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

Comparing AnyEvent/README (file contents):
Revision 1.16 by root, Sat Apr 19 04:58:14 2008 UTC vs.
Revision 1.45 by root, Fri Jul 17 14:57:03 2009 UTC

1NAME 1NAME
2 AnyEvent - provide framework for multiple event loops 2 AnyEvent - provide framework for multiple event loops
3 3
4 EV, Event, Coro::EV, Coro::Event, Glib, Tk, Perl - various supported 4 EV, Event, Glib, Tk, Perl, Event::Lib, Qt and POE are various supported
5 event loops 5 event loops.
6 6
7SYNOPSIS 7SYNOPSIS
8 use AnyEvent; 8 use AnyEvent;
9 9
10 # file descriptor readable
10 my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub { 11 my $w = AnyEvent->io (fh => $fh, poll => "r", cb => sub { ... });
12
13 # one-shot or repeating timers
14 my $w = AnyEvent->timer (after => $seconds, cb => sub { ... });
15 my $w = AnyEvent->timer (after => $seconds, interval => $seconds, cb => ...
16
17 print AnyEvent->now; # prints current event loop time
18 print AnyEvent->time; # think Time::HiRes::time or simply CORE::time.
19
20 # POSIX signal
21 my $w = AnyEvent->signal (signal => "TERM", cb => sub { ... });
22
23 # child process exit
24 my $w = AnyEvent->child (pid => $pid, cb => sub {
25 my ($pid, $status) = @_;
11 ... 26 ...
12 }); 27 });
13 28
14 my $w = AnyEvent->timer (after => $seconds, cb => sub { 29 # called when event loop idle (if applicable)
15 ... 30 my $w = AnyEvent->idle (cb => sub { ... });
16 });
17 31
18 my $w = AnyEvent->condvar; # stores whether a condition was flagged 32 my $w = AnyEvent->condvar; # stores whether a condition was flagged
33 $w->send; # wake up current and all future recv's
19 $w->wait; # enters "main loop" till $condvar gets ->broadcast 34 $w->recv; # enters "main loop" till $condvar gets ->send
20 $w->broadcast; # wake up current and all future wait's 35 # use a condvar in callback mode:
36 $w->cb (sub { $_[0]->recv });
37
38INTRODUCTION/TUTORIAL
39 This manpage is mainly a reference manual. If you are interested in a
40 tutorial or some gentle introduction, have a look at the AnyEvent::Intro
41 manpage.
21 42
22WHY YOU SHOULD USE THIS MODULE (OR NOT) 43WHY YOU SHOULD USE THIS MODULE (OR NOT)
23 Glib, POE, IO::Async, Event... CPAN offers event models by the dozen 44 Glib, POE, IO::Async, Event... CPAN offers event models by the dozen
24 nowadays. So what is different about AnyEvent? 45 nowadays. So what is different about AnyEvent?
25 46
26 Executive Summary: AnyEvent is *compatible*, AnyEvent is *free of 47 Executive Summary: AnyEvent is *compatible*, AnyEvent is *free of
27 policy* and AnyEvent is *small and efficient*. 48 policy* and AnyEvent is *small and efficient*.
28 49
29 First and foremost, *AnyEvent is not an event model* itself, it only 50 First and foremost, *AnyEvent is not an event model* itself, it only
30 interfaces to whatever event model the main program happens to use in a 51 interfaces to whatever event model the main program happens to use, in a
31 pragmatic way. For event models and certain classes of immortals alike, 52 pragmatic way. For event models and certain classes of immortals alike,
32 the statement "there can only be one" is a bitter reality: In general, 53 the statement "there can only be one" is a bitter reality: In general,
33 only one event loop can be active at the same time in a process. 54 only one event loop can be active at the same time in a process.
34 AnyEvent helps hiding the differences between those event loops. 55 AnyEvent cannot change this, but it can hide the differences between
56 those event loops.
35 57
36 The goal of AnyEvent is to offer module authors the ability to do event 58 The goal of AnyEvent is to offer module authors the ability to do event
37 programming (waiting for I/O or timer events) without subscribing to a 59 programming (waiting for I/O or timer events) without subscribing to a
38 religion, a way of living, and most importantly: without forcing your 60 religion, a way of living, and most importantly: without forcing your
39 module users into the same thing by forcing them to use the same event 61 module users into the same thing by forcing them to use the same event
40 model you use. 62 model you use.
41 63
42 For modules like POE or IO::Async (which is a total misnomer as it is 64 For modules like POE or IO::Async (which is a total misnomer as it is
43 actually doing all I/O *synchronously*...), using them in your module is 65 actually doing all I/O *synchronously*...), using them in your module is
44 like joining a cult: After you joined, you are dependent on them and you 66 like joining a cult: After you joined, you are dependent on them and you
45 cannot use anything else, as it is simply incompatible to everything 67 cannot use anything else, as they are simply incompatible to everything
46 that isn't itself. What's worse, all the potential users of your module 68 that isn't them. What's worse, all the potential users of your module
47 are *also* forced to use the same event loop you use. 69 are *also* forced to use the same event loop you use.
48 70
49 AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works 71 AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works
50 fine. AnyEvent + Tk works fine etc. etc. but none of these work together 72 fine. AnyEvent + Tk works fine etc. etc. but none of these work together
51 with the rest: POE + IO::Async? no go. Tk + Event? no go. Again: if your 73 with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if your
52 module uses one of those, every user of your module has to use it, too. 74 module uses one of those, every user of your module has to use it, too.
53 But if your module uses AnyEvent, it works transparently with all event 75 But if your module uses AnyEvent, it works transparently with all event
54 models it supports (including stuff like POE and IO::Async, as long as 76 models it supports (including stuff like IO::Async, as long as those use
55 those use one of the supported event loops. It is trivial to add new 77 one of the supported event loops. It is trivial to add new event loops
56 event loops to AnyEvent, too, so it is future-proof). 78 to AnyEvent, too, so it is future-proof).
57 79
58 In addition to being free of having to use *the one and only true event 80 In addition to being free of having to use *the one and only true event
59 model*, AnyEvent also is free of bloat and policy: with POE or similar 81 model*, AnyEvent also is free of bloat and policy: with POE or similar
60 modules, you get an enourmous amount of code and strict rules you have 82 modules, you get an enormous amount of code and strict rules you have to
61 to follow. AnyEvent, on the other hand, is lean and up to the point, by 83 follow. AnyEvent, on the other hand, is lean and up to the point, by
62 only offering the functionality that is necessary, in as thin as a 84 only offering the functionality that is necessary, in as thin as a
63 wrapper as technically possible. 85 wrapper as technically possible.
64 86
87 Of course, AnyEvent comes with a big (and fully optional!) toolbox of
88 useful functionality, such as an asynchronous DNS resolver, 100%
89 non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms
90 such as Windows) and lots of real-world knowledge and workarounds for
91 platform bugs and differences.
92
65 Of course, if you want lots of policy (this can arguably be somewhat 93 Now, if you *do want* lots of policy (this can arguably be somewhat
66 useful) and you want to force your users to use the one and only event 94 useful) and you want to force your users to use the one and only event
67 model, you should *not* use this module. 95 model, you should *not* use this module.
68 96
69DESCRIPTION 97DESCRIPTION
70 AnyEvent provides an identical interface to multiple event loops. This 98 AnyEvent provides an identical interface to multiple event loops. This
… …
75 The interface itself is vaguely similar, but not identical to the Event 103 The interface itself is vaguely similar, but not identical to the Event
76 module. 104 module.
77 105
78 During the first call of any watcher-creation method, the module tries 106 During the first call of any watcher-creation method, the module tries
79 to detect the currently loaded event loop by probing whether one of the 107 to detect the currently loaded event loop by probing whether one of the
80 following modules is already loaded: Coro::EV, Coro::Event, EV, Event, 108 following modules is already loaded: EV, Event, Glib,
81 Glib, Tk. The first one found is used. If none are found, the module 109 AnyEvent::Impl::Perl, Tk, Event::Lib, Qt, POE. The first one found is
82 tries to load these modules in the stated order. The first one that can 110 used. If none are found, the module tries to load these modules
111 (excluding Tk, Event::Lib, Qt and POE as the pure perl adaptor should
112 always succeed) in the order given. The first one that can be
83 be successfully loaded will be used. If, after this, still none could be 113 successfully loaded will be used. If, after this, still none could be
84 found, AnyEvent will fall back to a pure-perl event loop, which is not 114 found, AnyEvent will fall back to a pure-perl event loop, which is not
85 very efficient, but should work everywhere. 115 very efficient, but should work everywhere.
86 116
87 Because AnyEvent first checks for modules that are already loaded, 117 Because AnyEvent first checks for modules that are already loaded,
88 loading an event model explicitly before first using AnyEvent will 118 loading an event model explicitly before first using AnyEvent will
… …
97 starts using it, all bets are off. Maybe you should tell their authors 127 starts using it, all bets are off. Maybe you should tell their authors
98 to use AnyEvent so their modules work together with others seamlessly... 128 to use AnyEvent so their modules work together with others seamlessly...
99 129
100 The pure-perl implementation of AnyEvent is called 130 The pure-perl implementation of AnyEvent is called
101 "AnyEvent::Impl::Perl". Like other event modules you can load it 131 "AnyEvent::Impl::Perl". Like other event modules you can load it
102 explicitly. 132 explicitly and enjoy the high availability of that event loop :)
103 133
104WATCHERS 134WATCHERS
105 AnyEvent has the central concept of a *watcher*, which is an object that 135 AnyEvent has the central concept of a *watcher*, which is an object that
106 stores relevant data for each kind of event you are waiting for, such as 136 stores relevant data for each kind of event you are waiting for, such as
107 the callback to call, the filehandle to watch, etc. 137 the callback to call, the file handle to watch, etc.
108 138
109 These watchers are normal Perl objects with normal Perl lifetime. After 139 These watchers are normal Perl objects with normal Perl lifetime. After
110 creating a watcher it will immediately "watch" for events and invoke the 140 creating a watcher it will immediately "watch" for events and invoke the
111 callback when the event occurs (of course, only when the event model is 141 callback when the event occurs (of course, only when the event model is
112 in control). 142 in control).
113 143
144 Note that callbacks must not permanently change global variables
145 potentially in use by the event loop (such as $_ or $[) and that
146 callbacks must not "die". The former is good programming practise in
147 Perl and the latter stems from the fact that exception handling differs
148 widely between event loops.
149
114 To disable the watcher you have to destroy it (e.g. by setting the 150 To disable the watcher you have to destroy it (e.g. by setting the
115 variable you store it in to "undef" or otherwise deleting all references 151 variable you store it in to "undef" or otherwise deleting all references
116 to it). 152 to it).
117 153
118 All watchers are created by calling a method on the "AnyEvent" class. 154 All watchers are created by calling a method on the "AnyEvent" class.
… …
120 Many watchers either are used with "recursion" (repeating timers for 156 Many watchers either are used with "recursion" (repeating timers for
121 example), or need to refer to their watcher object in other ways. 157 example), or need to refer to their watcher object in other ways.
122 158
123 An any way to achieve that is this pattern: 159 An any way to achieve that is this pattern:
124 160
125 my $w; $w = AnyEvent->type (arg => value ..., cb => sub { 161 my $w; $w = AnyEvent->type (arg => value ..., cb => sub {
126 # you can use $w here, for example to undef it 162 # you can use $w here, for example to undef it
127 undef $w; 163 undef $w;
128 }); 164 });
129 165
130 Note that "my $w; $w =" combination. This is necessary because in Perl, 166 Note that "my $w; $w =" combination. This is necessary because in Perl,
131 my variables are only visible after the statement in which they are 167 my variables are only visible after the statement in which they are
132 declared. 168 declared.
133 169
134 IO WATCHERS 170 I/O WATCHERS
135 You can create an I/O watcher by calling the "AnyEvent->io" method with 171 You can create an I/O watcher by calling the "AnyEvent->io" method with
136 the following mandatory key-value pairs as arguments: 172 the following mandatory key-value pairs as arguments:
137 173
138 "fh" the Perl *file handle* (*not* file descriptor) to watch for events. 174 "fh" is the Perl *file handle* (or a naked file descriptor) to watch for
175 events (AnyEvent might or might not keep a reference to this file
176 handle). Note that only file handles pointing to things for which
177 non-blocking operation makes sense are allowed. This includes sockets,
178 most character devices, pipes, fifos and so on, but not for example
179 files or block devices.
180
139 "poll" must be a string that is either "r" or "w", which creates a 181 "poll" must be a string that is either "r" or "w", which creates a
140 watcher waiting for "r"eadable or "w"ritable events, respectively. "cb" 182 watcher waiting for "r"eadable or "w"ritable events, respectively.
183
141 is the callback to invoke each time the file handle becomes ready. 184 "cb" is the callback to invoke each time the file handle becomes ready.
142 185
143 File handles will be kept alive, so as long as the watcher exists, the 186 Although the callback might get passed parameters, their value and
144 file handle exists, too. 187 presence is undefined and you cannot rely on them. Portable AnyEvent
188 callbacks cannot use arguments passed to I/O watcher callbacks.
145 189
190 The I/O watcher might use the underlying file descriptor or a copy of
146 It is not allowed to close a file handle as long as any watcher is 191 it. You must not close a file handle as long as any watcher is active on
147 active on the underlying file descriptor. 192 the underlying file descriptor.
148 193
149 Some event loops issue spurious readyness notifications, so you should 194 Some event loops issue spurious readyness notifications, so you should
150 always use non-blocking calls when reading/writing from/to your file 195 always use non-blocking calls when reading/writing from/to your file
151 handles. 196 handles.
152 197
153 Example:
154
155 # wait for readability of STDIN, then read a line and disable the watcher 198 Example: wait for readability of STDIN, then read a line and disable the
199 watcher.
200
156 my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub { 201 my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
157 chomp (my $input = <STDIN>); 202 chomp (my $input = <STDIN>);
158 warn "read: $input\n"; 203 warn "read: $input\n";
159 undef $w; 204 undef $w;
160 }); 205 });
… …
162 TIME WATCHERS 207 TIME WATCHERS
163 You can create a time watcher by calling the "AnyEvent->timer" method 208 You can create a time watcher by calling the "AnyEvent->timer" method
164 with the following mandatory arguments: 209 with the following mandatory arguments:
165 210
166 "after" specifies after how many seconds (fractional values are 211 "after" specifies after how many seconds (fractional values are
167 supported) should the timer activate. "cb" the callback to invoke in 212 supported) the callback should be invoked. "cb" is the callback to
168 that case. 213 invoke in that case.
169 214
170 The timer callback will be invoked at most once: if you want a repeating 215 Although the callback might get passed parameters, their value and
171 timer you have to create a new watcher (this is a limitation by both Tk 216 presence is undefined and you cannot rely on them. Portable AnyEvent
172 and Glib). 217 callbacks cannot use arguments passed to time watcher callbacks.
173 218
174 Example: 219 The callback will normally be invoked once only. If you specify another
220 parameter, "interval", as a strictly positive number (> 0), then the
221 callback will be invoked regularly at that interval (in fractional
222 seconds) after the first invocation. If "interval" is specified with a
223 false value, then it is treated as if it were missing.
175 224
225 The callback will be rescheduled before invoking the callback, but no
226 attempt is done to avoid timer drift in most backends, so the interval
227 is only approximate.
228
176 # fire an event after 7.7 seconds 229 Example: fire an event after 7.7 seconds.
230
177 my $w = AnyEvent->timer (after => 7.7, cb => sub { 231 my $w = AnyEvent->timer (after => 7.7, cb => sub {
178 warn "timeout\n"; 232 warn "timeout\n";
179 }); 233 });
180 234
181 # to cancel the timer: 235 # to cancel the timer:
182 undef $w; 236 undef $w;
183 237
184 Example 2:
185
186 # fire an event after 0.5 seconds, then roughly every second 238 Example 2: fire an event after 0.5 seconds, then roughly every second.
187 my $w;
188 239
189 my $cb = sub {
190 # cancel the old timer while creating a new one
191 $w = AnyEvent->timer (after => 1, cb => $cb); 240 my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub {
241 warn "timeout\n";
192 }; 242 };
193
194 # start the "loop" by creating the first watcher
195 $w = AnyEvent->timer (after => 0.5, cb => $cb);
196 243
197 TIMING ISSUES 244 TIMING ISSUES
198 There are two ways to handle timers: based on real time (relative, "fire 245 There are two ways to handle timers: based on real time (relative, "fire
199 in 10 seconds") and based on wallclock time (absolute, "fire at 12 246 in 10 seconds") and based on wallclock time (absolute, "fire at 12
200 o'clock"). 247 o'clock").
201 248
202 While most event loops expect timers to specified in a relative way, 249 While most event loops expect timers to specified in a relative way,
203 they use absolute time internally. This makes a difference when your 250 they use absolute time internally. This makes a difference when your
204 clock "jumps", for example, when ntp decides to set your clock backwards 251 clock "jumps", for example, when ntp decides to set your clock backwards
205 from the wrong 2014-01-01 to 2008-01-01, a watcher that you created to 252 from the wrong date of 2014-01-01 to 2008-01-01, a watcher that is
206 fire "after" a second might actually take six years to finally fire. 253 supposed to fire "after" a second might actually take six years to
254 finally fire.
207 255
208 AnyEvent cannot compensate for this. The only event loop that is 256 AnyEvent cannot compensate for this. The only event loop that is
209 conscious about these issues is EV, which offers both relative 257 conscious about these issues is EV, which offers both relative
210 (ev_timer) and absolute (ev_periodic) timers. 258 (ev_timer, based on true relative time) and absolute (ev_periodic, based
259 on wallclock time) timers.
211 260
212 AnyEvent always prefers relative timers, if available, matching the 261 AnyEvent always prefers relative timers, if available, matching the
213 AnyEvent API. 262 AnyEvent API.
214 263
264 AnyEvent has two additional methods that return the "current time":
265
266 AnyEvent->time
267 This returns the "current wallclock time" as a fractional number of
268 seconds since the Epoch (the same thing as "time" or
269 "Time::HiRes::time" return, and the result is guaranteed to be
270 compatible with those).
271
272 It progresses independently of any event loop processing, i.e. each
273 call will check the system clock, which usually gets updated
274 frequently.
275
276 AnyEvent->now
277 This also returns the "current wallclock time", but unlike "time",
278 above, this value might change only once per event loop iteration,
279 depending on the event loop (most return the same time as "time",
280 above). This is the time that AnyEvent's timers get scheduled
281 against.
282
283 *In almost all cases (in all cases if you don't care), this is the
284 function to call when you want to know the current time.*
285
286 This function is also often faster then "AnyEvent->time", and thus
287 the preferred method if you want some timestamp (for example,
288 AnyEvent::Handle uses this to update it's activity timeouts).
289
290 The rest of this section is only of relevance if you try to be very
291 exact with your timing, you can skip it without bad conscience.
292
293 For a practical example of when these times differ, consider
294 Event::Lib and EV and the following set-up:
295
296 The event loop is running and has just invoked one of your callback
297 at time=500 (assume no other callbacks delay processing). In your
298 callback, you wait a second by executing "sleep 1" (blocking the
299 process for a second) and then (at time=501) you create a relative
300 timer that fires after three seconds.
301
302 With Event::Lib, "AnyEvent->time" and "AnyEvent->now" will both
303 return 501, because that is the current time, and the timer will be
304 scheduled to fire at time=504 (501 + 3).
305
306 With EV, "AnyEvent->time" returns 501 (as that is the current time),
307 but "AnyEvent->now" returns 500, as that is the time the last event
308 processing phase started. With EV, your timer gets scheduled to run
309 at time=503 (500 + 3).
310
311 In one sense, Event::Lib is more exact, as it uses the current time
312 regardless of any delays introduced by event processing. However,
313 most callbacks do not expect large delays in processing, so this
314 causes a higher drift (and a lot more system calls to get the
315 current time).
316
317 In another sense, EV is more exact, as your timer will be scheduled
318 at the same time, regardless of how long event processing actually
319 took.
320
321 In either case, if you care (and in most cases, you don't), then you
322 can get whatever behaviour you want with any event loop, by taking
323 the difference between "AnyEvent->time" and "AnyEvent->now" into
324 account.
325
326 AnyEvent->now_update
327 Some event loops (such as EV or AnyEvent::Impl::Perl) cache the
328 current time for each loop iteration (see the discussion of
329 AnyEvent->now, above).
330
331 When a callback runs for a long time (or when the process sleeps),
332 then this "current" time will differ substantially from the real
333 time, which might affect timers and time-outs.
334
335 When this is the case, you can call this method, which will update
336 the event loop's idea of "current time".
337
338 Note that updating the time *might* cause some events to be handled.
339
215 SIGNAL WATCHERS 340 SIGNAL WATCHERS
216 You can watch for signals using a signal watcher, "signal" is the signal 341 You can watch for signals using a signal watcher, "signal" is the signal
217 *name* without any "SIG" prefix, "cb" is the Perl callback to be invoked 342 *name* in uppercase and without any "SIG" prefix, "cb" is the Perl
218 whenever a signal occurs. 343 callback to be invoked whenever a signal occurs.
219 344
345 Although the callback might get passed parameters, their value and
346 presence is undefined and you cannot rely on them. Portable AnyEvent
347 callbacks cannot use arguments passed to signal watcher callbacks.
348
220 Multiple signals occurances can be clumped together into one callback 349 Multiple signal occurrences can be clumped together into one callback
221 invocation, and callback invocation will be synchronous. synchronous 350 invocation, and callback invocation will be synchronous. Synchronous
222 means that it might take a while until the signal gets handled by the 351 means that it might take a while until the signal gets handled by the
223 process, but it is guarenteed not to interrupt any other callbacks. 352 process, but it is guaranteed not to interrupt any other callbacks.
224 353
225 The main advantage of using these watchers is that you can share a 354 The main advantage of using these watchers is that you can share a
226 signal between multiple watchers. 355 signal between multiple watchers.
227 356
228 This watcher might use %SIG, so programs overwriting those signals 357 This watcher might use %SIG, so programs overwriting those signals
… …
234 363
235 CHILD PROCESS WATCHERS 364 CHILD PROCESS WATCHERS
236 You can also watch on a child process exit and catch its exit status. 365 You can also watch on a child process exit and catch its exit status.
237 366
238 The child process is specified by the "pid" argument (if set to 0, it 367 The child process is specified by the "pid" argument (if set to 0, it
239 watches for any child process exit). The watcher will trigger as often 368 watches for any child process exit). The watcher will triggered only
240 as status change for the child are received. This works by installing a 369 when the child process has finished and an exit status is available, not
241 signal handler for "SIGCHLD". The callback will be called with the pid 370 on any trace events (stopped/continued).
242 and exit status (as returned by waitpid).
243 371
244 Example: wait for pid 1333 372 The callback will be called with the pid and exit status (as returned by
373 waitpid), so unlike other watcher types, you *can* rely on child watcher
374 callback arguments.
245 375
376 This watcher type works by installing a signal handler for "SIGCHLD",
377 and since it cannot be shared, nothing else should use SIGCHLD or reap
378 random child processes (waiting for specific child processes, e.g.
379 inside "system", is just fine).
380
381 There is a slight catch to child watchers, however: you usually start
382 them *after* the child process was created, and this means the process
383 could have exited already (and no SIGCHLD will be sent anymore).
384
385 Not all event models handle this correctly (neither POE nor IO::Async
386 do, see their AnyEvent::Impl manpages for details), but even for event
387 models that *do* handle this correctly, they usually need to be loaded
388 before the process exits (i.e. before you fork in the first place).
389 AnyEvent's pure perl event loop handles all cases correctly regardless
390 of when you start the watcher.
391
392 This means you cannot create a child watcher as the very first thing in
393 an AnyEvent program, you *have* to create at least one watcher before
394 you "fork" the child (alternatively, you can call "AnyEvent::detect").
395
396 Example: fork a process and wait for it
397
398 my $done = AnyEvent->condvar;
399
400 my $pid = fork or exit 5;
401
246 my $w = AnyEvent->child ( 402 my $w = AnyEvent->child (
247 pid => 1333, 403 pid => $pid,
248 cb => sub { 404 cb => sub {
249 my ($pid, $status) = @_; 405 my ($pid, $status) = @_;
250 warn "pid $pid exited with status $status"; 406 warn "pid $pid exited with status $status";
407 $done->send;
251 }, 408 },
252 ); 409 );
410
411 # do something else, then wait for process exit
412 $done->recv;
413
414 IDLE WATCHERS
415 Sometimes there is a need to do something, but it is not so important to
416 do it instantly, but only when there is nothing better to do. This
417 "nothing better to do" is usually defined to be "no other events need
418 attention by the event loop".
419
420 Idle watchers ideally get invoked when the event loop has nothing better
421 to do, just before it would block the process to wait for new events.
422 Instead of blocking, the idle watcher is invoked.
423
424 Most event loops unfortunately do not really support idle watchers (only
425 EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent
426 will simply call the callback "from time to time".
427
428 Example: read lines from STDIN, but only process them when the program
429 is otherwise idle:
430
431 my @lines; # read data
432 my $idle_w;
433 my $io_w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub {
434 push @lines, scalar <STDIN>;
435
436 # start an idle watcher, if not already done
437 $idle_w ||= AnyEvent->idle (cb => sub {
438 # handle only one line, when there are lines left
439 if (my $line = shift @lines) {
440 print "handled when idle: $line";
441 } else {
442 # otherwise disable the idle watcher again
443 undef $idle_w;
444 }
445 });
446 });
253 447
254 CONDITION VARIABLES 448 CONDITION VARIABLES
449 If you are familiar with some event loops you will know that all of them
450 require you to run some blocking "loop", "run" or similar function that
451 will actively watch for new events and call your callbacks.
452
453 AnyEvent is slightly different: it expects somebody else to run the
454 event loop and will only block when necessary (usually when told by the
455 user).
456
457 The instrument to do that is called a "condition variable", so called
458 because they represent a condition that must become true.
459
460 Now is probably a good time to look at the examples further below.
461
255 Condition variables can be created by calling the "AnyEvent->condvar" 462 Condition variables can be created by calling the "AnyEvent->condvar"
256 method without any arguments. 463 method, usually without arguments. The only argument pair allowed is
464 "cb", which specifies a callback to be called when the condition
465 variable becomes true, with the condition variable as the first argument
466 (but not the results).
257 467
258 A condition variable waits for a condition - precisely that the 468 After creation, the condition variable is "false" until it becomes
259 "->broadcast" method has been called. 469 "true" by calling the "send" method (or calling the condition variable
470 as if it were a callback, read about the caveats in the description for
471 the "->send" method).
260 472
261 They are very useful to signal that a condition has been fulfilled, for 473 Condition variables are similar to callbacks, except that you can
474 optionally wait for them. They can also be called merge points - points
475 in time where multiple outstanding events have been processed. And yet
476 another way to call them is transactions - each condition variable can
477 be used to represent a transaction, which finishes at some point and
478 delivers a result.
479
480 Condition variables are very useful to signal that something has
262 example, if you write a module that does asynchronous http requests, 481 finished, for example, if you write a module that does asynchronous http
263 then a condition variable would be the ideal candidate to signal the 482 requests, then a condition variable would be the ideal candidate to
264 availability of results. 483 signal the availability of results. The user can either act when the
484 callback is called or can synchronously "->recv" for the results.
265 485
266 You can also use condition variables to block your main program until an 486 You can also use them to simulate traditional event loops - for example,
267 event occurs - for example, you could "->wait" in your main program 487 you can block your main program until an event occurs - for example, you
268 until the user clicks the Quit button in your app, which would 488 could "->recv" in your main program until the user clicks the Quit
269 "->broadcast" the "quit" event. 489 button of your app, which would "->send" the "quit" event.
270 490
271 Note that condition variables recurse into the event loop - if you have 491 Note that condition variables recurse into the event loop - if you have
272 two pirces of code that call "->wait" in a round-robbin fashion, you 492 two pieces of code that call "->recv" in a round-robin fashion, you
273 lose. Therefore, condition variables are good to export to your caller, 493 lose. Therefore, condition variables are good to export to your caller,
274 but you should avoid making a blocking wait yourself, at least in 494 but you should avoid making a blocking wait yourself, at least in
275 callbacks, as this asks for trouble. 495 callbacks, as this asks for trouble.
276 496
277 This object has two methods: 497 Condition variables are represented by hash refs in perl, and the keys
498 used by AnyEvent itself are all named "_ae_XXX" to make subclassing easy
499 (it is often useful to build your own transaction class on top of
500 AnyEvent). To subclass, use "AnyEvent::CondVar" as base class and call
501 it's "new" method in your own "new" method.
278 502
279 $cv->wait 503 There are two "sides" to a condition variable - the "producer side"
280 Wait (blocking if necessary) until the "->broadcast" method has been 504 which eventually calls "-> send", and the "consumer side", which waits
281 called on c<$cv>, while servicing other watchers normally. 505 for the send to occur.
282 506
283 You can only wait once on a condition - additional calls will return 507 Example: wait for a timer.
284 immediately.
285
286 Not all event models support a blocking wait - some die in that case
287 (programs might want to do that to stay interactive), so *if you are
288 using this from a module, never require a blocking wait*, but let
289 the caller decide whether the call will block or not (for example,
290 by coupling condition variables with some kind of request results
291 and supporting callbacks so the caller knows that getting the result
292 will not block, while still suppporting blocking waits if the caller
293 so desires).
294
295 Another reason *never* to "->wait" in a module is that you cannot
296 sensibly have two "->wait"'s in parallel, as that would require
297 multiple interpreters or coroutines/threads, none of which
298 "AnyEvent" can supply (the coroutine-aware backends
299 AnyEvent::Impl::CoroEV and AnyEvent::Impl::CoroEvent explicitly
300 support concurrent "->wait"'s from different coroutines, however).
301
302 $cv->broadcast
303 Flag the condition as ready - a running "->wait" and all further
304 calls to "wait" will (eventually) return after this method has been
305 called. If nobody is waiting the broadcast will be remembered..
306
307 Example:
308 508
309 # wait till the result is ready 509 # wait till the result is ready
310 my $result_ready = AnyEvent->condvar; 510 my $result_ready = AnyEvent->condvar;
311 511
312 # do something such as adding a timer 512 # do something such as adding a timer
313 # or socket watcher the calls $result_ready->broadcast 513 # or socket watcher the calls $result_ready->send
314 # when the "result" is ready. 514 # when the "result" is ready.
315 # in this case, we simply use a timer: 515 # in this case, we simply use a timer:
316 my $w = AnyEvent->timer ( 516 my $w = AnyEvent->timer (
317 after => 1, 517 after => 1,
318 cb => sub { $result_ready->broadcast }, 518 cb => sub { $result_ready->send },
319 ); 519 );
320 520
321 # this "blocks" (while handling events) till the watcher 521 # this "blocks" (while handling events) till the callback
322 # calls broadcast 522 # calls -<send
323 $result_ready->wait; 523 $result_ready->recv;
524
525 Example: wait for a timer, but take advantage of the fact that condition
526 variables are also callable directly.
527
528 my $done = AnyEvent->condvar;
529 my $delay = AnyEvent->timer (after => 5, cb => $done);
530 $done->recv;
531
532 Example: Imagine an API that returns a condvar and doesn't support
533 callbacks. This is how you make a synchronous call, for example from the
534 main program:
535
536 use AnyEvent::CouchDB;
537
538 ...
539
540 my @info = $couchdb->info->recv;
541
542 And this is how you would just set a callback to be called whenever the
543 results are available:
544
545 $couchdb->info->cb (sub {
546 my @info = $_[0]->recv;
547 });
548
549 METHODS FOR PRODUCERS
550 These methods should only be used by the producing side, i.e. the
551 code/module that eventually sends the signal. Note that it is also the
552 producer side which creates the condvar in most cases, but it isn't
553 uncommon for the consumer to create it as well.
554
555 $cv->send (...)
556 Flag the condition as ready - a running "->recv" and all further
557 calls to "recv" will (eventually) return after this method has been
558 called. If nobody is waiting the send will be remembered.
559
560 If a callback has been set on the condition variable, it is called
561 immediately from within send.
562
563 Any arguments passed to the "send" call will be returned by all
564 future "->recv" calls.
565
566 Condition variables are overloaded so one can call them directly (as
567 if they were a code reference). Calling them directly is the same as
568 calling "send".
569
570 $cv->croak ($error)
571 Similar to send, but causes all call's to "->recv" to invoke
572 "Carp::croak" with the given error message/object/scalar.
573
574 This can be used to signal any errors to the condition variable
575 user/consumer. Doing it this way instead of calling "croak" directly
576 delays the error detetcion, but has the overwhelmign advantage that
577 it diagnoses the error at the place where the result is expected,
578 and not deep in some event clalback without connection to the actual
579 code causing the problem.
580
581 $cv->begin ([group callback])
582 $cv->end
583 These two methods can be used to combine many transactions/events
584 into one. For example, a function that pings many hosts in parallel
585 might want to use a condition variable for the whole process.
586
587 Every call to "->begin" will increment a counter, and every call to
588 "->end" will decrement it. If the counter reaches 0 in "->end", the
589 (last) callback passed to "begin" will be executed. That callback is
590 *supposed* to call "->send", but that is not required. If no
591 callback was set, "send" will be called without any arguments.
592
593 You can think of "$cv->send" giving you an OR condition (one call
594 sends), while "$cv->begin" and "$cv->end" giving you an AND
595 condition (all "begin" calls must be "end"'ed before the condvar
596 sends).
597
598 Let's start with a simple example: you have two I/O watchers (for
599 example, STDOUT and STDERR for a program), and you want to wait for
600 both streams to close before activating a condvar:
601
602 my $cv = AnyEvent->condvar;
603
604 $cv->begin; # first watcher
605 my $w1 = AnyEvent->io (fh => $fh1, cb => sub {
606 defined sysread $fh1, my $buf, 4096
607 or $cv->end;
608 });
609
610 $cv->begin; # second watcher
611 my $w2 = AnyEvent->io (fh => $fh2, cb => sub {
612 defined sysread $fh2, my $buf, 4096
613 or $cv->end;
614 });
615
616 $cv->recv;
617
618 This works because for every event source (EOF on file handle),
619 there is one call to "begin", so the condvar waits for all calls to
620 "end" before sending.
621
622 The ping example mentioned above is slightly more complicated, as
623 the there are results to be passwd back, and the number of tasks
624 that are begung can potentially be zero:
625
626 my $cv = AnyEvent->condvar;
627
628 my %result;
629 $cv->begin (sub { $cv->send (\%result) });
630
631 for my $host (@list_of_hosts) {
632 $cv->begin;
633 ping_host_then_call_callback $host, sub {
634 $result{$host} = ...;
635 $cv->end;
636 };
637 }
638
639 $cv->end;
640
641 This code fragment supposedly pings a number of hosts and calls
642 "send" after results for all then have have been gathered - in any
643 order. To achieve this, the code issues a call to "begin" when it
644 starts each ping request and calls "end" when it has received some
645 result for it. Since "begin" and "end" only maintain a counter, the
646 order in which results arrive is not relevant.
647
648 There is an additional bracketing call to "begin" and "end" outside
649 the loop, which serves two important purposes: first, it sets the
650 callback to be called once the counter reaches 0, and second, it
651 ensures that "send" is called even when "no" hosts are being pinged
652 (the loop doesn't execute once).
653
654 This is the general pattern when you "fan out" into multiple (but
655 potentially none) subrequests: use an outer "begin"/"end" pair to
656 set the callback and ensure "end" is called at least once, and then,
657 for each subrequest you start, call "begin" and for each subrequest
658 you finish, call "end".
659
660 METHODS FOR CONSUMERS
661 These methods should only be used by the consuming side, i.e. the code
662 awaits the condition.
663
664 $cv->recv
665 Wait (blocking if necessary) until the "->send" or "->croak" methods
666 have been called on c<$cv>, while servicing other watchers normally.
667
668 You can only wait once on a condition - additional calls are valid
669 but will return immediately.
670
671 If an error condition has been set by calling "->croak", then this
672 function will call "croak".
673
674 In list context, all parameters passed to "send" will be returned,
675 in scalar context only the first one will be returned.
676
677 Note that doing a blocking wait in a callback is not supported by
678 any event loop, that is, recursive invocation of a blocking "->recv"
679 is not allowed, and the "recv" call will "croak" if such a condition
680 is detected. This condition can be slightly loosened by using
681 Coro::AnyEvent, which allows you to do a blocking "->recv" from any
682 thread that doesn't run the event loop itself.
683
684 Not all event models support a blocking wait - some die in that case
685 (programs might want to do that to stay interactive), so *if you are
686 using this from a module, never require a blocking wait*. Instead,
687 let the caller decide whether the call will block or not (for
688 example, by coupling condition variables with some kind of request
689 results and supporting callbacks so the caller knows that getting
690 the result will not block, while still supporting blocking waits if
691 the caller so desires).
692
693 You can ensure that "-recv" never blocks by setting a callback and
694 only calling "->recv" from within that callback (or at a later
695 time). This will work even when the event loop does not support
696 blocking waits otherwise.
697
698 $bool = $cv->ready
699 Returns true when the condition is "true", i.e. whether "send" or
700 "croak" have been called.
701
702 $cb = $cv->cb ($cb->($cv))
703 This is a mutator function that returns the callback set and
704 optionally replaces it before doing so.
705
706 The callback will be called when the condition becomes "true", i.e.
707 when "send" or "croak" are called, with the only argument being the
708 condition variable itself. Calling "recv" inside the callback or at
709 any later time is guaranteed not to block.
710
711SUPPORTED EVENT LOOPS/BACKENDS
712 The available backend classes are (every class has its own manpage):
713
714 Backends that are autoprobed when no other event loop can be found.
715 EV is the preferred backend when no other event loop seems to be in
716 use. If EV is not installed, then AnyEvent will try Event, and,
717 failing that, will fall back to its own pure-perl implementation,
718 which is available everywhere as it comes with AnyEvent itself.
719
720 AnyEvent::Impl::EV based on EV (interface to libev, best choice).
721 AnyEvent::Impl::Event based on Event, very stable, few glitches.
722 AnyEvent::Impl::Perl pure-perl implementation, fast and portable.
723
724 Backends that are transparently being picked up when they are used.
725 These will be used when they are currently loaded when the first
726 watcher is created, in which case it is assumed that the application
727 is using them. This means that AnyEvent will automatically pick the
728 right backend when the main program loads an event module before
729 anything starts to create watchers. Nothing special needs to be done
730 by the main program.
731
732 AnyEvent::Impl::Glib based on Glib, slow but very stable.
733 AnyEvent::Impl::Tk based on Tk, very broken.
734 AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse.
735 AnyEvent::Impl::POE based on POE, very slow, some limitations.
736
737 Backends with special needs.
738 Qt requires the Qt::Application to be instantiated first, but will
739 otherwise be picked up automatically. As long as the main program
740 instantiates the application before any AnyEvent watchers are
741 created, everything should just work.
742
743 AnyEvent::Impl::Qt based on Qt.
744
745 Support for IO::Async can only be partial, as it is too broken and
746 architecturally limited to even support the AnyEvent API. It also is
747 the only event loop that needs the loop to be set explicitly, so it
748 can only be used by a main program knowing about AnyEvent. See
749 AnyEvent::Impl::Async for the gory details.
750
751 AnyEvent::Impl::IOAsync based on IO::Async, cannot be autoprobed.
752
753 Event loops that are indirectly supported via other backends.
754 Some event loops can be supported via other modules:
755
756 There is no direct support for WxWidgets (Wx) or Prima.
757
758 WxWidgets has no support for watching file handles. However, you can
759 use WxWidgets through the POE adaptor, as POE has a Wx backend that
760 simply polls 20 times per second, which was considered to be too
761 horrible to even consider for AnyEvent.
762
763 Prima is not supported as nobody seems to be using it, but it has a
764 POE backend, so it can be supported through POE.
765
766 AnyEvent knows about both Prima and Wx, however, and will try to
767 load POE when detecting them, in the hope that POE will pick them
768 up, in which case everything will be automatic.
324 769
325GLOBAL VARIABLES AND FUNCTIONS 770GLOBAL VARIABLES AND FUNCTIONS
771 These are not normally required to use AnyEvent, but can be useful to
772 write AnyEvent extension modules.
773
326 $AnyEvent::MODEL 774 $AnyEvent::MODEL
327 Contains "undef" until the first watcher is being created. Then it 775 Contains "undef" until the first watcher is being created, before
776 the backend has been autodetected.
777
328 contains the event model that is being used, which is the name of 778 Afterwards it contains the event model that is being used, which is
329 the Perl class implementing the model. This class is usually one of 779 the name of the Perl class implementing the model. This class is
330 the "AnyEvent::Impl:xxx" modules, but can be any other class in the 780 usually one of the "AnyEvent::Impl:xxx" modules, but can be any
331 case AnyEvent has been extended at runtime (e.g. in *rxvt-unicode*). 781 other class in the case AnyEvent has been extended at runtime (e.g.
332 782 in *rxvt-unicode* it will be "urxvt::anyevent").
333 The known classes so far are:
334
335 AnyEvent::Impl::CoroEV based on Coro::EV, best choice.
336 AnyEvent::Impl::CoroEvent based on Coro::Event, second best choice.
337 AnyEvent::Impl::EV based on EV (an interface to libev, also best choice).
338 AnyEvent::Impl::Event based on Event, also second best choice :)
339 AnyEvent::Impl::Glib based on Glib, third-best choice.
340 AnyEvent::Impl::Tk based on Tk, very bad choice.
341 AnyEvent::Impl::Perl pure-perl implementation, inefficient but portable.
342 783
343 AnyEvent::detect 784 AnyEvent::detect
344 Returns $AnyEvent::MODEL, forcing autodetection of the event model 785 Returns $AnyEvent::MODEL, forcing autodetection of the event model
345 if necessary. You should only call this function right before you 786 if necessary. You should only call this function right before you
346 would have created an AnyEvent watcher anyway, that is, as late as 787 would have created an AnyEvent watcher anyway, that is, as late as
347 possible at runtime. 788 possible at runtime, and not e.g. while initialising of your module.
789
790 If you need to do some initialisation before AnyEvent watchers are
791 created, use "post_detect".
792
793 $guard = AnyEvent::post_detect { BLOCK }
794 Arranges for the code block to be executed as soon as the event
795 model is autodetected (or immediately if this has already happened).
796
797 The block will be executed *after* the actual backend has been
798 detected ($AnyEvent::MODEL is set), but *before* any watchers have
799 been created, so it is possible to e.g. patch @AnyEvent::ISA or do
800 other initialisations - see the sources of AnyEvent::Strict or
801 AnyEvent::AIO to see how this is used.
802
803 The most common usage is to create some global watchers, without
804 forcing event module detection too early, for example, AnyEvent::AIO
805 creates and installs the global IO::AIO watcher in a "post_detect"
806 block to avoid autodetecting the event module at load time.
807
808 If called in scalar or list context, then it creates and returns an
809 object that automatically removes the callback again when it is
810 destroyed. See Coro::BDB for a case where this is useful.
811
812 @AnyEvent::post_detect
813 If there are any code references in this array (you can "push" to it
814 before or after loading AnyEvent), then they will called directly
815 after the event loop has been chosen.
816
817 You should check $AnyEvent::MODEL before adding to this array,
818 though: if it is defined then the event loop has already been
819 detected, and the array will be ignored.
820
821 Best use "AnyEvent::post_detect { BLOCK }" when your application
822 allows it,as it takes care of these details.
823
824 This variable is mainly useful for modules that can do something
825 useful when AnyEvent is used and thus want to know when it is
826 initialised, but do not need to even load it by default. This array
827 provides the means to hook into AnyEvent passively, without loading
828 it.
348 829
349WHAT TO DO IN A MODULE 830WHAT TO DO IN A MODULE
350 As a module author, you should "use AnyEvent" and call AnyEvent methods 831 As a module author, you should "use AnyEvent" and call AnyEvent methods
351 freely, but you should not load a specific event module or rely on it. 832 freely, but you should not load a specific event module or rely on it.
352 833
353 Be careful when you create watchers in the module body - AnyEvent will 834 Be careful when you create watchers in the module body - AnyEvent will
354 decide which event module to use as soon as the first method is called, 835 decide which event module to use as soon as the first method is called,
355 so by calling AnyEvent in your module body you force the user of your 836 so by calling AnyEvent in your module body you force the user of your
356 module to load the event module first. 837 module to load the event module first.
357 838
358 Never call "->wait" on a condition variable unless you *know* that the 839 Never call "->recv" on a condition variable unless you *know* that the
359 "->broadcast" method has been called on it already. This is because it 840 "->send" method has been called on it already. This is because it will
360 will stall the whole program, and the whole point of using events is to 841 stall the whole program, and the whole point of using events is to stay
361 stay interactive. 842 interactive.
362 843
363 It is fine, however, to call "->wait" when the user of your module 844 It is fine, however, to call "->recv" when the user of your module
364 requests it (i.e. if you create a http request object ad have a method 845 requests it (i.e. if you create a http request object ad have a method
365 called "results" that returns the results, it should call "->wait" 846 called "results" that returns the results, it should call "->recv"
366 freely, as the user of your module knows what she is doing. always). 847 freely, as the user of your module knows what she is doing. always).
367 848
368WHAT TO DO IN THE MAIN PROGRAM 849WHAT TO DO IN THE MAIN PROGRAM
369 There will always be a single main program - the only place that should 850 There will always be a single main program - the only place that should
370 dictate which event model to use. 851 dictate which event model to use.
… …
372 If it doesn't care, it can just "use AnyEvent" and use it itself, or not 853 If it doesn't care, it can just "use AnyEvent" and use it itself, or not
373 do anything special (it does not need to be event-based) and let 854 do anything special (it does not need to be event-based) and let
374 AnyEvent decide which implementation to chose if some module relies on 855 AnyEvent decide which implementation to chose if some module relies on
375 it. 856 it.
376 857
377 If the main program relies on a specific event model. For example, in 858 If the main program relies on a specific event model - for example, in
378 Gtk2 programs you have to rely on the Glib module. You should load the 859 Gtk2 programs you have to rely on the Glib module - you should load the
379 event module before loading AnyEvent or any module that uses it: 860 event module before loading AnyEvent or any module that uses it:
380 generally speaking, you should load it as early as possible. The reason 861 generally speaking, you should load it as early as possible. The reason
381 is that modules might create watchers when they are loaded, and AnyEvent 862 is that modules might create watchers when they are loaded, and AnyEvent
382 will decide on the event model to use as soon as it creates watchers, 863 will decide on the event model to use as soon as it creates watchers,
383 and it might chose the wrong one unless you load the correct one 864 and it might chose the wrong one unless you load the correct one
384 yourself. 865 yourself.
385 866
386 You can chose to use a rather inefficient pure-perl implementation by 867 You can chose to use a pure-perl implementation by loading the
387 loading the "AnyEvent::Impl::Perl" module, which gives you similar 868 "AnyEvent::Impl::Perl" module, which gives you similar behaviour
388 behaviour everywhere, but letting AnyEvent chose is generally better. 869 everywhere, but letting AnyEvent chose the model is generally better.
870
871 MAINLOOP EMULATION
872 Sometimes (often for short test scripts, or even standalone programs who
873 only want to use AnyEvent), you do not want to run a specific event
874 loop.
875
876 In that case, you can use a condition variable like this:
877
878 AnyEvent->condvar->recv;
879
880 This has the effect of entering the event loop and looping forever.
881
882 Note that usually your program has some exit condition, in which case it
883 is better to use the "traditional" approach of storing a condition
884 variable somewhere, waiting for it, and sending it when the program
885 should exit cleanly.
886
887OTHER MODULES
888 The following is a non-exhaustive list of additional modules that use
889 AnyEvent as a client and can therefore be mixed easily with other
890 AnyEvent modules and other event loops in the same program. Some of the
891 modules come with AnyEvent, most are available via CPAN.
892
893 AnyEvent::Util
894 Contains various utility functions that replace often-used but
895 blocking functions such as "inet_aton" by event-/callback-based
896 versions.
897
898 AnyEvent::Socket
899 Provides various utility functions for (internet protocol) sockets,
900 addresses and name resolution. Also functions to create non-blocking
901 tcp connections or tcp servers, with IPv6 and SRV record support and
902 more.
903
904 AnyEvent::Handle
905 Provide read and write buffers, manages watchers for reads and
906 writes, supports raw and formatted I/O, I/O queued and fully
907 transparent and non-blocking SSL/TLS (via AnyEvent::TLS.
908
909 AnyEvent::DNS
910 Provides rich asynchronous DNS resolver capabilities.
911
912 AnyEvent::HTTP
913 A simple-to-use HTTP library that is capable of making a lot of
914 concurrent HTTP requests.
915
916 AnyEvent::HTTPD
917 Provides a simple web application server framework.
918
919 AnyEvent::FastPing
920 The fastest ping in the west.
921
922 AnyEvent::DBI
923 Executes DBI requests asynchronously in a proxy process.
924
925 AnyEvent::AIO
926 Truly asynchronous I/O, should be in the toolbox of every event
927 programmer. AnyEvent::AIO transparently fuses IO::AIO and AnyEvent
928 together.
929
930 AnyEvent::BDB
931 Truly asynchronous Berkeley DB access. AnyEvent::BDB transparently
932 fuses BDB and AnyEvent together.
933
934 AnyEvent::GPSD
935 A non-blocking interface to gpsd, a daemon delivering GPS
936 information.
937
938 AnyEvent::IRC
939 AnyEvent based IRC client module family (replacing the older
940 Net::IRC3).
941
942 AnyEvent::XMPP
943 AnyEvent based XMPP (Jabber protocol) module family (replacing the
944 older Net::XMPP2>.
945
946 AnyEvent::IGS
947 A non-blocking interface to the Internet Go Server protocol (used by
948 App::IGS).
949
950 Net::FCP
951 AnyEvent-based implementation of the Freenet Client Protocol,
952 birthplace of AnyEvent.
953
954 Event::ExecFlow
955 High level API for event-based execution flow control.
956
957 Coro
958 Has special support for AnyEvent via Coro::AnyEvent.
959
960ERROR AND EXCEPTION HANDLING
961 In general, AnyEvent does not do any error handling - it relies on the
962 caller to do that if required. The AnyEvent::Strict module (see also the
963 "PERL_ANYEVENT_STRICT" environment variable, below) provides strict
964 checking of all AnyEvent methods, however, which is highly useful during
965 development.
966
967 As for exception handling (i.e. runtime errors and exceptions thrown
968 while executing a callback), this is not only highly event-loop
969 specific, but also not in any way wrapped by this module, as this is the
970 job of the main program.
971
972 The pure perl event loop simply re-throws the exception (usually within
973 "condvar->recv"), the Event and EV modules call "$Event/EV::DIED->()",
974 Glib uses "install_exception_handler" and so on.
975
976ENVIRONMENT VARIABLES
977 The following environment variables are used by this module or its
978 submodules.
979
980 Note that AnyEvent will remove *all* environment variables starting with
981 "PERL_ANYEVENT_" from %ENV when it is loaded while taint mode is
982 enabled.
983
984 "PERL_ANYEVENT_VERBOSE"
985 By default, AnyEvent will be completely silent except in fatal
986 conditions. You can set this environment variable to make AnyEvent
987 more talkative.
988
989 When set to 1 or higher, causes AnyEvent to warn about unexpected
990 conditions, such as not being able to load the event model specified
991 by "PERL_ANYEVENT_MODEL".
992
993 When set to 2 or higher, cause AnyEvent to report to STDERR which
994 event model it chooses.
995
996 "PERL_ANYEVENT_STRICT"
997 AnyEvent does not do much argument checking by default, as thorough
998 argument checking is very costly. Setting this variable to a true
999 value will cause AnyEvent to load "AnyEvent::Strict" and then to
1000 thoroughly check the arguments passed to most method calls. If it
1001 finds any problems, it will croak.
1002
1003 In other words, enables "strict" mode.
1004
1005 Unlike "use strict", it is definitely recommended to keep it off in
1006 production. Keeping "PERL_ANYEVENT_STRICT=1" in your environment
1007 while developing programs can be very useful, however.
1008
1009 "PERL_ANYEVENT_MODEL"
1010 This can be used to specify the event model to be used by AnyEvent,
1011 before auto detection and -probing kicks in. It must be a string
1012 consisting entirely of ASCII letters. The string "AnyEvent::Impl::"
1013 gets prepended and the resulting module name is loaded and if the
1014 load was successful, used as event model. If it fails to load
1015 AnyEvent will proceed with auto detection and -probing.
1016
1017 This functionality might change in future versions.
1018
1019 For example, to force the pure perl model (AnyEvent::Impl::Perl) you
1020 could start your program like this:
1021
1022 PERL_ANYEVENT_MODEL=Perl perl ...
1023
1024 "PERL_ANYEVENT_PROTOCOLS"
1025 Used by both AnyEvent::DNS and AnyEvent::Socket to determine
1026 preferences for IPv4 or IPv6. The default is unspecified (and might
1027 change, or be the result of auto probing).
1028
1029 Must be set to a comma-separated list of protocols or address
1030 families, current supported: "ipv4" and "ipv6". Only protocols
1031 mentioned will be used, and preference will be given to protocols
1032 mentioned earlier in the list.
1033
1034 This variable can effectively be used for denial-of-service attacks
1035 against local programs (e.g. when setuid), although the impact is
1036 likely small, as the program has to handle conenction and other
1037 failures anyways.
1038
1039 Examples: "PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6" - prefer IPv4 over
1040 IPv6, but support both and try to use both.
1041 "PERL_ANYEVENT_PROTOCOLS=ipv4" - only support IPv4, never try to
1042 resolve or contact IPv6 addresses.
1043 "PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4" support either IPv4 or IPv6, but
1044 prefer IPv6 over IPv4.
1045
1046 "PERL_ANYEVENT_EDNS0"
1047 Used by AnyEvent::DNS to decide whether to use the EDNS0 extension
1048 for DNS. This extension is generally useful to reduce DNS traffic,
1049 but some (broken) firewalls drop such DNS packets, which is why it
1050 is off by default.
1051
1052 Setting this variable to 1 will cause AnyEvent::DNS to announce
1053 EDNS0 in its DNS requests.
1054
1055 "PERL_ANYEVENT_MAX_FORKS"
1056 The maximum number of child processes that
1057 "AnyEvent::Util::fork_call" will create in parallel.
1058
1059 "PERL_ANYEVENT_MAX_OUTSTANDING_DNS"
1060 The default value for the "max_outstanding" parameter for the
1061 default DNS resolver - this is the maximum number of parallel DNS
1062 requests that are sent to the DNS server.
1063
1064 "PERL_ANYEVENT_RESOLV_CONF"
1065 The file to use instead of /etc/resolv.conf (or OS-specific
1066 configuration) in the default resolver. When set to the empty
1067 string, no default config will be used.
1068
1069 "PERL_ANYEVENT_CA_FILE", "PERL_ANYEVENT_CA_PATH".
1070 When neither "ca_file" nor "ca_path" was specified during
1071 AnyEvent::TLS context creation, and either of these environment
1072 variables exist, they will be used to specify CA certificate
1073 locations instead of a system-dependent default.
389 1074
390SUPPLYING YOUR OWN EVENT MODEL INTERFACE 1075SUPPLYING YOUR OWN EVENT MODEL INTERFACE
391 This is an advanced topic that you do not normally need to use AnyEvent 1076 This is an advanced topic that you do not normally need to use AnyEvent
392 in a module. This section is only of use to event loop authors who want 1077 in a module. This section is only of use to event loop authors who want
393 to provide AnyEvent compatibility. 1078 to provide AnyEvent compatibility.
… …
428 *rxvt-unicode* also cheats a bit by not providing blocking access to 1113 *rxvt-unicode* also cheats a bit by not providing blocking access to
429 condition variables: code blocking while waiting for a condition will 1114 condition variables: code blocking while waiting for a condition will
430 "die". This still works with most modules/usages, and blocking calls 1115 "die". This still works with most modules/usages, and blocking calls
431 must not be done in an interactive application, so it makes sense. 1116 must not be done in an interactive application, so it makes sense.
432 1117
433ENVIRONMENT VARIABLES
434 The following environment variables are used by this module:
435
436 "PERL_ANYEVENT_VERBOSE" when set to 2 or higher, cause AnyEvent to
437 report to STDERR which event model it chooses.
438
439EXAMPLE PROGRAM 1118EXAMPLE PROGRAM
440 The following program uses an IO watcher to read data from STDIN, a 1119 The following program uses an I/O watcher to read data from STDIN, a
441 timer to display a message once per second, and a condition variable to 1120 timer to display a message once per second, and a condition variable to
442 quit the program when the user enters quit: 1121 quit the program when the user enters quit:
443 1122
444 use AnyEvent; 1123 use AnyEvent;
445 1124
… …
450 poll => 'r', 1129 poll => 'r',
451 cb => sub { 1130 cb => sub {
452 warn "io event <$_[0]>\n"; # will always output <r> 1131 warn "io event <$_[0]>\n"; # will always output <r>
453 chomp (my $input = <STDIN>); # read a line 1132 chomp (my $input = <STDIN>); # read a line
454 warn "read: $input\n"; # output what has been read 1133 warn "read: $input\n"; # output what has been read
455 $cv->broadcast if $input =~ /^q/i; # quit program if /^q/i 1134 $cv->send if $input =~ /^q/i; # quit program if /^q/i
456 }, 1135 },
457 ); 1136 );
458 1137
459 my $time_watcher; # can only be used once 1138 my $time_watcher; # can only be used once
460 1139
… …
465 }); 1144 });
466 } 1145 }
467 1146
468 new_timer; # create first timer 1147 new_timer; # create first timer
469 1148
470 $cv->wait; # wait until user enters /^q/i 1149 $cv->recv; # wait until user enters /^q/i
471 1150
472REAL-WORLD EXAMPLE 1151REAL-WORLD EXAMPLE
473 Consider the Net::FCP module. It features (among others) the following 1152 Consider the Net::FCP module. It features (among others) the following
474 API calls, which are to freenet what HTTP GET requests are to http: 1153 API calls, which are to freenet what HTTP GET requests are to http:
475 1154
… …
524 syswrite $txn->{fh}, $txn->{request} 1203 syswrite $txn->{fh}, $txn->{request}
525 or die "connection or write error"; 1204 or die "connection or write error";
526 $txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r }); 1205 $txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r });
527 1206
528 Again, "fh_ready_r" waits till all data has arrived, and then stores the 1207 Again, "fh_ready_r" waits till all data has arrived, and then stores the
529 result and signals any possible waiters that the request ahs finished: 1208 result and signals any possible waiters that the request has finished:
530 1209
531 sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf}; 1210 sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf};
532 1211
533 if (end-of-file or data complete) { 1212 if (end-of-file or data complete) {
534 $txn->{result} = $txn->{buf}; 1213 $txn->{result} = $txn->{buf};
535 $txn->{finished}->broadcast; 1214 $txn->{finished}->send;
536 $txb->{cb}->($txn) of $txn->{cb}; # also call callback 1215 $txb->{cb}->($txn) of $txn->{cb}; # also call callback
537 } 1216 }
538 1217
539 The "result" method, finally, just waits for the finished signal (if the 1218 The "result" method, finally, just waits for the finished signal (if the
540 request was already finished, it doesn't wait, of course, and returns 1219 request was already finished, it doesn't wait, of course, and returns
541 the data: 1220 the data:
542 1221
543 $txn->{finished}->wait; 1222 $txn->{finished}->recv;
544 return $txn->{result}; 1223 return $txn->{result};
545 1224
546 The actual code goes further and collects all errors ("die"s, 1225 The actual code goes further and collects all errors ("die"s,
547 exceptions) that occured during request processing. The "result" method 1226 exceptions) that occurred during request processing. The "result" method
548 detects whether an exception as thrown (it is stored inside the $txn 1227 detects whether an exception as thrown (it is stored inside the $txn
549 object) and just throws the exception, which means connection errors and 1228 object) and just throws the exception, which means connection errors and
550 other problems get reported tot he code that tries to use the result, 1229 other problems get reported tot he code that tries to use the result,
551 not in a random callback. 1230 not in a random callback.
552 1231
… …
583 1262
584 my $quit = AnyEvent->condvar; 1263 my $quit = AnyEvent->condvar;
585 1264
586 $fcp->txn_client_get ($url)->cb (sub { 1265 $fcp->txn_client_get ($url)->cb (sub {
587 ... 1266 ...
588 $quit->broadcast; 1267 $quit->send;
589 }); 1268 });
590 1269
591 $quit->wait; 1270 $quit->recv;
1271
1272BENCHMARKS
1273 To give you an idea of the performance and overheads that AnyEvent adds
1274 over the event loops themselves and to give you an impression of the
1275 speed of various event loops I prepared some benchmarks.
1276
1277 BENCHMARKING ANYEVENT OVERHEAD
1278 Here is a benchmark of various supported event models used natively and
1279 through AnyEvent. The benchmark creates a lot of timers (with a zero
1280 timeout) and I/O watchers (watching STDOUT, a pty, to become writable,
1281 which it is), lets them fire exactly once and destroys them again.
1282
1283 Source code for this benchmark is found as eg/bench in the AnyEvent
1284 distribution.
1285
1286 Explanation of the columns
1287 *watcher* is the number of event watchers created/destroyed. Since
1288 different event models feature vastly different performances, each event
1289 loop was given a number of watchers so that overall runtime is
1290 acceptable and similar between tested event loop (and keep them from
1291 crashing): Glib would probably take thousands of years if asked to
1292 process the same number of watchers as EV in this benchmark.
1293
1294 *bytes* is the number of bytes (as measured by the resident set size,
1295 RSS) consumed by each watcher. This method of measuring captures both C
1296 and Perl-based overheads.
1297
1298 *create* is the time, in microseconds (millionths of seconds), that it
1299 takes to create a single watcher. The callback is a closure shared
1300 between all watchers, to avoid adding memory overhead. That means
1301 closure creation and memory usage is not included in the figures.
1302
1303 *invoke* is the time, in microseconds, used to invoke a simple callback.
1304 The callback simply counts down a Perl variable and after it was invoked
1305 "watcher" times, it would "->send" a condvar once to signal the end of
1306 this phase.
1307
1308 *destroy* is the time, in microseconds, that it takes to destroy a
1309 single watcher.
1310
1311 Results
1312 name watchers bytes create invoke destroy comment
1313 EV/EV 400000 224 0.47 0.35 0.27 EV native interface
1314 EV/Any 100000 224 2.88 0.34 0.27 EV + AnyEvent watchers
1315 CoroEV/Any 100000 224 2.85 0.35 0.28 coroutines + Coro::Signal
1316 Perl/Any 100000 452 4.13 0.73 0.95 pure perl implementation
1317 Event/Event 16000 517 32.20 31.80 0.81 Event native interface
1318 Event/Any 16000 590 35.85 31.55 1.06 Event + AnyEvent watchers
1319 IOAsync/Any 16000 989 38.10 32.77 11.13 via IO::Async::Loop::IO_Poll
1320 IOAsync/Any 16000 990 37.59 29.50 10.61 via IO::Async::Loop::Epoll
1321 Glib/Any 16000 1357 102.33 12.31 51.00 quadratic behaviour
1322 Tk/Any 2000 1860 27.20 66.31 14.00 SEGV with >> 2000 watchers
1323 POE/Event 2000 6328 109.99 751.67 14.02 via POE::Loop::Event
1324 POE/Select 2000 6027 94.54 809.13 579.80 via POE::Loop::Select
1325
1326 Discussion
1327 The benchmark does *not* measure scalability of the event loop very
1328 well. For example, a select-based event loop (such as the pure perl one)
1329 can never compete with an event loop that uses epoll when the number of
1330 file descriptors grows high. In this benchmark, all events become ready
1331 at the same time, so select/poll-based implementations get an unnatural
1332 speed boost.
1333
1334 Also, note that the number of watchers usually has a nonlinear effect on
1335 overall speed, that is, creating twice as many watchers doesn't take
1336 twice the time - usually it takes longer. This puts event loops tested
1337 with a higher number of watchers at a disadvantage.
1338
1339 To put the range of results into perspective, consider that on the
1340 benchmark machine, handling an event takes roughly 1600 CPU cycles with
1341 EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000
1342 CPU cycles with POE.
1343
1344 "EV" is the sole leader regarding speed and memory use, which are both
1345 maximal/minimal, respectively. Even when going through AnyEvent, it uses
1346 far less memory than any other event loop and is still faster than Event
1347 natively.
1348
1349 The pure perl implementation is hit in a few sweet spots (both the
1350 constant timeout and the use of a single fd hit optimisations in the
1351 perl interpreter and the backend itself). Nevertheless this shows that
1352 it adds very little overhead in itself. Like any select-based backend
1353 its performance becomes really bad with lots of file descriptors (and
1354 few of them active), of course, but this was not subject of this
1355 benchmark.
1356
1357 The "Event" module has a relatively high setup and callback invocation
1358 cost, but overall scores in on the third place.
1359
1360 "IO::Async" performs admirably well, about on par with "Event", even
1361 when using its pure perl backend.
1362
1363 "Glib"'s memory usage is quite a bit higher, but it features a faster
1364 callback invocation and overall ends up in the same class as "Event".
1365 However, Glib scales extremely badly, doubling the number of watchers
1366 increases the processing time by more than a factor of four, making it
1367 completely unusable when using larger numbers of watchers (note that
1368 only a single file descriptor was used in the benchmark, so
1369 inefficiencies of "poll" do not account for this).
1370
1371 The "Tk" adaptor works relatively well. The fact that it crashes with
1372 more than 2000 watchers is a big setback, however, as correctness takes
1373 precedence over speed. Nevertheless, its performance is surprising, as
1374 the file descriptor is dup()ed for each watcher. This shows that the
1375 dup() employed by some adaptors is not a big performance issue (it does
1376 incur a hidden memory cost inside the kernel which is not reflected in
1377 the figures above).
1378
1379 "POE", regardless of underlying event loop (whether using its pure perl
1380 select-based backend or the Event module, the POE-EV backend couldn't be
1381 tested because it wasn't working) shows abysmal performance and memory
1382 usage with AnyEvent: Watchers use almost 30 times as much memory as EV
1383 watchers, and 10 times as much memory as Event (the high memory
1384 requirements are caused by requiring a session for each watcher).
1385 Watcher invocation speed is almost 900 times slower than with AnyEvent's
1386 pure perl implementation.
1387
1388 The design of the POE adaptor class in AnyEvent can not really account
1389 for the performance issues, though, as session creation overhead is
1390 small compared to execution of the state machine, which is coded pretty
1391 optimally within AnyEvent::Impl::POE (and while everybody agrees that
1392 using multiple sessions is not a good approach, especially regarding
1393 memory usage, even the author of POE could not come up with a faster
1394 design).
1395
1396 Summary
1397 * Using EV through AnyEvent is faster than any other event loop (even
1398 when used without AnyEvent), but most event loops have acceptable
1399 performance with or without AnyEvent.
1400
1401 * The overhead AnyEvent adds is usually much smaller than the overhead
1402 of the actual event loop, only with extremely fast event loops such
1403 as EV adds AnyEvent significant overhead.
1404
1405 * You should avoid POE like the plague if you want performance or
1406 reasonable memory usage.
1407
1408 BENCHMARKING THE LARGE SERVER CASE
1409 This benchmark actually benchmarks the event loop itself. It works by
1410 creating a number of "servers": each server consists of a socket pair, a
1411 timeout watcher that gets reset on activity (but never fires), and an
1412 I/O watcher waiting for input on one side of the socket. Each time the
1413 socket watcher reads a byte it will write that byte to a random other
1414 "server".
1415
1416 The effect is that there will be a lot of I/O watchers, only part of
1417 which are active at any one point (so there is a constant number of
1418 active fds for each loop iteration, but which fds these are is random).
1419 The timeout is reset each time something is read because that reflects
1420 how most timeouts work (and puts extra pressure on the event loops).
1421
1422 In this benchmark, we use 10000 socket pairs (20000 sockets), of which
1423 100 (1%) are active. This mirrors the activity of large servers with
1424 many connections, most of which are idle at any one point in time.
1425
1426 Source code for this benchmark is found as eg/bench2 in the AnyEvent
1427 distribution.
1428
1429 Explanation of the columns
1430 *sockets* is the number of sockets, and twice the number of "servers"
1431 (as each server has a read and write socket end).
1432
1433 *create* is the time it takes to create a socket pair (which is
1434 nontrivial) and two watchers: an I/O watcher and a timeout watcher.
1435
1436 *request*, the most important value, is the time it takes to handle a
1437 single "request", that is, reading the token from the pipe and
1438 forwarding it to another server. This includes deleting the old timeout
1439 and creating a new one that moves the timeout into the future.
1440
1441 Results
1442 name sockets create request
1443 EV 20000 69.01 11.16
1444 Perl 20000 73.32 35.87
1445 IOAsync 20000 157.00 98.14 epoll
1446 IOAsync 20000 159.31 616.06 poll
1447 Event 20000 212.62 257.32
1448 Glib 20000 651.16 1896.30
1449 POE 20000 349.67 12317.24 uses POE::Loop::Event
1450
1451 Discussion
1452 This benchmark *does* measure scalability and overall performance of the
1453 particular event loop.
1454
1455 EV is again fastest. Since it is using epoll on my system, the setup
1456 time is relatively high, though.
1457
1458 Perl surprisingly comes second. It is much faster than the C-based event
1459 loops Event and Glib.
1460
1461 IO::Async performs very well when using its epoll backend, and still
1462 quite good compared to Glib when using its pure perl backend.
1463
1464 Event suffers from high setup time as well (look at its code and you
1465 will understand why). Callback invocation also has a high overhead
1466 compared to the "$_->() for .."-style loop that the Perl event loop
1467 uses. Event uses select or poll in basically all documented
1468 configurations.
1469
1470 Glib is hit hard by its quadratic behaviour w.r.t. many watchers. It
1471 clearly fails to perform with many filehandles or in busy servers.
1472
1473 POE is still completely out of the picture, taking over 1000 times as
1474 long as EV, and over 100 times as long as the Perl implementation, even
1475 though it uses a C-based event loop in this case.
1476
1477 Summary
1478 * The pure perl implementation performs extremely well.
1479
1480 * Avoid Glib or POE in large projects where performance matters.
1481
1482 BENCHMARKING SMALL SERVERS
1483 While event loops should scale (and select-based ones do not...) even to
1484 large servers, most programs we (or I :) actually write have only a few
1485 I/O watchers.
1486
1487 In this benchmark, I use the same benchmark program as in the large
1488 server case, but it uses only eight "servers", of which three are active
1489 at any one time. This should reflect performance for a small server
1490 relatively well.
1491
1492 The columns are identical to the previous table.
1493
1494 Results
1495 name sockets create request
1496 EV 16 20.00 6.54
1497 Perl 16 25.75 12.62
1498 Event 16 81.27 35.86
1499 Glib 16 32.63 15.48
1500 POE 16 261.87 276.28 uses POE::Loop::Event
1501
1502 Discussion
1503 The benchmark tries to test the performance of a typical small server.
1504 While knowing how various event loops perform is interesting, keep in
1505 mind that their overhead in this case is usually not as important, due
1506 to the small absolute number of watchers (that is, you need efficiency
1507 and speed most when you have lots of watchers, not when you only have a
1508 few of them).
1509
1510 EV is again fastest.
1511
1512 Perl again comes second. It is noticeably faster than the C-based event
1513 loops Event and Glib, although the difference is too small to really
1514 matter.
1515
1516 POE also performs much better in this case, but is is still far behind
1517 the others.
1518
1519 Summary
1520 * C-based event loops perform very well with small number of watchers,
1521 as the management overhead dominates.
1522
1523 THE IO::Lambda BENCHMARK
1524 Recently I was told about the benchmark in the IO::Lambda manpage, which
1525 could be misinterpreted to make AnyEvent look bad. In fact, the
1526 benchmark simply compares IO::Lambda with POE, and IO::Lambda looks
1527 better (which shouldn't come as a surprise to anybody). As such, the
1528 benchmark is fine, and mostly shows that the AnyEvent backend from
1529 IO::Lambda isn't very optimal. But how would AnyEvent compare when used
1530 without the extra baggage? To explore this, I wrote the equivalent
1531 benchmark for AnyEvent.
1532
1533 The benchmark itself creates an echo-server, and then, for 500 times,
1534 connects to the echo server, sends a line, waits for the reply, and then
1535 creates the next connection. This is a rather bad benchmark, as it
1536 doesn't test the efficiency of the framework or much non-blocking I/O,
1537 but it is a benchmark nevertheless.
1538
1539 name runtime
1540 Lambda/select 0.330 sec
1541 + optimized 0.122 sec
1542 Lambda/AnyEvent 0.327 sec
1543 + optimized 0.138 sec
1544 Raw sockets/select 0.077 sec
1545 POE/select, components 0.662 sec
1546 POE/select, raw sockets 0.226 sec
1547 POE/select, optimized 0.404 sec
1548
1549 AnyEvent/select/nb 0.085 sec
1550 AnyEvent/EV/nb 0.068 sec
1551 +state machine 0.134 sec
1552
1553 The benchmark is also a bit unfair (my fault): the IO::Lambda/POE
1554 benchmarks actually make blocking connects and use 100% blocking I/O,
1555 defeating the purpose of an event-based solution. All of the newly
1556 written AnyEvent benchmarks use 100% non-blocking connects (using
1557 AnyEvent::Socket::tcp_connect and the asynchronous pure perl DNS
1558 resolver), so AnyEvent is at a disadvantage here, as non-blocking
1559 connects generally require a lot more bookkeeping and event handling
1560 than blocking connects (which involve a single syscall only).
1561
1562 The last AnyEvent benchmark additionally uses AnyEvent::Handle, which
1563 offers similar expressive power as POE and IO::Lambda, using
1564 conventional Perl syntax. This means that both the echo server and the
1565 client are 100% non-blocking, further placing it at a disadvantage.
1566
1567 As you can see, the AnyEvent + EV combination even beats the
1568 hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl
1569 backend easily beats IO::Lambda and POE.
1570
1571 And even the 100% non-blocking version written using the high-level (and
1572 slow :) AnyEvent::Handle abstraction beats both POE and IO::Lambda by a
1573 large margin, even though it does all of DNS, tcp-connect and socket I/O
1574 in a non-blocking way.
1575
1576 The two AnyEvent benchmarks programs can be found as eg/ae0.pl and
1577 eg/ae2.pl in the AnyEvent distribution, the remaining benchmarks are
1578 part of the IO::lambda distribution and were used without any changes.
1579
1580SIGNALS
1581 AnyEvent currently installs handlers for these signals:
1582
1583 SIGCHLD
1584 A handler for "SIGCHLD" is installed by AnyEvent's child watcher
1585 emulation for event loops that do not support them natively. Also,
1586 some event loops install a similar handler.
1587
1588 Additionally, when AnyEvent is loaded and SIGCHLD is set to IGNORE,
1589 then AnyEvent will reset it to default, to avoid losing child exit
1590 statuses.
1591
1592 SIGPIPE
1593 A no-op handler is installed for "SIGPIPE" when $SIG{PIPE} is
1594 "undef" when AnyEvent gets loaded.
1595
1596 The rationale for this is that AnyEvent users usually do not really
1597 depend on SIGPIPE delivery (which is purely an optimisation for
1598 shell use, or badly-written programs), but "SIGPIPE" can cause
1599 spurious and rare program exits as a lot of people do not expect
1600 "SIGPIPE" when writing to some random socket.
1601
1602 The rationale for installing a no-op handler as opposed to ignoring
1603 it is that this way, the handler will be restored to defaults on
1604 exec.
1605
1606 Feel free to install your own handler, or reset it to defaults.
1607
1608FORK
1609 Most event libraries are not fork-safe. The ones who are usually are
1610 because they rely on inefficient but fork-safe "select" or "poll" calls.
1611 Only EV is fully fork-aware.
1612
1613 If you have to fork, you must either do so *before* creating your first
1614 watcher OR you must not use AnyEvent at all in the child.
1615
1616SECURITY CONSIDERATIONS
1617 AnyEvent can be forced to load any event model via
1618 $ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used
1619 to execute arbitrary code or directly gain access, it can easily be used
1620 to make the program hang or malfunction in subtle ways, as AnyEvent
1621 watchers will not be active when the program uses a different event
1622 model than specified in the variable.
1623
1624 You can make AnyEvent completely ignore this variable by deleting it
1625 before the first watcher gets created, e.g. with a "BEGIN" block:
1626
1627 BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} }
1628
1629 use AnyEvent;
1630
1631 Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can
1632 be used to probe what backend is used and gain other information (which
1633 is probably even less useful to an attacker than PERL_ANYEVENT_MODEL),
1634 and $ENV{PERL_ANYEVENT_STRICT}.
1635
1636 Note that AnyEvent will remove *all* environment variables starting with
1637 "PERL_ANYEVENT_" from %ENV when it is loaded while taint mode is
1638 enabled.
1639
1640BUGS
1641 Perl 5.8 has numerous memleaks that sometimes hit this module and are
1642 hard to work around. If you suffer from memleaks, first upgrade to Perl
1643 5.10 and check wether the leaks still show up. (Perl 5.10.0 has other
1644 annoying memleaks, such as leaking on "map" and "grep" but it is usually
1645 not as pronounced).
592 1646
593SEE ALSO 1647SEE ALSO
594 Event modules: Coro::EV, EV, EV::Glib, Glib::EV, Coro::Event, Event, 1648 Utility functions: AnyEvent::Util.
595 Glib::Event, Glib, Coro, Tk.
596 1649
597 Implementations: AnyEvent::Impl::CoroEV, AnyEvent::Impl::EV, 1650 Event modules: EV, EV::Glib, Glib::EV, Event, Glib::Event, Glib, Tk,
598 AnyEvent::Impl::CoroEvent, AnyEvent::Impl::Event, AnyEvent::Impl::Glib, 1651 Event::Lib, Qt, POE.
1652
1653 Implementations: AnyEvent::Impl::EV, AnyEvent::Impl::Event,
599 AnyEvent::Impl::Tk, AnyEvent::Impl::Perl. 1654 AnyEvent::Impl::Glib, AnyEvent::Impl::Tk, AnyEvent::Impl::Perl,
1655 AnyEvent::Impl::EventLib, AnyEvent::Impl::Qt, AnyEvent::Impl::POE,
1656 AnyEvent::Impl::IOAsync.
600 1657
1658 Non-blocking file handles, sockets, TCP clients and servers:
1659 AnyEvent::Handle, AnyEvent::Socket, AnyEvent::TLS.
1660
1661 Asynchronous DNS: AnyEvent::DNS.
1662
1663 Coroutine support: Coro, Coro::AnyEvent, Coro::EV, Coro::Event,
1664
601 Nontrivial usage examples: Net::FCP, Net::XMPP2. 1665 Nontrivial usage examples: AnyEvent::GPSD, AnyEvent::XMPP,
1666 AnyEvent::HTTP.
602 1667
1668AUTHOR
1669 Marc Lehmann <schmorp@schmorp.de>
1670 http://home.schmorp.de/
603 1671

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines