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 1.155 by root, Fri Jun 6 07:07:01 2008 UTC vs.
Revision 1.356 by root, Fri Aug 12 18:41:25 2011 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 # if you prefer function calls, look at the AE manpage for
13 # an alternative API.
14
15 # file handle or descriptor readable
11 my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub { 16 my $w = AnyEvent->io (fh => $fh, poll => "r", cb => sub { ... });
17
18 # one-shot or repeating timers
19 my $w = AnyEvent->timer (after => $seconds, cb => sub { ... });
20 my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ...);
21
22 print AnyEvent->now; # prints current event loop time
23 print AnyEvent->time; # think Time::HiRes::time or simply CORE::time.
24
25 # POSIX signal
26 my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... });
27
28 # child process exit
29 my $w = AnyEvent->child (pid => $pid, cb => sub {
30 my ($pid, $status) = @_;
12 ... 31 ...
13 }); 32 });
14 33
15 my $w = AnyEvent->timer (after => $seconds, cb => sub { 34 # called when event loop idle (if applicable)
16 ... 35 my $w = AnyEvent->idle (cb => sub { ... });
17 });
18 36
19 my $w = AnyEvent->condvar; # stores whether a condition was flagged 37 my $w = AnyEvent->condvar; # stores whether a condition was flagged
20 $w->send; # wake up current and all future recv's 38 $w->send; # wake up current and all future recv's
21 $w->recv; # enters "main loop" till $condvar gets ->send 39 $w->recv; # enters "main loop" till $condvar gets ->send
40 # use a condvar in callback mode:
41 $w->cb (sub { $_[0]->recv });
22 42
23=head1 INTRODUCTION/TUTORIAL 43=head1 INTRODUCTION/TUTORIAL
24 44
25This manpage is mainly a reference manual. If you are interested 45This manpage is mainly a reference manual. If you are interested
26in a tutorial or some gentle introduction, have a look at the 46in a tutorial or some gentle introduction, have a look at the
27L<AnyEvent::Intro> manpage. 47L<AnyEvent::Intro> manpage.
28 48
49=head1 SUPPORT
50
51An FAQ document is available as L<AnyEvent::FAQ>.
52
53There also is a mailinglist for discussing all things AnyEvent, and an IRC
54channel, too.
55
56See the AnyEvent project page at the B<Schmorpforge Ta-Sa Software
57Repository>, at L<http://anyevent.schmorp.de>, for more info.
58
29=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT) 59=head1 WHY YOU SHOULD USE THIS MODULE (OR NOT)
30 60
31Glib, POE, IO::Async, Event... CPAN offers event models by the dozen 61Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
32nowadays. So what is different about AnyEvent? 62nowadays. So what is different about AnyEvent?
33 63
34Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of 64Executive Summary: AnyEvent is I<compatible>, AnyEvent is I<free of
35policy> and AnyEvent is I<small and efficient>. 65policy> and AnyEvent is I<small and efficient>.
36 66
37First and foremost, I<AnyEvent is not an event model> itself, it only 67First and foremost, I<AnyEvent is not an event model> itself, it only
38interfaces to whatever event model the main program happens to use in a 68interfaces to whatever event model the main program happens to use, in a
39pragmatic way. For event models and certain classes of immortals alike, 69pragmatic way. For event models and certain classes of immortals alike,
40the statement "there can only be one" is a bitter reality: In general, 70the statement "there can only be one" is a bitter reality: In general,
41only one event loop can be active at the same time in a process. AnyEvent 71only one event loop can be active at the same time in a process. AnyEvent
42helps hiding the differences between those event loops. 72cannot change this, but it can hide the differences between those event
73loops.
43 74
44The goal of AnyEvent is to offer module authors the ability to do event 75The goal of AnyEvent is to offer module authors the ability to do event
45programming (waiting for I/O or timer events) without subscribing to a 76programming (waiting for I/O or timer events) without subscribing to a
46religion, a way of living, and most importantly: without forcing your 77religion, a way of living, and most importantly: without forcing your
47module users into the same thing by forcing them to use the same event 78module users into the same thing by forcing them to use the same event
48model you use. 79model you use.
49 80
50For modules like POE or IO::Async (which is a total misnomer as it is 81For modules like POE or IO::Async (which is a total misnomer as it is
51actually doing all I/O I<synchronously>...), using them in your module is 82actually doing all I/O I<synchronously>...), using them in your module is
52like joining a cult: After you joined, you are dependent on them and you 83like joining a cult: After you join, you are dependent on them and you
53cannot use anything else, as it is simply incompatible to everything that 84cannot use anything else, as they are simply incompatible to everything
54isn't itself. What's worse, all the potential users of your module are 85that isn't them. What's worse, all the potential users of your
55I<also> forced to use the same event loop you use. 86module are I<also> forced to use the same event loop you use.
56 87
57AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works 88AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
58fine. AnyEvent + Tk works fine etc. etc. but none of these work together 89fine. AnyEvent + Tk works fine etc. etc. but none of these work together
59with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if 90with the rest: POE + EV? No go. Tk + Event? No go. Again: if your module
60your module uses one of those, every user of your module has to use it, 91uses one of those, every user of your module has to use it, too. But if
61too. But if your module uses AnyEvent, it works transparently with all 92your module uses AnyEvent, it works transparently with all event models it
62event models it supports (including stuff like POE and IO::Async, as long 93supports (including stuff like IO::Async, as long as those use one of the
63as those use one of the supported event loops. It is trivial to add new 94supported event loops. It is easy to add new event loops to AnyEvent, too,
64event loops to AnyEvent, too, so it is future-proof). 95so it is future-proof).
65 96
66In addition to being free of having to use I<the one and only true event 97In addition to being free of having to use I<the one and only true event
67model>, AnyEvent also is free of bloat and policy: with POE or similar 98model>, AnyEvent also is free of bloat and policy: with POE or similar
68modules, you get an enormous amount of code and strict rules you have to 99modules, you get an enormous amount of code and strict rules you have to
69follow. AnyEvent, on the other hand, is lean and up to the point, by only 100follow. AnyEvent, on the other hand, is lean and to the point, by only
70offering the functionality that is necessary, in as thin as a wrapper as 101offering the functionality that is necessary, in as thin as a wrapper as
71technically possible. 102technically possible.
72 103
73Of course, AnyEvent comes with a big (and fully optional!) toolbox 104Of course, AnyEvent comes with a big (and fully optional!) toolbox
74of useful functionality, such as an asynchronous DNS resolver, 100% 105of useful functionality, such as an asynchronous DNS resolver, 100%
80useful) and you want to force your users to use the one and only event 111useful) and you want to force your users to use the one and only event
81model, you should I<not> use this module. 112model, you should I<not> use this module.
82 113
83=head1 DESCRIPTION 114=head1 DESCRIPTION
84 115
85L<AnyEvent> provides an identical interface to multiple event loops. This 116L<AnyEvent> provides a uniform interface to various event loops. This
86allows module authors to utilise an event loop without forcing module 117allows module authors to use event loop functionality without forcing
87users to use the same event loop (as only a single event loop can coexist 118module users to use a specific event loop implementation (since more
88peacefully at any one time). 119than one event loop cannot coexist peacefully).
89 120
90The interface itself is vaguely similar, but not identical to the L<Event> 121The interface itself is vaguely similar, but not identical to the L<Event>
91module. 122module.
92 123
93During the first call of any watcher-creation method, the module tries 124During the first call of any watcher-creation method, the module tries
94to detect the currently loaded event loop by probing whether one of the 125to detect the currently loaded event loop by probing whether one of the
95following modules is already loaded: L<EV>, 126following modules is already loaded: L<EV>, L<AnyEvent::Loop>,
96L<Event>, L<Glib>, L<AnyEvent::Impl::Perl>, L<Tk>, L<Event::Lib>, L<Qt>, 127L<Event>, L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>. The first one
97L<POE>. The first one found is used. If none are found, the module tries 128found is used. If none are detected, the module tries to load the first
98to load these modules (excluding Tk, Event::Lib, Qt and POE as the pure perl 129four modules in the order given; but note that if L<EV> is not
99adaptor should always succeed) in the order given. The first one that can 130available, the pure-perl L<AnyEvent::Loop> should always work, so
100be successfully loaded will be used. If, after this, still none could be 131the other two are not normally tried.
101found, AnyEvent will fall back to a pure-perl event loop, which is not
102very efficient, but should work everywhere.
103 132
104Because AnyEvent first checks for modules that are already loaded, loading 133Because AnyEvent first checks for modules that are already loaded, loading
105an event model explicitly before first using AnyEvent will likely make 134an event model explicitly before first using AnyEvent will likely make
106that model the default. For example: 135that model the default. For example:
107 136
109 use AnyEvent; 138 use AnyEvent;
110 139
111 # .. AnyEvent will likely default to Tk 140 # .. AnyEvent will likely default to Tk
112 141
113The I<likely> means that, if any module loads another event model and 142The I<likely> means that, if any module loads another event model and
114starts using it, all bets are off. Maybe you should tell their authors to 143starts using it, all bets are off - this case should be very rare though,
115use AnyEvent so their modules work together with others seamlessly... 144as very few modules hardcode event loops without announcing this very
145loudly.
116 146
117The pure-perl implementation of AnyEvent is called 147The pure-perl implementation of AnyEvent is called C<AnyEvent::Loop>. Like
118C<AnyEvent::Impl::Perl>. Like other event modules you can load it 148other event modules you can load it explicitly and enjoy the high
119explicitly and enjoy the high availability of that event loop :) 149availability of that event loop :)
120 150
121=head1 WATCHERS 151=head1 WATCHERS
122 152
123AnyEvent has the central concept of a I<watcher>, which is an object that 153AnyEvent has the central concept of a I<watcher>, which is an object that
124stores relevant data for each kind of event you are waiting for, such as 154stores relevant data for each kind of event you are waiting for, such as
127These watchers are normal Perl objects with normal Perl lifetime. After 157These watchers are normal Perl objects with normal Perl lifetime. After
128creating a watcher it will immediately "watch" for events and invoke the 158creating a watcher it will immediately "watch" for events and invoke the
129callback when the event occurs (of course, only when the event model 159callback when the event occurs (of course, only when the event model
130is in control). 160is in control).
131 161
162Note that B<callbacks must not permanently change global variables>
163potentially in use by the event loop (such as C<$_> or C<$[>) and that B<<
164callbacks must not C<die> >>. The former is good programming practice in
165Perl and the latter stems from the fact that exception handling differs
166widely between event loops.
167
132To disable the watcher you have to destroy it (e.g. by setting the 168To disable a watcher you have to destroy it (e.g. by setting the
133variable you store it in to C<undef> or otherwise deleting all references 169variable you store it in to C<undef> or otherwise deleting all references
134to it). 170to it).
135 171
136All watchers are created by calling a method on the C<AnyEvent> class. 172All watchers are created by calling a method on the C<AnyEvent> class.
137 173
138Many watchers either are used with "recursion" (repeating timers for 174Many watchers either are used with "recursion" (repeating timers for
139example), or need to refer to their watcher object in other ways. 175example), or need to refer to their watcher object in other ways.
140 176
141An any way to achieve that is this pattern: 177One way to achieve that is this pattern:
142 178
143 my $w; $w = AnyEvent->type (arg => value ..., cb => sub { 179 my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
144 # you can use $w here, for example to undef it 180 # you can use $w here, for example to undef it
145 undef $w; 181 undef $w;
146 }); 182 });
149my variables are only visible after the statement in which they are 185my variables are only visible after the statement in which they are
150declared. 186declared.
151 187
152=head2 I/O WATCHERS 188=head2 I/O WATCHERS
153 189
190 $w = AnyEvent->io (
191 fh => <filehandle_or_fileno>,
192 poll => <"r" or "w">,
193 cb => <callback>,
194 );
195
154You can create an I/O watcher by calling the C<< AnyEvent->io >> method 196You can create an I/O watcher by calling the C<< AnyEvent->io >> method
155with the following mandatory key-value pairs as arguments: 197with the following mandatory key-value pairs as arguments:
156 198
157C<fh> the Perl I<file handle> (I<not> file descriptor) to watch 199C<fh> is the Perl I<file handle> (or a naked file descriptor) to watch
200for events (AnyEvent might or might not keep a reference to this file
201handle). Note that only file handles pointing to things for which
202non-blocking operation makes sense are allowed. This includes sockets,
203most character devices, pipes, fifos and so on, but not for example files
204or block devices.
205
158for events. C<poll> must be a string that is either C<r> or C<w>, 206C<poll> must be a string that is either C<r> or C<w>, which creates a
159which creates a watcher waiting for "r"eadable or "w"ritable events, 207watcher waiting for "r"eadable or "w"ritable events, respectively.
208
160respectively. C<cb> is the callback to invoke each time the file handle 209C<cb> is the callback to invoke each time the file handle becomes ready.
161becomes ready.
162 210
163Although the callback might get passed parameters, their value and 211Although the callback might get passed parameters, their value and
164presence is undefined and you cannot rely on them. Portable AnyEvent 212presence is undefined and you cannot rely on them. Portable AnyEvent
165callbacks cannot use arguments passed to I/O watcher callbacks. 213callbacks cannot use arguments passed to I/O watcher callbacks.
166 214
167The I/O watcher might use the underlying file descriptor or a copy of it. 215The I/O watcher might use the underlying file descriptor or a copy of it.
168You must not close a file handle as long as any watcher is active on the 216You must not close a file handle as long as any watcher is active on the
169underlying file descriptor. 217underlying file descriptor.
170 218
171Some event loops issue spurious readyness notifications, so you should 219Some event loops issue spurious readiness notifications, so you should
172always use non-blocking calls when reading/writing from/to your file 220always use non-blocking calls when reading/writing from/to your file
173handles. 221handles.
174 222
175Example:
176
177 # wait for readability of STDIN, then read a line and disable the watcher 223Example: wait for readability of STDIN, then read a line and disable the
224watcher.
225
178 my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub { 226 my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
179 chomp (my $input = <STDIN>); 227 chomp (my $input = <STDIN>);
180 warn "read: $input\n"; 228 warn "read: $input\n";
181 undef $w; 229 undef $w;
182 }); 230 });
183 231
184=head2 TIME WATCHERS 232=head2 TIME WATCHERS
185 233
234 $w = AnyEvent->timer (after => <seconds>, cb => <callback>);
235
236 $w = AnyEvent->timer (
237 after => <fractional_seconds>,
238 interval => <fractional_seconds>,
239 cb => <callback>,
240 );
241
186You can create a time watcher by calling the C<< AnyEvent->timer >> 242You can create a time watcher by calling the C<< AnyEvent->timer >>
187method with the following mandatory arguments: 243method with the following mandatory arguments:
188 244
189C<after> specifies after how many seconds (fractional values are 245C<after> specifies after how many seconds (fractional values are
190supported) the callback should be invoked. C<cb> is the callback to invoke 246supported) the callback should be invoked. C<cb> is the callback to invoke
192 248
193Although the callback might get passed parameters, their value and 249Although the callback might get passed parameters, their value and
194presence is undefined and you cannot rely on them. Portable AnyEvent 250presence is undefined and you cannot rely on them. Portable AnyEvent
195callbacks cannot use arguments passed to time watcher callbacks. 251callbacks cannot use arguments passed to time watcher callbacks.
196 252
197The timer callback will be invoked at most once: if you want a repeating 253The callback will normally be invoked only once. If you specify another
198timer you have to create a new watcher (this is a limitation by both Tk 254parameter, C<interval>, as a strictly positive number (> 0), then the
199and Glib). 255callback will be invoked regularly at that interval (in fractional
256seconds) after the first invocation. If C<interval> is specified with a
257false value, then it is treated as if it were not specified at all.
200 258
201Example: 259The callback will be rescheduled before invoking the callback, but no
260attempt is made to avoid timer drift in most backends, so the interval is
261only approximate.
202 262
203 # fire an event after 7.7 seconds 263Example: fire an event after 7.7 seconds.
264
204 my $w = AnyEvent->timer (after => 7.7, cb => sub { 265 my $w = AnyEvent->timer (after => 7.7, cb => sub {
205 warn "timeout\n"; 266 warn "timeout\n";
206 }); 267 });
207 268
208 # to cancel the timer: 269 # to cancel the timer:
209 undef $w; 270 undef $w;
210 271
211Example 2:
212
213 # fire an event after 0.5 seconds, then roughly every second 272Example 2: fire an event after 0.5 seconds, then roughly every second.
214 my $w;
215 273
216 my $cb = sub {
217 # cancel the old timer while creating a new one
218 $w = AnyEvent->timer (after => 1, cb => $cb); 274 my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
275 warn "timeout\n";
219 }; 276 };
220
221 # start the "loop" by creating the first watcher
222 $w = AnyEvent->timer (after => 0.5, cb => $cb);
223 277
224=head3 TIMING ISSUES 278=head3 TIMING ISSUES
225 279
226There are two ways to handle timers: based on real time (relative, "fire 280There are two ways to handle timers: based on real time (relative, "fire
227in 10 seconds") and based on wallclock time (absolute, "fire at 12 281in 10 seconds") and based on wallclock time (absolute, "fire at 12
229 283
230While most event loops expect timers to specified in a relative way, they 284While most event loops expect timers to specified in a relative way, they
231use absolute time internally. This makes a difference when your clock 285use absolute time internally. This makes a difference when your clock
232"jumps", for example, when ntp decides to set your clock backwards from 286"jumps", for example, when ntp decides to set your clock backwards from
233the wrong date of 2014-01-01 to 2008-01-01, a watcher that is supposed to 287the wrong date of 2014-01-01 to 2008-01-01, a watcher that is supposed to
234fire "after" a second might actually take six years to finally fire. 288fire "after a second" might actually take six years to finally fire.
235 289
236AnyEvent cannot compensate for this. The only event loop that is conscious 290AnyEvent cannot compensate for this. The only event loop that is conscious
237about these issues is L<EV>, which offers both relative (ev_timer, based 291of these issues is L<EV>, which offers both relative (ev_timer, based
238on true relative time) and absolute (ev_periodic, based on wallclock time) 292on true relative time) and absolute (ev_periodic, based on wallclock time)
239timers. 293timers.
240 294
241AnyEvent always prefers relative timers, if available, matching the 295AnyEvent always prefers relative timers, if available, matching the
242AnyEvent API. 296AnyEvent API.
264I<In almost all cases (in all cases if you don't care), this is the 318I<In almost all cases (in all cases if you don't care), this is the
265function to call when you want to know the current time.> 319function to call when you want to know the current time.>
266 320
267This function is also often faster then C<< AnyEvent->time >>, and 321This function is also often faster then C<< AnyEvent->time >>, and
268thus the preferred method if you want some timestamp (for example, 322thus the preferred method if you want some timestamp (for example,
269L<AnyEvent::Handle> uses this to update it's activity timeouts). 323L<AnyEvent::Handle> uses this to update its activity timeouts).
270 324
271The rest of this section is only of relevance if you try to be very exact 325The rest of this section is only of relevance if you try to be very exact
272with your timing, you can skip it without bad conscience. 326with your timing; you can skip it without a bad conscience.
273 327
274For a practical example of when these times differ, consider L<Event::Lib> 328For a practical example of when these times differ, consider L<Event::Lib>
275and L<EV> and the following set-up: 329and L<EV> and the following set-up:
276 330
277The event loop is running and has just invoked one of your callback at 331The event loop is running and has just invoked one of your callbacks at
278time=500 (assume no other callbacks delay processing). In your callback, 332time=500 (assume no other callbacks delay processing). In your callback,
279you wait a second by executing C<sleep 1> (blocking the process for a 333you wait a second by executing C<sleep 1> (blocking the process for a
280second) and then (at time=501) you create a relative timer that fires 334second) and then (at time=501) you create a relative timer that fires
281after three seconds. 335after three seconds.
282 336
300In either case, if you care (and in most cases, you don't), then you 354In either case, if you care (and in most cases, you don't), then you
301can get whatever behaviour you want with any event loop, by taking the 355can get whatever behaviour you want with any event loop, by taking the
302difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into 356difference between C<< AnyEvent->time >> and C<< AnyEvent->now >> into
303account. 357account.
304 358
359=item AnyEvent->now_update
360
361Some event loops (such as L<EV> or L<AnyEvent::Loop>) cache the current
362time for each loop iteration (see the discussion of L<< AnyEvent->now >>,
363above).
364
365When a callback runs for a long time (or when the process sleeps), then
366this "current" time will differ substantially from the real time, which
367might affect timers and time-outs.
368
369When this is the case, you can call this method, which will update the
370event loop's idea of "current time".
371
372A typical example would be a script in a web server (e.g. C<mod_perl>) -
373when mod_perl executes the script, then the event loop will have the wrong
374idea about the "current time" (being potentially far in the past, when the
375script ran the last time). In that case you should arrange a call to C<<
376AnyEvent->now_update >> each time the web server process wakes up again
377(e.g. at the start of your script, or in a handler).
378
379Note that updating the time I<might> cause some events to be handled.
380
305=back 381=back
306 382
307=head2 SIGNAL WATCHERS 383=head2 SIGNAL WATCHERS
308 384
385 $w = AnyEvent->signal (signal => <uppercase_signal_name>, cb => <callback>);
386
309You can watch for signals using a signal watcher, C<signal> is the signal 387You can watch for signals using a signal watcher, C<signal> is the signal
310I<name> without any C<SIG> prefix, C<cb> is the Perl callback to 388I<name> in uppercase and without any C<SIG> prefix, C<cb> is the Perl
311be invoked whenever a signal occurs. 389callback to be invoked whenever a signal occurs.
312 390
313Although the callback might get passed parameters, their value and 391Although the callback might get passed parameters, their value and
314presence is undefined and you cannot rely on them. Portable AnyEvent 392presence is undefined and you cannot rely on them. Portable AnyEvent
315callbacks cannot use arguments passed to signal watcher callbacks. 393callbacks cannot use arguments passed to signal watcher callbacks.
316 394
318invocation, and callback invocation will be synchronous. Synchronous means 396invocation, and callback invocation will be synchronous. Synchronous means
319that it might take a while until the signal gets handled by the process, 397that it might take a while until the signal gets handled by the process,
320but it is guaranteed not to interrupt any other callbacks. 398but it is guaranteed not to interrupt any other callbacks.
321 399
322The main advantage of using these watchers is that you can share a signal 400The main advantage of using these watchers is that you can share a signal
323between multiple watchers. 401between multiple watchers, and AnyEvent will ensure that signals will not
402interrupt your program at bad times.
324 403
325This watcher might use C<%SIG>, so programs overwriting those signals 404This watcher might use C<%SIG> (depending on the event loop used),
326directly will likely not work correctly. 405so programs overwriting those signals directly will likely not work
406correctly.
327 407
328Example: exit on SIGINT 408Example: exit on SIGINT
329 409
330 my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 }); 410 my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 });
331 411
412=head3 Restart Behaviour
413
414While restart behaviour is up to the event loop implementation, most will
415not restart syscalls (that includes L<Async::Interrupt> and AnyEvent's
416pure perl implementation).
417
418=head3 Safe/Unsafe Signals
419
420Perl signals can be either "safe" (synchronous to opcode handling) or
421"unsafe" (asynchronous) - the former might get delayed indefinitely, the
422latter might corrupt your memory.
423
424AnyEvent signal handlers are, in addition, synchronous to the event loop,
425i.e. they will not interrupt your running perl program but will only be
426called as part of the normal event handling (just like timer, I/O etc.
427callbacks, too).
428
429=head3 Signal Races, Delays and Workarounds
430
431Many event loops (e.g. Glib, Tk, Qt, IO::Async) do not support attaching
432callbacks to signals in a generic way, which is a pity, as you cannot
433do race-free signal handling in perl, requiring C libraries for
434this. AnyEvent will try to do its best, which means in some cases,
435signals will be delayed. The maximum time a signal might be delayed is
436specified in C<$AnyEvent::MAX_SIGNAL_LATENCY> (default: 10 seconds). This
437variable can be changed only before the first signal watcher is created,
438and should be left alone otherwise. This variable determines how often
439AnyEvent polls for signals (in case a wake-up was missed). Higher values
440will cause fewer spurious wake-ups, which is better for power and CPU
441saving.
442
443All these problems can be avoided by installing the optional
444L<Async::Interrupt> module, which works with most event loops. It will not
445work with inherently broken event loops such as L<Event> or L<Event::Lib>
446(and not with L<POE> currently, as POE does its own workaround with
447one-second latency). For those, you just have to suffer the delays.
448
332=head2 CHILD PROCESS WATCHERS 449=head2 CHILD PROCESS WATCHERS
333 450
451 $w = AnyEvent->child (pid => <process id>, cb => <callback>);
452
334You can also watch on a child process exit and catch its exit status. 453You can also watch for a child process exit and catch its exit status.
335 454
336The child process is specified by the C<pid> argument (if set to C<0>, it 455The child process is specified by the C<pid> argument (on some backends,
337watches for any child process exit). The watcher will trigger as often 456using C<0> watches for any child process exit, on others this will
338as status change for the child are received. This works by installing a 457croak). The watcher will be triggered only when the child process has
339signal handler for C<SIGCHLD>. The callback will be called with the pid 458finished and an exit status is available, not on any trace events
340and exit status (as returned by waitpid), so unlike other watcher types, 459(stopped/continued).
341you I<can> rely on child watcher callback arguments. 460
461The callback will be called with the pid and exit status (as returned by
462waitpid), so unlike other watcher types, you I<can> rely on child watcher
463callback arguments.
464
465This watcher type works by installing a signal handler for C<SIGCHLD>,
466and since it cannot be shared, nothing else should use SIGCHLD or reap
467random child processes (waiting for specific child processes, e.g. inside
468C<system>, is just fine).
342 469
343There is a slight catch to child watchers, however: you usually start them 470There is a slight catch to child watchers, however: you usually start them
344I<after> the child process was created, and this means the process could 471I<after> the child process was created, and this means the process could
345have exited already (and no SIGCHLD will be sent anymore). 472have exited already (and no SIGCHLD will be sent anymore).
346 473
347Not all event models handle this correctly (POE doesn't), but even for 474Not all event models handle this correctly (neither POE nor IO::Async do,
475see their AnyEvent::Impl manpages for details), but even for event models
348event models that I<do> handle this correctly, they usually need to be 476that I<do> handle this correctly, they usually need to be loaded before
349loaded before the process exits (i.e. before you fork in the first place). 477the process exits (i.e. before you fork in the first place). AnyEvent's
478pure perl event loop handles all cases correctly regardless of when you
479start the watcher.
350 480
351This means you cannot create a child watcher as the very first thing in an 481This means you cannot create a child watcher as the very first
352AnyEvent program, you I<have> to create at least one watcher before you 482thing in an AnyEvent program, you I<have> to create at least one
353C<fork> the child (alternatively, you can call C<AnyEvent::detect>). 483watcher before you C<fork> the child (alternatively, you can call
484C<AnyEvent::detect>).
485
486As most event loops do not support waiting for child events, they will be
487emulated by AnyEvent in most cases, in which case the latency and race
488problems mentioned in the description of signal watchers apply.
354 489
355Example: fork a process and wait for it 490Example: fork a process and wait for it
356 491
357 my $done = AnyEvent->condvar; 492 my $done = AnyEvent->condvar;
358 493
368 ); 503 );
369 504
370 # do something else, then wait for process exit 505 # do something else, then wait for process exit
371 $done->recv; 506 $done->recv;
372 507
508=head2 IDLE WATCHERS
509
510 $w = AnyEvent->idle (cb => <callback>);
511
512This will repeatedly invoke the callback after the process becomes idle,
513until either the watcher is destroyed or new events have been detected.
514
515Idle watchers are useful when there is a need to do something, but it
516is not so important (or wise) to do it instantly. The callback will be
517invoked only when there is "nothing better to do", which is usually
518defined as "all outstanding events have been handled and no new events
519have been detected". That means that idle watchers ideally get invoked
520when the event loop has just polled for new events but none have been
521detected. Instead of blocking to wait for more events, the idle watchers
522will be invoked.
523
524Unfortunately, most event loops do not really support idle watchers (only
525EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent
526will simply call the callback "from time to time".
527
528Example: read lines from STDIN, but only process them when the
529program is otherwise idle:
530
531 my @lines; # read data
532 my $idle_w;
533 my $io_w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
534 push @lines, scalar <STDIN>;
535
536 # start an idle watcher, if not already done
537 $idle_w ||= AnyEvent->idle (cb => sub {
538 # handle only one line, when there are lines left
539 if (my $line = shift @lines) {
540 print "handled when idle: $line";
541 } else {
542 # otherwise disable the idle watcher again
543 undef $idle_w;
544 }
545 });
546 });
547
373=head2 CONDITION VARIABLES 548=head2 CONDITION VARIABLES
549
550 $cv = AnyEvent->condvar;
551
552 $cv->send (<list>);
553 my @res = $cv->recv;
374 554
375If you are familiar with some event loops you will know that all of them 555If you are familiar with some event loops you will know that all of them
376require you to run some blocking "loop", "run" or similar function that 556require you to run some blocking "loop", "run" or similar function that
377will actively watch for new events and call your callbacks. 557will actively watch for new events and call your callbacks.
378 558
379AnyEvent is different, it expects somebody else to run the event loop and 559AnyEvent is slightly different: it expects somebody else to run the event
380will only block when necessary (usually when told by the user). 560loop and will only block when necessary (usually when told by the user).
381 561
382The instrument to do that is called a "condition variable", so called 562The tool to do that is called a "condition variable", so called because
383because they represent a condition that must become true. 563they represent a condition that must become true.
564
565Now is probably a good time to look at the examples further below.
384 566
385Condition variables can be created by calling the C<< AnyEvent->condvar 567Condition variables can be created by calling the C<< AnyEvent->condvar
386>> method, usually without arguments. The only argument pair allowed is 568>> method, usually without arguments. The only argument pair allowed is
387C<cb>, which specifies a callback to be called when the condition variable 569C<cb>, which specifies a callback to be called when the condition variable
388becomes true. 570becomes true, with the condition variable as the first argument (but not
571the results).
389 572
390After creation, the condition variable is "false" until it becomes "true" 573After creation, the condition variable is "false" until it becomes "true"
391by calling the C<send> method (or calling the condition variable as if it 574by calling the C<send> method (or calling the condition variable as if it
392were a callback, read about the caveats in the description for the C<< 575were a callback, read about the caveats in the description for the C<<
393->send >> method). 576->send >> method).
394 577
395Condition variables are similar to callbacks, except that you can 578Since condition variables are the most complex part of the AnyEvent API, here are
396optionally wait for them. They can also be called merge points - points 579some different mental models of what they are - pick the ones you can connect to:
397in time where multiple outstanding events have been processed. And yet 580
398another way to call them is transactions - each condition variable can be 581=over 4
399used to represent a transaction, which finishes at some point and delivers 582
400a result. 583=item * Condition variables are like callbacks - you can call them (and pass them instead
584of callbacks). Unlike callbacks however, you can also wait for them to be called.
585
586=item * Condition variables are signals - one side can emit or send them,
587the other side can wait for them, or install a handler that is called when
588the signal fires.
589
590=item * Condition variables are like "Merge Points" - points in your program
591where you merge multiple independent results/control flows into one.
592
593=item * Condition variables represent a transaction - functions that start
594some kind of transaction can return them, leaving the caller the choice
595between waiting in a blocking fashion, or setting a callback.
596
597=item * Condition variables represent future values, or promises to deliver
598some result, long before the result is available.
599
600=back
401 601
402Condition variables are very useful to signal that something has finished, 602Condition variables are very useful to signal that something has finished,
403for example, if you write a module that does asynchronous http requests, 603for example, if you write a module that does asynchronous http requests,
404then a condition variable would be the ideal candidate to signal the 604then a condition variable would be the ideal candidate to signal the
405availability of results. The user can either act when the callback is 605availability of results. The user can either act when the callback is
418 618
419Condition variables are represented by hash refs in perl, and the keys 619Condition variables are represented by hash refs in perl, and the keys
420used by AnyEvent itself are all named C<_ae_XXX> to make subclassing 620used by AnyEvent itself are all named C<_ae_XXX> to make subclassing
421easy (it is often useful to build your own transaction class on top of 621easy (it is often useful to build your own transaction class on top of
422AnyEvent). To subclass, use C<AnyEvent::CondVar> as base class and call 622AnyEvent). To subclass, use C<AnyEvent::CondVar> as base class and call
423it's C<new> method in your own C<new> method. 623its C<new> method in your own C<new> method.
424 624
425There are two "sides" to a condition variable - the "producer side" which 625There are two "sides" to a condition variable - the "producer side" which
426eventually calls C<< -> send >>, and the "consumer side", which waits 626eventually calls C<< -> send >>, and the "consumer side", which waits
427for the send to occur. 627for the send to occur.
428 628
429Example: wait for a timer. 629Example: wait for a timer.
430 630
431 # wait till the result is ready 631 # condition: "wait till the timer is fired"
432 my $result_ready = AnyEvent->condvar; 632 my $timer_fired = AnyEvent->condvar;
433 633
434 # do something such as adding a timer 634 # create the timer - we could wait for, say
435 # or socket watcher the calls $result_ready->send 635 # a handle becomign ready, or even an
436 # when the "result" is ready. 636 # AnyEvent::HTTP request to finish, but
437 # in this case, we simply use a timer: 637 # in this case, we simply use a timer:
438 my $w = AnyEvent->timer ( 638 my $w = AnyEvent->timer (
439 after => 1, 639 after => 1,
440 cb => sub { $result_ready->send }, 640 cb => sub { $timer_fired->send },
441 ); 641 );
442 642
443 # this "blocks" (while handling events) till the callback 643 # this "blocks" (while handling events) till the callback
444 # calls send 644 # calls ->send
445 $result_ready->recv; 645 $timer_fired->recv;
446 646
447Example: wait for a timer, but take advantage of the fact that 647Example: wait for a timer, but take advantage of the fact that condition
448condition variables are also code references. 648variables are also callable directly.
449 649
450 my $done = AnyEvent->condvar; 650 my $done = AnyEvent->condvar;
451 my $delay = AnyEvent->timer (after => 5, cb => $done); 651 my $delay = AnyEvent->timer (after => 5, cb => $done);
452 $done->recv; 652 $done->recv;
653
654Example: Imagine an API that returns a condvar and doesn't support
655callbacks. This is how you make a synchronous call, for example from
656the main program:
657
658 use AnyEvent::CouchDB;
659
660 ...
661
662 my @info = $couchdb->info->recv;
663
664And this is how you would just set a callback to be called whenever the
665results are available:
666
667 $couchdb->info->cb (sub {
668 my @info = $_[0]->recv;
669 });
453 670
454=head3 METHODS FOR PRODUCERS 671=head3 METHODS FOR PRODUCERS
455 672
456These methods should only be used by the producing side, i.e. the 673These methods should only be used by the producing side, i.e. the
457code/module that eventually sends the signal. Note that it is also 674code/module that eventually sends the signal. Note that it is also
470immediately from within send. 687immediately from within send.
471 688
472Any arguments passed to the C<send> call will be returned by all 689Any arguments passed to the C<send> call will be returned by all
473future C<< ->recv >> calls. 690future C<< ->recv >> calls.
474 691
475Condition variables are overloaded so one can call them directly 692Condition variables are overloaded so one can call them directly (as if
476(as a code reference). Calling them directly is the same as calling 693they were a code reference). Calling them directly is the same as calling
477C<send>. Note, however, that many C-based event loops do not handle 694C<send>.
478overloading, so as tempting as it may be, passing a condition variable
479instead of a callback does not work. Both the pure perl and EV loops
480support overloading, however, as well as all functions that use perl to
481invoke a callback (as in L<AnyEvent::Socket> and L<AnyEvent::DNS> for
482example).
483 695
484=item $cv->croak ($error) 696=item $cv->croak ($error)
485 697
486Similar to send, but causes all call's to C<< ->recv >> to invoke 698Similar to send, but causes all calls to C<< ->recv >> to invoke
487C<Carp::croak> with the given error message/object/scalar. 699C<Carp::croak> with the given error message/object/scalar.
488 700
489This can be used to signal any errors to the condition variable 701This can be used to signal any errors to the condition variable
490user/consumer. 702user/consumer. Doing it this way instead of calling C<croak> directly
703delays the error detection, but has the overwhelming advantage that it
704diagnoses the error at the place where the result is expected, and not
705deep in some event callback with no connection to the actual code causing
706the problem.
491 707
492=item $cv->begin ([group callback]) 708=item $cv->begin ([group callback])
493 709
494=item $cv->end 710=item $cv->end
495
496These two methods are EXPERIMENTAL and MIGHT CHANGE.
497 711
498These two methods can be used to combine many transactions/events into 712These two methods can be used to combine many transactions/events into
499one. For example, a function that pings many hosts in parallel might want 713one. For example, a function that pings many hosts in parallel might want
500to use a condition variable for the whole process. 714to use a condition variable for the whole process.
501 715
502Every call to C<< ->begin >> will increment a counter, and every call to 716Every call to C<< ->begin >> will increment a counter, and every call to
503C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end 717C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end
504>>, the (last) callback passed to C<begin> will be executed. That callback 718>>, the (last) callback passed to C<begin> will be executed, passing the
505is I<supposed> to call C<< ->send >>, but that is not required. If no 719condvar as first argument. That callback is I<supposed> to call C<< ->send
506callback was set, C<send> will be called without any arguments. 720>>, but that is not required. If no group callback was set, C<send> will
721be called without any arguments.
507 722
508Let's clarify this with the ping example: 723You can think of C<< $cv->send >> giving you an OR condition (one call
724sends), while C<< $cv->begin >> and C<< $cv->end >> giving you an AND
725condition (all C<begin> calls must be C<end>'ed before the condvar sends).
726
727Let's start with a simple example: you have two I/O watchers (for example,
728STDOUT and STDERR for a program), and you want to wait for both streams to
729close before activating a condvar:
509 730
510 my $cv = AnyEvent->condvar; 731 my $cv = AnyEvent->condvar;
511 732
733 $cv->begin; # first watcher
734 my $w1 = AnyEvent->io (fh => $fh1, cb => sub {
735 defined sysread $fh1, my $buf, 4096
736 or $cv->end;
737 });
738
739 $cv->begin; # second watcher
740 my $w2 = AnyEvent->io (fh => $fh2, cb => sub {
741 defined sysread $fh2, my $buf, 4096
742 or $cv->end;
743 });
744
745 $cv->recv;
746
747This works because for every event source (EOF on file handle), there is
748one call to C<begin>, so the condvar waits for all calls to C<end> before
749sending.
750
751The ping example mentioned above is slightly more complicated, as the
752there are results to be passwd back, and the number of tasks that are
753begun can potentially be zero:
754
755 my $cv = AnyEvent->condvar;
756
512 my %result; 757 my %result;
513 $cv->begin (sub { $cv->send (\%result) }); 758 $cv->begin (sub { shift->send (\%result) });
514 759
515 for my $host (@list_of_hosts) { 760 for my $host (@list_of_hosts) {
516 $cv->begin; 761 $cv->begin;
517 ping_host_then_call_callback $host, sub { 762 ping_host_then_call_callback $host, sub {
518 $result{$host} = ...; 763 $result{$host} = ...;
533loop, which serves two important purposes: first, it sets the callback 778loop, which serves two important purposes: first, it sets the callback
534to be called once the counter reaches C<0>, and second, it ensures that 779to be called once the counter reaches C<0>, and second, it ensures that
535C<send> is called even when C<no> hosts are being pinged (the loop 780C<send> is called even when C<no> hosts are being pinged (the loop
536doesn't execute once). 781doesn't execute once).
537 782
538This is the general pattern when you "fan out" into multiple subrequests: 783This is the general pattern when you "fan out" into multiple (but
539use an outer C<begin>/C<end> pair to set the callback and ensure C<end> 784potentially zero) subrequests: use an outer C<begin>/C<end> pair to set
540is called at least once, and then, for each subrequest you start, call 785the callback and ensure C<end> is called at least once, and then, for each
541C<begin> and for each subrequest you finish, call C<end>. 786subrequest you start, call C<begin> and for each subrequest you finish,
787call C<end>.
542 788
543=back 789=back
544 790
545=head3 METHODS FOR CONSUMERS 791=head3 METHODS FOR CONSUMERS
546 792
550=over 4 796=over 4
551 797
552=item $cv->recv 798=item $cv->recv
553 799
554Wait (blocking if necessary) until the C<< ->send >> or C<< ->croak 800Wait (blocking if necessary) until the C<< ->send >> or C<< ->croak
555>> methods have been called on c<$cv>, while servicing other watchers 801>> methods have been called on C<$cv>, while servicing other watchers
556normally. 802normally.
557 803
558You can only wait once on a condition - additional calls are valid but 804You can only wait once on a condition - additional calls are valid but
559will return immediately. 805will return immediately.
560 806
562function will call C<croak>. 808function will call C<croak>.
563 809
564In list context, all parameters passed to C<send> will be returned, 810In list context, all parameters passed to C<send> will be returned,
565in scalar context only the first one will be returned. 811in scalar context only the first one will be returned.
566 812
813Note that doing a blocking wait in a callback is not supported by any
814event loop, that is, recursive invocation of a blocking C<< ->recv
815>> is not allowed, and the C<recv> call will C<croak> if such a
816condition is detected. This condition can be slightly loosened by using
817L<Coro::AnyEvent>, which allows you to do a blocking C<< ->recv >> from
818any thread that doesn't run the event loop itself.
819
567Not all event models support a blocking wait - some die in that case 820Not all event models support a blocking wait - some die in that case
568(programs might want to do that to stay interactive), so I<if you are 821(programs might want to do that to stay interactive), so I<if you are
569using this from a module, never require a blocking wait>, but let the 822using this from a module, never require a blocking wait>. Instead, let the
570caller decide whether the call will block or not (for example, by coupling 823caller decide whether the call will block or not (for example, by coupling
571condition variables with some kind of request results and supporting 824condition variables with some kind of request results and supporting
572callbacks so the caller knows that getting the result will not block, 825callbacks so the caller knows that getting the result will not block,
573while still supporting blocking waits if the caller so desires). 826while still supporting blocking waits if the caller so desires).
574 827
575Another reason I<never> to C<< ->recv >> in a module is that you cannot
576sensibly have two C<< ->recv >>'s in parallel, as that would require
577multiple interpreters or coroutines/threads, none of which C<AnyEvent>
578can supply.
579
580The L<Coro> module, however, I<can> and I<does> supply coroutines and, in
581fact, L<Coro::AnyEvent> replaces AnyEvent's condvars by coroutine-safe
582versions and also integrates coroutines into AnyEvent, making blocking
583C<< ->recv >> calls perfectly safe as long as they are done from another
584coroutine (one that doesn't run the event loop).
585
586You can ensure that C<< -recv >> never blocks by setting a callback and 828You can ensure that C<< ->recv >> never blocks by setting a callback and
587only calling C<< ->recv >> from within that callback (or at a later 829only calling C<< ->recv >> from within that callback (or at a later
588time). This will work even when the event loop does not support blocking 830time). This will work even when the event loop does not support blocking
589waits otherwise. 831waits otherwise.
590 832
591=item $bool = $cv->ready 833=item $bool = $cv->ready
592 834
593Returns true when the condition is "true", i.e. whether C<send> or 835Returns true when the condition is "true", i.e. whether C<send> or
594C<croak> have been called. 836C<croak> have been called.
595 837
596=item $cb = $cv->cb ([new callback]) 838=item $cb = $cv->cb ($cb->($cv))
597 839
598This is a mutator function that returns the callback set and optionally 840This is a mutator function that returns the callback set and optionally
599replaces it before doing so. 841replaces it before doing so.
600 842
601The callback will be called when the condition becomes "true", i.e. when 843The callback will be called when the condition becomes "true", i.e. when
602C<send> or C<croak> are called, with the only argument being the condition 844C<send> or C<croak> are called, with the only argument being the
603variable itself. Calling C<recv> inside the callback or at any later time 845condition variable itself. If the condition is already true, the
604is guaranteed not to block. 846callback is called immediately when it is set. Calling C<recv> inside
847the callback or at any later time is guaranteed not to block.
605 848
606=back 849=back
607 850
851=head1 SUPPORTED EVENT LOOPS/BACKENDS
852
853The available backend classes are (every class has its own manpage):
854
855=over 4
856
857=item Backends that are autoprobed when no other event loop can be found.
858
859EV is the preferred backend when no other event loop seems to be in
860use. If EV is not installed, then AnyEvent will fall back to its own
861pure-perl implementation, which is available everywhere as it comes with
862AnyEvent itself.
863
864 AnyEvent::Impl::EV based on EV (interface to libev, best choice).
865 AnyEvent::Impl::Perl pure-perl AnyEvent::Loop, fast and portable.
866
867=item Backends that are transparently being picked up when they are used.
868
869These will be used if they are already loaded when the first watcher
870is created, in which case it is assumed that the application is using
871them. This means that AnyEvent will automatically pick the right backend
872when the main program loads an event module before anything starts to
873create watchers. Nothing special needs to be done by the main program.
874
875 AnyEvent::Impl::Event based on Event, very stable, few glitches.
876 AnyEvent::Impl::Glib based on Glib, slow but very stable.
877 AnyEvent::Impl::Tk based on Tk, very broken.
878 AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
879 AnyEvent::Impl::POE based on POE, very slow, some limitations.
880 AnyEvent::Impl::Irssi used when running within irssi.
881 AnyEvent::Impl::IOAsync based on IO::Async.
882 AnyEvent::Impl::Cocoa based on Cocoa::EventLoop.
883 AnyEvent::Impl::FLTK2 based on FLTK (fltk 2 binding).
884
885=item Backends with special needs.
886
887Qt requires the Qt::Application to be instantiated first, but will
888otherwise be picked up automatically. As long as the main program
889instantiates the application before any AnyEvent watchers are created,
890everything should just work.
891
892 AnyEvent::Impl::Qt based on Qt.
893
894=item Event loops that are indirectly supported via other backends.
895
896Some event loops can be supported via other modules:
897
898There is no direct support for WxWidgets (L<Wx>) or L<Prima>.
899
900B<WxWidgets> has no support for watching file handles. However, you can
901use WxWidgets through the POE adaptor, as POE has a Wx backend that simply
902polls 20 times per second, which was considered to be too horrible to even
903consider for AnyEvent.
904
905B<Prima> is not supported as nobody seems to be using it, but it has a POE
906backend, so it can be supported through POE.
907
908AnyEvent knows about both L<Prima> and L<Wx>, however, and will try to
909load L<POE> when detecting them, in the hope that POE will pick them up,
910in which case everything will be automatic.
911
912=back
913
608=head1 GLOBAL VARIABLES AND FUNCTIONS 914=head1 GLOBAL VARIABLES AND FUNCTIONS
609 915
916These are not normally required to use AnyEvent, but can be useful to
917write AnyEvent extension modules.
918
610=over 4 919=over 4
611 920
612=item $AnyEvent::MODEL 921=item $AnyEvent::MODEL
613 922
614Contains C<undef> until the first watcher is being created. Then it 923Contains C<undef> until the first watcher is being created, before the
924backend has been autodetected.
925
615contains the event model that is being used, which is the name of the 926Afterwards it contains the event model that is being used, which is the
616Perl class implementing the model. This class is usually one of the 927name of the Perl class implementing the model. This class is usually one
617C<AnyEvent::Impl:xxx> modules, but can be any other class in the case 928of the C<AnyEvent::Impl::xxx> modules, but can be any other class in the
618AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode>). 929case AnyEvent has been extended at runtime (e.g. in I<rxvt-unicode> it
619 930will be C<urxvt::anyevent>).
620The known classes so far are:
621
622 AnyEvent::Impl::EV based on EV (an interface to libev, best choice).
623 AnyEvent::Impl::Event based on Event, second best choice.
624 AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
625 AnyEvent::Impl::Glib based on Glib, third-best choice.
626 AnyEvent::Impl::Tk based on Tk, very bad choice.
627 AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs).
628 AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
629 AnyEvent::Impl::POE based on POE, not generic enough for full support.
630
631There is no support for WxWidgets, as WxWidgets has no support for
632watching file handles. However, you can use WxWidgets through the
633POE Adaptor, as POE has a Wx backend that simply polls 20 times per
634second, which was considered to be too horrible to even consider for
635AnyEvent. Likewise, other POE backends can be used by AnyEvent by using
636it's adaptor.
637
638AnyEvent knows about L<Prima> and L<Wx> and will try to use L<POE> when
639autodetecting them.
640 931
641=item AnyEvent::detect 932=item AnyEvent::detect
642 933
643Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model 934Returns C<$AnyEvent::MODEL>, forcing autodetection of the event model
644if necessary. You should only call this function right before you would 935if necessary. You should only call this function right before you would
645have created an AnyEvent watcher anyway, that is, as late as possible at 936have created an AnyEvent watcher anyway, that is, as late as possible at
646runtime. 937runtime, and not e.g. during initialisation of your module.
938
939If you need to do some initialisation before AnyEvent watchers are
940created, use C<post_detect>.
647 941
648=item $guard = AnyEvent::post_detect { BLOCK } 942=item $guard = AnyEvent::post_detect { BLOCK }
649 943
650Arranges for the code block to be executed as soon as the event model is 944Arranges for the code block to be executed as soon as the event model is
651autodetected (or immediately if this has already happened). 945autodetected (or immediately if that has already happened).
946
947The block will be executed I<after> the actual backend has been detected
948(C<$AnyEvent::MODEL> is set), but I<before> any watchers have been
949created, so it is possible to e.g. patch C<@AnyEvent::ISA> or do
950other initialisations - see the sources of L<AnyEvent::Strict> or
951L<AnyEvent::AIO> to see how this is used.
952
953The most common usage is to create some global watchers, without forcing
954event module detection too early, for example, L<AnyEvent::AIO> creates
955and installs the global L<IO::AIO> watcher in a C<post_detect> block to
956avoid autodetecting the event module at load time.
652 957
653If called in scalar or list context, then it creates and returns an object 958If called in scalar or list context, then it creates and returns an object
654that automatically removes the callback again when it is destroyed. See 959that automatically removes the callback again when it is destroyed (or
960C<undef> when the hook was immediately executed). See L<AnyEvent::AIO> for
655L<Coro::BDB> for a case where this is useful. 961a case where this is useful.
962
963Example: Create a watcher for the IO::AIO module and store it in
964C<$WATCHER>, but do so only do so after the event loop is initialised.
965
966 our WATCHER;
967
968 my $guard = AnyEvent::post_detect {
969 $WATCHER = AnyEvent->io (fh => IO::AIO::poll_fileno, poll => 'r', cb => \&IO::AIO::poll_cb);
970 };
971
972 # the ||= is important in case post_detect immediately runs the block,
973 # as to not clobber the newly-created watcher. assigning both watcher and
974 # post_detect guard to the same variable has the advantage of users being
975 # able to just C<undef $WATCHER> if the watcher causes them grief.
976
977 $WATCHER ||= $guard;
656 978
657=item @AnyEvent::post_detect 979=item @AnyEvent::post_detect
658 980
659If there are any code references in this array (you can C<push> to it 981If there are any code references in this array (you can C<push> to it
660before or after loading AnyEvent), then they will called directly after 982before or after loading AnyEvent), then they will be called directly
661the event loop has been chosen. 983after the event loop has been chosen.
662 984
663You should check C<$AnyEvent::MODEL> before adding to this array, though: 985You should check C<$AnyEvent::MODEL> before adding to this array, though:
664if it contains a true value then the event loop has already been detected, 986if it is defined then the event loop has already been detected, and the
665and the array will be ignored. 987array will be ignored.
666 988
667Best use C<AnyEvent::post_detect { BLOCK }> instead. 989Best use C<AnyEvent::post_detect { BLOCK }> when your application allows
990it, as it takes care of these details.
991
992This variable is mainly useful for modules that can do something useful
993when AnyEvent is used and thus want to know when it is initialised, but do
994not need to even load it by default. This array provides the means to hook
995into AnyEvent passively, without loading it.
996
997Example: To load Coro::AnyEvent whenever Coro and AnyEvent are used
998together, you could put this into Coro (this is the actual code used by
999Coro to accomplish this):
1000
1001 if (defined $AnyEvent::MODEL) {
1002 # AnyEvent already initialised, so load Coro::AnyEvent
1003 require Coro::AnyEvent;
1004 } else {
1005 # AnyEvent not yet initialised, so make sure to load Coro::AnyEvent
1006 # as soon as it is
1007 push @AnyEvent::post_detect, sub { require Coro::AnyEvent };
1008 }
1009
1010=item AnyEvent::postpone { BLOCK }
1011
1012Arranges for the block to be executed as soon as possible, but not before
1013the call itself returns. In practise, the block will be executed just
1014before the event loop polls for new events, or shortly afterwards.
1015
1016This function never returns anything (to make the C<return postpone { ...
1017}> idiom more useful.
1018
1019To understand the usefulness of this function, consider a function that
1020asynchronously does something for you and returns some transaction
1021object or guard to let you cancel the operation. For example,
1022C<AnyEvent::Socket::tcp_connect>:
1023
1024 # start a conenction attempt unless one is active
1025 $self->{connect_guard} ||= AnyEvent::Socket::tcp_connect "www.example.net", 80, sub {
1026 delete $self->{connect_guard};
1027 ...
1028 };
1029
1030Imagine that this function could instantly call the callback, for
1031example, because it detects an obvious error such as a negative port
1032number. Invoking the callback before the function returns causes problems
1033however: the callback will be called and will try to delete the guard
1034object. But since the function hasn't returned yet, there is nothing to
1035delete. When the function eventually returns it will assign the guard
1036object to C<< $self->{connect_guard} >>, where it will likely never be
1037deleted, so the program thinks it is still trying to connect.
1038
1039This is where C<AnyEvent::postpone> should be used. Instead of calling the
1040callback directly on error:
1041
1042 $cb->(undef), return # signal error to callback, BAD!
1043 if $some_error_condition;
1044
1045It should use C<postpone>:
1046
1047 AnyEvent::postpone { $cb->(undef) }, return # signal error to callback, later
1048 if $some_error_condition;
668 1049
669=back 1050=back
670 1051
671=head1 WHAT TO DO IN A MODULE 1052=head1 WHAT TO DO IN A MODULE
672 1053
683because it will stall the whole program, and the whole point of using 1064because it will stall the whole program, and the whole point of using
684events is to stay interactive. 1065events is to stay interactive.
685 1066
686It is fine, however, to call C<< ->recv >> when the user of your module 1067It is fine, however, to call C<< ->recv >> when the user of your module
687requests it (i.e. if you create a http request object ad have a method 1068requests it (i.e. if you create a http request object ad have a method
688called C<results> that returns the results, it should call C<< ->recv >> 1069called C<results> that returns the results, it may call C<< ->recv >>
689freely, as the user of your module knows what she is doing. always). 1070freely, as the user of your module knows what she is doing. Always).
690 1071
691=head1 WHAT TO DO IN THE MAIN PROGRAM 1072=head1 WHAT TO DO IN THE MAIN PROGRAM
692 1073
693There will always be a single main program - the only place that should 1074There will always be a single main program - the only place that should
694dictate which event model to use. 1075dictate which event model to use.
695 1076
696If it doesn't care, it can just "use AnyEvent" and use it itself, or not 1077If the program is not event-based, it need not do anything special, even
697do anything special (it does not need to be event-based) and let AnyEvent 1078when it depends on a module that uses an AnyEvent. If the program itself
698decide which implementation to chose if some module relies on it. 1079uses AnyEvent, but does not care which event loop is used, all it needs
1080to do is C<use AnyEvent>. In either case, AnyEvent will choose the best
1081available loop implementation.
699 1082
700If the main program relies on a specific event model - for example, in 1083If the main program relies on a specific event model - for example, in
701Gtk2 programs you have to rely on the Glib module - you should load the 1084Gtk2 programs you have to rely on the Glib module - you should load the
702event module before loading AnyEvent or any module that uses it: generally 1085event module before loading AnyEvent or any module that uses it: generally
703speaking, you should load it as early as possible. The reason is that 1086speaking, you should load it as early as possible. The reason is that
704modules might create watchers when they are loaded, and AnyEvent will 1087modules might create watchers when they are loaded, and AnyEvent will
705decide on the event model to use as soon as it creates watchers, and it 1088decide on the event model to use as soon as it creates watchers, and it
706might chose the wrong one unless you load the correct one yourself. 1089might choose the wrong one unless you load the correct one yourself.
707 1090
708You can chose to use a pure-perl implementation by loading the 1091You can chose to use a pure-perl implementation by loading the
709C<AnyEvent::Impl::Perl> module, which gives you similar behaviour 1092C<AnyEvent::Loop> module, which gives you similar behaviour
710everywhere, but letting AnyEvent chose the model is generally better. 1093everywhere, but letting AnyEvent chose the model is generally better.
711 1094
712=head2 MAINLOOP EMULATION 1095=head2 MAINLOOP EMULATION
713 1096
714Sometimes (often for short test scripts, or even standalone programs who 1097Sometimes (often for short test scripts, or even standalone programs who
727 1110
728 1111
729=head1 OTHER MODULES 1112=head1 OTHER MODULES
730 1113
731The following is a non-exhaustive list of additional modules that use 1114The following is a non-exhaustive list of additional modules that use
732AnyEvent and can therefore be mixed easily with other AnyEvent modules 1115AnyEvent as a client and can therefore be mixed easily with other AnyEvent
733in the same program. Some of the modules come with AnyEvent, some are 1116modules and other event loops in the same program. Some of the modules
734available via CPAN. 1117come as part of AnyEvent, the others are available via CPAN.
735 1118
736=over 4 1119=over 4
737 1120
738=item L<AnyEvent::Util> 1121=item L<AnyEvent::Util>
739 1122
740Contains various utility functions that replace often-used but blocking 1123Contains various utility functions that replace often-used blocking
741functions such as C<inet_aton> by event-/callback-based versions. 1124functions such as C<inet_aton> with event/callback-based versions.
742
743=item L<AnyEvent::Handle>
744
745Provide read and write buffers and manages watchers for reads and writes.
746 1125
747=item L<AnyEvent::Socket> 1126=item L<AnyEvent::Socket>
748 1127
749Provides various utility functions for (internet protocol) sockets, 1128Provides various utility functions for (internet protocol) sockets,
750addresses and name resolution. Also functions to create non-blocking tcp 1129addresses and name resolution. Also functions to create non-blocking tcp
751connections or tcp servers, with IPv6 and SRV record support and more. 1130connections or tcp servers, with IPv6 and SRV record support and more.
752 1131
1132=item L<AnyEvent::Handle>
1133
1134Provide read and write buffers, manages watchers for reads and writes,
1135supports raw and formatted I/O, I/O queued and fully transparent and
1136non-blocking SSL/TLS (via L<AnyEvent::TLS>).
1137
753=item L<AnyEvent::DNS> 1138=item L<AnyEvent::DNS>
754 1139
755Provides rich asynchronous DNS resolver capabilities. 1140Provides rich asynchronous DNS resolver capabilities.
756 1141
1142=item L<AnyEvent::HTTP>, L<AnyEvent::IRC>, L<AnyEvent::XMPP>, L<AnyEvent::GPSD>, L<AnyEvent::IGS>, L<AnyEvent::FCP>
1143
1144Implement event-based interfaces to the protocols of the same name (for
1145the curious, IGS is the International Go Server and FCP is the Freenet
1146Client Protocol).
1147
1148=item L<AnyEvent::Handle::UDP>
1149
1150Here be danger!
1151
1152As Pauli would put it, "Not only is it not right, it's not even wrong!" -
1153there are so many things wrong with AnyEvent::Handle::UDP, most notably
1154its use of a stream-based API with a protocol that isn't streamable, that
1155the only way to improve it is to delete it.
1156
1157It features data corruption (but typically only under load) and general
1158confusion. On top, the author is not only clueless about UDP but also
1159fact-resistant - some gems of his understanding: "connect doesn't work
1160with UDP", "UDP packets are not IP packets", "UDP only has datagrams, not
1161packets", "I don't need to implement proper error checking as UDP doesn't
1162support error checking" and so on - he doesn't even understand what's
1163wrong with his module when it is explained to him.
1164
757=item L<AnyEvent::HTTP> 1165=item L<AnyEvent::DBI>
758 1166
759A simple-to-use HTTP library that is capable of making a lot of concurrent 1167Executes L<DBI> requests asynchronously in a proxy process for you,
760HTTP requests. 1168notifying you in an event-based way when the operation is finished.
1169
1170=item L<AnyEvent::AIO>
1171
1172Truly asynchronous (as opposed to non-blocking) I/O, should be in the
1173toolbox of every event programmer. AnyEvent::AIO transparently fuses
1174L<IO::AIO> and AnyEvent together, giving AnyEvent access to event-based
1175file I/O, and much more.
761 1176
762=item L<AnyEvent::HTTPD> 1177=item L<AnyEvent::HTTPD>
763 1178
764Provides a simple web application server framework. 1179A simple embedded webserver.
765 1180
766=item L<AnyEvent::FastPing> 1181=item L<AnyEvent::FastPing>
767 1182
768The fastest ping in the west. 1183The fastest ping in the west.
769 1184
770=item L<Net::IRC3>
771
772AnyEvent based IRC client module family.
773
774=item L<Net::XMPP2>
775
776AnyEvent based XMPP (Jabber protocol) module family.
777
778=item L<Net::FCP>
779
780AnyEvent-based implementation of the Freenet Client Protocol, birthplace
781of AnyEvent.
782
783=item L<Event::ExecFlow>
784
785High level API for event-based execution flow control.
786
787=item L<Coro> 1185=item L<Coro>
788 1186
789Has special support for AnyEvent via L<Coro::AnyEvent>. 1187Has special support for AnyEvent via L<Coro::AnyEvent>.
790 1188
791=item L<AnyEvent::AIO>, L<IO::AIO>
792
793Truly asynchronous I/O, should be in the toolbox of every event
794programmer. AnyEvent::AIO transparently fuses IO::AIO and AnyEvent
795together.
796
797=item L<AnyEvent::BDB>, L<BDB>
798
799Truly asynchronous Berkeley DB access. AnyEvent::AIO transparently fuses
800IO::AIO and AnyEvent together.
801
802=item L<IO::Lambda>
803
804The lambda approach to I/O - don't ask, look there. Can use AnyEvent.
805
806=back 1189=back
807 1190
808=cut 1191=cut
809 1192
810package AnyEvent; 1193package AnyEvent;
811 1194
812no warnings; 1195# basically a tuned-down version of common::sense
813use strict; 1196sub common_sense {
1197 # from common:.sense 3.4
1198 ${^WARNING_BITS} ^= ${^WARNING_BITS} ^ "\x3c\x3f\x33\x00\x0f\xf0\x0f\xc0\xf0\xfc\x33\x00";
1199 # use strict vars subs - NO UTF-8, as Util.pm doesn't like this atm. (uts46data.pl)
1200 $^H |= 0x00000600;
1201}
814 1202
1203BEGIN { AnyEvent::common_sense }
1204
815use Carp; 1205use Carp ();
816 1206
817our $VERSION = 4.14; 1207our $VERSION = '5.34';
818our $MODEL; 1208our $MODEL;
819 1209
820our $AUTOLOAD; 1210our $AUTOLOAD;
821our @ISA; 1211our @ISA;
822 1212
823our @REGISTRY; 1213our @REGISTRY;
824 1214
825our $WIN32; 1215our $VERBOSE;
826 1216
827BEGIN { 1217BEGIN {
828 my $win32 = ! ! ($^O =~ /mswin32/i); 1218 require "AnyEvent/constants.pl";
829 eval "sub WIN32(){ $win32 }";
830}
831 1219
1220 eval "sub TAINT (){" . (${^TAINT}*1) . "}";
1221
1222 delete @ENV{grep /^PERL_ANYEVENT_/, keys %ENV}
1223 if ${^TAINT};
1224
832our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1; 1225 $VERBOSE = $ENV{PERL_ANYEVENT_VERBOSE}*1;
1226
1227}
1228
1229our $MAX_SIGNAL_LATENCY = 10;
833 1230
834our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred 1231our %PROTOCOL; # (ipv4|ipv6) => (1|2), higher numbers are preferred
835 1232
836{ 1233{
837 my $idx; 1234 my $idx;
838 $PROTOCOL{$_} = ++$idx 1235 $PROTOCOL{$_} = ++$idx
839 for reverse split /\s*,\s*/, 1236 for reverse split /\s*,\s*/,
840 $ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6"; 1237 $ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6";
841} 1238}
842 1239
1240our @post_detect;
1241
1242sub post_detect(&) {
1243 my ($cb) = @_;
1244
1245 push @post_detect, $cb;
1246
1247 defined wantarray
1248 ? bless \$cb, "AnyEvent::Util::postdetect"
1249 : ()
1250}
1251
1252sub AnyEvent::Util::postdetect::DESTROY {
1253 @post_detect = grep $_ != ${$_[0]}, @post_detect;
1254}
1255
1256our $POSTPONE_W;
1257our @POSTPONE;
1258
1259sub _postpone_exec {
1260 undef $POSTPONE_W;
1261
1262 &{ shift @POSTPONE }
1263 while @POSTPONE;
1264}
1265
1266sub postpone(&) {
1267 push @POSTPONE, shift;
1268
1269 $POSTPONE_W ||= AE::timer (0, 0, \&_postpone_exec);
1270
1271 ()
1272}
1273
843my @models = ( 1274our @models = (
844 [EV:: => AnyEvent::Impl::EV::], 1275 [EV:: => AnyEvent::Impl::EV:: , 1],
845 [Event:: => AnyEvent::Impl::Event::], 1276 [AnyEvent::Loop:: => AnyEvent::Impl::Perl:: , 1],
846 [AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl::],
847 # everything below here will not be autoprobed 1277 # everything below here will not (normally) be autoprobed
848 # as the pureperl backend should work everywhere 1278 # as the pure perl backend should work everywhere
849 # and is usually faster 1279 # and is usually faster
1280 [Event:: => AnyEvent::Impl::Event::, 1],
1281 [Glib:: => AnyEvent::Impl::Glib:: , 1], # becomes extremely slow with many watchers
1282 [Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
1283 [Irssi:: => AnyEvent::Impl::Irssi::], # Irssi has a bogus "Event" package
850 [Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles 1284 [Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles
851 [Glib:: => AnyEvent::Impl::Glib::], # becomes extremely slow with many watchers
852 [Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy
853 [Qt:: => AnyEvent::Impl::Qt::], # requires special main program 1285 [Qt:: => AnyEvent::Impl::Qt::], # requires special main program
854 [POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza 1286 [POE::Kernel:: => AnyEvent::Impl::POE::], # lasciate ogni speranza
855 [Wx:: => AnyEvent::Impl::POE::], 1287 [Wx:: => AnyEvent::Impl::POE::],
856 [Prima:: => AnyEvent::Impl::POE::], 1288 [Prima:: => AnyEvent::Impl::POE::],
1289 [IO::Async::Loop:: => AnyEvent::Impl::IOAsync::], # a bitch to autodetect
1290 [Cocoa::EventLoop:: => AnyEvent::Impl::Cocoa::],
1291 [FLTK:: => AnyEvent::Impl::FLTK2::],
857); 1292);
858 1293
859our %method = map +($_ => 1), qw(io timer time now signal child condvar one_event DESTROY);
860
861our @post_detect;
862
863sub post_detect(&) { 1294sub detect() {
864 my ($cb) = @_; 1295 # free some memory
1296 *detect = sub () { $MODEL };
865 1297
866 if ($MODEL) { 1298 local $!; # for good measure
867 $cb->(); 1299 local $SIG{__DIE__};
868 1300
869 1 1301 if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z0-9:]+)$/) {
1302 my $model = $1;
1303 $model = "AnyEvent::Impl::$model" unless $model =~ s/::$//;
1304 if (eval "require $model") {
1305 $MODEL = $model;
1306 warn "AnyEvent: loaded model '$model' (forced by \$ENV{PERL_ANYEVENT_MODEL}), using it.\n" if $VERBOSE >= 2;
870 } else { 1307 } else {
871 push @post_detect, $cb; 1308 warn "AnyEvent: unable to load model '$model' (from \$ENV{PERL_ANYEVENT_MODEL}):\n$@" if $VERBOSE;
872 1309 }
873 defined wantarray
874 ? bless \$cb, "AnyEvent::Util::PostDetect"
875 : ()
876 } 1310 }
877}
878 1311
879sub AnyEvent::Util::PostDetect::DESTROY { 1312 # check for already loaded models
880 @post_detect = grep $_ != ${$_[0]}, @post_detect;
881}
882
883sub detect() {
884 unless ($MODEL) { 1313 unless ($MODEL) {
885 no strict 'refs'; 1314 for (@REGISTRY, @models) {
886 local $SIG{__DIE__}; 1315 my ($package, $model) = @$_;
887 1316 if (${"$package\::VERSION"} > 0) {
888 if ($ENV{PERL_ANYEVENT_MODEL} =~ /^([a-zA-Z]+)$/) {
889 my $model = "AnyEvent::Impl::$1";
890 if (eval "require $model") { 1317 if (eval "require $model") {
891 $MODEL = $model; 1318 $MODEL = $model;
892 warn "AnyEvent: loaded model '$model' (forced by \$PERL_ANYEVENT_MODEL), using it.\n" if $verbose > 1; 1319 warn "AnyEvent: autodetected model '$model', using it.\n" if $VERBOSE >= 2;
893 } else { 1320 last;
894 warn "AnyEvent: unable to load model '$model' (from \$PERL_ANYEVENT_MODEL):\n$@" if $verbose; 1321 }
895 } 1322 }
896 } 1323 }
897 1324
898 # check for already loaded models
899 unless ($MODEL) { 1325 unless ($MODEL) {
1326 # try to autoload a model
900 for (@REGISTRY, @models) { 1327 for (@REGISTRY, @models) {
901 my ($package, $model) = @$_; 1328 my ($package, $model, $autoload) = @$_;
1329 if (
1330 $autoload
1331 and eval "require $package"
902 if (${"$package\::VERSION"} > 0) { 1332 and ${"$package\::VERSION"} > 0
903 if (eval "require $model") { 1333 and eval "require $model"
1334 ) {
904 $MODEL = $model; 1335 $MODEL = $model;
905 warn "AnyEvent: autodetected model '$model', using it.\n" if $verbose > 1; 1336 warn "AnyEvent: autoloaded model '$model', using it.\n" if $VERBOSE >= 2;
906 last; 1337 last;
907 }
908 } 1338 }
909 } 1339 }
910 1340
911 unless ($MODEL) {
912 # try to load a model
913
914 for (@REGISTRY, @models) {
915 my ($package, $model) = @$_;
916 if (eval "require $package"
917 and ${"$package\::VERSION"} > 0
918 and eval "require $model") {
919 $MODEL = $model;
920 warn "AnyEvent: autoprobed model '$model', using it.\n" if $verbose > 1;
921 last;
922 }
923 }
924
925 $MODEL 1341 $MODEL
926 or die "No event module selected for AnyEvent and autodetect failed. Install any one of these modules: EV, Event or Glib."; 1342 or die "AnyEvent: backend autodetection failed - did you properly install AnyEvent?\n";
927 }
928 } 1343 }
929
930 unshift @ISA, $MODEL;
931 push @{"$MODEL\::ISA"}, "AnyEvent::Base";
932
933 (shift @post_detect)->() while @post_detect;
934 } 1344 }
935 1345
1346 # free memory only needed for probing
1347 undef @models;
1348 undef @REGISTRY;
1349
1350 push @{"$MODEL\::ISA"}, "AnyEvent::Base";
1351 unshift @ISA, $MODEL;
1352
1353 # now nuke some methods that are overridden by the backend.
1354 # SUPER usage is not allowed in these.
1355 for (qw(time signal child idle)) {
1356 undef &{"AnyEvent::Base::$_"}
1357 if defined &{"$MODEL\::$_"};
1358 }
1359
1360 if ($ENV{PERL_ANYEVENT_STRICT}) {
1361 eval { require AnyEvent::Strict };
1362 warn "AnyEvent: cannot load AnyEvent::Strict: $@"
1363 if $@ && $VERBOSE;
1364 }
1365
1366 (shift @post_detect)->() while @post_detect;
1367 undef @post_detect;
1368
1369 *post_detect = sub(&) {
1370 shift->();
1371
1372 undef
1373 };
1374
1375 # recover a few more bytes
1376 postpone {
1377 undef &AUTOLOAD;
1378 };
1379
936 $MODEL 1380 $MODEL
937} 1381}
1382
1383our %method = map +($_ => 1),
1384 qw(io timer time now now_update signal child idle condvar DESTROY);
938 1385
939sub AUTOLOAD { 1386sub AUTOLOAD {
940 (my $func = $AUTOLOAD) =~ s/.*://; 1387 (my $func = $AUTOLOAD) =~ s/.*://;
941 1388
942 $method{$func} 1389 $method{$func}
943 or croak "$func: not a valid method for AnyEvent objects"; 1390 or Carp::croak "$func: not a valid AnyEvent class method";
944 1391
945 detect unless $MODEL; 1392 # free some memory
1393 undef %method;
1394
1395 detect;
946 1396
947 my $class = shift; 1397 my $class = shift;
948 $class->$func (@_); 1398 $class->$func (@_);
949} 1399}
950 1400
1401# utility function to dup a filehandle. this is used by many backends
1402# to support binding more than one watcher per filehandle (they usually
1403# allow only one watcher per fd, so we dup it to get a different one).
1404sub _dupfh($$;$$) {
1405 my ($poll, $fh, $r, $w) = @_;
1406
1407 # cygwin requires the fh mode to be matching, unix doesn't
1408 my ($rw, $mode) = $poll eq "r" ? ($r, "<&") : ($w, ">&");
1409
1410 open my $fh2, $mode, $fh
1411 or die "AnyEvent->io: cannot dup() filehandle in mode '$poll': $!,";
1412
1413 # we assume CLOEXEC is already set by perl in all important cases
1414
1415 ($fh2, $rw)
1416}
1417
1418=head1 SIMPLIFIED AE API
1419
1420Starting with version 5.0, AnyEvent officially supports a second, much
1421simpler, API that is designed to reduce the calling, typing and memory
1422overhead by using function call syntax and a fixed number of parameters.
1423
1424See the L<AE> manpage for details.
1425
1426=cut
1427
1428package AE;
1429
1430our $VERSION = $AnyEvent::VERSION;
1431
1432
1433sub _reset() {
1434 eval q{
1435 # fall back to the main API by default - backends and AnyEvent::Base
1436 # implementations can overwrite these.
1437
1438 sub io($$$) {
1439 AnyEvent->io (fh => $_[0], poll => $_[1] ? "w" : "r", cb => $_[2])
1440 }
1441
1442 sub timer($$$) {
1443 AnyEvent->timer (after => $_[0], interval => $_[1], cb => $_[2])
1444 }
1445
1446 sub signal($$) {
1447 AnyEvent->signal (signal => $_[0], cb => $_[1])
1448 }
1449
1450 sub child($$) {
1451 AnyEvent->child (pid => $_[0], cb => $_[1])
1452 }
1453
1454 sub idle($) {
1455 AnyEvent->idle (cb => $_[0])
1456 }
1457
1458 sub cv(;&) {
1459 AnyEvent->condvar (@_ ? (cb => $_[0]) : ())
1460 }
1461
1462 sub now() {
1463 AnyEvent->now
1464 }
1465
1466 sub now_update() {
1467 AnyEvent->now_update
1468 }
1469
1470 sub time() {
1471 AnyEvent->time
1472 }
1473
1474 *postpone = \&AnyEvent::postpone;
1475 };
1476 die if $@;
1477}
1478
1479BEGIN { _reset }
1480
951package AnyEvent::Base; 1481package AnyEvent::Base;
952 1482
953# default implementation for now and time 1483# default implementations for many methods
954 1484
955use Time::HiRes (); 1485sub time {
1486 eval q{ # poor man's autoloading {}
1487 # probe for availability of Time::HiRes
1488 if (eval "use Time::HiRes (); Time::HiRes::time (); 1") {
1489 warn "AnyEvent: using Time::HiRes for sub-second timing accuracy.\n" if $VERBOSE >= 8;
1490 *AE::time = \&Time::HiRes::time;
1491 # if (eval "use POSIX (); (POSIX::times())...
1492 } else {
1493 warn "AnyEvent: using built-in time(), WARNING, no sub-second resolution!\n" if $VERBOSE;
1494 *AE::time = sub (){ time }; # epic fail
1495 }
956 1496
957sub time { Time::HiRes::time } 1497 *time = sub { AE::time }; # different prototypes
958sub now { Time::HiRes::time } 1498 };
1499 die if $@;
1500
1501 &time
1502}
1503
1504*now = \&time;
1505
1506sub now_update { }
1507
1508sub _poll {
1509 Carp::croak "$AnyEvent::MODEL does not support blocking waits. Caught";
1510}
959 1511
960# default implementation for ->condvar 1512# default implementation for ->condvar
1513# in fact, the default should not be overwritten
961 1514
962sub condvar { 1515sub condvar {
1516 eval q{ # poor man's autoloading {}
1517 *condvar = sub {
963 bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, AnyEvent::CondVar:: 1518 bless { @_ == 3 ? (_ae_cb => $_[2]) : () }, "AnyEvent::CondVar"
1519 };
1520
1521 *AE::cv = sub (;&) {
1522 bless { @_ ? (_ae_cb => shift) : () }, "AnyEvent::CondVar"
1523 };
1524 };
1525 die if $@;
1526
1527 &condvar
964} 1528}
965 1529
966# default implementation for ->signal 1530# default implementation for ->signal
967 1531
968our %SIG_CB; 1532our $HAVE_ASYNC_INTERRUPT;
1533
1534sub _have_async_interrupt() {
1535 $HAVE_ASYNC_INTERRUPT = 1*(!$ENV{PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT}
1536 && eval "use Async::Interrupt 1.02 (); 1")
1537 unless defined $HAVE_ASYNC_INTERRUPT;
1538
1539 $HAVE_ASYNC_INTERRUPT
1540}
1541
1542our ($SIGPIPE_R, $SIGPIPE_W, %SIG_CB, %SIG_EV, $SIG_IO);
1543our (%SIG_ASY, %SIG_ASY_W);
1544our ($SIG_COUNT, $SIG_TW);
1545
1546# install a dummy wakeup watcher to reduce signal catching latency
1547# used by Impls
1548sub _sig_add() {
1549 unless ($SIG_COUNT++) {
1550 # try to align timer on a full-second boundary, if possible
1551 my $NOW = AE::now;
1552
1553 $SIG_TW = AE::timer
1554 $MAX_SIGNAL_LATENCY - ($NOW - int $NOW),
1555 $MAX_SIGNAL_LATENCY,
1556 sub { } # just for the PERL_ASYNC_CHECK
1557 ;
1558 }
1559}
1560
1561sub _sig_del {
1562 undef $SIG_TW
1563 unless --$SIG_COUNT;
1564}
1565
1566our $_sig_name_init; $_sig_name_init = sub {
1567 eval q{ # poor man's autoloading {}
1568 undef $_sig_name_init;
1569
1570 if (_have_async_interrupt) {
1571 *sig2num = \&Async::Interrupt::sig2num;
1572 *sig2name = \&Async::Interrupt::sig2name;
1573 } else {
1574 require Config;
1575
1576 my %signame2num;
1577 @signame2num{ split ' ', $Config::Config{sig_name} }
1578 = split ' ', $Config::Config{sig_num};
1579
1580 my @signum2name;
1581 @signum2name[values %signame2num] = keys %signame2num;
1582
1583 *sig2num = sub($) {
1584 $_[0] > 0 ? shift : $signame2num{+shift}
1585 };
1586 *sig2name = sub ($) {
1587 $_[0] > 0 ? $signum2name[+shift] : shift
1588 };
1589 }
1590 };
1591 die if $@;
1592};
1593
1594sub sig2num ($) { &$_sig_name_init; &sig2num }
1595sub sig2name($) { &$_sig_name_init; &sig2name }
969 1596
970sub signal { 1597sub signal {
1598 eval q{ # poor man's autoloading {}
1599 # probe for availability of Async::Interrupt
1600 if (_have_async_interrupt) {
1601 warn "AnyEvent: using Async::Interrupt for race-free signal handling.\n" if $VERBOSE >= 8;
1602
1603 $SIGPIPE_R = new Async::Interrupt::EventPipe;
1604 $SIG_IO = AE::io $SIGPIPE_R->fileno, 0, \&_signal_exec;
1605
1606 } else {
1607 warn "AnyEvent: using emulated perl signal handling with latency timer.\n" if $VERBOSE >= 8;
1608
1609 if (AnyEvent::WIN32) {
1610 require AnyEvent::Util;
1611
1612 ($SIGPIPE_R, $SIGPIPE_W) = AnyEvent::Util::portable_pipe ();
1613 AnyEvent::Util::fh_nonblocking ($SIGPIPE_R, 1) if $SIGPIPE_R;
1614 AnyEvent::Util::fh_nonblocking ($SIGPIPE_W, 1) if $SIGPIPE_W; # just in case
1615 } else {
1616 pipe $SIGPIPE_R, $SIGPIPE_W;
1617 fcntl $SIGPIPE_R, AnyEvent::F_SETFL, AnyEvent::O_NONBLOCK if $SIGPIPE_R;
1618 fcntl $SIGPIPE_W, AnyEvent::F_SETFL, AnyEvent::O_NONBLOCK if $SIGPIPE_W; # just in case
1619
1620 # not strictly required, as $^F is normally 2, but let's make sure...
1621 fcntl $SIGPIPE_R, AnyEvent::F_SETFD, AnyEvent::FD_CLOEXEC;
1622 fcntl $SIGPIPE_W, AnyEvent::F_SETFD, AnyEvent::FD_CLOEXEC;
1623 }
1624
1625 $SIGPIPE_R
1626 or Carp::croak "AnyEvent: unable to create a signal reporting pipe: $!\n";
1627
1628 $SIG_IO = AE::io $SIGPIPE_R, 0, \&_signal_exec;
1629 }
1630
1631 *signal = $HAVE_ASYNC_INTERRUPT
1632 ? sub {
971 my (undef, %arg) = @_; 1633 my (undef, %arg) = @_;
972 1634
1635 # async::interrupt
973 my $signal = uc $arg{signal} 1636 my $signal = sig2num $arg{signal};
974 or Carp::croak "required option 'signal' is missing";
975
976 $SIG_CB{$signal}{$arg{cb}} = $arg{cb}; 1637 $SIG_CB{$signal}{$arg{cb}} = $arg{cb};
1638
1639 $SIG_ASY{$signal} ||= new Async::Interrupt
1640 cb => sub { undef $SIG_EV{$signal} },
1641 signal => $signal,
1642 pipe => [$SIGPIPE_R->filenos],
1643 pipe_autodrain => 0,
1644 ;
1645
1646 bless [$signal, $arg{cb}], "AnyEvent::Base::signal"
1647 }
1648 : sub {
1649 my (undef, %arg) = @_;
1650
1651 # pure perl
1652 my $signal = sig2name $arg{signal};
1653 $SIG_CB{$signal}{$arg{cb}} = $arg{cb};
1654
977 $SIG{$signal} ||= sub { 1655 $SIG{$signal} ||= sub {
1656 local $!;
1657 syswrite $SIGPIPE_W, "\x00", 1 unless %SIG_EV;
1658 undef $SIG_EV{$signal};
1659 };
1660
1661 # can't do signal processing without introducing races in pure perl,
1662 # so limit the signal latency.
1663 _sig_add;
1664
1665 bless [$signal, $arg{cb}], "AnyEvent::Base::signal"
1666 }
1667 ;
1668
1669 *AnyEvent::Base::signal::DESTROY = sub {
1670 my ($signal, $cb) = @{$_[0]};
1671
1672 _sig_del;
1673
1674 delete $SIG_CB{$signal}{$cb};
1675
1676 $HAVE_ASYNC_INTERRUPT
1677 ? delete $SIG_ASY{$signal}
1678 : # delete doesn't work with older perls - they then
1679 # print weird messages, or just unconditionally exit
1680 # instead of getting the default action.
1681 undef $SIG{$signal}
1682 unless keys %{ $SIG_CB{$signal} };
1683 };
1684
1685 *_signal_exec = sub {
1686 $HAVE_ASYNC_INTERRUPT
1687 ? $SIGPIPE_R->drain
1688 : sysread $SIGPIPE_R, (my $dummy), 9;
1689
1690 while (%SIG_EV) {
1691 for (keys %SIG_EV) {
1692 delete $SIG_EV{$_};
978 $_->() for values %{ $SIG_CB{$signal} || {} }; 1693 &$_ for values %{ $SIG_CB{$_} || {} };
1694 }
1695 }
1696 };
979 }; 1697 };
1698 die if $@;
980 1699
981 bless [$signal, $arg{cb}], "AnyEvent::Base::Signal" 1700 &signal
982}
983
984sub AnyEvent::Base::Signal::DESTROY {
985 my ($signal, $cb) = @{$_[0]};
986
987 delete $SIG_CB{$signal}{$cb};
988
989 $SIG{$signal} = 'DEFAULT' unless keys %{ $SIG_CB{$signal} };
990} 1701}
991 1702
992# default implementation for ->child 1703# default implementation for ->child
993 1704
994our %PID_CB; 1705our %PID_CB;
995our $CHLD_W; 1706our $CHLD_W;
996our $CHLD_DELAY_W; 1707our $CHLD_DELAY_W;
997our $PID_IDLE;
998our $WNOHANG;
999 1708
1000sub _child_wait { 1709# used by many Impl's
1001 while (0 < (my $pid = waitpid -1, $WNOHANG)) { 1710sub _emit_childstatus($$) {
1711 my (undef, $rpid, $rstatus) = @_;
1712
1713 $_->($rpid, $rstatus)
1002 $_->($pid, $?) for (values %{ $PID_CB{$pid} || {} }), 1714 for values %{ $PID_CB{$rpid} || {} },
1003 (values %{ $PID_CB{0} || {} }); 1715 values %{ $PID_CB{0} || {} };
1004 }
1005
1006 undef $PID_IDLE;
1007}
1008
1009sub _sigchld {
1010 # make sure we deliver these changes "synchronous" with the event loop.
1011 $CHLD_DELAY_W ||= AnyEvent->timer (after => 0, cb => sub {
1012 undef $CHLD_DELAY_W;
1013 &_child_wait;
1014 });
1015} 1716}
1016 1717
1017sub child { 1718sub child {
1719 eval q{ # poor man's autoloading {}
1720 *_sigchld = sub {
1721 my $pid;
1722
1723 AnyEvent->_emit_childstatus ($pid, $?)
1724 while ($pid = waitpid -1, WNOHANG) > 0;
1725 };
1726
1727 *child = sub {
1018 my (undef, %arg) = @_; 1728 my (undef, %arg) = @_;
1019 1729
1020 defined (my $pid = $arg{pid} + 0) 1730 my $pid = $arg{pid};
1021 or Carp::croak "required option 'pid' is missing"; 1731 my $cb = $arg{cb};
1022 1732
1023 $PID_CB{$pid}{$arg{cb}} = $arg{cb}; 1733 $PID_CB{$pid}{$cb+0} = $cb;
1024 1734
1025 unless ($WNOHANG) {
1026 $WNOHANG = eval { local $SIG{__DIE__}; require POSIX; &POSIX::WNOHANG } || 1;
1027 }
1028
1029 unless ($CHLD_W) { 1735 unless ($CHLD_W) {
1030 $CHLD_W = AnyEvent->signal (signal => 'CHLD', cb => \&_sigchld); 1736 $CHLD_W = AE::signal CHLD => \&_sigchld;
1031 # child could be a zombie already, so make at least one round 1737 # child could be a zombie already, so make at least one round
1032 &_sigchld; 1738 &_sigchld;
1033 } 1739 }
1034 1740
1035 bless [$pid, $arg{cb}], "AnyEvent::Base::Child" 1741 bless [$pid, $cb+0], "AnyEvent::Base::child"
1036} 1742 };
1037 1743
1038sub AnyEvent::Base::Child::DESTROY { 1744 *AnyEvent::Base::child::DESTROY = sub {
1039 my ($pid, $cb) = @{$_[0]}; 1745 my ($pid, $icb) = @{$_[0]};
1040 1746
1041 delete $PID_CB{$pid}{$cb}; 1747 delete $PID_CB{$pid}{$icb};
1042 delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} }; 1748 delete $PID_CB{$pid} unless keys %{ $PID_CB{$pid} };
1043 1749
1044 undef $CHLD_W unless keys %PID_CB; 1750 undef $CHLD_W unless keys %PID_CB;
1751 };
1752 };
1753 die if $@;
1754
1755 &child
1756}
1757
1758# idle emulation is done by simply using a timer, regardless
1759# of whether the process is idle or not, and not letting
1760# the callback use more than 50% of the time.
1761sub idle {
1762 eval q{ # poor man's autoloading {}
1763 *idle = sub {
1764 my (undef, %arg) = @_;
1765
1766 my ($cb, $w, $rcb) = $arg{cb};
1767
1768 $rcb = sub {
1769 if ($cb) {
1770 $w = AE::time;
1771 &$cb;
1772 $w = AE::time - $w;
1773
1774 # never use more then 50% of the time for the idle watcher,
1775 # within some limits
1776 $w = 0.0001 if $w < 0.0001;
1777 $w = 5 if $w > 5;
1778
1779 $w = AE::timer $w, 0, $rcb;
1780 } else {
1781 # clean up...
1782 undef $w;
1783 undef $rcb;
1784 }
1785 };
1786
1787 $w = AE::timer 0.05, 0, $rcb;
1788
1789 bless \\$cb, "AnyEvent::Base::idle"
1790 };
1791
1792 *AnyEvent::Base::idle::DESTROY = sub {
1793 undef $${$_[0]};
1794 };
1795 };
1796 die if $@;
1797
1798 &idle
1045} 1799}
1046 1800
1047package AnyEvent::CondVar; 1801package AnyEvent::CondVar;
1048 1802
1049our @ISA = AnyEvent::CondVar::Base::; 1803our @ISA = AnyEvent::CondVar::Base::;
1050 1804
1805# only to be used for subclassing
1806sub new {
1807 my $class = shift;
1808 bless AnyEvent->condvar (@_), $class
1809}
1810
1051package AnyEvent::CondVar::Base; 1811package AnyEvent::CondVar::Base;
1052 1812
1053use overload 1813#use overload
1054 '&{}' => sub { my $self = shift; sub { $self->send (@_) } }, 1814# '&{}' => sub { my $self = shift; sub { $self->send (@_) } },
1055 fallback => 1; 1815# fallback => 1;
1816
1817# save 300+ kilobytes by dirtily hardcoding overloading
1818${"AnyEvent::CondVar::Base::OVERLOAD"}{dummy}++; # Register with magic by touching.
1819*{'AnyEvent::CondVar::Base::()'} = sub { }; # "Make it findable via fetchmethod."
1820*{'AnyEvent::CondVar::Base::(&{}'} = sub { my $self = shift; sub { $self->send (@_) } }; # &{}
1821${'AnyEvent::CondVar::Base::()'} = 1; # fallback
1822
1823our $WAITING;
1056 1824
1057sub _send { 1825sub _send {
1058 # nop 1826 # nop
1827}
1828
1829sub _wait {
1830 AnyEvent->_poll until $_[0]{_ae_sent};
1059} 1831}
1060 1832
1061sub send { 1833sub send {
1062 my $cv = shift; 1834 my $cv = shift;
1063 $cv->{_ae_sent} = [@_]; 1835 $cv->{_ae_sent} = [@_];
1072 1844
1073sub ready { 1845sub ready {
1074 $_[0]{_ae_sent} 1846 $_[0]{_ae_sent}
1075} 1847}
1076 1848
1077sub _wait {
1078 AnyEvent->one_event while !$_[0]{_ae_sent};
1079}
1080
1081sub recv { 1849sub recv {
1850 unless ($_[0]{_ae_sent}) {
1851 $WAITING
1852 and Carp::croak "AnyEvent::CondVar: recursive blocking wait attempted";
1853
1854 local $WAITING = 1;
1082 $_[0]->_wait; 1855 $_[0]->_wait;
1856 }
1083 1857
1084 Carp::croak $_[0]{_ae_croak} if $_[0]{_ae_croak}; 1858 $_[0]{_ae_croak}
1085 wantarray ? @{ $_[0]{_ae_sent} } : $_[0]{_ae_sent}[0] 1859 and Carp::croak $_[0]{_ae_croak};
1860
1861 wantarray
1862 ? @{ $_[0]{_ae_sent} }
1863 : $_[0]{_ae_sent}[0]
1086} 1864}
1087 1865
1088sub cb { 1866sub cb {
1089 $_[0]{_ae_cb} = $_[1] if @_ > 1; 1867 my $cv = shift;
1868
1869 @_
1870 and $cv->{_ae_cb} = shift
1871 and $cv->{_ae_sent}
1872 and (delete $cv->{_ae_cb})->($cv);
1873
1090 $_[0]{_ae_cb} 1874 $cv->{_ae_cb}
1091} 1875}
1092 1876
1093sub begin { 1877sub begin {
1094 ++$_[0]{_ae_counter}; 1878 ++$_[0]{_ae_counter};
1095 $_[0]{_ae_end_cb} = $_[1] if @_ > 1; 1879 $_[0]{_ae_end_cb} = $_[1] if @_ > 1;
1100 &{ $_[0]{_ae_end_cb} || sub { $_[0]->send } }; 1884 &{ $_[0]{_ae_end_cb} || sub { $_[0]->send } };
1101} 1885}
1102 1886
1103# undocumented/compatibility with pre-3.4 1887# undocumented/compatibility with pre-3.4
1104*broadcast = \&send; 1888*broadcast = \&send;
1105*wait = \&_wait; 1889*wait = \&recv;
1890
1891=head1 ERROR AND EXCEPTION HANDLING
1892
1893In general, AnyEvent does not do any error handling - it relies on the
1894caller to do that if required. The L<AnyEvent::Strict> module (see also
1895the C<PERL_ANYEVENT_STRICT> environment variable, below) provides strict
1896checking of all AnyEvent methods, however, which is highly useful during
1897development.
1898
1899As for exception handling (i.e. runtime errors and exceptions thrown while
1900executing a callback), this is not only highly event-loop specific, but
1901also not in any way wrapped by this module, as this is the job of the main
1902program.
1903
1904The pure perl event loop simply re-throws the exception (usually
1905within C<< condvar->recv >>), the L<Event> and L<EV> modules call C<<
1906$Event/EV::DIED->() >>, L<Glib> uses C<< install_exception_handler >> and
1907so on.
1908
1909=head1 ENVIRONMENT VARIABLES
1910
1911The following environment variables are used by this module or its
1912submodules.
1913
1914Note that AnyEvent will remove I<all> environment variables starting with
1915C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
1916enabled.
1917
1918=over 4
1919
1920=item C<PERL_ANYEVENT_VERBOSE>
1921
1922By default, AnyEvent will be completely silent except in fatal
1923conditions. You can set this environment variable to make AnyEvent more
1924talkative.
1925
1926When set to C<1> or higher, causes AnyEvent to warn about unexpected
1927conditions, such as not being able to load the event model specified by
1928C<PERL_ANYEVENT_MODEL>.
1929
1930When set to C<2> or higher, cause AnyEvent to report to STDERR which event
1931model it chooses.
1932
1933When set to C<8> or higher, then AnyEvent will report extra information on
1934which optional modules it loads and how it implements certain features.
1935
1936=item C<PERL_ANYEVENT_STRICT>
1937
1938AnyEvent does not do much argument checking by default, as thorough
1939argument checking is very costly. Setting this variable to a true value
1940will cause AnyEvent to load C<AnyEvent::Strict> and then to thoroughly
1941check the arguments passed to most method calls. If it finds any problems,
1942it will croak.
1943
1944In other words, enables "strict" mode.
1945
1946Unlike C<use strict> (or its modern cousin, C<< use L<common::sense>
1947>>, it is definitely recommended to keep it off in production. Keeping
1948C<PERL_ANYEVENT_STRICT=1> in your environment while developing programs
1949can be very useful, however.
1950
1951=item C<PERL_ANYEVENT_MODEL>
1952
1953This can be used to specify the event model to be used by AnyEvent, before
1954auto detection and -probing kicks in.
1955
1956It normally is a string consisting entirely of ASCII letters (e.g. C<EV>
1957or C<IOAsync>). The string C<AnyEvent::Impl::> gets prepended and the
1958resulting module name is loaded and - if the load was successful - used as
1959event model backend. If it fails to load then AnyEvent will proceed with
1960auto detection and -probing.
1961
1962If the string ends with C<::> instead (e.g. C<AnyEvent::Impl::EV::>) then
1963nothing gets prepended and the module name is used as-is (hint: C<::> at
1964the end of a string designates a module name and quotes it appropriately).
1965
1966For example, to force the pure perl model (L<AnyEvent::Loop::Perl>) you
1967could start your program like this:
1968
1969 PERL_ANYEVENT_MODEL=Perl perl ...
1970
1971=item C<PERL_ANYEVENT_PROTOCOLS>
1972
1973Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
1974for IPv4 or IPv6. The default is unspecified (and might change, or be the result
1975of auto probing).
1976
1977Must be set to a comma-separated list of protocols or address families,
1978current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
1979used, and preference will be given to protocols mentioned earlier in the
1980list.
1981
1982This variable can effectively be used for denial-of-service attacks
1983against local programs (e.g. when setuid), although the impact is likely
1984small, as the program has to handle conenction and other failures anyways.
1985
1986Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
1987but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
1988- only support IPv4, never try to resolve or contact IPv6
1989addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
1990IPv6, but prefer IPv6 over IPv4.
1991
1992=item C<PERL_ANYEVENT_EDNS0>
1993
1994Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension
1995for DNS. This extension is generally useful to reduce DNS traffic, but
1996some (broken) firewalls drop such DNS packets, which is why it is off by
1997default.
1998
1999Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
2000EDNS0 in its DNS requests.
2001
2002=item C<PERL_ANYEVENT_MAX_FORKS>
2003
2004The maximum number of child processes that C<AnyEvent::Util::fork_call>
2005will create in parallel.
2006
2007=item C<PERL_ANYEVENT_MAX_OUTSTANDING_DNS>
2008
2009The default value for the C<max_outstanding> parameter for the default DNS
2010resolver - this is the maximum number of parallel DNS requests that are
2011sent to the DNS server.
2012
2013=item C<PERL_ANYEVENT_RESOLV_CONF>
2014
2015The file to use instead of F</etc/resolv.conf> (or OS-specific
2016configuration) in the default resolver. When set to the empty string, no
2017default config will be used.
2018
2019=item C<PERL_ANYEVENT_CA_FILE>, C<PERL_ANYEVENT_CA_PATH>.
2020
2021When neither C<ca_file> nor C<ca_path> was specified during
2022L<AnyEvent::TLS> context creation, and either of these environment
2023variables exist, they will be used to specify CA certificate locations
2024instead of a system-dependent default.
2025
2026=item C<PERL_ANYEVENT_AVOID_GUARD> and C<PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT>
2027
2028When these are set to C<1>, then the respective modules are not
2029loaded. Mostly good for testing AnyEvent itself.
2030
2031=back
1106 2032
1107=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE 2033=head1 SUPPLYING YOUR OWN EVENT MODEL INTERFACE
1108 2034
1109This is an advanced topic that you do not normally need to use AnyEvent in 2035This is an advanced topic that you do not normally need to use AnyEvent in
1110a module. This section is only of use to event loop authors who want to 2036a module. This section is only of use to event loop authors who want to
1144 2070
1145I<rxvt-unicode> also cheats a bit by not providing blocking access to 2071I<rxvt-unicode> also cheats a bit by not providing blocking access to
1146condition variables: code blocking while waiting for a condition will 2072condition variables: code blocking while waiting for a condition will
1147C<die>. This still works with most modules/usages, and blocking calls must 2073C<die>. This still works with most modules/usages, and blocking calls must
1148not be done in an interactive application, so it makes sense. 2074not be done in an interactive application, so it makes sense.
1149
1150=head1 ENVIRONMENT VARIABLES
1151
1152The following environment variables are used by this module:
1153
1154=over 4
1155
1156=item C<PERL_ANYEVENT_VERBOSE>
1157
1158By default, AnyEvent will be completely silent except in fatal
1159conditions. You can set this environment variable to make AnyEvent more
1160talkative.
1161
1162When set to C<1> or higher, causes AnyEvent to warn about unexpected
1163conditions, such as not being able to load the event model specified by
1164C<PERL_ANYEVENT_MODEL>.
1165
1166When set to C<2> or higher, cause AnyEvent to report to STDERR which event
1167model it chooses.
1168
1169=item C<PERL_ANYEVENT_MODEL>
1170
1171This can be used to specify the event model to be used by AnyEvent, before
1172auto detection and -probing kicks in. It must be a string consisting
1173entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended
1174and the resulting module name is loaded and if the load was successful,
1175used as event model. If it fails to load AnyEvent will proceed with
1176auto detection and -probing.
1177
1178This functionality might change in future versions.
1179
1180For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you
1181could start your program like this:
1182
1183 PERL_ANYEVENT_MODEL=Perl perl ...
1184
1185=item C<PERL_ANYEVENT_PROTOCOLS>
1186
1187Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences
1188for IPv4 or IPv6. The default is unspecified (and might change, or be the result
1189of auto probing).
1190
1191Must be set to a comma-separated list of protocols or address families,
1192current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be
1193used, and preference will be given to protocols mentioned earlier in the
1194list.
1195
1196This variable can effectively be used for denial-of-service attacks
1197against local programs (e.g. when setuid), although the impact is likely
1198small, as the program has to handle connection errors already-
1199
1200Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6,
1201but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4>
1202- only support IPv4, never try to resolve or contact IPv6
1203addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or
1204IPv6, but prefer IPv6 over IPv4.
1205
1206=item C<PERL_ANYEVENT_EDNS0>
1207
1208Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension
1209for DNS. This extension is generally useful to reduce DNS traffic, but
1210some (broken) firewalls drop such DNS packets, which is why it is off by
1211default.
1212
1213Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce
1214EDNS0 in its DNS requests.
1215
1216=item C<PERL_ANYEVENT_MAX_FORKS>
1217
1218The maximum number of child processes that C<AnyEvent::Util::fork_call>
1219will create in parallel.
1220
1221=back
1222 2075
1223=head1 EXAMPLE PROGRAM 2076=head1 EXAMPLE PROGRAM
1224 2077
1225The following program uses an I/O watcher to read data from STDIN, a timer 2078The following program uses an I/O watcher to read data from STDIN, a timer
1226to display a message once per second, and a condition variable to quit the 2079to display a message once per second, and a condition variable to quit the
1239 warn "read: $input\n"; # output what has been read 2092 warn "read: $input\n"; # output what has been read
1240 $cv->send if $input =~ /^q/i; # quit program if /^q/i 2093 $cv->send if $input =~ /^q/i; # quit program if /^q/i
1241 }, 2094 },
1242 ); 2095 );
1243 2096
1244 my $time_watcher; # can only be used once
1245
1246 sub new_timer {
1247 $timer = AnyEvent->timer (after => 1, cb => sub { 2097 my $time_watcher = AnyEvent->timer (after => 1, interval => 1, cb => sub {
1248 warn "timeout\n"; # print 'timeout' about every second 2098 warn "timeout\n"; # print 'timeout' at most every second
1249 &new_timer; # and restart the time
1250 }); 2099 });
1251 }
1252
1253 new_timer; # create first timer
1254 2100
1255 $cv->recv; # wait until user enters /^q/i 2101 $cv->recv; # wait until user enters /^q/i
1256 2102
1257=head1 REAL-WORLD EXAMPLE 2103=head1 REAL-WORLD EXAMPLE
1258 2104
1331 2177
1332The actual code goes further and collects all errors (C<die>s, exceptions) 2178The actual code goes further and collects all errors (C<die>s, exceptions)
1333that occurred during request processing. The C<result> method detects 2179that occurred during request processing. The C<result> method detects
1334whether an exception as thrown (it is stored inside the $txn object) 2180whether an exception as thrown (it is stored inside the $txn object)
1335and just throws the exception, which means connection errors and other 2181and just throws the exception, which means connection errors and other
1336problems get reported tot he code that tries to use the result, not in a 2182problems get reported to the code that tries to use the result, not in a
1337random callback. 2183random callback.
1338 2184
1339All of this enables the following usage styles: 2185All of this enables the following usage styles:
1340 2186
13411. Blocking: 21871. Blocking:
1389through AnyEvent. The benchmark creates a lot of timers (with a zero 2235through AnyEvent. The benchmark creates a lot of timers (with a zero
1390timeout) and I/O watchers (watching STDOUT, a pty, to become writable, 2236timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
1391which it is), lets them fire exactly once and destroys them again. 2237which it is), lets them fire exactly once and destroys them again.
1392 2238
1393Source code for this benchmark is found as F<eg/bench> in the AnyEvent 2239Source code for this benchmark is found as F<eg/bench> in the AnyEvent
1394distribution. 2240distribution. It uses the L<AE> interface, which makes a real difference
2241for the EV and Perl backends only.
1395 2242
1396=head3 Explanation of the columns 2243=head3 Explanation of the columns
1397 2244
1398I<watcher> is the number of event watchers created/destroyed. Since 2245I<watcher> is the number of event watchers created/destroyed. Since
1399different event models feature vastly different performances, each event 2246different event models feature vastly different performances, each event
1420watcher. 2267watcher.
1421 2268
1422=head3 Results 2269=head3 Results
1423 2270
1424 name watchers bytes create invoke destroy comment 2271 name watchers bytes create invoke destroy comment
1425 EV/EV 400000 244 0.56 0.46 0.31 EV native interface 2272 EV/EV 100000 223 0.47 0.43 0.27 EV native interface
1426 EV/Any 100000 244 2.50 0.46 0.29 EV + AnyEvent watchers 2273 EV/Any 100000 223 0.48 0.42 0.26 EV + AnyEvent watchers
1427 CoroEV/Any 100000 244 2.49 0.44 0.29 coroutines + Coro::Signal 2274 Coro::EV/Any 100000 223 0.47 0.42 0.26 coroutines + Coro::Signal
1428 Perl/Any 100000 513 4.92 0.87 1.12 pure perl implementation 2275 Perl/Any 100000 431 2.70 0.74 0.92 pure perl implementation
1429 Event/Event 16000 516 31.88 31.30 0.85 Event native interface 2276 Event/Event 16000 516 31.16 31.84 0.82 Event native interface
1430 Event/Any 16000 590 35.75 31.42 1.08 Event + AnyEvent watchers 2277 Event/Any 16000 1203 42.61 34.79 1.80 Event + AnyEvent watchers
2278 IOAsync/Any 16000 1911 41.92 27.45 16.81 via IO::Async::Loop::IO_Poll
2279 IOAsync/Any 16000 1726 40.69 26.37 15.25 via IO::Async::Loop::Epoll
1431 Glib/Any 16000 1357 98.22 12.41 54.00 quadratic behaviour 2280 Glib/Any 16000 1118 89.00 12.57 51.17 quadratic behaviour
1432 Tk/Any 2000 1860 26.97 67.98 14.00 SEGV with >> 2000 watchers 2281 Tk/Any 2000 1346 20.96 10.75 8.00 SEGV with >> 2000 watchers
1433 POE/Event 2000 6644 108.64 736.02 14.73 via POE::Loop::Event 2282 POE/Any 2000 6951 108.97 795.32 14.24 via POE::Loop::Event
1434 POE/Select 2000 6343 94.13 809.12 565.96 via POE::Loop::Select 2283 POE/Any 2000 6648 94.79 774.40 575.51 via POE::Loop::Select
1435 2284
1436=head3 Discussion 2285=head3 Discussion
1437 2286
1438The benchmark does I<not> measure scalability of the event loop very 2287The benchmark does I<not> measure scalability of the event loop very
1439well. For example, a select-based event loop (such as the pure perl one) 2288well. For example, a select-based event loop (such as the pure perl one)
1451benchmark machine, handling an event takes roughly 1600 CPU cycles with 2300benchmark machine, handling an event takes roughly 1600 CPU cycles with
1452EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU 2301EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU
1453cycles with POE. 2302cycles with POE.
1454 2303
1455C<EV> is the sole leader regarding speed and memory use, which are both 2304C<EV> is the sole leader regarding speed and memory use, which are both
1456maximal/minimal, respectively. Even when going through AnyEvent, it uses 2305maximal/minimal, respectively. When using the L<AE> API there is zero
2306overhead (when going through the AnyEvent API create is about 5-6 times
2307slower, with other times being equal, so still uses far less memory than
1457far less memory than any other event loop and is still faster than Event 2308any other event loop and is still faster than Event natively).
1458natively.
1459 2309
1460The pure perl implementation is hit in a few sweet spots (both the 2310The pure perl implementation is hit in a few sweet spots (both the
1461constant timeout and the use of a single fd hit optimisations in the perl 2311constant timeout and the use of a single fd hit optimisations in the perl
1462interpreter and the backend itself). Nevertheless this shows that it 2312interpreter and the backend itself). Nevertheless this shows that it
1463adds very little overhead in itself. Like any select-based backend its 2313adds very little overhead in itself. Like any select-based backend its
1464performance becomes really bad with lots of file descriptors (and few of 2314performance becomes really bad with lots of file descriptors (and few of
1465them active), of course, but this was not subject of this benchmark. 2315them active), of course, but this was not subject of this benchmark.
1466 2316
1467The C<Event> module has a relatively high setup and callback invocation 2317The C<Event> module has a relatively high setup and callback invocation
1468cost, but overall scores in on the third place. 2318cost, but overall scores in on the third place.
2319
2320C<IO::Async> performs admirably well, about on par with C<Event>, even
2321when using its pure perl backend.
1469 2322
1470C<Glib>'s memory usage is quite a bit higher, but it features a 2323C<Glib>'s memory usage is quite a bit higher, but it features a
1471faster callback invocation and overall ends up in the same class as 2324faster callback invocation and overall ends up in the same class as
1472C<Event>. However, Glib scales extremely badly, doubling the number of 2325C<Event>. However, Glib scales extremely badly, doubling the number of
1473watchers increases the processing time by more than a factor of four, 2326watchers increases the processing time by more than a factor of four,
1534In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100 2387In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100
1535(1%) are active. This mirrors the activity of large servers with many 2388(1%) are active. This mirrors the activity of large servers with many
1536connections, most of which are idle at any one point in time. 2389connections, most of which are idle at any one point in time.
1537 2390
1538Source code for this benchmark is found as F<eg/bench2> in the AnyEvent 2391Source code for this benchmark is found as F<eg/bench2> in the AnyEvent
1539distribution. 2392distribution. It uses the L<AE> interface, which makes a real difference
2393for the EV and Perl backends only.
1540 2394
1541=head3 Explanation of the columns 2395=head3 Explanation of the columns
1542 2396
1543I<sockets> is the number of sockets, and twice the number of "servers" (as 2397I<sockets> is the number of sockets, and twice the number of "servers" (as
1544each server has a read and write socket end). 2398each server has a read and write socket end).
1551it to another server. This includes deleting the old timeout and creating 2405it to another server. This includes deleting the old timeout and creating
1552a new one that moves the timeout into the future. 2406a new one that moves the timeout into the future.
1553 2407
1554=head3 Results 2408=head3 Results
1555 2409
1556 name sockets create request 2410 name sockets create request
1557 EV 20000 69.01 11.16 2411 EV 20000 62.66 7.99
1558 Perl 20000 73.32 35.87 2412 Perl 20000 68.32 32.64
1559 Event 20000 212.62 257.32 2413 IOAsync 20000 174.06 101.15 epoll
1560 Glib 20000 651.16 1896.30 2414 IOAsync 20000 174.67 610.84 poll
2415 Event 20000 202.69 242.91
2416 Glib 20000 557.01 1689.52
1561 POE 20000 349.67 12317.24 uses POE::Loop::Event 2417 POE 20000 341.54 12086.32 uses POE::Loop::Event
1562 2418
1563=head3 Discussion 2419=head3 Discussion
1564 2420
1565This benchmark I<does> measure scalability and overall performance of the 2421This benchmark I<does> measure scalability and overall performance of the
1566particular event loop. 2422particular event loop.
1568EV is again fastest. Since it is using epoll on my system, the setup time 2424EV is again fastest. Since it is using epoll on my system, the setup time
1569is relatively high, though. 2425is relatively high, though.
1570 2426
1571Perl surprisingly comes second. It is much faster than the C-based event 2427Perl surprisingly comes second. It is much faster than the C-based event
1572loops Event and Glib. 2428loops Event and Glib.
2429
2430IO::Async performs very well when using its epoll backend, and still quite
2431good compared to Glib when using its pure perl backend.
1573 2432
1574Event suffers from high setup time as well (look at its code and you will 2433Event suffers from high setup time as well (look at its code and you will
1575understand why). Callback invocation also has a high overhead compared to 2434understand why). Callback invocation also has a high overhead compared to
1576the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event 2435the C<< $_->() for .. >>-style loop that the Perl event loop uses. Event
1577uses select or poll in basically all documented configurations. 2436uses select or poll in basically all documented configurations.
1640=item * C-based event loops perform very well with small number of 2499=item * C-based event loops perform very well with small number of
1641watchers, as the management overhead dominates. 2500watchers, as the management overhead dominates.
1642 2501
1643=back 2502=back
1644 2503
2504=head2 THE IO::Lambda BENCHMARK
2505
2506Recently I was told about the benchmark in the IO::Lambda manpage, which
2507could be misinterpreted to make AnyEvent look bad. In fact, the benchmark
2508simply compares IO::Lambda with POE, and IO::Lambda looks better (which
2509shouldn't come as a surprise to anybody). As such, the benchmark is
2510fine, and mostly shows that the AnyEvent backend from IO::Lambda isn't
2511very optimal. But how would AnyEvent compare when used without the extra
2512baggage? To explore this, I wrote the equivalent benchmark for AnyEvent.
2513
2514The benchmark itself creates an echo-server, and then, for 500 times,
2515connects to the echo server, sends a line, waits for the reply, and then
2516creates the next connection. This is a rather bad benchmark, as it doesn't
2517test the efficiency of the framework or much non-blocking I/O, but it is a
2518benchmark nevertheless.
2519
2520 name runtime
2521 Lambda/select 0.330 sec
2522 + optimized 0.122 sec
2523 Lambda/AnyEvent 0.327 sec
2524 + optimized 0.138 sec
2525 Raw sockets/select 0.077 sec
2526 POE/select, components 0.662 sec
2527 POE/select, raw sockets 0.226 sec
2528 POE/select, optimized 0.404 sec
2529
2530 AnyEvent/select/nb 0.085 sec
2531 AnyEvent/EV/nb 0.068 sec
2532 +state machine 0.134 sec
2533
2534The benchmark is also a bit unfair (my fault): the IO::Lambda/POE
2535benchmarks actually make blocking connects and use 100% blocking I/O,
2536defeating the purpose of an event-based solution. All of the newly
2537written AnyEvent benchmarks use 100% non-blocking connects (using
2538AnyEvent::Socket::tcp_connect and the asynchronous pure perl DNS
2539resolver), so AnyEvent is at a disadvantage here, as non-blocking connects
2540generally require a lot more bookkeeping and event handling than blocking
2541connects (which involve a single syscall only).
2542
2543The last AnyEvent benchmark additionally uses L<AnyEvent::Handle>, which
2544offers similar expressive power as POE and IO::Lambda, using conventional
2545Perl syntax. This means that both the echo server and the client are 100%
2546non-blocking, further placing it at a disadvantage.
2547
2548As you can see, the AnyEvent + EV combination even beats the
2549hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl
2550backend easily beats IO::Lambda and POE.
2551
2552And even the 100% non-blocking version written using the high-level (and
2553slow :) L<AnyEvent::Handle> abstraction beats both POE and IO::Lambda
2554higher level ("unoptimised") abstractions by a large margin, even though
2555it does all of DNS, tcp-connect and socket I/O in a non-blocking way.
2556
2557The two AnyEvent benchmarks programs can be found as F<eg/ae0.pl> and
2558F<eg/ae2.pl> in the AnyEvent distribution, the remaining benchmarks are
2559part of the IO::Lambda distribution and were used without any changes.
2560
2561
2562=head1 SIGNALS
2563
2564AnyEvent currently installs handlers for these signals:
2565
2566=over 4
2567
2568=item SIGCHLD
2569
2570A handler for C<SIGCHLD> is installed by AnyEvent's child watcher
2571emulation for event loops that do not support them natively. Also, some
2572event loops install a similar handler.
2573
2574Additionally, when AnyEvent is loaded and SIGCHLD is set to IGNORE, then
2575AnyEvent will reset it to default, to avoid losing child exit statuses.
2576
2577=item SIGPIPE
2578
2579A no-op handler is installed for C<SIGPIPE> when C<$SIG{PIPE}> is C<undef>
2580when AnyEvent gets loaded.
2581
2582The rationale for this is that AnyEvent users usually do not really depend
2583on SIGPIPE delivery (which is purely an optimisation for shell use, or
2584badly-written programs), but C<SIGPIPE> can cause spurious and rare
2585program exits as a lot of people do not expect C<SIGPIPE> when writing to
2586some random socket.
2587
2588The rationale for installing a no-op handler as opposed to ignoring it is
2589that this way, the handler will be restored to defaults on exec.
2590
2591Feel free to install your own handler, or reset it to defaults.
2592
2593=back
2594
2595=cut
2596
2597undef $SIG{CHLD}
2598 if $SIG{CHLD} eq 'IGNORE';
2599
2600$SIG{PIPE} = sub { }
2601 unless defined $SIG{PIPE};
2602
2603=head1 RECOMMENDED/OPTIONAL MODULES
2604
2605One of AnyEvent's main goals is to be 100% Pure-Perl(tm): only perl (and
2606its built-in modules) are required to use it.
2607
2608That does not mean that AnyEvent won't take advantage of some additional
2609modules if they are installed.
2610
2611This section explains which additional modules will be used, and how they
2612affect AnyEvent's operation.
2613
2614=over 4
2615
2616=item L<Async::Interrupt>
2617
2618This slightly arcane module is used to implement fast signal handling: To
2619my knowledge, there is no way to do completely race-free and quick
2620signal handling in pure perl. To ensure that signals still get
2621delivered, AnyEvent will start an interval timer to wake up perl (and
2622catch the signals) with some delay (default is 10 seconds, look for
2623C<$AnyEvent::MAX_SIGNAL_LATENCY>).
2624
2625If this module is available, then it will be used to implement signal
2626catching, which means that signals will not be delayed, and the event loop
2627will not be interrupted regularly, which is more efficient (and good for
2628battery life on laptops).
2629
2630This affects not just the pure-perl event loop, but also other event loops
2631that have no signal handling on their own (e.g. Glib, Tk, Qt).
2632
2633Some event loops (POE, Event, Event::Lib) offer signal watchers natively,
2634and either employ their own workarounds (POE) or use AnyEvent's workaround
2635(using C<$AnyEvent::MAX_SIGNAL_LATENCY>). Installing L<Async::Interrupt>
2636does nothing for those backends.
2637
2638=item L<EV>
2639
2640This module isn't really "optional", as it is simply one of the backend
2641event loops that AnyEvent can use. However, it is simply the best event
2642loop available in terms of features, speed and stability: It supports
2643the AnyEvent API optimally, implements all the watcher types in XS, does
2644automatic timer adjustments even when no monotonic clock is available,
2645can take avdantage of advanced kernel interfaces such as C<epoll> and
2646C<kqueue>, and is the fastest backend I<by far>. You can even embed
2647L<Glib>/L<Gtk2> in it (or vice versa, see L<EV::Glib> and L<Glib::EV>).
2648
2649If you only use backends that rely on another event loop (e.g. C<Tk>),
2650then this module will do nothing for you.
2651
2652=item L<Guard>
2653
2654The guard module, when used, will be used to implement
2655C<AnyEvent::Util::guard>. This speeds up guards considerably (and uses a
2656lot less memory), but otherwise doesn't affect guard operation much. It is
2657purely used for performance.
2658
2659=item L<JSON> and L<JSON::XS>
2660
2661One of these modules is required when you want to read or write JSON data
2662via L<AnyEvent::Handle>. L<JSON> is also written in pure-perl, but can take
2663advantage of the ultra-high-speed L<JSON::XS> module when it is installed.
2664
2665=item L<Net::SSLeay>
2666
2667Implementing TLS/SSL in Perl is certainly interesting, but not very
2668worthwhile: If this module is installed, then L<AnyEvent::Handle> (with
2669the help of L<AnyEvent::TLS>), gains the ability to do TLS/SSL.
2670
2671=item L<Time::HiRes>
2672
2673This module is part of perl since release 5.008. It will be used when the
2674chosen event library does not come with a timing source of its own. The
2675pure-perl event loop (L<AnyEvent::Loop>) will additionally load it to
2676try to use a monotonic clock for timing stability.
2677
2678=back
2679
1645 2680
1646=head1 FORK 2681=head1 FORK
1647 2682
1648Most event libraries are not fork-safe. The ones who are usually are 2683Most event libraries are not fork-safe. The ones who are usually are
1649because they rely on inefficient but fork-safe C<select> or C<poll> 2684because they rely on inefficient but fork-safe C<select> or C<poll> calls
1650calls. Only L<EV> is fully fork-aware. 2685- higher performance APIs such as BSD's kqueue or the dreaded Linux epoll
2686are usually badly thought-out hacks that are incompatible with fork in
2687one way or another. Only L<EV> is fully fork-aware and ensures that you
2688continue event-processing in both parent and child (or both, if you know
2689what you are doing).
2690
2691This means that, in general, you cannot fork and do event processing in
2692the child if the event library was initialised before the fork (which
2693usually happens when the first AnyEvent watcher is created, or the library
2694is loaded).
1651 2695
1652If you have to fork, you must either do so I<before> creating your first 2696If you have to fork, you must either do so I<before> creating your first
1653watcher OR you must not use AnyEvent at all in the child. 2697watcher OR you must not use AnyEvent at all in the child OR you must do
2698something completely out of the scope of AnyEvent.
2699
2700The problem of doing event processing in the parent I<and> the child
2701is much more complicated: even for backends that I<are> fork-aware or
2702fork-safe, their behaviour is not usually what you want: fork clones all
2703watchers, that means all timers, I/O watchers etc. are active in both
2704parent and child, which is almost never what you want. USing C<exec>
2705to start worker children from some kind of manage rprocess is usually
2706preferred, because it is much easier and cleaner, at the expense of having
2707to have another binary.
1654 2708
1655 2709
1656=head1 SECURITY CONSIDERATIONS 2710=head1 SECURITY CONSIDERATIONS
1657 2711
1658AnyEvent can be forced to load any event model via 2712AnyEvent can be forced to load any event model via
1669 2723
1670 use AnyEvent; 2724 use AnyEvent;
1671 2725
1672Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can 2726Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
1673be used to probe what backend is used and gain other information (which is 2727be used to probe what backend is used and gain other information (which is
1674probably even less useful to an attacker than PERL_ANYEVENT_MODEL). 2728probably even less useful to an attacker than PERL_ANYEVENT_MODEL), and
2729$ENV{PERL_ANYEVENT_STRICT}.
2730
2731Note that AnyEvent will remove I<all> environment variables starting with
2732C<PERL_ANYEVENT_> from C<%ENV> when it is loaded while taint mode is
2733enabled.
2734
2735
2736=head1 BUGS
2737
2738Perl 5.8 has numerous memleaks that sometimes hit this module and are hard
2739to work around. If you suffer from memleaks, first upgrade to Perl 5.10
2740and check wether the leaks still show up. (Perl 5.10.0 has other annoying
2741memleaks, such as leaking on C<map> and C<grep> but it is usually not as
2742pronounced).
1675 2743
1676 2744
1677=head1 SEE ALSO 2745=head1 SEE ALSO
1678 2746
2747Tutorial/Introduction: L<AnyEvent::Intro>.
2748
2749FAQ: L<AnyEvent::FAQ>.
2750
1679Utility functions: L<AnyEvent::Util>. 2751Utility functions: L<AnyEvent::Util>.
1680 2752
1681Event modules: L<EV>, L<EV::Glib>, L<Glib::EV>, L<Event>, L<Glib::Event>, 2753Event modules: L<AnyEvent::Loop>, L<EV>, L<EV::Glib>, L<Glib::EV>,
1682L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>. 2754L<Event>, L<Glib::Event>, L<Glib>, L<Tk>, L<Event::Lib>, L<Qt>, L<POE>.
1683 2755
1684Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>, 2756Implementations: L<AnyEvent::Impl::EV>, L<AnyEvent::Impl::Event>,
1685L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>, 2757L<AnyEvent::Impl::Glib>, L<AnyEvent::Impl::Tk>, L<AnyEvent::Impl::Perl>,
1686L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>, 2758L<AnyEvent::Impl::EventLib>, L<AnyEvent::Impl::Qt>,
1687L<AnyEvent::Impl::POE>. 2759L<AnyEvent::Impl::POE>, L<AnyEvent::Impl::IOAsync>, L<Anyevent::Impl::Irssi>.
1688 2760
1689Non-blocking file handles, sockets, TCP clients and 2761Non-blocking file handles, sockets, TCP clients and
1690servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>. 2762servers: L<AnyEvent::Handle>, L<AnyEvent::Socket>, L<AnyEvent::TLS>.
1691 2763
1692Asynchronous DNS: L<AnyEvent::DNS>. 2764Asynchronous DNS: L<AnyEvent::DNS>.
1693 2765
1694Coroutine support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>, 2766Thread support: L<Coro>, L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>.
1695 2767
1696Nontrivial usage examples: L<Net::FCP>, L<Net::XMPP2>, L<AnyEvent::DNS>. 2768Nontrivial usage examples: L<AnyEvent::GPSD>, L<AnyEvent::IRC>,
2769L<AnyEvent::HTTP>.
1697 2770
1698 2771
1699=head1 AUTHOR 2772=head1 AUTHOR
1700 2773
1701 Marc Lehmann <schmorp@schmorp.de> 2774 Marc Lehmann <schmorp@schmorp.de>

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines