AnyEvent->child (pid => $pid, cb => sub {
22 my ($pid, $status) = @_; 27 my ($pid, $status) = @_;
23 ... 28 ...
24 }); 29 });
30
31 # called when event loop idle (if applicable)
32 my $w = AnyEvent->idle (cb => sub { ... });
25 33
26 my $w = AnyEvent->condvar; # stores whether a condition was flagged 34 my $w = AnyEvent->condvar; # stores whether a condition was flagged
27 $w->send; # wake up current and all future recv's 35 $w->send; # wake up current and all future recv's
28 $w->recv; # enters "main loop" till $condvar gets ->send 36 $w->recv; # enters "main loop" till $condvar gets ->send
29 # use a condvar in callback mode: 37 # use a condvar in callback mode:
32=head1 INTRODUCTION/TUTORIAL 40=head1 INTRODUCTION/TUTORIAL
33 41
34This manpage is mainly a reference manual. If you are interested 42This manpage is mainly a reference manual. If you are interested
35in a tutorial or some gentle introduction, have a look at the 43in a tutorial or some gentle introduction, have a look at the
36L<AnyEvent::Intro> manpage. 44L<AnyEvent::Intro> manpage.
45
46=head1 SUPPORT
47
48There is a mailinglist for discussing all things AnyEvent, and an IRC
49channel, too.
50
51See the AnyEvent project page at the B<Schmorpforge Ta-Sa Software
52Repository>, at L<http://anyevent.schmorp.de>, for more info.
37 53
38=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT) 54=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)
39 55
40Glib, POE, IO::Async, Event... CPAN offers event models by the dozen 56Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
41nowadays. So what is different about AnyEvent? 57nowadays. So what is different about AnyEvent?
137These watchers are normal Perl objects with normal Perl lifetime. After 153These watchers are normal Perl objects with normal Perl lifetime. After
138creating a watcher it will immediately "watch" for events and invoke the 154creating a watcher it will immediately "watch" for events and invoke the
139callback when the event occurs (of course, only when the event model 155callback when the event occurs (of course, only when the event model
140is in control). 156is in control).
141 157
158Note that B<callbacks must not permanently change global variables>
159potentially in use by the event loop (such as C<$_> or C<$[>) and that B<<
160callbacks must not C<die> >>. The former is good programming practise in
161Perl and the latter stems from the fact that exception handling differs
162widely between event loops.
163
142To disable the watcher you have to destroy it (e.g. by setting the 164To disable the watcher you have to destroy it (e.g. by setting the
143variable you store it in to C<undef> or otherwise deleting all references 165variable you store it in to C<undef> or otherwise deleting all references
144to it). 166to it).
145 167
146All watchers are created by calling a method on the C<AnyEvent> class. 168All watchers are created by calling a method on the C<AnyEvent> class.
159my variables are only visible after the statement in which they are 181my variables are only visible after the statement in which they are
160declared. 182declared.
161 183
162=head2 I/O WATCHERS 184=head2 I/O WATCHERS
163 185
186 $w = AnyEvent->io (
187 fh => <filehandle_or_fileno>,
188 poll => <"r" or "w">,
189 cb => <callback>,
190 );
191
164You can create an I/O watcher by calling the C<< AnyEvent->io >> method 192You can create an I/O watcher by calling the C<< AnyEvent->io >> method
165with the following mandatory key-value pairs as arguments: 193with the following mandatory key-value pairs as arguments:
166 194
167C<fh> the Perl I<file handle> (I<not> file descriptor) to watch for events 195C<fh> is the Perl I<file handle> (or a naked file descriptor) to watch
168(AnyEvent might or might not keep a reference to this file handle). C<poll> 196for events (AnyEvent might or might not keep a reference to this file
197handle). Note that only file handles pointing to things for which
198non-blocking operation makes sense are allowed. This includes sockets,
199most character devices, pipes, fifos and so on, but not for example files
200or block devices.
201
169must be a string that is either C<r> or C<w>, which creates a watcher 202C<poll> must be a string that is either C<r> or C<w>, which creates a
170waiting for "r"eadable or "w"ritable events, respectively. C<cb> is the 203watcher waiting for "r"eadable or "w"ritable events, respectively.
204
171callback to invoke each time the file handle becomes ready. 205C<cb> is the callback to invoke each time the file handle becomes ready.
172 206
173Although the callback might get passed parameters, their value and 207Although the callback might get passed parameters, their value and
174presence is undefined and you cannot rely on them. Portable AnyEvent 208presence is undefined and you cannot rely on them. Portable AnyEvent
175callbacks cannot use arguments passed to I/O watcher callbacks. 209callbacks cannot use arguments passed to I/O watcher callbacks.
176 210
191 undef $w; 225 undef $w;
192 }); 226 });
193 227
194=head2 TIME WATCHERS 228=head2 TIME WATCHERS
195 229
230 $w = AnyEvent->timer (after => <seconds>, cb => <callback>);
231
232 $w = AnyEvent->timer (
233 after => <fractional_seconds>,
234 interval => <fractional_seconds>,
235 cb => <callback>,
236 );
237
196You can create a time watcher by calling the C<< AnyEvent->timer >> 238You can create a time watcher by calling the C<< AnyEvent->timer >>
197method with the following mandatory arguments: 239method with the following mandatory arguments:
198 240
199C<after> specifies after how many seconds (fractional values are 241C<after> specifies after how many seconds (fractional values are
200supported) the callback should be invoked. C<cb> is the callback to invoke 242supported) the callback should be invoked. C<cb> is the callback to invoke
308In either case, if you care (and in most cases, you don't), then you 350In either case, if you care (and in most cases, you don't), then you
309can get whatever behaviour you want with any event loop, by taking the 351can get whatever behaviour you want with any event loop, by taking the
310difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into 352difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into
311account. 353account.
312 354
355=item AnyEvent->now_update
356
357Some event loops (such as L<EV> or L<AnyEvent::Impl::Perl>) cache
358the current time for each loop iteration (see the discussion of L<<
359AnyEvent->now >>, above).
360
361When a callback runs for a long time (or when the process sleeps), then
362this "current" time will differ substantially from the real time, which
363might affect timers and time-outs.
364
365When this is the case, you can call this method, which will update the
366event loop's idea of "current time".
367
368A typical example would be a script in a web server (e.g. C<mod_perl>) -
369when mod_perl executes the script, then the event loop will have the wrong
370idea about the "current time" (being potentially far in the past, when the
371script ran the last time). In that case you should arrange a call to C<<
372AnyEvent->now_update >> each time the web server process wakes up again
373(e.g. at the start of your script, or in a handler).
374
375Note that updating the time I<might> cause some events to be handled.
376
313=back 377=back
314 378
315=head2 SIGNAL WATCHERS 379=head2 SIGNAL WATCHERS
380
381 $w = AnyEvent->signal (signal => <uppercase_signal_name>, cb => <callback>);
316 382
317You can watch for signals using a signal watcher, C<signal> is the signal 383You can watch for signals using a signal watcher, C<signal> is the signal
318I<name> in uppercase and without any C<SIG> prefix, C<cb> is the Perl 384I<name> in uppercase and without any C<SIG> prefix, C<cb> is the Perl
319callback to be invoked whenever a signal occurs. 385callback to be invoked whenever a signal occurs.
320 386
326invocation, and callback invocation will be synchronous. Synchronous means 392invocation, and callback invocation will be synchronous. Synchronous means
327that it might take a while until the signal gets handled by the process, 393that it might take a while until the signal gets handled by the process,
328but it is guaranteed not to interrupt any other callbacks. 394but it is guaranteed not to interrupt any other callbacks.
329 395
330The main advantage of using these watchers is that you can share a signal 396The main advantage of using these watchers is that you can share a signal
331between multiple watchers. 397between multiple watchers, and AnyEvent will ensure that signals will not
398interrupt your program at bad times.
332 399
333This watcher might use C<%SIG>, so programs overwriting those signals 400This watcher might use C<%SIG> (depending on the event loop used),
334directly will likely not work correctly. 401so programs overwriting those signals directly will likely not work
402correctly.
335 403
336Example: exit on SIGINT 404Example: exit on SIGINT
337 405
338 my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 }); 406 my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
339 407
408=head3 Restart Behaviour
409
410While restart behaviour is up to the event loop implementation, most will
411not restart syscalls (that includes L<Async::Interrupt> and AnyEvent's
412pure perl implementation).
413
414=head3 Safe/Unsafe Signals
415
416Perl signals can be either "safe" (synchronous to opcode handling) or
417"unsafe" (asynchronous) - the former might get delayed indefinitely, the
418latter might corrupt your memory.
419
420AnyEvent signal handlers are, in addition, synchronous to the event loop,
421i.e. they will not interrupt your running perl program but will only be
422called as part of the normal event handling (just like timer, I/O etc.
423callbacks, too).
424
425=head3 Signal Races, Delays and Workarounds
426
427Many event loops (e.g. Glib, Tk, Qt, IO::Async) do not support attaching
428callbacks to signals in a generic way, which is a pity, as you cannot
429do race-free signal handling in perl, requiring C libraries for
430this. AnyEvent will try to do it's best, which means in some cases,
431signals will be delayed. The maximum time a signal might be delayed is
432specified in C<$AnyEvent::MAX_SIGNAL_LATENCY> (default: 10 seconds). This
433variable can be changed only before the first signal watcher is created,
434and should be left alone otherwise. This variable determines how often
435AnyEvent polls for signals (in case a wake-up was missed). Higher values
436will cause fewer spurious wake-ups, which is better for power and CPU
437saving.
438
439All these problems can be avoided by installing the optional
440L<Async::Interrupt> module, which works with most event loops. It will not
441work with inherently broken event loops such as L<Event> or L<Event::Lib>
442(and not with L<POE> currently, as POE does it's own workaround with
443one-second latency). For those, you just have to suffer the delays.
444
340=head2 CHILD PROCESS WATCHERS 445=head2 CHILD PROCESS WATCHERS
341 446
447 $w = AnyEvent->child (pid => <process id>, cb => <callback>);
448
342You can also watch on a child process exit and catch its exit status. 449You can also watch on a child process exit and catch its exit status.
343 450
344The child process is specified by the C<pid> argument (if set to C<0>, it 451The child process is specified by the C<pid> argument (one some backends,
345watches for any child process exit). The watcher will triggered only when 452using C<0> watches for any child process exit, on others this will
346the child process has finished and an exit status is available, not on 453croak). The watcher will be triggered only when the child process has
347any trace events (stopped/continued). 454finished and an exit status is available, not on any trace events
455(stopped/continued).
348 456
349The callback will be called with the pid and exit status (as returned by 457The callback will be called with the pid and exit status (as returned by
350waitpid), so unlike other watcher types, you I<can> rely on child watcher 458waitpid), so unlike other watcher types, you I<can> rely on child watcher
351callback arguments. 459callback arguments.
352 460
357 465
358There is a slight catch to child watchers, however: you usually start them 466There is a slight catch to child watchers, however: you usually start them
359I<after> the child process was created, and this means the process could 467I<after> the child process was created, and this means the process could
360have exited already (and no SIGCHLD will be sent anymore). 468have exited already (and no SIGCHLD will be sent anymore).
361 469
362Not all event models handle this correctly (POE doesn't), but even for 470Not all event models handle this correctly (neither POE nor IO::Async do,
471see their AnyEvent::Impl manpages for details), but even for event models
363event models that I<do> handle this correctly, they usually need to be 472that I<do> handle this correctly, they usually need to be loaded before
364loaded before the process exits (i.e. before you fork in the first place). 473the process exits (i.e. before you fork in the first place). AnyEvent's
474pure perl event loop handles all cases correctly regardless of when you
475start the watcher.
365 476
366This means you cannot create a child watcher as the very first thing in an 477This means you cannot create a child watcher as the very first
367AnyEvent program, you I<have> to create at least one watcher before you 478thing in an AnyEvent program, you I<have> to create at least one
368C<fork> the child (alternatively, you can call C<AnyEvent::detect>). 479watcher before you C<fork> the child (alternatively, you can call
480C<AnyEvent::detect>).
481
482As most event loops do not support waiting for child events, they will be
483emulated by AnyEvent in most cases, in which the latency and race problems
484mentioned in the description of signal watchers apply.
369 485
370Example: fork a process and wait for it 486Example: fork a process and wait for it
371 487
372 my $done = AnyEvent->condvar; 488 my $done = AnyEvent->condvar;
373 489
383 ); 499 );
384 500
385 # do something else, then wait for process exit 501 # do something else, then wait for process exit
386 $done->recv; 502 $done->recv;
387 503
504=head2 IDLE WATCHERS
505
506 $w = AnyEvent->idle (cb => <callback>);
507
508Sometimes there is a need to do something, but it is not so important
509to do it instantly, but only when there is nothing better to do. This
510"nothing better to do" is usually defined to be "no other events need
511attention by the event loop".
512
513Idle watchers ideally get invoked when the event loop has nothing
514better to do, just before it would block the process to wait for new
515events. Instead of blocking, the idle watcher is invoked.
516
517Most event loops unfortunately do not really support idle watchers (only
518EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent
519will simply call the callback "from time to time".
520
521Example: read lines from STDIN, but only process them when the
522program is otherwise idle:
523
524 my @lines; # read data
525 my $idle_w;
526 my $io_w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
527 push @lines, scalar <STDIN>;
528
529 # start an idle watcher, if not already done
530 $idle_w ||= AnyEvent->idle (cb => sub {
531 # handle only one line, when there are lines left
532 if (my $line = shift @lines) {
533 print "handled when idle: $line";
534 } else {
535 # otherwise disable the idle watcher again
536 undef $idle_w;
537 }
538 });
539 });
540
388=head2 CONDITION VARIABLES 541=head2 CONDITION VARIABLES
542
543 $cv = AnyEvent->condvar;
544
545 $cv->send (<list>);
546 my @res = $cv->recv;
389 547
390If you are familiar with some event loops you will know that all of them 548If you are familiar with some event loops you will know that all of them
391require you to run some blocking "loop", "run" or similar function that 549require you to run some blocking "loop", "run" or similar function that
392will actively watch for new events and call your callbacks. 550will actively watch for new events and call your callbacks.
393 551
394AnyEvent is different, it expects somebody else to run the event loop and 552AnyEvent is slightly different: it expects somebody else to run the event
395will only block when necessary (usually when told by the user). 553loop and will only block when necessary (usually when told by the user).
396 554
397The instrument to do that is called a "condition variable", so called 555The instrument to do that is called a "condition variable", so called
398because they represent a condition that must become true. 556because they represent a condition that must become true.
399 557
558Now is probably a good time to look at the examples further below.
559
400Condition variables can be created by calling the C<< AnyEvent->condvar 560Condition variables can be created by calling the C<< AnyEvent->condvar
401>> method, usually without arguments. The only argument pair allowed is 561>> method, usually without arguments. The only argument pair allowed is
402
403C<cb>, which specifies a callback to be called when the condition variable 562C<cb>, which specifies a callback to be called when the condition variable
404becomes true, with the condition variable as the first argument (but not 563becomes true, with the condition variable as the first argument (but not
405the results). 564the results).
406 565
407After creation, the condition variable is "false" until it becomes "true" 566After creation, the condition variable is "false" until it becomes "true"
412Condition variables are similar to callbacks, except that you can 571Condition variables are similar to callbacks, except that you can
413optionally wait for them. They can also be called merge points - points 572optionally wait for them. They can also be called merge points - points
414in time where multiple outstanding events have been processed. And yet 573in time where multiple outstanding events have been processed. And yet
415another way to call them is transactions - each condition variable can be 574another way to call them is transactions - each condition variable can be
416used to represent a transaction, which finishes at some point and delivers 575used to represent a transaction, which finishes at some point and delivers
417a result. 576a result. And yet some people know them as "futures" - a promise to
577compute/deliver something that you can wait for.
418 578
419Condition variables are very useful to signal that something has finished, 579Condition variables are very useful to signal that something has finished,
420for example, if you write a module that does asynchronous http requests, 580for example, if you write a module that does asynchronous http requests,
421then a condition variable would be the ideal candidate to signal the 581then a condition variable would be the ideal candidate to signal the
422availability of results. The user can either act when the callback is 582availability of results. The user can either act when the callback is
456 after => 1, 616 after => 1,
457 cb => sub { $result_ready->send }, 617 cb => sub { $result_ready->send },
458 ); 618 );
459 619
460 # this "blocks" (while handling events) till the callback 620 # this "blocks" (while handling events) till the callback
461 # calls send 621 # calls ->send
462 $result_ready->recv; 622 $result_ready->recv;
463 623
464Example: wait for a timer, but take advantage of the fact that 624Example: wait for a timer, but take advantage of the fact that condition
465condition variables are also code references. 625variables are also callable directly.
466 626
467 my $done = AnyEvent->condvar; 627 my $done = AnyEvent->condvar;
468 my $delay = AnyEvent->timer (after => 5, cb => $done); 628 my $delay = AnyEvent->timer (after => 5, cb => $done);
469 $done->recv; 629 $done->recv;
470 630
476 636
477 ... 637 ...
478 638
479 my @info = $couchdb->info->recv; 639 my @info = $couchdb->info->recv;
480 640
481And this is how you would just ste a callback to be called whenever the 641And this is how you would just set a callback to be called whenever the
482results are available: 642results are available:
483 643
484 $couchdb->info->cb (sub { 644 $couchdb->info->cb (sub {
485 my @info = $_[0]->recv; 645 my @info = $_[0]->recv;
486 }); 646 });
504immediately from within send. 664immediately from within send.
505 665
506Any arguments passed to the C<send> call will be returned by all 666Any arguments passed to the C<send> call will be returned by all
507future C<< ->recv >> calls. 667future C<< ->recv >> calls.
508 668
509Condition variables are overloaded so one can call them directly 669Condition variables are overloaded so one can call them directly (as if
510(as a code reference). Calling them directly is the same as calling 670they were a code reference). Calling them directly is the same as calling
511C<send>. Note, however, that many C-based event loops do not handle 671C<send>.
512overloading, so as tempting as it may be, passing a condition variable
513instead of a callback does not work. Both the pure perl and EV loops
514support overloading, however, as well as all functions that use perl to
515invoke a callback (as in L<AnyEvent::Socket> and L<AnyEvent::DNS> for
516example).
517 672
518=item $cv->croak ($error) 673=item $cv->croak ($error)
519 674
520Similar to send, but causes all call's to C<< ->recv >> to invoke 675Similar to send, but causes all call's to C<< ->recv >> to invoke
521C<Carp::croak> with the given error message/object/scalar. 676C<Carp::croak> with the given error message/object/scalar.
522 677
523This can be used to signal any errors to the condition variable 678This can be used to signal any errors to the condition variable
524user/consumer. 679user/consumer. Doing it this way instead of calling C<croak> directly
680delays the error detetcion, but has the overwhelmign advantage that it
681diagnoses the error at the place where the result is expected, and not
682deep in some event clalback without connection to the actual code causing
683the problem.
525 684
526=item $cv->begin ([group callback]) 685=item $cv->begin ([group callback])
527 686
528=item $cv->end 687=item $cv->end
529
530These two methods are EXPERIMENTAL and MIGHT CHANGE.
531 688
532These two methods can be used to combine many transactions/events into 689These two methods can be used to combine many transactions/events into
533one. For example, a function that pings many hosts in parallel might want 690one. For example, a function that pings many hosts in parallel might want
534to use a condition variable for the whole process. 691to use a condition variable for the whole process.
535 692
536Every call to C<< ->begin >> will increment a counter, and every call to 693Every call to C<< ->begin >> will increment a counter, and every call to
537C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end 694C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end
538>>, the (last) callback passed to C<begin> will be executed. That callback 695>>, the (last) callback passed to C<begin> will be executed, passing the
539is I<supposed> to call C<< ->send >>, but that is not required. If no 696condvar as first argument. That callback is I<supposed> to call C<< ->send
540callback was set, C<send> will be called without any arguments. 697>>, but that is not required. If no group callback was set, C<send> will
698be called without any arguments.
541 699
542Let's clarify this with the ping example: 700You can think of C<< $cv->send >> giving you an OR condition (one call
701sends), while C<< $cv->begin >> and C<< $cv->end >> giving you an AND
702condition (all C<begin> calls must be C<end>'ed before the condvar sends).
703
704Let's start with a simple example: you have two I/O watchers (for example,
705STDOUT and STDERR for a program), and you want to wait for both streams to
706close before activating a condvar:
543 707
544 my $cv = AnyEvent->condvar; 708 my $cv = AnyEvent->condvar;
545 709
710 $cv->begin; # first watcher
711 my $w1 = AnyEvent->io (fh => $fh1, cb => sub {
712 defined sysread $fh1, my $buf, 4096
713 or $cv->end;
714 });
715
716 $cv->begin; # second watcher
717 my $w2 = AnyEvent->io (fh => $fh2, cb => sub {
718 defined sysread $fh2, my $buf, 4096
719 or $cv->end;
720 });
721
722 $cv->recv;
723
724This works because for every event source (EOF on file handle), there is
725one call to C<begin>, so the condvar waits for all calls to C<end> before
726sending.
727
728The ping example mentioned above is slightly more complicated, as the
729there are results to be passwd back, and the number of tasks that are
730begung can potentially be zero:
731
732 my $cv = AnyEvent->condvar;
733
546 my %result; 734 my %result;
547 $cv->begin (sub { $cv->send (\%result) }); 735 $cv->begin (sub { shift->send (\%result) });
548 736
549 for my $host (@list_of_hosts) { 737 for my $host (@list_of_hosts) {
550 $cv->begin; 738 $cv->begin;
551 ping_host_then_call_callback $host, sub { 739 ping_host_then_call_callback $host, sub {
552 $result{$host} = ...; 740 $result{$host} = ...;
567loop, which serves two important purposes: first, it sets the callback 755loop, which serves two important purposes: first, it sets the callback
568to be called once the counter reaches C<0>, and second, it ensures that 756to be called once the counter reaches C<0>, and second, it ensures that
569C<send> is called even when C<no> hosts are being pinged (the loop 757C<send> is called even when C<no> hosts are being pinged (the loop
570doesn't execute once). 758doesn't execute once).
571 759
572This is the general pattern when you "fan out" into multiple subrequests: 760This is the general pattern when you "fan out" into multiple (but
573use an outer C<begin>/C<end> pair to set the callback and ensure C<end> 761potentially none) subrequests: use an outer C<begin>/C<end> pair to set
574is called at least once, and then, for each subrequest you start, call 762the callback and ensure C<end> is called at least once, and then, for each
575C<begin> and for each subrequest you finish, call C<end>. 763subrequest you start, call C<begin> and for each subrequest you finish,
764call C<end>.
576 765
577=back 766=back
578 767
579=head3 METHODS FOR CONSUMERS 768=head3 METHODS FOR CONSUMERS
580 769
596function will call C<croak>. 785function will call C<croak>.
597 786
598In list context, all parameters passed to C<send> will be returned, 787In list context, all parameters passed to C<send> will be returned,
599in scalar context only the first one will be returned. 788in scalar context only the first one will be returned.
600 789
790Note that doing a blocking wait in a callback is not supported by any
791event loop, that is, recursive invocation of a blocking C<< ->recv
792>> is not allowed, and the C<recv> call will C<croak> if such a
793condition is detected. This condition can be slightly loosened by using
794L<Coro::AnyEvent>, which allows you to do a blocking C<< ->recv >> from
795any thread that doesn't run the event loop itself.
796
601Not all event models support a blocking wait - some die in that case 797Not all event models support a blocking wait - some die in that case
602(programs might want to do that to stay interactive), so I<if you are 798(programs might want to do that to stay interactive), so I<if you are
603using this from a module, never require a blocking wait>, but let the 799using this from a module, never require a blocking wait>. Instead, let the
604caller decide whether the call will block or not (for example, by coupling 800caller decide whether the call will block or not (for example, by coupling
605condition variables with some kind of request results and supporting 801condition variables with some kind of request results and supporting
606callbacks so the caller knows that getting the result will not block, 802callbacks so the caller knows that getting the result will not block,
607while still supporting blocking waits if the caller so desires). 803while still supporting blocking waits if the caller so desires).
608 804
609Another reason I<never> to C<< ->recv >> in a module is that you cannot
610sensibly have two C<< ->recv >>'s in parallel, as that would require
611multiple interpreters or coroutines/threads, none of which C<AnyEvent>
612can supply.
613
614The L<Coro> module, however, I<can> and I<does> supply coroutines and, in
615fact, L<Coro::AnyEvent> replaces AnyEvent's condvars by coroutine-safe
616versions and also integrates coroutines into AnyEvent, making blocking
617C<< ->recv >> calls perfectly safe as long as they are done from another
618coroutine (one that doesn't run the event loop).
619
620You can ensure that C<< -recv >> never blocks by setting a callback and 805You can ensure that C<< -recv >> never blocks by setting a callback and
621only calling C<< ->recv >> from within that callback (or at a later 806only calling C<< ->recv >> from within that callback (or at a later
622time). This will work even when the event loop does not support blocking 807time). This will work even when the event loop does not support blocking
623waits otherwise. 808waits otherwise.
624 809
630=item $cb = $cv->cb ($cb->($cv)) 815=item $cb = $cv->cb ($cb->($cv))
631 816
632This is a mutator function that returns the callback set and optionally 817This is a mutator function that returns the callback set and optionally
633replaces it before doing so. 818replaces it before doing so.
634 819
635The callback will be called when the condition becomes "true", i.e. when 820The callback will be called when the condition becomes (or already was)
636C<send> or C<croak> are called, with the only argument being the condition 821"true", i.e. when C<send> or C<croak> are called (or were called), with
637variable itself. Calling C<recv> inside the callback or at any later time 822the only argument being the condition variable itself. Calling C<recv>
638is guaranteed not to block. 823inside the callback or at any later time is guaranteed not to block.
639 824
640=back 825=back
641 826
827=head1 SUPPORTED EVENT LOOPS/BACKENDS
828
829The available backend classes are (every class has its own manpage):
830
831=over 4
832
833=item Backends that are autoprobed when no other event loop can be found.
834
835EV is the preferred backend when no other event loop seems to be in
836use. If EV is not installed, then AnyEvent will fall back to its own
837pure-perl implementation, which is available everywhere as it comes with
838AnyEvent itself.
839
840 AnyEvent::Impl::EV based on EV (interface to libev, best choice).
841 AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
842
843=item Backends that are transparently being picked up when they are used.
844
845These will be used when they are currently loaded when the first watcher
846is created, in which case it is assumed that the application is using
847them. This means that AnyEvent will automatically pick the right backend
848when the main program loads an event module before anything starts to
849create watchers. Nothing special needs to be done by the main program.
850
851 AnyEvent::Impl::Event based on Event, very stable, few glitches.
852 AnyEvent::Impl::Glib based on Glib, slow but very stable.
853 AnyEvent::Impl::Tk based on Tk, very broken.
854 AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
855 AnyEvent::Impl::POE based on POE, very slow, some limitations.
856 AnyEvent::Impl::Irssi used when running within irssi.
857
858=item Backends with special needs.
859
860Qt requires the Qt::Application to be instantiated first, but will
861otherwise be picked up automatically. As long as the main program
862instantiates the application before any AnyEvent watchers are created,
863everything should just work.
864
865 AnyEvent::Impl::Qt based on Qt.
866
867Support for IO::Async can only be partial, as it is too broken and
868architecturally limited to even support the AnyEvent API. It also
869is the only event loop that needs the loop to be set explicitly, so
870it can only be used by a main program knowing about AnyEvent. See
871L<AnyEvent::Impl::Async> for the gory details.
872
873 AnyEvent::Impl::IOAsync based on IO::Async, cannot be autoprobed.
874
875=item Event loops that are indirectly supported via other backends.
876
877Some event loops can be supported via other modules:
878
879There is no direct support for WxWidgets (L<Wx>) or L<Prima>.
880
881B<WxWidgets> has no support for watching file handles. However, you can
882use WxWidgets through the POE adaptor, as POE has a Wx backend that simply
883polls 20 times per second, which was considered to be too horrible to even
884consider for AnyEvent.
885
886B<Prima> is not supported as nobody seems to be using it, but it has a POE
887backend, so it can be supported through POE.
888
889AnyEvent knows about both L<Prima> and L<Wx>, however, and will try to
890load L<POE> when detecting them, in the hope that POE will pick them up,
891in which case everything will be automatic.
892
893=back
894
642=head1 GLOBAL VARIABLES AND FUNCTIONS 895=head1 GLOBAL VARIABLES AND FUNCTIONS
643 896
897These are not normally required to use AnyEvent, but can be useful to
898write AnyEvent extension modules.
899
644=over 4 900=over 4
645 901
646=item $AnyEvent::MODEL 902=item $AnyEvent::MODEL
647 903
648Contains C<undef> until the first watcher is being created. Then it 904Contains C<undef> until the first watcher is being created, before the
905backend has been autodetected.
906
649contains the event model that is being used, which is the name of the 907Afterwards it contains the event model that is being used, which is the
650Perl class implementing the model. This class is usually one of the 908name of the Perl class implementing the model. This class is usually one
651C<AnyEvent::Impl:xxx> modules, but can be any other class in the case 909of the C<AnyEvent::Impl:xxx> modules, but can be any other class in the
652AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>). 910case AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode> it
653 911will be C<urxvt::anyevent>).
654The known classes so far are:
655
656 AnyEvent::Impl::EV based on EV (an interface to libev, best choice).
657 AnyEvent::Impl::Event based on Event, second best choice.
658 AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
659 AnyEvent::Impl::Glib based on Glib, third-best choice.
660 AnyEvent::Impl::Tk based on Tk, very bad choice.
661 AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs).
662 AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
663 AnyEvent::Impl::POE based on POE, not generic enough for full support.
664
665There is no support for WxWidgets, as WxWidgets has no support for
666watching file handles. However, you can use WxWidgets through the
667POE Adaptor, as POE has a Wx backend that simply polls 20 times per
668second, which was considered to be too horrible to even consider for
669AnyEvent. Likewise, other POE backends can be used by AnyEvent by using
670it's adaptor.
671
672AnyEvent knows about L<Prima> and L<Wx> and will try to use L<POE> when
673autodetecting them.
674 912
675=item AnyEvent::detect 913=item AnyEvent::detect
676 914
677Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model 915Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model
678if necessary. You should only call this function right before you would 916if necessary. You should only call this function right before you would
679have created an AnyEvent watcher anyway, that is, as late as possible at 917have created an AnyEvent watcher anyway, that is, as late as possible at
680runtime. 918runtime, and not e.g. while initialising of your module.
919
920If you need to do some initialisation before AnyEvent watchers are
921created, use C<post_detect>.
681 922
682=item $guard = AnyEvent::post_detect { BLOCK } 923=item $guard = AnyEvent::post_detect { BLOCK }
683 924
684Arranges for the code block to be executed as soon as the event model is 925Arranges for the code block to be executed as soon as the event model is
685autodetected (or immediately if this has already happened). 926autodetected (or immediately if this has already happened).
686 927
928The block will be executed I<after> the actual backend has been detected
929(C<$AnyEvent::MODEL> is set), but I<before> any watchers have been
930created, so it is possible to e.g. patch C<@AnyEvent::ISA> or do
931other initialisations - see the sources of L<AnyEvent::Strict> or
932L<AnyEvent::AIO> to see how this is used.
933
934The most common usage is to create some global watchers, without forcing
935event module detection too early, for example, L<AnyEvent::AIO> creates
936and installs the global L<IO::AIO> watcher in a C<post_detect> block to
937avoid autodetecting the event module at load time.
938
687If called in scalar or list context, then it creates and returns an object 939If called in scalar or list context, then it creates and returns an object
688that automatically removes the callback again when it is destroyed. See 940that automatically removes the callback again when it is destroyed (or
941C<undef> when the hook was immediately executed). See L<AnyEvent::AIO> for
689L<Coro::BDB> for a case where this is useful. 942a case where this is useful.
943
944Example: Create a watcher for the IO::AIO module and store it in
945C<$WATCHER>. Only do so after the event loop is initialised, though.
946
947 our WATCHER;
948
949 my $guard = AnyEvent::post_detect {
950 $WATCHER = AnyEvent->io (fh => IO::AIO::poll_fileno, poll => 'r', cb => \&IO::AIO::poll_cb);
951 };
952
953 # the ||= is important in case post_detect immediately runs the block,
954 # as to not clobber the newly-created watcher. assigning both watcher and
955 # post_detect guard to the same variable has the advantage of users being
956 # able to just C<undef $WATCHER> if the watcher causes them grief.
957
958 $WATCHER ||= $guard;
690 959
691=item @AnyEvent::post_detect 960=item @AnyEvent::post_detect
692 961
693If there are any code references in this array (you can C<push> to it 962If there are any code references in this array (you can C<push> to it
694before or after loading AnyEvent), then they will called directly after 963before or after loading AnyEvent), then they will called directly after
695the event loop has been chosen. 964the event loop has been chosen.
696 965
697You should check C<$AnyEvent::MODEL> before adding to this array, though: 966You should check C<$AnyEvent::MODEL> before adding to this array, though:
698if it contains a true value then the event loop has already been detected, 967if it is defined then the event loop has already been detected, and the
699and the array will be ignored. 968array will be ignored.
700 969
701Best use C<AnyEvent::post_detect { BLOCK }> instead. 970Best use C<AnyEvent::post_detect { BLOCK }> when your application allows
971it, as it takes care of these details.
972
973This variable is mainly useful for modules that can do something useful
974when AnyEvent is used and thus want to know when it is initialised, but do
975not need to even load it by default. This array provides the means to hook
976into AnyEvent passively, without loading it.
977
978Example: To load Coro::AnyEvent whenever Coro and AnyEvent are used
979together, you could put this into Coro (this is the actual code used by
980Coro to accomplish this):
981
982 if (defined $AnyEvent::MODEL) {
983 # AnyEvent already initialised, so load Coro::AnyEvent
984 require Coro::AnyEvent;
985 } else {
986 # AnyEvent not yet initialised, so make sure to load Coro::AnyEvent
987 # as soon as it is
988 push @AnyEvent::post_detect, sub { require Coro::AnyEvent };
989 }
702 990
703=back 991=back
704 992
705=head1 WHAT TO DO IN A MODULE 993=head1 WHAT TO DO IN A MODULE
706 994
761 1049
762 1050
763=head1 OTHER MODULES 1051=head1 OTHER MODULES
764 1052
765The following is a non-exhaustive list of additional modules that use 1053The following is a non-exhaustive list of additional modules that use
766AnyEvent and can therefore be mixed easily with other AnyEvent modules 1054AnyEvent as a client and can therefore be mixed easily with other AnyEvent
767in the same program. Some of the modules come with AnyEvent, some are 1055modules and other event loops in the same program. Some of the modules
768available via CPAN. 1056come with AnyEvent, most are available via CPAN.
769 1057
770=over 4 1058=over 4
771 1059
772=item L<AnyEvent::Util> 1060=item L<AnyEvent::Util>
773 1061
782 1070
783=item L<AnyEvent::Handle> 1071=item L<AnyEvent::Handle>
784 1072
785Provide read and write buffers, manages watchers for reads and writes, 1073Provide read and write buffers, manages watchers for reads and writes,
786supports raw and formatted I/O, I/O queued and fully transparent and 1074supports raw and formatted I/O, I/O queued and fully transparent and
787non-blocking SSL/TLS. 1075non-blocking SSL/TLS (via L<AnyEvent::TLS>.
788 1076
789=item L<AnyEvent::DNS> 1077=item L<AnyEvent::DNS>
790 1078
791Provides rich asynchronous DNS resolver capabilities. 1079Provides rich asynchronous DNS resolver capabilities.
792 1080
820 1108
821=item L<AnyEvent::GPSD> 1109=item L<AnyEvent::GPSD>
822 1110
823A non-blocking interface to gpsd, a daemon delivering GPS information. 1111A non-blocking interface to gpsd, a daemon delivering GPS information.
824 1112
1113=item L<AnyEvent::IRC>
1114
1115AnyEvent based IRC client module family (replacing the older Net::IRC3).
1116
1117=item L<AnyEvent::XMPP>
1118
1119AnyEvent based XMPP (Jabber protocol) module family (replacing the older
1120Net::XMPP2>.
1121
825=item L<AnyEvent::IGS> 1122=item L<AnyEvent::IGS>
826 1123
827A non-blocking interface to the Internet Go Server protocol (used by 1124A non-blocking interface to the Internet Go Server protocol (used by
828L<App::IGS>). 1125L<App::IGS>).
829 1126
830=item L<AnyEvent::IRC>
831
832AnyEvent based IRC client module family (replacing the older Net::IRC3).
833
834=item L<Net::XMPP2>
835
836AnyEvent based XMPP (Jabber protocol) module family.
837
838=item L<Net::FCP> 1127=item L<Net::FCP>
839 1128
840AnyEvent-based implementation of the Freenet Client Protocol, birthplace 1129AnyEvent-based implementation of the Freenet Client Protocol, birthplace
841of AnyEvent. 1130of AnyEvent.
842 1131
846 1135
847=item L<Coro> 1136=item L<Coro>
848 1137
849Has special support for AnyEvent via L<Coro::AnyEvent>. 1138Has special support for AnyEvent via L<Coro::AnyEvent>.
850 1139
851=item L<IO::Lambda>
852
853The lambda approach to I/O - don't ask, look there. Can use AnyEvent.
854
855=back 1140=back
856 1141
857=cut 1142=cut
858 1143
859package AnyEvent; 1144package AnyEvent;
860 1145
861no warnings; 1146# basically a tuned-down version of common::sense
862use strict qw(vars subs); 1147sub common_sense {
1148 # from common:.sense 1.0
1149 ${^WARNING_BITS} = "\xfc\x3f\x33\x00\x0f\xf3\xcf\xc0\xf3\xfc\x33\x00";
1150 # use strict vars subs - NO UTF-8, as Util.pm doesn't like this atm. (uts46data.pl)
1151 $^H |= 0x00000600;
1152}
863 1153
1154BEGIN { AnyEvent::common_sense }
1155
864use Carp; 1156use Carp ();
865 1157
866our $VERSION = 4.34; 1158our $VERSION = '5.23';
867our $MODEL; 1159our $MODEL;
868 1160
869our $AUTOLOAD; 1161our $AUTOLOAD;
870our @ISA; 1162our @ISA;
871 1163
872our @REGISTRY; 1164our @REGISTRY;
873 1165
874our $WIN32; 1166our $VERBOSE;
875 1167
876BEGIN { 1168BEGIN {
877 my $win32 = ! ! ($^O =~ /mswin32/i); 1169 eval "sub WIN32(){ " . (($^O =~ /mswin32/i)*1) ." }";
878 eval "sub WIN32(){ $win32 }"; 1170 eval "sub TAINT(){ " . (${^TAINT}*1) . " }";
879}
880 1171
1172 delete @ENV{grep /^PERL_ANYEVENT_/, keys %ENV}
1173 if ${^TAINT};
1174
881our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1; 1175 $VERBOSE = $ENV{PERL_ANYEVENT_VERBOSE}*1;
1176
1177}
1178
1179our $MAX_SIGNAL_LATENCY = 10;
882 1180
883our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred 1181our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
884 1182
885{ 1183{
886 my $idx; 1184 my $idx;
888 for reverse split /\s*,\s*/, 1186 for reverse split /\s*,\s*/,
889 $ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6"; 1187 $ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6";
890} 1188}
891 1189
892my @models = ( 1190my @models = (
893 [EV:: => AnyEvent::Impl::EV::], 1191 [EV:: => AnyEvent::Impl::EV:: , 1],
894 [Event:: => AnyEvent::Impl::Event::],
895 [AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::], 1192 [AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl:: , 1],
896 # everything below here will not be autoprobed 1193 # everything below here will not (normally) be autoprobed
897 # as the pureperl backend should work everywhere 1194 # as the pureperl backend should work everywhere
898 # and is usually faster 1195 # and is usually faster
1196 [Event:: => AnyEvent::Impl::Event::, 1],
1197 [Glib:: => AnyEvent::Impl::Glib:: , 1], # becomes extremely slow with many watchers
1198 [Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
1199 [Irssi:: => AnyEvent::Impl::Irssi::], # Irssi has a bogus "Event" package
899 [Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles 1200 [Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles
900 [Glib:: => AnyEvent::Impl::Glib::], # becomes extremely slow with many watchers
901 [Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
902 [Qt:: => AnyEvent::Impl::Qt::], # requires special main program 1201 [Qt:: => AnyEvent::Impl::Qt::], # requires special main program
903 [POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza 1202 [POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
904 [Wx:: => AnyEvent::Impl::POE::], 1203 [Wx:: => AnyEvent::Impl::POE::],
905 [Prima:: => AnyEvent::Impl::POE::], 1204 [Prima:: => AnyEvent::Impl::POE::],
1205 # IO::Async is just too broken - we would need workarounds for its
1206 # byzantine signal and broken child handling, among others.
1207 # IO::Async is rather hard to detect, as it doesn't have any
1208 # obvious default class.
1209 [IO::Async:: => AnyEvent::Impl::IOAsync::], # requires special main program
1210 [IO::Async::Loop:: => AnyEvent::Impl::IOAsync::], # requires special main program
1211 [IO::Async::Notifier:: => AnyEvent::Impl::IOAsync::], # requires special main program
1212 [AnyEvent::Impl::IOAsync:: => AnyEvent::Impl::IOAsync::], # requires special main program
906); 1213);
907 1214
908our %method = map +($_ => 1), qw(io timer time now signal child condvar one_event DESTROY); 1215our %method = map +($_ => 1),
1216 qw(io timer time now now_update signal child idle condvar one_event DESTROY);
909 1217
910our @post_detect; 1218our @post_detect;
911 1219
912sub post_detect(&) { 1220sub post_detect(&) {
913 my ($cb) = @_; 1221 my ($cb) = @_;
914 1222
915 if ($MODEL) { 1223 if ($MODEL) {
916 $cb->(); 1224 $cb->();
917 1225
918 1 1226 undef
919 } else { 1227 } else {
920 push @post_detect, $cb; 1228 push @post_detect, $cb;
921 1229
922 defined wantarray 1230 defined wantarray
923 ? bless \$cb, "AnyEvent::Util::PostDetect" 1231 ? bless \$cb, "AnyEvent::Util::postdetect"
924 : () 1232 : ()
925 } 1233 }
926} 1234}
927 1235
928sub AnyEvent::Util::PostDetect::DESTROY { 1236sub AnyEvent::Util::postdetect::DESTROY {
929 @post_detect = grep $_ != ${$_[0]}, @post_detect; 1237 @post_detect = grep $_ != ${$_[0]}, @post_detect;
930} 1238}
931 1239
932sub detect() { 1240sub detect() {
933 unless ($MODEL) { 1241 unless ($MODEL) {
934 no strict 'refs';
935 local $SIG{__DIE__}; 1242 local $SIG{__DIE__};
936 1243
937 if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) { 1244 if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) {
938 my $model = "AnyEvent::Impl::$1"; 1245 my $model = "AnyEvent::Impl::$1";
939 if (eval "require $model") { 1246 if (eval "require $model") {
940 $MODEL = $model; 1247 $MODEL = $model;
941 warn "AnyEvent: loaded model '$model' (forced by \$PERL_ANYEVENT_MODEL), using it.\n" if $verbose > 1; 1248 warn "AnyEvent: loaded model '$model' (forced by \$ENV{PERL_ANYEVENT_MODEL}), using it.\n" if $VERBOSE >= 2;
942 } else { 1249 } else {
943 warn "AnyEvent: unable to load model '$model' (from \$PERL_ANYEVENT_MODEL):\n$@" if $verbose; 1250 warn "AnyEvent: unable to load model '$model' (from \$ENV{PERL_ANYEVENT_MODEL}):\n$@" if $VERBOSE;
944 } 1251 }
945 } 1252 }
946 1253
947 # check for already loaded models 1254 # check for already loaded models
948 unless ($MODEL) { 1255 unless ($MODEL) {
949 for (@REGISTRY, @models) { 1256 for (@REGISTRY, @models) {
950 my ($package, $model) = @$_; 1257 my ($package, $model) = @$_;
951 if (${"$package\::VERSION"} > 0) { 1258 if (${"$package\::VERSION"} > 0) {
952 if (eval "require $model") { 1259 if (eval "require $model") {
953 $MODEL = $model; 1260 $MODEL = $model;
954 warn "AnyEvent: autodetected model '$model', using it.\n" if $verbose > 1; 1261 warn "AnyEvent: autodetected model '$model', using it.\n" if $VERBOSE >= 2;
955 last; 1262 last;
956 } 1263 }
957 } 1264 }
958 } 1265 }
959 1266
960 unless ($MODEL) { 1267 unless ($MODEL) {
961 # try to load a model 1268 # try to autoload a model
962
963 for (@REGISTRY, @models) { 1269 for (@REGISTRY, @models) {
964 my ($package, $model) = @$_; 1270 my ($package, $model, $autoload) = @$_;
1271 if (
1272 $autoload
965 if (eval "require $package" 1273 and eval "require $package"
966 and ${"$package\::VERSION"} > 0 1274 and ${"$package\::VERSION"} > 0
967 and eval "require $model") { 1275 and eval "require $model"
1276 ) {
968 $MODEL = $model; 1277 $MODEL = $model;
969 warn "AnyEvent: autoprobed model '$model', using it.\n" if $verbose > 1; 1278 warn "AnyEvent: autoloaded model '$model', using it.\n" if $VERBOSE >= 2;
970 last; 1279 last;
971 } 1280 }
972 } 1281 }
973 1282
974 $MODEL 1283 $MODEL
975 or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib."; 1284 or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib.\n";
976 } 1285 }
977 } 1286 }
978 1287
979 push @{"$MODEL\::ISA"}, "AnyEvent::Base"; 1288 push @{"$MODEL\::ISA"}, "AnyEvent::Base";
980 1289
990 1299
991sub AUTOLOAD { 1300sub AUTOLOAD {
992 (my $func = $AUTOLOAD) =~ s/.*://; 1301 (my $func = $AUTOLOAD) =~ s/.*://;
993 1302
994 $method{$func} 1303 $method{$func}
995 or croak "$func: not a valid method for AnyEvent objects"; 1304 or Carp::croak "$func: not a valid method for AnyEvent objects";
996 1305
997 detect unless $MODEL; 1306 detect unless $MODEL;
998 1307
999 my $class = shift; 1308 my $class = shift;
1000 $class->$func (@_); 1309 $class->$func (@_);
1001} 1310}
1002 1311
1003# utility function to dup a filehandle. this is used by many backends 1312# utility function to dup a filehandle. this is used by many backends
1004# to support binding more than one watcher per filehandle (they usually 1313# to support binding more than one watcher per filehandle (they usually
1005# allow only one watcher per fd, so we dup it to get a different one). 1314# allow only one watcher per fd, so we dup it to get a different one).
1006sub _dupfh($$$$) { 1315sub _dupfh($$;$$) {
1007 my ($poll, $fh, $r, $w) = @_; 1316 my ($poll, $fh, $r, $w) = @_;
1008 1317
1009 require Fcntl;
1010
1011 # cygwin requires the fh mode to be matching, unix doesn't 1318 # cygwin requires the fh mode to be matching, unix doesn't
1012 my ($rw, $mode) = $poll eq "r" ? ($r, "<") 1319 my ($rw, $mode) = $poll eq "r" ? ($r, "<&") : ($w, ">&");
1013 : $poll eq "w" ? ($w, ">")
1014 : Carp::croak "AnyEvent->io requires poll set to either 'r' or 'w'";
1015 1320
1016 open my $fh2, "$mode&" . fileno $fh 1321 open my $fh2, $mode, $fh
1017 or die "cannot dup() filehandle: $!"; 1322 or die "AnyEvent->io: cannot dup() filehandle in mode '$poll': $!,";
1018 1323
1019 # we assume CLOEXEC is already set by perl in all important cases 1324 # we assume CLOEXEC is already set by perl in all important cases
1020 1325
1021 ($fh2, $rw) 1326 ($fh2, $rw)
1022} 1327}
1023 1328
1329=head1 SIMPLIFIED AE API
1330
1331Starting with version 5.0, AnyEvent officially supports a second, much
1332simpler, API that is designed to reduce the calling, typing and memory
1333overhead.
1334
1335See the L<AE> manpage for details.
1336
1337=cut
1338
1339package AE;
1340
1341our $VERSION = $AnyEvent::VERSION;
1342
1343sub io($$$) {
1344 AnyEvent->io (fh => $_[0], poll => $_[1] ? "w" : "r", cb => $_[2])
1345}
1346
1347sub timer($$$) {
1348 AnyEvent->timer (after => $_[0], interval => $_[1], cb => $_[2])
1349}
1350
1351sub signal($$) {
1352 AnyEvent->signal (signal => $_[0], cb => $_[1])
1353}
1354
1355sub child($$) {
1356 AnyEvent->child (pid => $_[0], cb => $_[1])
1357}
1358
1359sub idle($) {
1360 AnyEvent->idle (cb => $_[0])
1361}
1362
1363sub cv(;&) {
1364 AnyEvent->condvar (@_ ? (cb => $_[0]) : ())
1365}
1366
1367sub now() {
1368 AnyEvent->now
1369}
1370
1371sub now_update() {
1372 AnyEvent->now_update
1373}
1374
1375sub time() {
1376 AnyEvent->time
1377}
1378
1024package AnyEvent::Base; 1379package AnyEvent::Base;
1025 1380
1026# default implementation for now and time 1381# default implementations for many methods
1027 1382
1028BEGIN { 1383sub _time() {
1384 # probe for availability of Time::HiRes
1029 if (eval "use Time::HiRes (); time (); 1") { 1385 if (eval "use Time::HiRes (); Time::HiRes::time (); 1") {
1386 warn "AnyEvent: using Time::HiRes for sub-second timing accuracy.\n" if $VERBOSE >= 8;
1030 *_time = \&Time::HiRes::time; 1387 *_time = \&Time::HiRes::time;
1031 # if (eval "use POSIX (); (POSIX::times())... 1388 # if (eval "use POSIX (); (POSIX::times())...
1032 } else { 1389 } else {
1390 warn "AnyEvent: using built-in time(), WARNING, no sub-second resolution!\n" if $VERBOSE;
1033 *_time = sub { time }; # epic fail 1391 *_time = sub { time }; # epic fail
1034 } 1392 }
1393
1394 &_time
1035} 1395}
1036 1396
1037sub time { _time } 1397sub time { _time }
1038sub now { _time } 1398sub now { _time }
1399sub now_update { }
1039 1400
1040# default implementation for ->condvar 1401# default implementation for ->condvar
1041 1402
1042sub condvar { 1403sub condvar {
1043 bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, AnyEvent::CondVar:: 1404 bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, "AnyEvent::CondVar"
1044} 1405}
1045 1406
1046# default implementation for ->signal 1407# default implementation for ->signal
1047 1408
1048our %SIG_CB; 1409our $HAVE_ASYNC_INTERRUPT;
1410
1411sub _have_async_interrupt() {
1412 $HAVE_ASYNC_INTERRUPT = 1*(!$ENV{PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT}
1413 && eval "use Async::Interrupt 1.02 (); 1")
1414 unless defined $HAVE_ASYNC_INTERRUPT;
1415
1416 $HAVE_ASYNC_INTERRUPT
1417}
1418
1419our ($SIGPIPE_R, $SIGPIPE_W, %SIG_CB, %SIG_EV, $SIG_IO);
1420our (%SIG_ASY, %SIG_ASY_W);
1421our ($SIG_COUNT, $SIG_TW);
1422
1423sub _signal_exec {
1424 $HAVE_ASYNC_INTERRUPT
1425 ? $SIGPIPE_R->drain
1426 : sysread $SIGPIPE_R, (my $dummy), 9;
1427
1428 while (%SIG_EV) {
1429 for (keys %SIG_EV) {
1430 delete $SIG_EV{$_};
1431 $_->() for values %{ $SIG_CB{$_} || {} };
1432 }
1433 }
1434}
1435
1436# install a dummy wakeup watcher to reduce signal catching latency
1437sub _sig_add() {
1438 unless ($SIG_COUNT++) {
1439 # try to align timer on a full-second boundary, if possible
1440 my $NOW = AE::now;
1441
1442 $SIG_TW = AE::timer
1443 $MAX_SIGNAL_LATENCY - ($NOW - int $NOW),
1444 $MAX_SIGNAL_LATENCY,
1445 sub { } # just for the PERL_ASYNC_CHECK
1446 ;
1447 }
1448}
1449
1450sub _sig_del {
1451 undef $SIG_TW
1452 unless --$SIG_COUNT;
1453}
1454
1455our $_sig_name_init; $_sig_name_init = sub {
1456 eval q{ # poor man's autoloading
1457 undef $_sig_name_init;
1458
1459 if (_have_async_interrupt) {
1460 *sig2num = \&Async::Interrupt::sig2num;
1461 *sig2name = \&Async::Interrupt::sig2name;
1462 } else {
1463 require Config;
1464
1465 my %signame2num;
1466 @signame2num{ split ' ', $Config::Config{sig_name} }
1467 = split ' ', $Config::Config{sig_num};
1468
1469 my @signum2name;
1470 @signum2name[values %signame2num] = keys %signame2num;
1471
1472 *sig2num = sub($) {
1473 $_[0] > 0 ? shift : $signame2num{+shift}
1474 };
1475 *sig2name = sub ($) {
1476 $_[0] > 0 ? $signum2name[+shift] : shift
1477 };
1478 }
1479 };
1480 die if $@;
1481};
1482
1483sub sig2num ($) { &$_sig_name_init; &sig2num }
1484sub sig2name($) { &$_sig_name_init; &sig2name }
1049 1485
1050sub signal { 1486sub signal {
1487 eval q{ # poor man's autoloading {}
1488 # probe for availability of Async::Interrupt
1489 if (_have_async_interrupt) {
1490 warn "AnyEvent: using Async::Interrupt for race-free signal handling.\n" if $VERBOSE >= 8;
1491
1492 $SIGPIPE_R = new Async::Interrupt::EventPipe;
1493 $SIG_IO = AE::io $SIGPIPE_R->fileno, 0, \&_signal_exec;
1494
1495 } else {
1496 warn "AnyEvent: using emulated perl signal handling with latency timer.\n" if $VERBOSE >= 8;
1497
1498 require Fcntl;
1499
1500 if (AnyEvent::WIN32) {
1501 require AnyEvent::Util;
1502
1503 ($SIGPIPE_R, $SIGPIPE_W) = AnyEvent::Util::portable_pipe ();
1504 AnyEvent::Util::fh_nonblocking ($SIGPIPE_R, 1) if $SIGPIPE_R;
1505 AnyEvent::Util::fh_nonblocking ($SIGPIPE_W, 1) if $SIGPIPE_W; # just in case
1506 } else {
1507 pipe $SIGPIPE_R, $SIGPIPE_W;
1508 fcntl $SIGPIPE_R, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_R;
1509 fcntl $SIGPIPE_W, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_W; # just in case
1510
1511 # not strictly required, as $^F is normally 2, but let's make sure...
1512 fcntl $SIGPIPE_R, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
1513 fcntl $SIGPIPE_W, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
1514 }
1515
1516 $SIGPIPE_R
1517 or Carp::croak "AnyEvent: unable to create a signal reporting pipe: $!\n";
1518
1519 $SIG_IO = AE::io $SIGPIPE_R, 0, \&_signal_exec;
1520 }
1521
1522 *signal = sub {
1051 my (undef, %arg) = @_; 1523 my (undef, %arg) = @_;
1052 1524
1053 my $signal = uc $arg{signal} 1525 my $signal = uc $arg{signal}
1054 or Carp::croak "required option 'signal' is missing"; 1526 or Carp::croak "required option 'signal' is missing";
1055 1527
1528 if ($HAVE_ASYNC_INTERRUPT) {
1529 # async::interrupt
1530
1531 $signal = sig2num $signal;
1056 $SIG_CB{$signal}{$arg{cb}} = $arg{cb}; 1532 $SIG_CB{$signal}{$arg{cb}} = $arg{cb};
1533
1534 $SIG_ASY{$signal} ||= new Async::Interrupt
1535 cb => sub { undef $SIG_EV{$signal} },
1536 signal => $signal,
1537 pipe => [$SIGPIPE_R->filenos],
1538 pipe_autodrain => 0,
1539 ;
1540
1541 } else {
1542 # pure perl
1543
1544 # AE::Util has been loaded in signal
1545 $signal = sig2name $signal;
1546 $SIG_CB{$signal}{$arg{cb}} = $arg{cb};
1547
1057 $SIG{$signal} ||= sub { 1548 $SIG{$signal} ||= sub {
1058 $_->() for values %{ $SIG_CB{$signal} || {} }; 1549 local $!;
1550 syswrite $SIGPIPE_W, "\x00", 1 unless %SIG_EV;
1551 undef $SIG_EV{$signal};
1552 };
1553
1554 # can't do signal processing without introducing races in pure perl,
1555 # so limit the signal latency.
1556 _sig_add;
1557 }
1558
1559 bless [$signal, $arg{cb}], "AnyEvent::Base::signal"
1560 };
1561
1562 *AnyEvent::Base::signal::DESTROY = sub {
1563 my ($signal, $cb) = @{$_[0]};
1564
1565 _sig_del;
1566
1567 delete $SIG_CB{$signal}{$cb};
1568
1569 $HAVE_ASYNC_INTERRUPT
1570 ? delete $SIG_ASY{$signal}
1571 : # delete doesn't work with older perls - they then
1572 # print weird messages, or just unconditionally exit
1573 # instead of getting the default action.
1574 undef $SIG{$signal}
1575 unless keys %{ $SIG_CB{$signal} };
1576 };
1059 }; 1577 };
1060 1578 die if $@;
1061 bless [$signal, $arg{cb}], "AnyEvent::Base::Signal" 1579 &signal
1062}
1063
1064sub AnyEvent::Base::Signal::DESTROY {
1065 my ($signal, $cb) = @{$_[0]};
1066
1067 delete $SIG_CB{$signal}{$cb};
1068
1069 delete $SIG{$signal} unless keys %{ $SIG_CB{$signal} };
1070} 1580}
1071 1581
1072# default implementation for ->child 1582# default implementation for ->child
1073 1583
1074our %PID_CB; 1584our %PID_CB;
1075our $CHLD_W; 1585our $CHLD_W;
1076our $CHLD_DELAY_W; 1586our $CHLD_DELAY_W;
1077our $PID_IDLE;
1078our $WNOHANG; 1587our $WNOHANG;
1079 1588
1080sub _child_wait { 1589sub _emit_childstatus($$) {
1081 while (0 < (my $pid = waitpid -1, $WNOHANG)) { 1590 my (undef, $rpid, $rstatus) = @_;
1591
1592 $_->($rpid, $rstatus)
1082 $_->($pid, $?) for (values %{ $PID_CB{$pid} || {} }), 1593 for values %{ $PID_CB{$rpid} || {} },
1083 (values %{ $PID_CB{0} || {} }); 1594 values %{ $PID_CB{0} || {} };
1084 }
1085
1086 undef $PID_IDLE;
1087} 1595}
1088 1596
1089sub _sigchld { 1597sub _sigchld {
1090 # make sure we deliver these changes "synchronous" with the event loop. 1598 my $pid;
1091 $CHLD_DELAY_W ||= AnyEvent->timer (after => 0, cb => sub { 1599
1092 undef $CHLD_DELAY_W; 1600 AnyEvent->_emit_childstatus ($pid, $?)
1093 &_child_wait; 1601 while ($pid = waitpid -1, $WNOHANG) > 0;
1094 });
1095} 1602}
1096 1603
1097sub child { 1604sub child {
1098 my (undef, %arg) = @_; 1605 my (undef, %arg) = @_;
1099 1606
1100 defined (my $pid = $arg{pid} + 0) 1607 defined (my $pid = $arg{pid} + 0)
1101 or Carp::croak "required option 'pid' is missing"; 1608 or Carp::croak "required option 'pid' is missing";
1102 1609
1103 $PID_CB{$pid}{$arg{cb}} = $arg{cb}; 1610 $PID_CB{$pid}{$arg{cb}} = $arg{cb};
1104 1611
1105 unless ($WNOHANG) { 1612 # WNOHANG is almost cetrainly 1 everywhere
1613 $WNOHANG ||= $^O =~ /^(?:openbsd|netbsd|linux|freebsd|cygwin|MSWin32)$/
1614 ? 1
1106 $WNOHANG = eval { local $SIG{__DIE__}; require POSIX; &POSIX::WNOHANG } || 1; 1615 : eval { local $SIG{__DIE__}; require POSIX; &POSIX::WNOHANG } || 1;
1107 }
1108 1616
1109 unless ($CHLD_W) { 1617 unless ($CHLD_W) {
1110 $CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_sigchld); 1618 $CHLD_W = AE::signal CHLD => \&_sigchld;
1111 # child could be a zombie already, so make at least one round 1619 # child could be a zombie already, so make at least one round
1112 &_sigchld; 1620 &_sigchld;
1113 } 1621 }
1114 1622
1115 bless [$pid, $arg{cb}], "AnyEvent::Base::Child" 1623 bless [$pid, $arg{cb}], "AnyEvent::Base::child"
1116} 1624}
1117 1625
1118sub AnyEvent::Base::Child::DESTROY { 1626sub AnyEvent::Base::child::DESTROY {
1119 my ($pid, $cb) = @{$_[0]}; 1627 my ($pid, $cb) = @{$_[0]};
1120 1628
1121 delete $PID_CB{$pid}{$cb}; 1629 delete $PID_CB{$pid}{$cb};
1122 delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} }; 1630 delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
1123 1631
1124 undef $CHLD_W unless keys %PID_CB; 1632 undef $CHLD_W unless keys %PID_CB;
1125} 1633}
1126 1634
1635# idle emulation is done by simply using a timer, regardless
1636# of whether the process is idle or not, and not letting
1637# the callback use more than 50% of the time.
1638sub idle {
1639 my (undef, %arg) = @_;
1640
1641 my ($cb, $w, $rcb) = $arg{cb};
1642
1643 $rcb = sub {
1644 if ($cb) {
1645 $w = _time;
1646 &$cb;
1647 $w = _time - $w;
1648
1649 # never use more then 50% of the time for the idle watcher,
1650 # within some limits
1651 $w = 0.0001 if $w < 0.0001;
1652 $w = 5 if $w > 5;
1653
1654 $w = AE::timer $w, 0, $rcb;
1655 } else {
1656 # clean up...
1657 undef $w;
1658 undef $rcb;
1659 }
1660 };
1661
1662 $w = AE::timer 0.05, 0, $rcb;
1663
1664 bless \\$cb, "AnyEvent::Base::idle"
1665}
1666
1667sub AnyEvent::Base::idle::DESTROY {
1668 undef $${$_[0]};
1669}
1670
1127package AnyEvent::CondVar; 1671package AnyEvent::CondVar;
1128 1672
1129our @ISA = AnyEvent::CondVar::Base::; 1673our @ISA = AnyEvent::CondVar::Base::;
1130 1674
1131package AnyEvent::CondVar::Base; 1675package AnyEvent::CondVar::Base;
1132 1676
1133use overload 1677#use overload
1134 '&{}' => sub { my $self = shift; sub { $self->send (@_) } }, 1678# '&{}' => sub { my $self = shift; sub { $self->send (@_) } },
1135 fallback => 1; 1679# fallback => 1;
1680
1681# save 300+ kilobytes by dirtily hardcoding overloading
1682${"AnyEvent::CondVar::Base::OVERLOAD"}{dummy}++; # Register with magic by touching.
1683*{'AnyEvent::CondVar::Base::()'} = sub { }; # "Make it findable via fetchmethod."
1684*{'AnyEvent::CondVar::Base::(&{}'} = sub { my $self = shift; sub { $self->send (@_) } }; # &{}
1685${'AnyEvent::CondVar::Base::()'} = 1; # fallback
1686
1687our $WAITING;
1136 1688
1137sub _send { 1689sub _send {
1138 # nop 1690 # nop
1139} 1691}
1140 1692
1153sub ready { 1705sub ready {
1154 $_[0]{_ae_sent} 1706 $_[0]{_ae_sent}
1155} 1707}
1156 1708
1157sub _wait { 1709sub _wait {
1710 $WAITING
1711 and !$_[0]{_ae_sent}
1712 and Carp::croak "AnyEvent::CondVar: recursive blocking wait detected";
1713
1714 local $WAITING = 1;
1158 AnyEvent->one_event while !$_[0]{_ae_sent}; 1715 AnyEvent->one_event while !$_[0]{_ae_sent};
1159} 1716}
1160 1717
1161sub recv { 1718sub recv {
1162 $_[0]->_wait; 1719 $_[0]->_wait;
1164 Carp::croak $_[0]{_ae_croak} if $_[0]{_ae_croak}; 1721 Carp::croak $_[0]{_ae_croak} if $_[0]{_ae_croak};
1165 wantarray ? @{ $_[0]{_ae_sent} } : $_[0]{_ae_sent}[0] 1722 wantarray ? @{ $_[0]{_ae_sent} } : $_[0]{_ae_sent}[0]
1166} 1723}
1167 1724
1168sub cb { 1725sub cb {
1169 $_[0]{_ae_cb} = $_[1] if @_ > 1; 1726 my $cv = shift;
1727
1728 @_
1729 and $cv->{_ae_cb} = shift
1730 and $cv->{_ae_sent}
1731 and (delete $cv->{_ae_cb})->($cv);
1732
1170 $_[0]{_ae_cb} 1733 $cv->{_ae_cb}
1171} 1734}
1172 1735
1173sub begin { 1736sub begin {
1174 ++$_[0]{_ae_counter}; 1737 ++$_[0]{_ae_counter};
1175 $_[0]{_ae_end_cb} = $_[1] if @_ > 1; 1738 $_[0]{_ae_end_cb} = $_[1] if @_ > 1;
1203so on. 1766so on.
1204 1767
1205=head1 ENVIRONMENT VARIABLES 1768=head1 ENVIRONMENT VARIABLES
1206 1769
1207The following environment variables are used by this module or its 1770The following environment variables are used by this module or its
1208submodules: 1771submodules.
1772
1773Note that AnyEvent will remove I<all> environment variables starting with
1774C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
1775enabled.
1209 1776
1210=over 4 1777=over 4
1211 1778
1212=item C<PERL_ANYEVENT_VERBOSE> 1779=item C<PERL_ANYEVENT_VERBOSE>
1213 1780
1220C<PERL_ANYEVENT_MODEL>. 1787C<PERL_ANYEVENT_MODEL>.
1221 1788
1222When set to C<2> or higher, cause AnyEvent to report to STDERR which event 1789When set to C<2> or higher, cause AnyEvent to report to STDERR which event
1223model it chooses. 1790model it chooses.
1224 1791
1792When set to C<8> or higher, then AnyEvent will report extra information on
1793which optional modules it loads and how it implements certain features.
1794
1225=item C<PERL_ANYEVENT_STRICT> 1795=item C<PERL_ANYEVENT_STRICT>
1226 1796
1227AnyEvent does not do much argument checking by default, as thorough 1797AnyEvent does not do much argument checking by default, as thorough
1228argument checking is very costly. Setting this variable to a true value 1798argument checking is very costly. Setting this variable to a true value
1229will cause AnyEvent to load C<AnyEvent::Strict> and then to thoroughly 1799will cause AnyEvent to load C<AnyEvent::Strict> and then to thoroughly
1230check the arguments passed to most method calls. If it finds any problems 1800check the arguments passed to most method calls. If it finds any problems,
1231it will croak. 1801it will croak.
1232 1802
1233In other words, enables "strict" mode. 1803In other words, enables "strict" mode.
1234 1804
1235Unlike C<use strict>, it is definitely recommended ot keep it off in 1805Unlike C<use strict> (or it's modern cousin, C<< use L<common::sense>
1236production. Keeping C<PERL_ANYEVENT_STRICT=1> in your environment while 1806>>, it is definitely recommended to keep it off in production. Keeping
1237developing programs can be very useful, however. 1807C<PERL_ANYEVENT_STRICT=1> in your environment while developing programs
1808can be very useful, however.
1238 1809
1239=item C<PERL_ANYEVENT_MODEL> 1810=item C<PERL_ANYEVENT_MODEL>
1240 1811
1241This can be used to specify the event model to be used by AnyEvent, before 1812This can be used to specify the event model to be used by AnyEvent, before
1242auto detection and -probing kicks in. It must be a string consisting 1813auto detection and -probing kicks in. It must be a string consisting
1263used, and preference will be given to protocols mentioned earlier in the 1834used, and preference will be given to protocols mentioned earlier in the
1264list. 1835list.
1265 1836
1266This variable can effectively be used for denial-of-service attacks 1837This variable can effectively be used for denial-of-service attacks
1267against local programs (e.g. when setuid), although the impact is likely 1838against local programs (e.g. when setuid), although the impact is likely
1268small, as the program has to handle connection errors already- 1839small, as the program has to handle conenction and other failures anyways.
1269 1840
1270Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6, 1841Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
1271but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4> 1842but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
1272- only support IPv4, never try to resolve or contact IPv6 1843- only support IPv4, never try to resolve or contact IPv6
1273addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or 1844addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
1285 1856
1286=item C<PERL_ANYEVENT_MAX_FORKS> 1857=item C<PERL_ANYEVENT_MAX_FORKS>
1287 1858
1288The maximum number of child processes that C<AnyEvent::Util::fork_call> 1859The maximum number of child processes that C<AnyEvent::Util::fork_call>
1289will create in parallel. 1860will create in parallel.
1861
1862=item C<PERL_ANYEVENT_MAX_OUTSTANDING_DNS>
1863
1864The default value for the C<max_outstanding> parameter for the default DNS
1865resolver - this is the maximum number of parallel DNS requests that are
1866sent to the DNS server.
1867
1868=item C<PERL_ANYEVENT_RESOLV_CONF>
1869
1870The file to use instead of F</etc/resolv.conf> (or OS-specific
1871configuration) in the default resolver. When set to the empty string, no
1872default config will be used.
1873
1874=item C<PERL_ANYEVENT_CA_FILE>, C<PERL_ANYEVENT_CA_PATH>.
1875
1876When neither C<ca_file> nor C<ca_path> was specified during
1877L<AnyEvent::TLS> context creation, and either of these environment
1878variables exist, they will be used to specify CA certificate locations
1879instead of a system-dependent default.
1880
1881=item C<PERL_ANYEVENT_AVOID_GUARD> and C<PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT>
1882
1883When these are set to C<1>, then the respective modules are not
1884loaded. Mostly good for testing AnyEvent itself.
1290 1885
1291=back 1886=back
1292 1887
1293=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE 1888=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE
1294 1889
1352 warn "read: $input\n"; # output what has been read 1947 warn "read: $input\n"; # output what has been read
1353 $cv->send if $input =~ /^q/i; # quit program if /^q/i 1948 $cv->send if $input =~ /^q/i; # quit program if /^q/i
1354 }, 1949 },
1355 ); 1950 );
1356 1951
1357 my $time_watcher; # can only be used once
1358
1359 sub new_timer {
1360 $timer = AnyEvent->timer (after => 1, cb => sub { 1952 my $time_watcher = AnyEvent->timer (after => 1, interval => 1, cb => sub {
1361 warn "timeout\n"; # print 'timeout' about every second 1953 warn "timeout\n"; # print 'timeout' at most every second
1362 &new_timer; # and restart the time
1363 }); 1954 });
1364 }
1365
1366 new_timer; # create first timer
1367 1955
1368 $cv->recv; # wait until user enters /^q/i 1956 $cv->recv; # wait until user enters /^q/i
1369 1957
1370=head1 REAL-WORLD EXAMPLE 1958=head1 REAL-WORLD EXAMPLE
1371 1959
1502through AnyEvent. The benchmark creates a lot of timers (with a zero 2090through AnyEvent. The benchmark creates a lot of timers (with a zero
1503timeout) and I/O watchers (watching STDOUT, a pty, to become writable, 2091timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
1504which it is), lets them fire exactly once and destroys them again. 2092which it is), lets them fire exactly once and destroys them again.
1505 2093
1506Source code for this benchmark is found as F<eg/bench> in the AnyEvent 2094Source code for this benchmark is found as F<eg/bench> in the AnyEvent
1507distribution. 2095distribution. It uses the L<AE> interface, which makes a real difference
2096for the EV and Perl backends only.
1508 2097
1509=head3 Explanation of the columns 2098=head3 Explanation of the columns
1510 2099
1511I<watcher> is the number of event watchers created/destroyed. Since 2100I<watcher> is the number of event watchers created/destroyed. Since
1512different event models feature vastly different performances, each event 2101different event models feature vastly different performances, each event
1533watcher. 2122watcher.
1534 2123
1535=head3 Results 2124=head3 Results
1536 2125
1537 name watchers bytes create invoke destroy comment 2126 name watchers bytes create invoke destroy comment
1538 EV/EV 400000 224 0.47 0.35 0.27 EV native interface 2127 EV/EV 100000 223 0.47 0.43 0.27 EV native interface
1539 EV/Any 100000 224 2.88 0.34 0.27 EV + AnyEvent watchers 2128 EV/Any 100000 223 0.48 0.42 0.26 EV + AnyEvent watchers
1540 CoroEV/Any 100000 224 2.85 0.35 0.28 coroutines + Coro::Signal 2129 Coro::EV/Any 100000 223 0.47 0.42 0.26 coroutines + Coro::Signal
1541 Perl/Any 100000 452 4.13 0.73 0.95 pure perl implementation 2130 Perl/Any 100000 431 2.70 0.74 0.92 pure perl implementation
1542 Event/Event 16000 517 32.20 31.80 0.81 Event native interface 2131 Event/Event 16000 516 31.16 31.84 0.82 Event native interface
1543 Event/Any 16000 590 35.85 31.55 1.06 Event + AnyEvent watchers 2132 Event/Any 16000 1203 42.61 34.79 1.80 Event + AnyEvent watchers
2133 IOAsync/Any 16000 1911 41.92 27.45 16.81 via IO::Async::Loop::IO_Poll
2134 IOAsync/Any 16000 1726 40.69 26.37 15.25 via IO::Async::Loop::Epoll
1544 Glib/Any 16000 1357 102.33 12.31 51.00 quadratic behaviour 2135 Glib/Any 16000 1118 89.00 12.57 51.17 quadratic behaviour
1545 Tk/Any 2000 1860 27.20 66.31 14.00 SEGV with >> 2000 watchers 2136 Tk/Any 2000 1346 20.96 10.75 8.00 SEGV with >> 2000 watchers
1546 POE/Event 2000 6328 109.99 751.67 14.02 via POE::Loop::Event 2137 POE/Any 2000 6951 108.97 795.32 14.24 via POE::Loop::Event
1547 POE/Select 2000 6027 94.54 809.13 579.80 via POE::Loop::Select 2138 POE/Any 2000 6648 94.79 774.40 575.51 via POE::Loop::Select
1548 2139
1549=head3 Discussion 2140=head3 Discussion
1550 2141
1551The benchmark does I<not> measure scalability of the event loop very 2142The benchmark does I<not> measure scalability of the event loop very
1552well. For example, a select-based event loop (such as the pure perl one) 2143well. For example, a select-based event loop (such as the pure perl one)
1564benchmark machine, handling an event takes roughly 1600 CPU cycles with 2155benchmark machine, handling an event takes roughly 1600 CPU cycles with
1565EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU 2156EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU
1566cycles with POE. 2157cycles with POE.
1567 2158
1568C<EV> is the sole leader regarding speed and memory use, which are both 2159C<EV> is the sole leader regarding speed and memory use, which are both
1569maximal/minimal, respectively. Even when going through AnyEvent, it uses 2160maximal/minimal, respectively. When using the L<AE> API there is zero
2161overhead (when going through the AnyEvent API create is about 5-6 times
2162slower, with other times being equal, so still uses far less memory than
1570far less memory than any other event loop and is still faster than Event 2163any other event loop and is still faster than Event natively).
1571natively.
1572 2164
1573The pure perl implementation is hit in a few sweet spots (both the 2165The pure perl implementation is hit in a few sweet spots (both the
1574constant timeout and the use of a single fd hit optimisations in the perl 2166constant timeout and the use of a single fd hit optimisations in the perl
1575interpreter and the backend itself). Nevertheless this shows that it 2167interpreter and the backend itself). Nevertheless this shows that it
1576adds very little overhead in itself. Like any select-based backend its 2168adds very little overhead in itself. Like any select-based backend its
1577performance becomes really bad with lots of file descriptors (and few of 2169performance becomes really bad with lots of file descriptors (and few of
1578them active), of course, but this was not subject of this benchmark. 2170them active), of course, but this was not subject of this benchmark.
1579 2171
1580The C<Event> module has a relatively high setup and callback invocation 2172The C<Event> module has a relatively high setup and callback invocation
1581cost, but overall scores in on the third place. 2173cost, but overall scores in on the third place.
2174
2175C<IO::Async> performs admirably well, about on par with C<Event>, even
2176when using its pure perl backend.
1582 2177
1583C<Glib>'s memory usage is quite a bit higher, but it features a 2178C<Glib>'s memory usage is quite a bit higher, but it features a
1584faster callback invocation and overall ends up in the same class as 2179faster callback invocation and overall ends up in the same class as
1585C<Event>. However, Glib scales extremely badly, doubling the number of 2180C<Event>. However, Glib scales extremely badly, doubling the number of
1586watchers increases the processing time by more than a factor of four, 2181watchers increases the processing time by more than a factor of four,
1647In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100 2242In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100
1648(1%) are active. This mirrors the activity of large servers with many 2243(1%) are active. This mirrors the activity of large servers with many
1649connections, most of which are idle at any one point in time. 2244connections, most of which are idle at any one point in time.
1650 2245
1651Source code for this benchmark is found as F<eg/bench2> in the AnyEvent 2246Source code for this benchmark is found as F<eg/bench2> in the AnyEvent
1652distribution. 2247distribution. It uses the L<AE> interface, which makes a real difference
2248for the EV and Perl backends only.
1653 2249
1654=head3 Explanation of the columns 2250=head3 Explanation of the columns
1655 2251
1656I<sockets> is the number of sockets, and twice the number of "servers" (as 2252I<sockets> is the number of sockets, and twice the number of "servers" (as
1657each server has a read and write socket end). 2253each server has a read and write socket end).
1664it to another server. This includes deleting the old timeout and creating 2260it to another server. This includes deleting the old timeout and creating
1665a new one that moves the timeout into the future. 2261a new one that moves the timeout into the future.
1666 2262
1667=head3 Results 2263=head3 Results
1668 2264
1669 name sockets create request 2265 name sockets create request
1670 EV 20000 69.01 11.16 2266 EV 20000 62.66 7.99
1671 Perl 20000 73.32 35.87 2267 Perl 20000 68.32 32.64
1672 Event 20000 212.62 257.32 2268 IOAsync 20000 174.06 101.15 epoll
1673 Glib 20000 651.16 1896.30 2269 IOAsync 20000 174.67 610.84 poll
2270 Event 20000 202.69 242.91
2271 Glib 20000 557.01 1689.52
1674 POE 20000 349.67 12317.24 uses POE::Loop::Event 2272 POE 20000 341.54 12086.32 uses POE::Loop::Event
1675 2273
1676=head3 Discussion 2274=head3 Discussion
1677 2275
1678This benchmark I<does> measure scalability and overall performance of the 2276This benchmark I<does> measure scalability and overall performance of the
1679particular event loop. 2277particular event loop.
1681EV is again fastest. Since it is using epoll on my system, the setup time 2279EV is again fastest. Since it is using epoll on my system, the setup time
1682is relatively high, though. 2280is relatively high, though.
1683 2281
1684Perl surprisingly comes second. It is much faster than the C-based event 2282Perl surprisingly comes second. It is much faster than the C-based event
1685loops Event and Glib. 2283loops Event and Glib.
2284
2285IO::Async performs very well when using its epoll backend, and still quite
2286good compared to Glib when using its pure perl backend.
1686 2287
1687Event suffers from high setup time as well (look at its code and you will 2288Event suffers from high setup time as well (look at its code and you will
1688understand why). Callback invocation also has a high overhead compared to 2289understand why). Callback invocation also has a high overhead compared to
1689the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event 2290the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event
1690uses select or poll in basically all documented configurations. 2291uses select or poll in basically all documented configurations.
1753=item * C-based event loops perform very well with small number of 2354=item * C-based event loops perform very well with small number of
1754watchers, as the management overhead dominates. 2355watchers, as the management overhead dominates.
1755 2356
1756=back 2357=back
1757 2358
2359=head2 THE IO::Lambda BENCHMARK
2360
2361Recently I was told about the benchmark in the IO::Lambda manpage, which
2362could be misinterpreted to make AnyEvent look bad. In fact, the benchmark
2363simply compares IO::Lambda with POE, and IO::Lambda looks better (which
2364shouldn't come as a surprise to anybody). As such, the benchmark is
2365fine, and mostly shows that the AnyEvent backend from IO::Lambda isn't
2366very optimal. But how would AnyEvent compare when used without the extra
2367baggage? To explore this, I wrote the equivalent benchmark for AnyEvent.
2368
2369The benchmark itself creates an echo-server, and then, for 500 times,
2370connects to the echo server, sends a line, waits for the reply, and then
2371creates the next connection. This is a rather bad benchmark, as it doesn't
2372test the efficiency of the framework or much non-blocking I/O, but it is a
2373benchmark nevertheless.
2374
2375 name runtime
2376 Lambda/select 0.330 sec
2377 + optimized 0.122 sec
2378 Lambda/AnyEvent 0.327 sec
2379 + optimized 0.138 sec
2380 Raw sockets/select 0.077 sec
2381 POE/select, components 0.662 sec
2382 POE/select, raw sockets 0.226 sec
2383 POE/select, optimized 0.404 sec
2384
2385 AnyEvent/select/nb 0.085 sec
2386 AnyEvent/EV/nb 0.068 sec
2387 +state machine 0.134 sec
2388
2389The benchmark is also a bit unfair (my fault): the IO::Lambda/POE
2390benchmarks actually make blocking connects and use 100% blocking I/O,
2391defeating the purpose of an event-based solution. All of the newly
2392written AnyEvent benchmarks use 100% non-blocking connects (using
2393AnyEvent::Socket::tcp_connect and the asynchronous pure perl DNS
2394resolver), so AnyEvent is at a disadvantage here, as non-blocking connects
2395generally require a lot more bookkeeping and event handling than blocking
2396connects (which involve a single syscall only).
2397
2398The last AnyEvent benchmark additionally uses L<AnyEvent::Handle>, which
2399offers similar expressive power as POE and IO::Lambda, using conventional
2400Perl syntax. This means that both the echo server and the client are 100%
2401non-blocking, further placing it at a disadvantage.
2402
2403As you can see, the AnyEvent + EV combination even beats the
2404hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl
2405backend easily beats IO::Lambda and POE.
2406
2407And even the 100% non-blocking version written using the high-level (and
2408slow :) L<AnyEvent::Handle> abstraction beats both POE and IO::Lambda
2409higher level ("unoptimised") abstractions by a large margin, even though
2410it does all of DNS, tcp-connect and socket I/O in a non-blocking way.
2411
2412The two AnyEvent benchmarks programs can be found as F<eg/ae0.pl> and
2413F<eg/ae2.pl> in the AnyEvent distribution, the remaining benchmarks are
2414part of the IO::Lambda distribution and were used without any changes.
2415
1758 2416
1759=head1 SIGNALS 2417=head1 SIGNALS
1760 2418
1761AnyEvent currently installs handlers for these signals: 2419AnyEvent currently installs handlers for these signals:
1762 2420
1765=item SIGCHLD 2423=item SIGCHLD
1766 2424
1767A handler for C<SIGCHLD> is installed by AnyEvent's child watcher 2425A handler for C<SIGCHLD> is installed by AnyEvent's child watcher
1768emulation for event loops that do not support them natively. Also, some 2426emulation for event loops that do not support them natively. Also, some
1769event loops install a similar handler. 2427event loops install a similar handler.
2428
2429Additionally, when AnyEvent is loaded and SIGCHLD is set to IGNORE, then
2430AnyEvent will reset it to default, to avoid losing child exit statuses.
1770 2431
1771=item SIGPIPE 2432=item SIGPIPE
1772 2433
1773A no-op handler is installed for C<SIGPIPE> when C<$SIG{PIPE}> is C<undef> 2434A no-op handler is installed for C<SIGPIPE> when C<$SIG{PIPE}> is C<undef>
1774when AnyEvent gets loaded. 2435when AnyEvent gets loaded.
1786 2447
1787=back 2448=back
1788 2449
1789=cut 2450=cut
1790 2451
2452undef $SIG{CHLD}
2453 if $SIG{CHLD} eq 'IGNORE';
2454
1791$SIG{PIPE} = sub { } 2455$SIG{PIPE} = sub { }
1792 unless defined $SIG{PIPE}; 2456 unless defined $SIG{PIPE};
1793 2457
2458=head1 RECOMMENDED/OPTIONAL MODULES
2459
2460One of AnyEvent's main goals is to be 100% Pure-Perl(tm): only perl (and
2461it's built-in modules) are required to use it.
2462
2463That does not mean that AnyEvent won't take advantage of some additional
2464modules if they are installed.
2465
2466This section explains which additional modules will be used, and how they
2467affect AnyEvent's operation.
2468
2469=over 4
2470
2471=item L<Async::Interrupt>
2472
2473This slightly arcane module is used to implement fast signal handling: To
2474my knowledge, there is no way to do completely race-free and quick
2475signal handling in pure perl. To ensure that signals still get
2476delivered, AnyEvent will start an interval timer to wake up perl (and
2477catch the signals) with some delay (default is 10 seconds, look for
2478C<$AnyEvent::MAX_SIGNAL_LATENCY>).
2479
2480If this module is available, then it will be used to implement signal
2481catching, which means that signals will not be delayed, and the event loop
2482will not be interrupted regularly, which is more efficient (and good for
2483battery life on laptops).
2484
2485This affects not just the pure-perl event loop, but also other event loops
2486that have no signal handling on their own (e.g. Glib, Tk, Qt).
2487
2488Some event loops (POE, Event, Event::Lib) offer signal watchers natively,
2489and either employ their own workarounds (POE) or use AnyEvent's workaround
2490(using C<$AnyEvent::MAX_SIGNAL_LATENCY>). Installing L<Async::Interrupt>
2491does nothing for those backends.
2492
2493=item L<EV>
2494
2495This module isn't really "optional", as it is simply one of the backend
2496event loops that AnyEvent can use. However, it is simply the best event
2497loop available in terms of features, speed and stability: It supports
2498the AnyEvent API optimally, implements all the watcher types in XS, does
2499automatic timer adjustments even when no monotonic clock is available,
2500can take avdantage of advanced kernel interfaces such as C<epoll> and
2501C<kqueue>, and is the fastest backend I<by far>. You can even embed
2502L<Glib>/L<Gtk2> in it (or vice versa, see L<EV::Glib> and L<Glib::EV>).
2503
2504=item L<Guard>
2505
2506The guard module, when used, will be used to implement
2507C<AnyEvent::Util::guard>. This speeds up guards considerably (and uses a
2508lot less memory), but otherwise doesn't affect guard operation much. It is
2509purely used for performance.
2510
2511=item L<JSON> and L<JSON::XS>
2512
2513One of these modules is required when you want to read or write JSON data
2514via L<AnyEvent::Handle>. It is also written in pure-perl, but can take
2515advantage of the ultra-high-speed L<JSON::XS> module when it is installed.
2516
2517In fact, L<AnyEvent::Handle> will use L<JSON::XS> by default if it is
2518installed.
2519
2520=item L<Net::SSLeay>
2521
2522Implementing TLS/SSL in Perl is certainly interesting, but not very
2523worthwhile: If this module is installed, then L<AnyEvent::Handle> (with
2524the help of L<AnyEvent::TLS>), gains the ability to do TLS/SSL.
2525
2526=item L<Time::HiRes>
2527
2528This module is part of perl since release 5.008. It will be used when the
2529chosen event library does not come with a timing source on it's own. The
2530pure-perl event loop (L<AnyEvent::Impl::Perl>) will additionally use it to
2531try to use a monotonic clock for timing stability.
2532
2533=back
2534
1794 2535
1795=head1 FORK 2536=head1 FORK
1796 2537
1797Most event libraries are not fork-safe. The ones who are usually are 2538Most event libraries are not fork-safe. The ones who are usually are
1798because they rely on inefficient but fork-safe C<select> or C<poll> 2539because they rely on inefficient but fork-safe C<select> or C<poll> calls
1799calls. Only L<EV> is fully fork-aware. 2540- higher performance APIs such as BSD's kqueue or the dreaded Linux epoll
2541are usually badly thought-out hacks that are incompatible with fork in
2542one way or another. Only L<EV> is fully fork-aware and ensures that you
2543continue event-processing in both parent and child (or both, if you know
2544what you are doing).
2545
2546This means that, in general, you cannot fork and do event processing in
2547the child if the event library was initialised before the fork (which
2548usually happens when the first AnyEvent watcher is created, or the library
2549is loaded).
1800 2550
1801If you have to fork, you must either do so I<before> creating your first 2551If you have to fork, you must either do so I<before> creating your first
1802watcher OR you must not use AnyEvent at all in the child. 2552watcher OR you must not use AnyEvent at all in the child OR you must do
2553something completely out of the scope of AnyEvent.
2554
2555The problem of doing event processing in the parent I<and> the child
2556is much more complicated: even for backends that I<are> fork-aware or
2557fork-safe, their behaviour is not usually what you want: fork clones all
2558watchers, that means all timers, I/O watchers etc. are active in both
2559parent and child, which is almost never what you want. USing C<exec>
2560to start worker children from some kind of manage rprocess is usually
2561preferred, because it is much easier and cleaner, at the expense of having
2562to have another binary.
1803 2563
1804 2564
1805=head1 SECURITY CONSIDERATIONS 2565=head1 SECURITY CONSIDERATIONS
1806 2566
1807AnyEvent can be forced to load any event model via 2567AnyEvent can be forced to load any event model via
1819 use AnyEvent; 2579 use AnyEvent;
1820 2580
1821Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can 2581Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
1822be used to probe what backend is used and gain other information (which is 2582be used to probe what backend is used and gain other information (which is
1823probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and 2583probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and
1824$ENV{PERL_ANYEGENT_STRICT}. 2584$ENV{PERL_ANYEVENT_STRICT}.
2585
2586Note that AnyEvent will remove I<all> environment variables starting with
2587C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
2588enabled.
1825 2589
1826 2590
1827=head1 BUGS 2591=head1 BUGS
1828 2592
1829Perl 5.8 has numerous memleaks that sometimes hit this module and are hard 2593Perl 5.8 has numerous memleaks that sometimes hit this module and are hard
1830to work around. If you suffer from memleaks, first upgrade to Perl 5.10 2594to work around. If you suffer from memleaks, first upgrade to Perl 5.10
1831and check wether the leaks still show up. (Perl 5.10.0 has other annoying 2595and check wether the leaks still show up. (Perl 5.10.0 has other annoying
1832mamleaks, such as leaking on C<map> and C<grep> but it is usually not as 2596memleaks, such as leaking on C<map> and C<grep> but it is usually not as
1833pronounced). 2597pronounced).
1834 2598
1835 2599
1836=head1 SEE ALSO 2600=head1 SEE ALSO
1837 2601
1841L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>. 2605L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>.
1842 2606
1843Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>, 2607Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>,
1844L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>, 2608L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>,
1845L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>, 2609L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>,
1846L<AnyEvent::Impl::POE>. 2610L<AnyEvent::Impl::POE>, L<AnyEvent::Impl::IOAsync>, L<Anyevent::Impl::Irssi>.
1847 2611
1848Non-blocking file handles, sockets, TCP clients and 2612Non-blocking file handles, sockets, TCP clients and
1849servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>. 2613servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>, L<AnyEvent::TLS>.
1850 2614
1851Asynchronous DNS: L<AnyEvent::DNS>. 2615Asynchronous DNS: L<AnyEvent::DNS>.
1852 2616
1853Coroutine support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>, 2617Coroutine support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>,
2618L<Coro::Event>,
1854 2619
1855Nontrivial usage examples: L<Net::FCP>, L<Net::XMPP2>, L<AnyEvent::DNS>. 2620Nontrivial usage examples: L<AnyEvent::GPSD>, L<AnyEvent::XMPP>,
2621L<AnyEvent::HTTP>.
1856 2622
1857 2623
1858=head1 AUTHOR 2624=head1 AUTHOR
1859 2625
1860 Marc Lehmann <schmorp@schmorp.de> 2626 Marc Lehmann <schmorp@schmorp.de>

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines