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 1.132 by root, Sun May 25 01:05:27 2008 UTC vs.
Revision 1.219 by root, Thu Jun 25 11:16:08 2009 UTC

1=head1 => NAME 1=head1 NAME
2 2
3AnyEvent - provide framework for multiple event loops 3AnyEvent - provide framework for multiple event loops
4 4
5EV, Event, Glib, Tk, Perl, Event::Lib, Qt, POE - various supported event loops 5EV, Event, Glib, Tk, Perl, Event::Lib, Qt and POE are various supported
6event loops.
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 { ... });
14
15 # one-shot or repeating timers
16 my $w = AnyEvent->timer (after => $seconds, cb => sub { ... });
17 my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ...
18
19 print AnyEvent->now; # prints current event loop time
20 print AnyEvent->time; # think Time::HiRes::time or simply CORE::time.
21
22 # POSIX signal
23 my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... });
24
25 # child process exit
26 my $w = AnyEvent->child (pid => $pid, cb => sub {
27 my ($pid, $status) = @_;
12 ... 28 ...
13 }); 29 });
14 30
15 my $w = AnyEvent->timer (after => $seconds, cb => sub { 31 # called when event loop idle (if applicable)
16 ... 32 my $w = AnyEvent->idle (cb => sub { ... });
17 });
18 33
19 my $w = AnyEvent->condvar; # stores whether a condition was flagged 34 my $w = AnyEvent->condvar; # stores whether a condition was flagged
20 $w->send; # wake up current and all future recv's 35 $w->send; # wake up current and all future recv's
21 $w->recv; # enters "main loop" till $condvar gets ->send 36 $w->recv; # enters "main loop" till $condvar gets ->send
37 # use a condvar in callback mode:
38 $w->cb (sub { $_[0]->recv });
39
40=head1 INTRODUCTION/TUTORIAL
41
42This manpage is mainly a reference manual. If you are interested
43in a tutorial or some gentle introduction, have a look at the
44L<AnyEvent::Intro> manpage.
22 45
23=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT) 46=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)
24 47
25Glib, POE, IO::Async, Event... CPAN offers event models by the dozen 48Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
26nowadays. So what is different about AnyEvent? 49nowadays. So what is different about AnyEvent?
27 50
28Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of 51Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of
29policy> and AnyEvent is I<small and efficient>. 52policy> and AnyEvent is I<small and efficient>.
30 53
31First and foremost, I<AnyEvent is not an event model> itself, it only 54First and foremost, I<AnyEvent is not an event model> itself, it only
32interfaces to whatever event model the main program happens to use in a 55interfaces to whatever event model the main program happens to use, in a
33pragmatic way. For event models and certain classes of immortals alike, 56pragmatic way. For event models and certain classes of immortals alike,
34the statement "there can only be one" is a bitter reality: In general, 57the statement "there can only be one" is a bitter reality: In general,
35only one event loop can be active at the same time in a process. AnyEvent 58only one event loop can be active at the same time in a process. AnyEvent
36helps hiding the differences between those event loops. 59cannot change this, but it can hide the differences between those event
60loops.
37 61
38The goal of AnyEvent is to offer module authors the ability to do event 62The goal of AnyEvent is to offer module authors the ability to do event
39programming (waiting for I/O or timer events) without subscribing to a 63programming (waiting for I/O or timer events) without subscribing to a
40religion, a way of living, and most importantly: without forcing your 64religion, a way of living, and most importantly: without forcing your
41module users into the same thing by forcing them to use the same event 65module users into the same thing by forcing them to use the same event
42model you use. 66model you use.
43 67
44For modules like POE or IO::Async (which is a total misnomer as it is 68For modules like POE or IO::Async (which is a total misnomer as it is
45actually doing all I/O I<synchronously>...), using them in your module is 69actually doing all I/O I<synchronously>...), using them in your module is
46like joining a cult: After you joined, you are dependent on them and you 70like joining a cult: After you joined, you are dependent on them and you
47cannot use anything else, as it is simply incompatible to everything that 71cannot use anything else, as they are simply incompatible to everything
48isn't itself. What's worse, all the potential users of your module are 72that isn't them. What's worse, all the potential users of your
49I<also> forced to use the same event loop you use. 73module are I<also> forced to use the same event loop you use.
50 74
51AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works 75AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
52fine. AnyEvent + Tk works fine etc. etc. but none of these work together 76fine. AnyEvent + Tk works fine etc. etc. but none of these work together
53with the rest: POE + IO::Async? no go. Tk + Event? no go. Again: if 77with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if
54your module uses one of those, every user of your module has to use it, 78your module uses one of those, every user of your module has to use it,
55too. But if your module uses AnyEvent, it works transparently with all 79too. But if your module uses AnyEvent, it works transparently with all
56event models it supports (including stuff like POE and IO::Async, as long 80event models it supports (including stuff like IO::Async, as long as those
57as those use one of the supported event loops. It is trivial to add new 81use one of the supported event loops. It is trivial to add new event loops
58event loops to AnyEvent, too, so it is future-proof). 82to AnyEvent, too, so it is future-proof).
59 83
60In addition to being free of having to use I<the one and only true event 84In addition to being free of having to use I<the one and only true event
61model>, AnyEvent also is free of bloat and policy: with POE or similar 85model>, AnyEvent also is free of bloat and policy: with POE or similar
62modules, you get an enormous amount of code and strict rules you have to 86modules, you get an enormous amount of code and strict rules you have to
63follow. AnyEvent, on the other hand, is lean and up to the point, by only 87follow. AnyEvent, on the other hand, is lean and up to the point, by only
64offering the functionality that is necessary, in as thin as a wrapper as 88offering the functionality that is necessary, in as thin as a wrapper as
65technically possible. 89technically possible.
66 90
91Of course, AnyEvent comes with a big (and fully optional!) toolbox
92of useful functionality, such as an asynchronous DNS resolver, 100%
93non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
94such as Windows) and lots of real-world knowledge and workarounds for
95platform bugs and differences.
96
67Of course, if you want lots of policy (this can arguably be somewhat 97Now, if you I<do want> lots of policy (this can arguably be somewhat
68useful) and you want to force your users to use the one and only event 98useful) and you want to force your users to use the one and only event
69model, you should I<not> use this module. 99model, you should I<not> use this module.
70 100
71=head1 DESCRIPTION 101=head1 DESCRIPTION
72 102
102starts using it, all bets are off. Maybe you should tell their authors to 132starts using it, all bets are off. Maybe you should tell their authors to
103use AnyEvent so their modules work together with others seamlessly... 133use AnyEvent so their modules work together with others seamlessly...
104 134
105The pure-perl implementation of AnyEvent is called 135The pure-perl implementation of AnyEvent is called
106C<AnyEvent::Impl::Perl>. Like other event modules you can load it 136C<AnyEvent::Impl::Perl>. Like other event modules you can load it
107explicitly. 137explicitly and enjoy the high availability of that event loop :)
108 138
109=head1 WATCHERS 139=head1 WATCHERS
110 140
111AnyEvent has the central concept of a I<watcher>, which is an object that 141AnyEvent has the central concept of a I<watcher>, which is an object that
112stores relevant data for each kind of event you are waiting for, such as 142stores relevant data for each kind of event you are waiting for, such as
115These watchers are normal Perl objects with normal Perl lifetime. After 145These watchers are normal Perl objects with normal Perl lifetime. After
116creating a watcher it will immediately "watch" for events and invoke the 146creating a watcher it will immediately "watch" for events and invoke the
117callback when the event occurs (of course, only when the event model 147callback when the event occurs (of course, only when the event model
118is in control). 148is in control).
119 149
150Note that B<callbacks must not permanently change global variables>
151potentially in use by the event loop (such as C<$_> or C<$[>) and that B<<
152callbacks must not C<die> >>. The former is good programming practise in
153Perl and the latter stems from the fact that exception handling differs
154widely between event loops.
155
120To disable the watcher you have to destroy it (e.g. by setting the 156To disable the watcher you have to destroy it (e.g. by setting the
121variable you store it in to C<undef> or otherwise deleting all references 157variable you store it in to C<undef> or otherwise deleting all references
122to it). 158to it).
123 159
124All watchers are created by calling a method on the C<AnyEvent> class. 160All watchers are created by calling a method on the C<AnyEvent> class.
126Many watchers either are used with "recursion" (repeating timers for 162Many watchers either are used with "recursion" (repeating timers for
127example), or need to refer to their watcher object in other ways. 163example), or need to refer to their watcher object in other ways.
128 164
129An any way to achieve that is this pattern: 165An any way to achieve that is this pattern:
130 166
131 my $w; $w = AnyEvent->type (arg => value ..., cb => sub { 167 my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
132 # you can use $w here, for example to undef it 168 # you can use $w here, for example to undef it
133 undef $w; 169 undef $w;
134 }); 170 });
135 171
136Note that C<my $w; $w => combination. This is necessary because in Perl, 172Note that C<my $w; $w => combination. This is necessary because in Perl,
137my variables are only visible after the statement in which they are 173my variables are only visible after the statement in which they are
138declared. 174declared.
139 175
140=head2 I/O WATCHERS 176=head2 I/O WATCHERS
141 177
142You can create an I/O watcher by calling the C<< AnyEvent->io >> method 178You can create an I/O watcher by calling the C<< AnyEvent->io >> method
143with the following mandatory key-value pairs as arguments: 179with the following mandatory key-value pairs as arguments:
144 180
145C<fh> the Perl I<file handle> (I<not> file descriptor) to watch 181C<fh> is the Perl I<file handle> (I<not> file descriptor) to watch
182for events (AnyEvent might or might not keep a reference to this file
183handle). Note that only file handles pointing to things for which
184non-blocking operation makes sense are allowed. This includes sockets,
185most character devices, pipes, fifos and so on, but not for example files
186or block devices.
187
146for events. C<poll> must be a string that is either C<r> or C<w>, 188C<poll> must be a string that is either C<r> or C<w>, which creates a
147which creates a watcher waiting for "r"eadable or "w"ritable events, 189watcher waiting for "r"eadable or "w"ritable events, respectively.
190
148respectively. C<cb> is the callback to invoke each time the file handle 191C<cb> is the callback to invoke each time the file handle becomes ready.
149becomes ready.
150 192
151Although the callback might get passed parameters, their value and 193Although the callback might get passed parameters, their value and
152presence is undefined and you cannot rely on them. Portable AnyEvent 194presence is undefined and you cannot rely on them. Portable AnyEvent
153callbacks cannot use arguments passed to I/O watcher callbacks. 195callbacks cannot use arguments passed to I/O watcher callbacks.
154 196
158 200
159Some event loops issue spurious readyness notifications, so you should 201Some event loops issue spurious readyness notifications, so you should
160always use non-blocking calls when reading/writing from/to your file 202always use non-blocking calls when reading/writing from/to your file
161handles. 203handles.
162 204
163Example:
164
165 # wait for readability of STDIN, then read a line and disable the watcher 205Example: wait for readability of STDIN, then read a line and disable the
206watcher.
207
166 my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub { 208 my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
167 chomp (my $input = <STDIN>); 209 chomp (my $input = <STDIN>);
168 warn "read: $input\n"; 210 warn "read: $input\n";
169 undef $w; 211 undef $w;
170 }); 212 });
180 222
181Although the callback might get passed parameters, their value and 223Although the callback might get passed parameters, their value and
182presence is undefined and you cannot rely on them. Portable AnyEvent 224presence is undefined and you cannot rely on them. Portable AnyEvent
183callbacks cannot use arguments passed to time watcher callbacks. 225callbacks cannot use arguments passed to time watcher callbacks.
184 226
185The timer callback will be invoked at most once: if you want a repeating 227The callback will normally be invoked once only. If you specify another
186timer you have to create a new watcher (this is a limitation by both Tk 228parameter, C<interval>, as a strictly positive number (> 0), then the
187and Glib). 229callback will be invoked regularly at that interval (in fractional
230seconds) after the first invocation. If C<interval> is specified with a
231false value, then it is treated as if it were missing.
188 232
189Example: 233The callback will be rescheduled before invoking the callback, but no
234attempt is done to avoid timer drift in most backends, so the interval is
235only approximate.
190 236
191 # fire an event after 7.7 seconds 237Example: fire an event after 7.7 seconds.
238
192 my $w = AnyEvent->timer (after => 7.7, cb => sub { 239 my $w = AnyEvent->timer (after => 7.7, cb => sub {
193 warn "timeout\n"; 240 warn "timeout\n";
194 }); 241 });
195 242
196 # to cancel the timer: 243 # to cancel the timer:
197 undef $w; 244 undef $w;
198 245
199Example 2:
200
201 # fire an event after 0.5 seconds, then roughly every second 246Example 2: fire an event after 0.5 seconds, then roughly every second.
202 my $w;
203 247
204 my $cb = sub {
205 # cancel the old timer while creating a new one
206 $w = AnyEvent->timer (after => 1, cb => $cb); 248 my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
249 warn "timeout\n";
207 }; 250 };
208
209 # start the "loop" by creating the first watcher
210 $w = AnyEvent->timer (after => 0.5, cb => $cb);
211 251
212=head3 TIMING ISSUES 252=head3 TIMING ISSUES
213 253
214There are two ways to handle timers: based on real time (relative, "fire 254There are two ways to handle timers: based on real time (relative, "fire
215in 10 seconds") and based on wallclock time (absolute, "fire at 12 255in 10 seconds") and based on wallclock time (absolute, "fire at 12
227timers. 267timers.
228 268
229AnyEvent always prefers relative timers, if available, matching the 269AnyEvent always prefers relative timers, if available, matching the
230AnyEvent API. 270AnyEvent API.
231 271
272AnyEvent has two additional methods that return the "current time":
273
274=over 4
275
276=item AnyEvent->time
277
278This returns the "current wallclock time" as a fractional number of
279seconds since the Epoch (the same thing as C<time> or C<Time::HiRes::time>
280return, and the result is guaranteed to be compatible with those).
281
282It progresses independently of any event loop processing, i.e. each call
283will check the system clock, which usually gets updated frequently.
284
285=item AnyEvent->now
286
287This also returns the "current wallclock time", but unlike C<time>, above,
288this value might change only once per event loop iteration, depending on
289the event loop (most return the same time as C<time>, above). This is the
290time that AnyEvent's timers get scheduled against.
291
292I<In almost all cases (in all cases if you don't care), this is the
293function to call when you want to know the current time.>
294
295This function is also often faster then C<< AnyEvent->time >>, and
296thus the preferred method if you want some timestamp (for example,
297L<AnyEvent::Handle> uses this to update it's activity timeouts).
298
299The rest of this section is only of relevance if you try to be very exact
300with your timing, you can skip it without bad conscience.
301
302For a practical example of when these times differ, consider L<Event::Lib>
303and L<EV> and the following set-up:
304
305The event loop is running and has just invoked one of your callback at
306time=500 (assume no other callbacks delay processing). In your callback,
307you wait a second by executing C<sleep 1> (blocking the process for a
308second) and then (at time=501) you create a relative timer that fires
309after three seconds.
310
311With L<Event::Lib>, C<< AnyEvent->time >> and C<< AnyEvent->now >> will
312both return C<501>, because that is the current time, and the timer will
313be scheduled to fire at time=504 (C<501> + C<3>).
314
315With L<EV>, C<< AnyEvent->time >> returns C<501> (as that is the current
316time), but C<< AnyEvent->now >> returns C<500>, as that is the time the
317last event processing phase started. With L<EV>, your timer gets scheduled
318to run at time=503 (C<500> + C<3>).
319
320In one sense, L<Event::Lib> is more exact, as it uses the current time
321regardless of any delays introduced by event processing. However, most
322callbacks do not expect large delays in processing, so this causes a
323higher drift (and a lot more system calls to get the current time).
324
325In another sense, L<EV> is more exact, as your timer will be scheduled at
326the same time, regardless of how long event processing actually took.
327
328In either case, if you care (and in most cases, you don't), then you
329can get whatever behaviour you want with any event loop, by taking the
330difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into
331account.
332
333=item AnyEvent->now_update
334
335Some event loops (such as L<EV> or L<AnyEvent::Impl::Perl>) cache
336the current time for each loop iteration (see the discussion of L<<
337AnyEvent->now >>, above).
338
339When a callback runs for a long time (or when the process sleeps), then
340this "current" time will differ substantially from the real time, which
341might affect timers and time-outs.
342
343When this is the case, you can call this method, which will update the
344event loop's idea of "current time".
345
346Note that updating the time I<might> cause some events to be handled.
347
348=back
349
232=head2 SIGNAL WATCHERS 350=head2 SIGNAL WATCHERS
233 351
234You can watch for signals using a signal watcher, C<signal> is the signal 352You can watch for signals using a signal watcher, C<signal> is the signal
235I<name> without any C<SIG> prefix, C<cb> is the Perl callback to 353I<name> in uppercase and without any C<SIG> prefix, C<cb> is the Perl
236be invoked whenever a signal occurs. 354callback to be invoked whenever a signal occurs.
237 355
238Although the callback might get passed parameters, their value and 356Although the callback might get passed parameters, their value and
239presence is undefined and you cannot rely on them. Portable AnyEvent 357presence is undefined and you cannot rely on them. Portable AnyEvent
240callbacks cannot use arguments passed to signal watcher callbacks. 358callbacks cannot use arguments passed to signal watcher callbacks.
241 359
257=head2 CHILD PROCESS WATCHERS 375=head2 CHILD PROCESS WATCHERS
258 376
259You can also watch on a child process exit and catch its exit status. 377You can also watch on a child process exit and catch its exit status.
260 378
261The child process is specified by the C<pid> argument (if set to C<0>, it 379The child process is specified by the C<pid> argument (if set to C<0>, it
262watches for any child process exit). The watcher will trigger as often 380watches for any child process exit). The watcher will triggered only when
263as status change for the child are received. This works by installing a 381the child process has finished and an exit status is available, not on
264signal handler for C<SIGCHLD>. The callback will be called with the pid 382any trace events (stopped/continued).
265and exit status (as returned by waitpid), so unlike other watcher types, 383
266you I<can> rely on child watcher callback arguments. 384The callback will be called with the pid and exit status (as returned by
385waitpid), so unlike other watcher types, you I<can> rely on child watcher
386callback arguments.
387
388This watcher type works by installing a signal handler for C<SIGCHLD>,
389and since it cannot be shared, nothing else should use SIGCHLD or reap
390random child processes (waiting for specific child processes, e.g. inside
391C<system>, is just fine).
267 392
268There is a slight catch to child watchers, however: you usually start them 393There is a slight catch to child watchers, however: you usually start them
269I<after> the child process was created, and this means the process could 394I<after> the child process was created, and this means the process could
270have exited already (and no SIGCHLD will be sent anymore). 395have exited already (and no SIGCHLD will be sent anymore).
271 396
272Not all event models handle this correctly (POE doesn't), but even for 397Not all event models handle this correctly (neither POE nor IO::Async do,
398see their AnyEvent::Impl manpages for details), but even for event models
273event models that I<do> handle this correctly, they usually need to be 399that I<do> handle this correctly, they usually need to be loaded before
274loaded before the process exits (i.e. before you fork in the first place). 400the process exits (i.e. before you fork in the first place). AnyEvent's
401pure perl event loop handles all cases correctly regardless of when you
402start the watcher.
275 403
276This means you cannot create a child watcher as the very first thing in an 404This means you cannot create a child watcher as the very first
277AnyEvent program, you I<have> to create at least one watcher before you 405thing in an AnyEvent program, you I<have> to create at least one
278C<fork> the child (alternatively, you can call C<AnyEvent::detect>). 406watcher before you C<fork> the child (alternatively, you can call
407C<AnyEvent::detect>).
279 408
280Example: fork a process and wait for it 409Example: fork a process and wait for it
281 410
282 my $done = AnyEvent->condvar; 411 my $done = AnyEvent->condvar;
283 412
284 my $pid = fork or exit 5; 413 my $pid = fork or exit 5;
285 414
286 my $w = AnyEvent->child ( 415 my $w = AnyEvent->child (
287 pid => $pid, 416 pid => $pid,
288 cb => sub { 417 cb => sub {
289 my ($pid, $status) = @_; 418 my ($pid, $status) = @_;
290 warn "pid $pid exited with status $status"; 419 warn "pid $pid exited with status $status";
291 $done->send; 420 $done->send;
292 }, 421 },
293 ); 422 );
294 423
295 # do something else, then wait for process exit 424 # do something else, then wait for process exit
296 $done->recv; 425 $done->recv;
426
427=head2 IDLE WATCHERS
428
429Sometimes there is a need to do something, but it is not so important
430to do it instantly, but only when there is nothing better to do. This
431"nothing better to do" is usually defined to be "no other events need
432attention by the event loop".
433
434Idle watchers ideally get invoked when the event loop has nothing
435better to do, just before it would block the process to wait for new
436events. Instead of blocking, the idle watcher is invoked.
437
438Most event loops unfortunately do not really support idle watchers (only
439EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent
440will simply call the callback "from time to time".
441
442Example: read lines from STDIN, but only process them when the
443program is otherwise idle:
444
445 my @lines; # read data
446 my $idle_w;
447 my $io_w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
448 push @lines, scalar <STDIN>;
449
450 # start an idle watcher, if not already done
451 $idle_w ||= AnyEvent->idle (cb => sub {
452 # handle only one line, when there are lines left
453 if (my $line = shift @lines) {
454 print "handled when idle: $line";
455 } else {
456 # otherwise disable the idle watcher again
457 undef $idle_w;
458 }
459 });
460 });
297 461
298=head2 CONDITION VARIABLES 462=head2 CONDITION VARIABLES
299 463
300If you are familiar with some event loops you will know that all of them 464If you are familiar with some event loops you will know that all of them
301require you to run some blocking "loop", "run" or similar function that 465require you to run some blocking "loop", "run" or similar function that
307The instrument to do that is called a "condition variable", so called 471The instrument to do that is called a "condition variable", so called
308because they represent a condition that must become true. 472because they represent a condition that must become true.
309 473
310Condition variables can be created by calling the C<< AnyEvent->condvar 474Condition variables can be created by calling the C<< AnyEvent->condvar
311>> method, usually without arguments. The only argument pair allowed is 475>> method, usually without arguments. The only argument pair allowed is
476
312C<cb>, which specifies a callback to be called when the condition variable 477C<cb>, which specifies a callback to be called when the condition variable
313becomes true. 478becomes true, with the condition variable as the first argument (but not
479the results).
314 480
315After creation, the condition variable is "false" until it becomes "true" 481After creation, the condition variable is "false" until it becomes "true"
316by calling the C<send> method (or calling the condition variable as if it 482by calling the C<send> method (or calling the condition variable as if it
317were a callback). 483were a callback, read about the caveats in the description for the C<<
484->send >> method).
318 485
319Condition variables are similar to callbacks, except that you can 486Condition variables are similar to callbacks, except that you can
320optionally wait for them. They can also be called merge points - points 487optionally wait for them. They can also be called merge points - points
321in time where multiple outstanding events have been processed. And yet 488in time where multiple outstanding events have been processed. And yet
322another way to call them is transactions - each condition variable can be 489another way to call them is transactions - each condition variable can be
373 540
374 my $done = AnyEvent->condvar; 541 my $done = AnyEvent->condvar;
375 my $delay = AnyEvent->timer (after => 5, cb => $done); 542 my $delay = AnyEvent->timer (after => 5, cb => $done);
376 $done->recv; 543 $done->recv;
377 544
545Example: Imagine an API that returns a condvar and doesn't support
546callbacks. This is how you make a synchronous call, for example from
547the main program:
548
549 use AnyEvent::CouchDB;
550
551 ...
552
553 my @info = $couchdb->info->recv;
554
555And this is how you would just ste a callback to be called whenever the
556results are available:
557
558 $couchdb->info->cb (sub {
559 my @info = $_[0]->recv;
560 });
561
378=head3 METHODS FOR PRODUCERS 562=head3 METHODS FOR PRODUCERS
379 563
380These methods should only be used by the producing side, i.e. the 564These methods should only be used by the producing side, i.e. the
381code/module that eventually sends the signal. Note that it is also 565code/module that eventually sends the signal. Note that it is also
382the producer side which creates the condvar in most cases, but it isn't 566the producer side which creates the condvar in most cases, but it isn't
394immediately from within send. 578immediately from within send.
395 579
396Any arguments passed to the C<send> call will be returned by all 580Any arguments passed to the C<send> call will be returned by all
397future C<< ->recv >> calls. 581future C<< ->recv >> calls.
398 582
399Condition variables are overloaded so one can call them directly (as a 583Condition variables are overloaded so one can call them directly
400code reference). Calling them directly is the same as calling C<send>. 584(as a code reference). Calling them directly is the same as calling
585C<send>. Note, however, that many C-based event loops do not handle
586overloading, so as tempting as it may be, passing a condition variable
587instead of a callback does not work. Both the pure perl and EV loops
588support overloading, however, as well as all functions that use perl to
589invoke a callback (as in L<AnyEvent::Socket> and L<AnyEvent::DNS> for
590example).
401 591
402=item $cv->croak ($error) 592=item $cv->croak ($error)
403 593
404Similar to send, but causes all call's to C<< ->recv >> to invoke 594Similar to send, but causes all call's to C<< ->recv >> to invoke
405C<Carp::croak> with the given error message/object/scalar. 595C<Carp::croak> with the given error message/object/scalar.
509=item $bool = $cv->ready 699=item $bool = $cv->ready
510 700
511Returns true when the condition is "true", i.e. whether C<send> or 701Returns true when the condition is "true", i.e. whether C<send> or
512C<croak> have been called. 702C<croak> have been called.
513 703
514=item $cb = $cv->cb ([new callback]) 704=item $cb = $cv->cb ($cb->($cv))
515 705
516This is a mutator function that returns the callback set and optionally 706This is a mutator function that returns the callback set and optionally
517replaces it before doing so. 707replaces it before doing so.
518 708
519The callback will be called when the condition becomes "true", i.e. when 709The callback will be called when the condition becomes "true", i.e. when
520C<send> or C<croak> are called. Calling C<recv> inside the callback 710C<send> or C<croak> are called, with the only argument being the condition
521or at any later time is guaranteed not to block. 711variable itself. Calling C<recv> inside the callback or at any later time
712is guaranteed not to block.
522 713
523=back 714=back
524 715
525=head1 GLOBAL VARIABLES AND FUNCTIONS 716=head1 GLOBAL VARIABLES AND FUNCTIONS
526 717
543 AnyEvent::Impl::Tk based on Tk, very bad choice. 734 AnyEvent::Impl::Tk based on Tk, very bad choice.
544 AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs). 735 AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs).
545 AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse. 736 AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
546 AnyEvent::Impl::POE based on POE, not generic enough for full support. 737 AnyEvent::Impl::POE based on POE, not generic enough for full support.
547 738
739 # warning, support for IO::Async is only partial, as it is too broken
740 # and limited toe ven support the AnyEvent API. See AnyEvent::Impl::Async.
741 AnyEvent::Impl::IOAsync based on IO::Async, cannot be autoprobed (see its docs).
742
548There is no support for WxWidgets, as WxWidgets has no support for 743There is no support for WxWidgets, as WxWidgets has no support for
549watching file handles. However, you can use WxWidgets through the 744watching file handles. However, you can use WxWidgets through the
550POE Adaptor, as POE has a Wx backend that simply polls 20 times per 745POE Adaptor, as POE has a Wx backend that simply polls 20 times per
551second, which was considered to be too horrible to even consider for 746second, which was considered to be too horrible to even consider for
552AnyEvent. Likewise, other POE backends can be used by AnyEvent by using 747AnyEvent. Likewise, other POE backends can be used by AnyEvent by using
612 807
613If it doesn't care, it can just "use AnyEvent" and use it itself, or not 808If it doesn't care, it can just "use AnyEvent" and use it itself, or not
614do anything special (it does not need to be event-based) and let AnyEvent 809do anything special (it does not need to be event-based) and let AnyEvent
615decide which implementation to chose if some module relies on it. 810decide which implementation to chose if some module relies on it.
616 811
617If the main program relies on a specific event model. For example, in 812If the main program relies on a specific event model - for example, in
618Gtk2 programs you have to rely on the Glib module. You should load the 813Gtk2 programs you have to rely on the Glib module - you should load the
619event module before loading AnyEvent or any module that uses it: generally 814event module before loading AnyEvent or any module that uses it: generally
620speaking, you should load it as early as possible. The reason is that 815speaking, you should load it as early as possible. The reason is that
621modules might create watchers when they are loaded, and AnyEvent will 816modules might create watchers when they are loaded, and AnyEvent will
622decide on the event model to use as soon as it creates watchers, and it 817decide on the event model to use as soon as it creates watchers, and it
623might chose the wrong one unless you load the correct one yourself. 818might chose the wrong one unless you load the correct one yourself.
624 819
625You can chose to use a rather inefficient pure-perl implementation by 820You can chose to use a pure-perl implementation by loading the
626loading the C<AnyEvent::Impl::Perl> module, which gives you similar 821C<AnyEvent::Impl::Perl> module, which gives you similar behaviour
627behaviour everywhere, but letting AnyEvent chose is generally better. 822everywhere, but letting AnyEvent chose the model is generally better.
823
824=head2 MAINLOOP EMULATION
825
826Sometimes (often for short test scripts, or even standalone programs who
827only want to use AnyEvent), you do not want to run a specific event loop.
828
829In that case, you can use a condition variable like this:
830
831 AnyEvent->condvar->recv;
832
833This has the effect of entering the event loop and looping forever.
834
835Note that usually your program has some exit condition, in which case
836it is better to use the "traditional" approach of storing a condition
837variable somewhere, waiting for it, and sending it when the program should
838exit cleanly.
839
628 840
629=head1 OTHER MODULES 841=head1 OTHER MODULES
630 842
631The following is a non-exhaustive list of additional modules that use 843The following is a non-exhaustive list of additional modules that use
632AnyEvent and can therefore be mixed easily with other AnyEvent modules 844AnyEvent and can therefore be mixed easily with other AnyEvent modules
638=item L<AnyEvent::Util> 850=item L<AnyEvent::Util>
639 851
640Contains various utility functions that replace often-used but blocking 852Contains various utility functions that replace often-used but blocking
641functions such as C<inet_aton> by event-/callback-based versions. 853functions such as C<inet_aton> by event-/callback-based versions.
642 854
643=item L<AnyEvent::Handle>
644
645Provide read and write buffers and manages watchers for reads and writes.
646
647=item L<AnyEvent::Socket> 855=item L<AnyEvent::Socket>
648 856
649Provides various utility functions for (internet protocol) sockets, 857Provides various utility functions for (internet protocol) sockets,
650addresses and name resolution. Also functions to create non-blocking tcp 858addresses and name resolution. Also functions to create non-blocking tcp
651connections or tcp servers, with IPv6 and SRV record support and more. 859connections or tcp servers, with IPv6 and SRV record support and more.
652 860
861=item L<AnyEvent::Handle>
862
863Provide read and write buffers, manages watchers for reads and writes,
864supports raw and formatted I/O, I/O queued and fully transparent and
865non-blocking SSL/TLS.
866
867=item L<AnyEvent::DNS>
868
869Provides rich asynchronous DNS resolver capabilities.
870
871=item L<AnyEvent::HTTP>
872
873A simple-to-use HTTP library that is capable of making a lot of concurrent
874HTTP requests.
875
653=item L<AnyEvent::HTTPD> 876=item L<AnyEvent::HTTPD>
654 877
655Provides a simple web application server framework. 878Provides a simple web application server framework.
656 879
657=item L<AnyEvent::DNS>
658
659Provides rich asynchronous DNS resolver capabilities.
660
661=item L<AnyEvent::FastPing> 880=item L<AnyEvent::FastPing>
662 881
663The fastest ping in the west. 882The fastest ping in the west.
664 883
884=item L<AnyEvent::DBI>
885
886Executes L<DBI> requests asynchronously in a proxy process.
887
888=item L<AnyEvent::AIO>
889
890Truly asynchronous I/O, should be in the toolbox of every event
891programmer. AnyEvent::AIO transparently fuses L<IO::AIO> and AnyEvent
892together.
893
894=item L<AnyEvent::BDB>
895
896Truly asynchronous Berkeley DB access. AnyEvent::BDB transparently fuses
897L<BDB> and AnyEvent together.
898
899=item L<AnyEvent::GPSD>
900
901A non-blocking interface to gpsd, a daemon delivering GPS information.
902
903=item L<AnyEvent::IGS>
904
905A non-blocking interface to the Internet Go Server protocol (used by
906L<App::IGS>).
907
665=item L<Net::IRC3> 908=item L<AnyEvent::IRC>
666 909
667AnyEvent based IRC client module family. 910AnyEvent based IRC client module family (replacing the older Net::IRC3).
668 911
669=item L<Net::XMPP2> 912=item L<Net::XMPP2>
670 913
671AnyEvent based XMPP (Jabber protocol) module family. 914AnyEvent based XMPP (Jabber protocol) module family.
672 915
681 924
682=item L<Coro> 925=item L<Coro>
683 926
684Has special support for AnyEvent via L<Coro::AnyEvent>. 927Has special support for AnyEvent via L<Coro::AnyEvent>.
685 928
686=item L<AnyEvent::AIO>, L<IO::AIO>
687
688Truly asynchronous I/O, should be in the toolbox of every event
689programmer. AnyEvent::AIO transparently fuses IO::AIO and AnyEvent
690together.
691
692=item L<AnyEvent::BDB>, L<BDB>
693
694Truly asynchronous Berkeley DB access. AnyEvent::AIO transparently fuses
695IO::AIO and AnyEvent together.
696
697=item L<IO::Lambda> 929=item L<IO::Lambda>
698 930
699The lambda approach to I/O - don't ask, look there. Can use AnyEvent. 931The lambda approach to I/O - don't ask, look there. Can use AnyEvent.
700 932
701=back 933=back
703=cut 935=cut
704 936
705package AnyEvent; 937package AnyEvent;
706 938
707no warnings; 939no warnings;
708use strict; 940use strict qw(vars subs);
709 941
710use Carp; 942use Carp;
711 943
712our $VERSION = '4.03'; 944our $VERSION = 4.412;
713our $MODEL; 945our $MODEL;
714 946
715our $AUTOLOAD; 947our $AUTOLOAD;
716our @ISA; 948our @ISA;
717 949
950our @REGISTRY;
951
952our $WIN32;
953
954BEGIN {
955 eval "sub WIN32(){ " . (($^O =~ /mswin32/i)*1) ." }";
956 eval "sub TAINT(){ " . (${^TAINT}*1) . " }";
957
958 delete @ENV{grep /^PERL_ANYEVENT_/, keys %ENV}
959 if ${^TAINT};
960}
961
718our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1; 962our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1;
719 963
720our @REGISTRY; 964our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
721
722our %PROTOCOL; # (ipv4|ipv6) => (1|2)
723 965
724{ 966{
725 my $idx; 967 my $idx;
726 $PROTOCOL{$_} = ++$idx 968 $PROTOCOL{$_} = ++$idx
969 for reverse split /\s*,\s*/,
727 for split /\s*,\s*/, $ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6"; 970 $ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6";
728} 971}
729 972
730my @models = ( 973my @models = (
731 [EV:: => AnyEvent::Impl::EV::], 974 [EV:: => AnyEvent::Impl::EV::],
732 [Event:: => AnyEvent::Impl::Event::], 975 [Event:: => AnyEvent::Impl::Event::],
733 [Tk:: => AnyEvent::Impl::Tk::],
734 [Wx:: => AnyEvent::Impl::POE::],
735 [Prima:: => AnyEvent::Impl::POE::],
736 [AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::], 976 [AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],
737 # everything below here will not be autoprobed as the pureperl backend should work everywhere 977 # everything below here will not be autoprobed
738 [Glib:: => AnyEvent::Impl::Glib::], 978 # as the pureperl backend should work everywhere
979 # and is usually faster
980 [Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles
981 [Glib:: => AnyEvent::Impl::Glib::], # becomes extremely slow with many watchers
739 [Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy 982 [Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
740 [Qt:: => AnyEvent::Impl::Qt::], # requires special main program 983 [Qt:: => AnyEvent::Impl::Qt::], # requires special main program
741 [POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza 984 [POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
985 [Wx:: => AnyEvent::Impl::POE::],
986 [Prima:: => AnyEvent::Impl::POE::],
987 # IO::Async is just too broken - we would need workaorunds for its
988 # byzantine signal and broken child handling, among others.
989 # IO::Async is rather hard to detect, as it doesn't have any
990 # obvious default class.
991# [IO::Async:: => AnyEvent::Impl::IOAsync::], # requires special main program
992# [IO::Async::Loop:: => AnyEvent::Impl::IOAsync::], # requires special main program
993# [IO::Async::Notifier:: => AnyEvent::Impl::IOAsync::], # requires special main program
742); 994);
743 995
744our %method = map +($_ => 1), qw(io timer signal child condvar one_event DESTROY); 996our %method = map +($_ => 1),
997 qw(io timer time now now_update signal child idle condvar one_event DESTROY);
745 998
746our @post_detect; 999our @post_detect;
747 1000
748sub post_detect(&) { 1001sub post_detect(&) {
749 my ($cb) = @_; 1002 my ($cb) = @_;
754 1 1007 1
755 } else { 1008 } else {
756 push @post_detect, $cb; 1009 push @post_detect, $cb;
757 1010
758 defined wantarray 1011 defined wantarray
759 ? bless \$cb, "AnyEvent::Util::PostDetect" 1012 ? bless \$cb, "AnyEvent::Util::postdetect"
760 : () 1013 : ()
761 } 1014 }
762} 1015}
763 1016
764sub AnyEvent::Util::PostDetect::DESTROY { 1017sub AnyEvent::Util::postdetect::DESTROY {
765 @post_detect = grep $_ != ${$_[0]}, @post_detect; 1018 @post_detect = grep $_ != ${$_[0]}, @post_detect;
766} 1019}
767 1020
768sub detect() { 1021sub detect() {
769 unless ($MODEL) { 1022 unless ($MODEL) {
770 no strict 'refs'; 1023 no strict 'refs';
1024 local $SIG{__DIE__};
771 1025
772 if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) { 1026 if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) {
773 my $model = "AnyEvent::Impl::$1"; 1027 my $model = "AnyEvent::Impl::$1";
774 if (eval "require $model") { 1028 if (eval "require $model") {
775 $MODEL = $model; 1029 $MODEL = $model;
805 last; 1059 last;
806 } 1060 }
807 } 1061 }
808 1062
809 $MODEL 1063 $MODEL
810 or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib."; 1064 or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib.\n";
811 } 1065 }
812 } 1066 }
813 1067
1068 push @{"$MODEL\::ISA"}, "AnyEvent::Base";
1069
814 unshift @ISA, $MODEL; 1070 unshift @ISA, $MODEL;
815 push @{"$MODEL\::ISA"}, "AnyEvent::Base"; 1071
1072 require AnyEvent::Strict if $ENV{PERL_ANYEVENT_STRICT};
816 1073
817 (shift @post_detect)->() while @post_detect; 1074 (shift @post_detect)->() while @post_detect;
818 } 1075 }
819 1076
820 $MODEL 1077 $MODEL
830 1087
831 my $class = shift; 1088 my $class = shift;
832 $class->$func (@_); 1089 $class->$func (@_);
833} 1090}
834 1091
1092# utility function to dup a filehandle. this is used by many backends
1093# to support binding more than one watcher per filehandle (they usually
1094# allow only one watcher per fd, so we dup it to get a different one).
1095sub _dupfh($$;$$) {
1096 my ($poll, $fh, $r, $w) = @_;
1097
1098 # cygwin requires the fh mode to be matching, unix doesn't
1099 my ($rw, $mode) = $poll eq "r" ? ($r, "<")
1100 : $poll eq "w" ? ($w, ">")
1101 : Carp::croak "AnyEvent->io requires poll set to either 'r' or 'w'";
1102
1103 open my $fh2, "$mode&" . fileno $fh
1104 or die "cannot dup() filehandle: $!,";
1105
1106 # we assume CLOEXEC is already set by perl in all important cases
1107
1108 ($fh2, $rw)
1109}
1110
835package AnyEvent::Base; 1111package AnyEvent::Base;
836 1112
1113# default implementations for many methods
1114
1115BEGIN {
1116 if (eval "use Time::HiRes (); Time::HiRes::time (); 1") {
1117 *_time = \&Time::HiRes::time;
1118 # if (eval "use POSIX (); (POSIX::times())...
1119 } else {
1120 *_time = sub { time }; # epic fail
1121 }
1122}
1123
1124sub time { _time }
1125sub now { _time }
1126sub now_update { }
1127
837# default implementation for ->condvar 1128# default implementation for ->condvar
838 1129
839sub condvar { 1130sub condvar {
840 bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, AnyEvent::CondVar:: 1131 bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, "AnyEvent::CondVar"
841} 1132}
842 1133
843# default implementation for ->signal 1134# default implementation for ->signal
844 1135
845our %SIG_CB; 1136our ($SIGPIPE_R, $SIGPIPE_W, %SIG_CB, %SIG_EV, $SIG_IO);
1137
1138sub _signal_exec {
1139 sysread $SIGPIPE_R, my $dummy, 4;
1140
1141 while (%SIG_EV) {
1142 for (keys %SIG_EV) {
1143 delete $SIG_EV{$_};
1144 $_->() for values %{ $SIG_CB{$_} || {} };
1145 }
1146 }
1147}
846 1148
847sub signal { 1149sub signal {
848 my (undef, %arg) = @_; 1150 my (undef, %arg) = @_;
849 1151
1152 unless ($SIGPIPE_R) {
1153 require Fcntl;
1154
1155 if (AnyEvent::WIN32) {
1156 require AnyEvent::Util;
1157
1158 ($SIGPIPE_R, $SIGPIPE_W) = AnyEvent::Util::portable_pipe ();
1159 AnyEvent::Util::fh_nonblocking ($SIGPIPE_R) if $SIGPIPE_R;
1160 AnyEvent::Util::fh_nonblocking ($SIGPIPE_W) if $SIGPIPE_W; # just in case
1161 } else {
1162 pipe $SIGPIPE_R, $SIGPIPE_W;
1163 fcntl $SIGPIPE_R, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_R;
1164 fcntl $SIGPIPE_W, &Fcntl::F_SETFL, &Fcntl::O_NONBLOCK if $SIGPIPE_W; # just in case
1165
1166 # not strictly required, as $^F is normally 2, but let's make sure...
1167 fcntl $SIGPIPE_R, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
1168 fcntl $SIGPIPE_W, &Fcntl::F_SETFD, &Fcntl::FD_CLOEXEC;
1169 }
1170
1171 $SIGPIPE_R
1172 or Carp::croak "AnyEvent: unable to create a signal reporting pipe: $!\n";
1173
1174 $SIG_IO = AnyEvent->io (fh => $SIGPIPE_R, poll => "r", cb => \&_signal_exec);
1175 }
1176
850 my $signal = uc $arg{signal} 1177 my $signal = uc $arg{signal}
851 or Carp::croak "required option 'signal' is missing"; 1178 or Carp::croak "required option 'signal' is missing";
852 1179
853 $SIG_CB{$signal}{$arg{cb}} = $arg{cb}; 1180 $SIG_CB{$signal}{$arg{cb}} = $arg{cb};
854 $SIG{$signal} ||= sub { 1181 $SIG{$signal} ||= sub {
855 $_->() for values %{ $SIG_CB{$signal} || {} }; 1182 local $!;
1183 syswrite $SIGPIPE_W, "\x00", 1 unless %SIG_EV;
1184 undef $SIG_EV{$signal};
856 }; 1185 };
857 1186
858 bless [$signal, $arg{cb}], "AnyEvent::Base::Signal" 1187 bless [$signal, $arg{cb}], "AnyEvent::Base::signal"
859} 1188}
860 1189
861sub AnyEvent::Base::Signal::DESTROY { 1190sub AnyEvent::Base::signal::DESTROY {
862 my ($signal, $cb) = @{$_[0]}; 1191 my ($signal, $cb) = @{$_[0]};
863 1192
864 delete $SIG_CB{$signal}{$cb}; 1193 delete $SIG_CB{$signal}{$cb};
865 1194
1195 # delete doesn't work with older perls - they then
1196 # print weird messages, or just unconditionally exit
1197 # instead of getting the default action.
866 $SIG{$signal} = 'DEFAULT' unless keys %{ $SIG_CB{$signal} }; 1198 undef $SIG{$signal} unless keys %{ $SIG_CB{$signal} };
867} 1199}
868 1200
869# default implementation for ->child 1201# default implementation for ->child
870 1202
871our %PID_CB; 1203our %PID_CB;
872our $CHLD_W; 1204our $CHLD_W;
873our $CHLD_DELAY_W; 1205our $CHLD_DELAY_W;
874our $PID_IDLE;
875our $WNOHANG; 1206our $WNOHANG;
876 1207
877sub _child_wait { 1208sub _sigchld {
878 while (0 < (my $pid = waitpid -1, $WNOHANG)) { 1209 while (0 < (my $pid = waitpid -1, $WNOHANG)) {
879 $_->($pid, $?) for (values %{ $PID_CB{$pid} || {} }), 1210 $_->($pid, $?) for (values %{ $PID_CB{$pid} || {} }),
880 (values %{ $PID_CB{0} || {} }); 1211 (values %{ $PID_CB{0} || {} });
881 } 1212 }
882
883 undef $PID_IDLE;
884}
885
886sub _sigchld {
887 # make sure we deliver these changes "synchronous" with the event loop.
888 $CHLD_DELAY_W ||= AnyEvent->timer (after => 0, cb => sub {
889 undef $CHLD_DELAY_W;
890 &_child_wait;
891 });
892} 1213}
893 1214
894sub child { 1215sub child {
895 my (undef, %arg) = @_; 1216 my (undef, %arg) = @_;
896 1217
897 defined (my $pid = $arg{pid} + 0) 1218 defined (my $pid = $arg{pid} + 0)
898 or Carp::croak "required option 'pid' is missing"; 1219 or Carp::croak "required option 'pid' is missing";
899 1220
900 $PID_CB{$pid}{$arg{cb}} = $arg{cb}; 1221 $PID_CB{$pid}{$arg{cb}} = $arg{cb};
901 1222
902 unless ($WNOHANG) {
903 $WNOHANG = eval { require POSIX; &POSIX::WNOHANG } || 1; 1223 $WNOHANG ||= eval { local $SIG{__DIE__}; require POSIX; &POSIX::WNOHANG } || 1;
904 }
905 1224
906 unless ($CHLD_W) { 1225 unless ($CHLD_W) {
907 $CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_sigchld); 1226 $CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_sigchld);
908 # child could be a zombie already, so make at least one round 1227 # child could be a zombie already, so make at least one round
909 &_sigchld; 1228 &_sigchld;
910 } 1229 }
911 1230
912 bless [$pid, $arg{cb}], "AnyEvent::Base::Child" 1231 bless [$pid, $arg{cb}], "AnyEvent::Base::child"
913} 1232}
914 1233
915sub AnyEvent::Base::Child::DESTROY { 1234sub AnyEvent::Base::child::DESTROY {
916 my ($pid, $cb) = @{$_[0]}; 1235 my ($pid, $cb) = @{$_[0]};
917 1236
918 delete $PID_CB{$pid}{$cb}; 1237 delete $PID_CB{$pid}{$cb};
919 delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} }; 1238 delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
920 1239
921 undef $CHLD_W unless keys %PID_CB; 1240 undef $CHLD_W unless keys %PID_CB;
1241}
1242
1243# idle emulation is done by simply using a timer, regardless
1244# of whether the process is idle or not, and not letting
1245# the callback use more than 50% of the time.
1246sub idle {
1247 my (undef, %arg) = @_;
1248
1249 my ($cb, $w, $rcb) = $arg{cb};
1250
1251 $rcb = sub {
1252 if ($cb) {
1253 $w = _time;
1254 &$cb;
1255 $w = _time - $w;
1256
1257 # never use more then 50% of the time for the idle watcher,
1258 # within some limits
1259 $w = 0.0001 if $w < 0.0001;
1260 $w = 5 if $w > 5;
1261
1262 $w = AnyEvent->timer (after => $w, cb => $rcb);
1263 } else {
1264 # clean up...
1265 undef $w;
1266 undef $rcb;
1267 }
1268 };
1269
1270 $w = AnyEvent->timer (after => 0.05, cb => $rcb);
1271
1272 bless \\$cb, "AnyEvent::Base::idle"
1273}
1274
1275sub AnyEvent::Base::idle::DESTROY {
1276 undef $${$_[0]};
922} 1277}
923 1278
924package AnyEvent::CondVar; 1279package AnyEvent::CondVar;
925 1280
926our @ISA = AnyEvent::CondVar::Base::; 1281our @ISA = AnyEvent::CondVar::Base::;
978} 1333}
979 1334
980# undocumented/compatibility with pre-3.4 1335# undocumented/compatibility with pre-3.4
981*broadcast = \&send; 1336*broadcast = \&send;
982*wait = \&_wait; 1337*wait = \&_wait;
1338
1339=head1 ERROR AND EXCEPTION HANDLING
1340
1341In general, AnyEvent does not do any error handling - it relies on the
1342caller to do that if required. The L<AnyEvent::Strict> module (see also
1343the C<PERL_ANYEVENT_STRICT> environment variable, below) provides strict
1344checking of all AnyEvent methods, however, which is highly useful during
1345development.
1346
1347As for exception handling (i.e. runtime errors and exceptions thrown while
1348executing a callback), this is not only highly event-loop specific, but
1349also not in any way wrapped by this module, as this is the job of the main
1350program.
1351
1352The pure perl event loop simply re-throws the exception (usually
1353within C<< condvar->recv >>), the L<Event> and L<EV> modules call C<<
1354$Event/EV::DIED->() >>, L<Glib> uses C<< install_exception_handler >> and
1355so on.
1356
1357=head1 ENVIRONMENT VARIABLES
1358
1359The following environment variables are used by this module or its
1360submodules.
1361
1362Note that AnyEvent will remove I<all> environment variables starting with
1363C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
1364enabled.
1365
1366=over 4
1367
1368=item C<PERL_ANYEVENT_VERBOSE>
1369
1370By default, AnyEvent will be completely silent except in fatal
1371conditions. You can set this environment variable to make AnyEvent more
1372talkative.
1373
1374When set to C<1> or higher, causes AnyEvent to warn about unexpected
1375conditions, such as not being able to load the event model specified by
1376C<PERL_ANYEVENT_MODEL>.
1377
1378When set to C<2> or higher, cause AnyEvent to report to STDERR which event
1379model it chooses.
1380
1381=item C<PERL_ANYEVENT_STRICT>
1382
1383AnyEvent does not do much argument checking by default, as thorough
1384argument checking is very costly. Setting this variable to a true value
1385will cause AnyEvent to load C<AnyEvent::Strict> and then to thoroughly
1386check the arguments passed to most method calls. If it finds any problems,
1387it will croak.
1388
1389In other words, enables "strict" mode.
1390
1391Unlike C<use strict>, it is definitely recommended to keep it off in
1392production. Keeping C<PERL_ANYEVENT_STRICT=1> in your environment while
1393developing programs can be very useful, however.
1394
1395=item C<PERL_ANYEVENT_MODEL>
1396
1397This can be used to specify the event model to be used by AnyEvent, before
1398auto detection and -probing kicks in. It must be a string consisting
1399entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended
1400and the resulting module name is loaded and if the load was successful,
1401used as event model. If it fails to load AnyEvent will proceed with
1402auto detection and -probing.
1403
1404This functionality might change in future versions.
1405
1406For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you
1407could start your program like this:
1408
1409 PERL_ANYEVENT_MODEL=Perl perl ...
1410
1411=item C<PERL_ANYEVENT_PROTOCOLS>
1412
1413Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
1414for IPv4 or IPv6. The default is unspecified (and might change, or be the result
1415of auto probing).
1416
1417Must be set to a comma-separated list of protocols or address families,
1418current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
1419used, and preference will be given to protocols mentioned earlier in the
1420list.
1421
1422This variable can effectively be used for denial-of-service attacks
1423against local programs (e.g. when setuid), although the impact is likely
1424small, as the program has to handle conenction and other failures anyways.
1425
1426Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
1427but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
1428- only support IPv4, never try to resolve or contact IPv6
1429addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
1430IPv6, but prefer IPv6 over IPv4.
1431
1432=item C<PERL_ANYEVENT_EDNS0>
1433
1434Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension
1435for DNS. This extension is generally useful to reduce DNS traffic, but
1436some (broken) firewalls drop such DNS packets, which is why it is off by
1437default.
1438
1439Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
1440EDNS0 in its DNS requests.
1441
1442=item C<PERL_ANYEVENT_MAX_FORKS>
1443
1444The maximum number of child processes that C<AnyEvent::Util::fork_call>
1445will create in parallel.
1446
1447=back
983 1448
984=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE 1449=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE
985 1450
986This is an advanced topic that you do not normally need to use AnyEvent in 1451This is an advanced topic that you do not normally need to use AnyEvent in
987a module. This section is only of use to event loop authors who want to 1452a module. This section is only of use to event loop authors who want to
1021 1486
1022I<rxvt-unicode> also cheats a bit by not providing blocking access to 1487I<rxvt-unicode> also cheats a bit by not providing blocking access to
1023condition variables: code blocking while waiting for a condition will 1488condition variables: code blocking while waiting for a condition will
1024C<die>. This still works with most modules/usages, and blocking calls must 1489C<die>. This still works with most modules/usages, and blocking calls must
1025not be done in an interactive application, so it makes sense. 1490not be done in an interactive application, so it makes sense.
1026
1027=head1 ENVIRONMENT VARIABLES
1028
1029The following environment variables are used by this module:
1030
1031=over 4
1032
1033=item C<PERL_ANYEVENT_VERBOSE>
1034
1035By default, AnyEvent will be completely silent except in fatal
1036conditions. You can set this environment variable to make AnyEvent more
1037talkative.
1038
1039When set to C<1> or higher, causes AnyEvent to warn about unexpected
1040conditions, such as not being able to load the event model specified by
1041C<PERL_ANYEVENT_MODEL>.
1042
1043When set to C<2> or higher, cause AnyEvent to report to STDERR which event
1044model it chooses.
1045
1046=item C<PERL_ANYEVENT_MODEL>
1047
1048This can be used to specify the event model to be used by AnyEvent, before
1049auto detection and -probing kicks in. It must be a string consisting
1050entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended
1051and the resulting module name is loaded and if the load was successful,
1052used as event model. If it fails to load AnyEvent will proceed with
1053auto detection and -probing.
1054
1055This functionality might change in future versions.
1056
1057For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you
1058could start your program like this:
1059
1060 PERL_ANYEVENT_MODEL=Perl perl ...
1061
1062=item C<PERL_ANYEVENT_PROTOCOLS>
1063
1064Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
1065for IPv4 or IPv6. The default is unspecified (and might change, or be the result
1066of auto probing).
1067
1068Must be set to a comma-separated list of protocols or address families,
1069current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
1070used, and preference will be given to protocols mentioned earlier in the
1071list.
1072
1073This variable can effectively be used for denial-of-service attacks
1074against local programs (e.g. when setuid), although the impact is likely
1075small, as the program has to handle connection errors already-
1076
1077Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
1078but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
1079- only support IPv4, never try to resolve or contact IPv6
1080addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
1081IPv6, but prefer IPv6 over IPv4.
1082
1083=item C<PERL_ANYEVENT_EDNS0>
1084
1085Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension
1086for DNS. This extension is generally useful to reduce DNS traffic, but
1087some (broken) firewalls drop such DNS packets, which is why it is off by
1088default.
1089
1090Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
1091EDNS0 in its DNS requests.
1092
1093=back
1094 1491
1095=head1 EXAMPLE PROGRAM 1492=head1 EXAMPLE PROGRAM
1096 1493
1097The following program uses an I/O watcher to read data from STDIN, a timer 1494The following program uses an I/O watcher to read data from STDIN, a timer
1098to display a message once per second, and a condition variable to quit the 1495to display a message once per second, and a condition variable to quit the
1292watcher. 1689watcher.
1293 1690
1294=head3 Results 1691=head3 Results
1295 1692
1296 name watchers bytes create invoke destroy comment 1693 name watchers bytes create invoke destroy comment
1297 EV/EV 400000 244 0.56 0.46 0.31 EV native interface 1694 EV/EV 400000 224 0.47 0.35 0.27 EV native interface
1298 EV/Any 100000 244 2.50 0.46 0.29 EV + AnyEvent watchers 1695 EV/Any 100000 224 2.88 0.34 0.27 EV + AnyEvent watchers
1299 CoroEV/Any 100000 244 2.49 0.44 0.29 coroutines + Coro::Signal 1696 CoroEV/Any 100000 224 2.85 0.35 0.28 coroutines + Coro::Signal
1300 Perl/Any 100000 513 4.92 0.87 1.12 pure perl implementation 1697 Perl/Any 100000 452 4.13 0.73 0.95 pure perl implementation
1301 Event/Event 16000 516 31.88 31.30 0.85 Event native interface 1698 Event/Event 16000 517 32.20 31.80 0.81 Event native interface
1302 Event/Any 16000 590 35.75 31.42 1.08 Event + AnyEvent watchers 1699 Event/Any 16000 590 35.85 31.55 1.06 Event + AnyEvent watchers
1303 Glib/Any 16000 1357 98.22 12.41 54.00 quadratic behaviour 1700 Glib/Any 16000 1357 102.33 12.31 51.00 quadratic behaviour
1304 Tk/Any 2000 1860 26.97 67.98 14.00 SEGV with >> 2000 watchers 1701 Tk/Any 2000 1860 27.20 66.31 14.00 SEGV with >> 2000 watchers
1305 POE/Event 2000 6644 108.64 736.02 14.73 via POE::Loop::Event 1702 POE/Event 2000 6328 109.99 751.67 14.02 via POE::Loop::Event
1306 POE/Select 2000 6343 94.13 809.12 565.96 via POE::Loop::Select 1703 POE/Select 2000 6027 94.54 809.13 579.80 via POE::Loop::Select
1307 1704
1308=head3 Discussion 1705=head3 Discussion
1309 1706
1310The benchmark does I<not> measure scalability of the event loop very 1707The benchmark does I<not> measure scalability of the event loop very
1311well. For example, a select-based event loop (such as the pure perl one) 1708well. For example, a select-based event loop (such as the pure perl one)
1512=item * C-based event loops perform very well with small number of 1909=item * C-based event loops perform very well with small number of
1513watchers, as the management overhead dominates. 1910watchers, as the management overhead dominates.
1514 1911
1515=back 1912=back
1516 1913
1914=head2 THE IO::Lambda BENCHMARK
1915
1916Recently I was told about the benchmark in the IO::Lambda manpage, which
1917could be misinterpreted to make AnyEvent look bad. In fact, the benchmark
1918simply compares IO::Lambda with POE, and IO::Lambda looks better (which
1919shouldn't come as a surprise to anybody). As such, the benchmark is
1920fine, and mostly shows that the AnyEvent backend from IO::Lambda isn't
1921very optimal. But how would AnyEvent compare when used without the extra
1922baggage? To explore this, I wrote the equivalent benchmark for AnyEvent.
1923
1924The benchmark itself creates an echo-server, and then, for 500 times,
1925connects to the echo server, sends a line, waits for the reply, and then
1926creates the next connection. This is a rather bad benchmark, as it doesn't
1927test the efficiency of the framework or much non-blocking I/O, but it is a
1928benchmark nevertheless.
1929
1930 name runtime
1931 Lambda/select 0.330 sec
1932 + optimized 0.122 sec
1933 Lambda/AnyEvent 0.327 sec
1934 + optimized 0.138 sec
1935 Raw sockets/select 0.077 sec
1936 POE/select, components 0.662 sec
1937 POE/select, raw sockets 0.226 sec
1938 POE/select, optimized 0.404 sec
1939
1940 AnyEvent/select/nb 0.085 sec
1941 AnyEvent/EV/nb 0.068 sec
1942 +state machine 0.134 sec
1943
1944The benchmark is also a bit unfair (my fault): the IO::Lambda/POE
1945benchmarks actually make blocking connects and use 100% blocking I/O,
1946defeating the purpose of an event-based solution. All of the newly
1947written AnyEvent benchmarks use 100% non-blocking connects (using
1948AnyEvent::Socket::tcp_connect and the asynchronous pure perl DNS
1949resolver), so AnyEvent is at a disadvantage here, as non-blocking connects
1950generally require a lot more bookkeeping and event handling than blocking
1951connects (which involve a single syscall only).
1952
1953The last AnyEvent benchmark additionally uses L<AnyEvent::Handle>, which
1954offers similar expressive power as POE and IO::Lambda, using conventional
1955Perl syntax. This means that both the echo server and the client are 100%
1956non-blocking, further placing it at a disadvantage.
1957
1958As you can see, the AnyEvent + EV combination even beats the
1959hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl
1960backend easily beats IO::Lambda and POE.
1961
1962And even the 100% non-blocking version written using the high-level (and
1963slow :) L<AnyEvent::Handle> abstraction beats both POE and IO::Lambda by a
1964large margin, even though it does all of DNS, tcp-connect and socket I/O
1965in a non-blocking way.
1966
1967The two AnyEvent benchmarks programs can be found as F<eg/ae0.pl> and
1968F<eg/ae2.pl> in the AnyEvent distribution, the remaining benchmarks are
1969part of the IO::lambda distribution and were used without any changes.
1970
1971
1972=head1 SIGNALS
1973
1974AnyEvent currently installs handlers for these signals:
1975
1976=over 4
1977
1978=item SIGCHLD
1979
1980A handler for C<SIGCHLD> is installed by AnyEvent's child watcher
1981emulation for event loops that do not support them natively. Also, some
1982event loops install a similar handler.
1983
1984If, when AnyEvent is loaded, SIGCHLD is set to IGNORE, then AnyEvent will
1985reset it to default, to avoid losing child exit statuses.
1986
1987=item SIGPIPE
1988
1989A no-op handler is installed for C<SIGPIPE> when C<$SIG{PIPE}> is C<undef>
1990when AnyEvent gets loaded.
1991
1992The rationale for this is that AnyEvent users usually do not really depend
1993on SIGPIPE delivery (which is purely an optimisation for shell use, or
1994badly-written programs), but C<SIGPIPE> can cause spurious and rare
1995program exits as a lot of people do not expect C<SIGPIPE> when writing to
1996some random socket.
1997
1998The rationale for installing a no-op handler as opposed to ignoring it is
1999that this way, the handler will be restored to defaults on exec.
2000
2001Feel free to install your own handler, or reset it to defaults.
2002
2003=back
2004
2005=cut
2006
2007undef $SIG{CHLD}
2008 if $SIG{CHLD} eq 'IGNORE';
2009
2010$SIG{PIPE} = sub { }
2011 unless defined $SIG{PIPE};
1517 2012
1518=head1 FORK 2013=head1 FORK
1519 2014
1520Most event libraries are not fork-safe. The ones who are usually are 2015Most event libraries are not fork-safe. The ones who are usually are
1521because they rely on inefficient but fork-safe C<select> or C<poll> 2016because they rely on inefficient but fork-safe C<select> or C<poll>
1535specified in the variable. 2030specified in the variable.
1536 2031
1537You can make AnyEvent completely ignore this variable by deleting it 2032You can make AnyEvent completely ignore this variable by deleting it
1538before the first watcher gets created, e.g. with a C<BEGIN> block: 2033before the first watcher gets created, e.g. with a C<BEGIN> block:
1539 2034
1540 BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} } 2035 BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
1541 2036
1542 use AnyEvent; 2037 use AnyEvent;
1543 2038
1544Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can 2039Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
1545be used to probe what backend is used and gain other information (which is 2040be used to probe what backend is used and gain other information (which is
1546probably even less useful to an attacker than PERL_ANYEVENT_MODEL). 2041probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and
2042$ENV{PERL_ANYEVENT_STRICT}.
2043
2044Note that AnyEvent will remove I<all> environment variables starting with
2045C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
2046enabled.
2047
2048
2049=head1 BUGS
2050
2051Perl 5.8 has numerous memleaks that sometimes hit this module and are hard
2052to work around. If you suffer from memleaks, first upgrade to Perl 5.10
2053and check wether the leaks still show up. (Perl 5.10.0 has other annoying
2054memleaks, such as leaking on C<map> and C<grep> but it is usually not as
2055pronounced).
1547 2056
1548 2057
1549=head1 SEE ALSO 2058=head1 SEE ALSO
1550 2059
1551Utility functions: L<AnyEvent::Util>. 2060Utility functions: L<AnyEvent::Util>.
1568Nontrivial usage examples: L<Net::FCP>, L<Net::XMPP2>, L<AnyEvent::DNS>. 2077Nontrivial usage examples: L<Net::FCP>, L<Net::XMPP2>, L<AnyEvent::DNS>.
1569 2078
1570 2079
1571=head1 AUTHOR 2080=head1 AUTHOR
1572 2081
1573 Marc Lehmann <schmorp@schmorp.de> 2082 Marc Lehmann <schmorp@schmorp.de>
1574 http://home.schmorp.de/ 2083 http://home.schmorp.de/
1575 2084
1576=cut 2085=cut
1577 2086
15781 20871
1579 2088

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines