1 | NAME |
1 | NAME |
2 | AnyEvent - provide framework for multiple event loops |
2 | AnyEvent - provide framework for multiple event loops |
3 | |
3 | |
4 | Event, Coro, Glib, Tk, Perl - various supported event loops |
4 | EV, Event, Glib, Tk, Perl, Event::Lib, Qt, POE - various supported event |
|
|
5 | loops |
5 | |
6 | |
6 | SYNOPSIS |
7 | SYNOPSIS |
7 | use AnyEvent; |
8 | use AnyEvent; |
8 | |
9 | |
9 | my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub { |
10 | my $w = AnyEvent->io (fh => $fh, poll => "r|w", cb => sub { |
… | |
… | |
12 | |
13 | |
13 | my $w = AnyEvent->timer (after => $seconds, cb => sub { |
14 | my $w = AnyEvent->timer (after => $seconds, cb => sub { |
14 | ... |
15 | ... |
15 | }); |
16 | }); |
16 | |
17 | |
17 | my $w = AnyEvent->condvar; # stores wether a condition was flagged |
18 | my $w = AnyEvent->condvar; # stores whether a condition was flagged |
|
|
19 | $w->send; # wake up current and all future recv's |
18 | $w->wait; # enters "main loop" till $condvar gets ->broadcast |
20 | $w->recv; # enters "main loop" till $condvar gets ->send |
19 | $w->broadcast; # wake up current and all future wait's |
21 | |
|
|
22 | INTRODUCTION/TUTORIAL |
|
|
23 | This manpage is mainly a reference manual. If you are interested in a |
|
|
24 | tutorial or some gentle introduction, have a look at the AnyEvent::Intro |
|
|
25 | manpage. |
|
|
26 | |
|
|
27 | WHY YOU SHOULD USE THIS MODULE (OR NOT) |
|
|
28 | Glib, POE, IO::Async, Event... CPAN offers event models by the dozen |
|
|
29 | nowadays. So what is different about AnyEvent? |
|
|
30 | |
|
|
31 | Executive Summary: AnyEvent is *compatible*, AnyEvent is *free of |
|
|
32 | policy* and AnyEvent is *small and efficient*. |
|
|
33 | |
|
|
34 | First and foremost, *AnyEvent is not an event model* itself, it only |
|
|
35 | interfaces to whatever event model the main program happens to use, in a |
|
|
36 | pragmatic way. For event models and certain classes of immortals alike, |
|
|
37 | the statement "there can only be one" is a bitter reality: In general, |
|
|
38 | only one event loop can be active at the same time in a process. |
|
|
39 | AnyEvent cannot change this, but it can hide the differences between |
|
|
40 | those event loops. |
|
|
41 | |
|
|
42 | The goal of AnyEvent is to offer module authors the ability to do event |
|
|
43 | programming (waiting for I/O or timer events) without subscribing to a |
|
|
44 | religion, a way of living, and most importantly: without forcing your |
|
|
45 | module users into the same thing by forcing them to use the same event |
|
|
46 | model you use. |
|
|
47 | |
|
|
48 | For modules like POE or IO::Async (which is a total misnomer as it is |
|
|
49 | actually doing all I/O *synchronously*...), using them in your module is |
|
|
50 | like joining a cult: After you joined, you are dependent on them and you |
|
|
51 | cannot use anything else, as they are simply incompatible to everything |
|
|
52 | that isn't them. What's worse, all the potential users of your module |
|
|
53 | are *also* forced to use the same event loop you use. |
|
|
54 | |
|
|
55 | AnyEvent is different: AnyEvent + POE works fine. AnyEvent + Glib works |
|
|
56 | fine. AnyEvent + Tk works fine etc. etc. but none of these work together |
|
|
57 | with the rest: POE + IO::Async? No go. Tk + Event? No go. Again: if your |
|
|
58 | module uses one of those, every user of your module has to use it, too. |
|
|
59 | But if your module uses AnyEvent, it works transparently with all event |
|
|
60 | models it supports (including stuff like IO::Async, as long as those use |
|
|
61 | one of the supported event loops. It is trivial to add new event loops |
|
|
62 | to AnyEvent, too, so it is future-proof). |
|
|
63 | |
|
|
64 | In addition to being free of having to use *the one and only true event |
|
|
65 | model*, AnyEvent also is free of bloat and policy: with POE or similar |
|
|
66 | modules, you get an enormous amount of code and strict rules you have to |
|
|
67 | follow. AnyEvent, on the other hand, is lean and up to the point, by |
|
|
68 | only offering the functionality that is necessary, in as thin as a |
|
|
69 | wrapper as technically possible. |
|
|
70 | |
|
|
71 | Of course, AnyEvent comes with a big (and fully optional!) toolbox of |
|
|
72 | useful functionality, such as an asynchronous DNS resolver, 100% |
|
|
73 | non-blocking connects (even with TLS/SSL, IPv6 and on broken platforms |
|
|
74 | such as Windows) and lots of real-world knowledge and workarounds for |
|
|
75 | platform bugs and differences. |
|
|
76 | |
|
|
77 | Now, if you *do want* lots of policy (this can arguably be somewhat |
|
|
78 | useful) and you want to force your users to use the one and only event |
|
|
79 | model, you should *not* use this module. |
20 | |
80 | |
21 | DESCRIPTION |
81 | DESCRIPTION |
22 | AnyEvent provides an identical interface to multiple event loops. This |
82 | AnyEvent provides an identical interface to multiple event loops. This |
23 | allows module authors to utilise an event loop without forcing module |
83 | allows module authors to utilise an event loop without forcing module |
24 | users to use the same event loop (as only a single event loop can |
84 | users to use the same event loop (as only a single event loop can |
25 | coexist peacefully at any one time). |
85 | coexist peacefully at any one time). |
26 | |
86 | |
27 | The interface itself is vaguely similar but not identical to the Event |
87 | The interface itself is vaguely similar, but not identical to the Event |
28 | module. |
88 | module. |
29 | |
89 | |
30 | On the first call of any method, the module tries to detect the |
90 | During the first call of any watcher-creation method, the module tries |
31 | currently loaded event loop by probing wether any of the following |
91 | to detect the currently loaded event loop by probing whether one of the |
32 | modules is loaded: Coro::Event, Event, Glib, Tk. The first one found is |
92 | following modules is already loaded: EV, Event, Glib, |
|
|
93 | AnyEvent::Impl::Perl, Tk, Event::Lib, Qt, POE. The first one found is |
33 | used. If none is found, the module tries to load these modules in the |
94 | used. If none are found, the module tries to load these modules |
34 | order given. The first one that could be successfully loaded will be |
95 | (excluding Tk, Event::Lib, Qt and POE as the pure perl adaptor should |
35 | used. If still none could be found, AnyEvent will fall back to a |
96 | always succeed) in the order given. The first one that can be |
36 | pure-perl event loop, which is also not very efficient. |
97 | successfully loaded will be used. If, after this, still none could be |
|
|
98 | found, AnyEvent will fall back to a pure-perl event loop, which is not |
|
|
99 | very efficient, but should work everywhere. |
37 | |
100 | |
38 | Because AnyEvent first checks for modules that are already loaded, |
101 | Because AnyEvent first checks for modules that are already loaded, |
39 | loading an Event model explicitly before first using AnyEvent will |
102 | loading an event model explicitly before first using AnyEvent will |
40 | likely make that model the default. For example: |
103 | likely make that model the default. For example: |
41 | |
104 | |
42 | use Tk; |
105 | use Tk; |
43 | use AnyEvent; |
106 | use AnyEvent; |
44 | |
107 | |
45 | # .. AnyEvent will likely default to Tk |
108 | # .. AnyEvent will likely default to Tk |
46 | |
109 | |
|
|
110 | The *likely* means that, if any module loads another event model and |
|
|
111 | starts using it, all bets are off. Maybe you should tell their authors |
|
|
112 | to use AnyEvent so their modules work together with others seamlessly... |
|
|
113 | |
47 | The pure-perl implementation of AnyEvent is called |
114 | The pure-perl implementation of AnyEvent is called |
48 | "AnyEvent::Impl::Perl". Like other event modules you can load it |
115 | "AnyEvent::Impl::Perl". Like other event modules you can load it |
49 | explicitly. |
116 | explicitly and enjoy the high availability of that event loop :) |
50 | |
117 | |
51 | WATCHERS |
118 | WATCHERS |
52 | AnyEvent has the central concept of a *watcher*, which is an object that |
119 | AnyEvent has the central concept of a *watcher*, which is an object that |
53 | stores relevant data for each kind of event you are waiting for, such as |
120 | stores relevant data for each kind of event you are waiting for, such as |
54 | the callback to call, the filehandle to watch, etc. |
121 | the callback to call, the file handle to watch, etc. |
55 | |
122 | |
56 | These watchers are normal Perl objects with normal Perl lifetime. After |
123 | These watchers are normal Perl objects with normal Perl lifetime. After |
57 | creating a watcher it will immediately "watch" for events and invoke the |
124 | creating a watcher it will immediately "watch" for events and invoke the |
|
|
125 | callback when the event occurs (of course, only when the event model is |
|
|
126 | in control). |
|
|
127 | |
58 | callback. To disable the watcher you have to destroy it (e.g. by setting |
128 | To disable the watcher you have to destroy it (e.g. by setting the |
59 | the variable that stores it to "undef" or otherwise deleting all |
129 | variable you store it in to "undef" or otherwise deleting all references |
60 | references to it). |
130 | to it). |
61 | |
131 | |
62 | All watchers are created by calling a method on the "AnyEvent" class. |
132 | All watchers are created by calling a method on the "AnyEvent" class. |
63 | |
133 | |
|
|
134 | Many watchers either are used with "recursion" (repeating timers for |
|
|
135 | example), or need to refer to their watcher object in other ways. |
|
|
136 | |
|
|
137 | An any way to achieve that is this pattern: |
|
|
138 | |
|
|
139 | my $w; $w = AnyEvent->type (arg => value ..., cb => sub { |
|
|
140 | # you can use $w here, for example to undef it |
|
|
141 | undef $w; |
|
|
142 | }); |
|
|
143 | |
|
|
144 | Note that "my $w; $w =" combination. This is necessary because in Perl, |
|
|
145 | my variables are only visible after the statement in which they are |
|
|
146 | declared. |
|
|
147 | |
64 | IO WATCHERS |
148 | I/O WATCHERS |
65 | You can create I/O watcher by calling the "AnyEvent->io" method with the |
149 | You can create an I/O watcher by calling the "AnyEvent->io" method with |
66 | following mandatory arguments: |
150 | the following mandatory key-value pairs as arguments: |
67 | |
151 | |
68 | "fh" the Perl *filehandle* (not filedescriptor) to watch for events. |
152 | "fh" the Perl *file handle* (*not* file descriptor) to watch for events |
|
|
153 | (AnyEvent might or might not keep a reference to this file handle). |
69 | "poll" must be a string that is either "r" or "w", that creates a |
154 | "poll" must be a string that is either "r" or "w", which creates a |
70 | watcher waiting for "r"eadable or "w"ritable events. "cb" teh callback |
155 | watcher waiting for "r"eadable or "w"ritable events, respectively. "cb" |
71 | to invoke everytime the filehandle becomes ready. |
156 | is the callback to invoke each time the file handle becomes ready. |
72 | |
157 | |
73 | Only one io watcher per "fh" and "poll" combination is allowed (i.e. on |
158 | Although the callback might get passed parameters, their value and |
74 | a socket you can have one r + one w, not any more (limitation comes from |
159 | presence is undefined and you cannot rely on them. Portable AnyEvent |
75 | Tk - if you are sure you are not using Tk this limitation is gone). |
160 | callbacks cannot use arguments passed to I/O watcher callbacks. |
76 | |
161 | |
77 | Filehandles will be kept alive, so as long as the watcher exists, the |
162 | The I/O watcher might use the underlying file descriptor or a copy of |
78 | filehandle exists, too. |
163 | it. You must not close a file handle as long as any watcher is active on |
|
|
164 | the underlying file descriptor. |
79 | |
165 | |
80 | Example: |
166 | Some event loops issue spurious readyness notifications, so you should |
|
|
167 | always use non-blocking calls when reading/writing from/to your file |
|
|
168 | handles. |
81 | |
169 | |
82 | # wait for readability of STDIN, then read a line and disable the watcher |
170 | Example: wait for readability of STDIN, then read a line and disable the |
|
|
171 | watcher. |
|
|
172 | |
83 | my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub { |
173 | my $w; $w = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub { |
84 | chomp (my $input = <STDIN>); |
174 | chomp (my $input = <STDIN>); |
85 | warn "read: $input\n"; |
175 | warn "read: $input\n"; |
86 | undef $w; |
176 | undef $w; |
87 | }); |
177 | }); |
88 | |
178 | |
89 | TIME WATCHERS |
179 | TIME WATCHERS |
90 | You can create a time watcher by calling the "AnyEvent->timer" method |
180 | You can create a time watcher by calling the "AnyEvent->timer" method |
91 | with the following mandatory arguments: |
181 | with the following mandatory arguments: |
92 | |
182 | |
93 | "after" after how many seconds (fractions are supported) should the |
183 | "after" specifies after how many seconds (fractional values are |
94 | timer activate. "cb" the callback to invoke. |
184 | supported) the callback should be invoked. "cb" is the callback to |
|
|
185 | invoke in that case. |
95 | |
186 | |
96 | The timer callback will be invoked at most once: if you want a repeating |
187 | Although the callback might get passed parameters, their value and |
97 | timer you have to create a new watcher (this is a limitation by both Tk |
188 | presence is undefined and you cannot rely on them. Portable AnyEvent |
98 | and Glib). |
189 | callbacks cannot use arguments passed to time watcher callbacks. |
99 | |
190 | |
100 | Example: |
191 | The callback will normally be invoked once only. If you specify another |
|
|
192 | parameter, "interval", as a strictly positive number (> 0), then the |
|
|
193 | callback will be invoked regularly at that interval (in fractional |
|
|
194 | seconds) after the first invocation. If "interval" is specified with a |
|
|
195 | false value, then it is treated as if it were missing. |
101 | |
196 | |
|
|
197 | The callback will be rescheduled before invoking the callback, but no |
|
|
198 | attempt is done to avoid timer drift in most backends, so the interval |
|
|
199 | is only approximate. |
|
|
200 | |
102 | # fire an event after 7.7 seconds |
201 | Example: fire an event after 7.7 seconds. |
|
|
202 | |
103 | my $w = AnyEvent->timer (after => 7.7, cb => sub { |
203 | my $w = AnyEvent->timer (after => 7.7, cb => sub { |
104 | warn "timeout\n"; |
204 | warn "timeout\n"; |
105 | }); |
205 | }); |
106 | |
206 | |
107 | # to cancel the timer: |
207 | # to cancel the timer: |
108 | undef $w |
208 | undef $w; |
109 | |
209 | |
110 | CONDITION WATCHERS |
210 | Example 2: fire an event after 0.5 seconds, then roughly every second. |
111 | Condition watchers can be created by calling the "AnyEvent->condvar" |
|
|
112 | method without any arguments. |
|
|
113 | |
211 | |
114 | A condition watcher watches for a condition - precisely that the |
212 | my $w = AnyEvent->timer (after => 0.5, interval => 1, cb => sub { |
115 | "->broadcast" method has been called. |
213 | warn "timeout\n"; |
|
|
214 | }; |
116 | |
215 | |
117 | The watcher has only two methods: |
216 | TIMING ISSUES |
|
|
217 | There are two ways to handle timers: based on real time (relative, "fire |
|
|
218 | in 10 seconds") and based on wallclock time (absolute, "fire at 12 |
|
|
219 | o'clock"). |
118 | |
220 | |
119 | $cv->wait |
221 | While most event loops expect timers to specified in a relative way, |
120 | Wait (blocking if necessary) until the "->broadcast" method has been |
222 | they use absolute time internally. This makes a difference when your |
121 | called on c<$cv>, while servicing other watchers normally. |
223 | clock "jumps", for example, when ntp decides to set your clock backwards |
|
|
224 | from the wrong date of 2014-01-01 to 2008-01-01, a watcher that is |
|
|
225 | supposed to fire "after" a second might actually take six years to |
|
|
226 | finally fire. |
122 | |
227 | |
123 | Not all event models support a blocking wait - some die in that |
228 | AnyEvent cannot compensate for this. The only event loop that is |
124 | case, so if you are using this from a module, never require a |
229 | conscious about these issues is EV, which offers both relative |
125 | blocking wait, but let the caller decide wether the call will block |
230 | (ev_timer, based on true relative time) and absolute (ev_periodic, based |
126 | or not (for example, by coupling condition variables with some kind |
231 | on wallclock time) timers. |
127 | of request results and supporting callbacks so the caller knows that |
|
|
128 | getting the result will not block, while still suppporting blockign |
|
|
129 | waits if the caller so desires). |
|
|
130 | |
232 | |
131 | You can only wait once on a condition - additional calls will return |
233 | AnyEvent always prefers relative timers, if available, matching the |
132 | immediately. |
234 | AnyEvent API. |
133 | |
235 | |
134 | $cv->broadcast |
236 | AnyEvent has two additional methods that return the "current time": |
135 | Flag the condition as ready - a running "->wait" and all further |
|
|
136 | calls to "wait" will return after this method has been called. If |
|
|
137 | nobody is waiting the broadcast will be remembered.. |
|
|
138 | |
237 | |
139 | Example: |
238 | AnyEvent->time |
|
|
239 | This returns the "current wallclock time" as a fractional number of |
|
|
240 | seconds since the Epoch (the same thing as "time" or |
|
|
241 | "Time::HiRes::time" return, and the result is guaranteed to be |
|
|
242 | compatible with those). |
140 | |
243 | |
141 | # wait till the result is ready |
244 | It progresses independently of any event loop processing, i.e. each |
142 | my $result_ready = AnyEvent->condvar; |
245 | call will check the system clock, which usually gets updated |
|
|
246 | frequently. |
143 | |
247 | |
144 | # do something such as adding a timer |
248 | AnyEvent->now |
145 | # or socket watcher the calls $result_ready->broadcast |
249 | This also returns the "current wallclock time", but unlike "time", |
146 | # when the "result" is ready. |
250 | above, this value might change only once per event loop iteration, |
|
|
251 | depending on the event loop (most return the same time as "time", |
|
|
252 | above). This is the time that AnyEvent's timers get scheduled |
|
|
253 | against. |
147 | |
254 | |
148 | $result_ready->wait; |
255 | *In almost all cases (in all cases if you don't care), this is the |
|
|
256 | function to call when you want to know the current time.* |
|
|
257 | |
|
|
258 | This function is also often faster then "AnyEvent->time", and thus |
|
|
259 | the preferred method if you want some timestamp (for example, |
|
|
260 | AnyEvent::Handle uses this to update it's activity timeouts). |
|
|
261 | |
|
|
262 | The rest of this section is only of relevance if you try to be very |
|
|
263 | exact with your timing, you can skip it without bad conscience. |
|
|
264 | |
|
|
265 | For a practical example of when these times differ, consider |
|
|
266 | Event::Lib and EV and the following set-up: |
|
|
267 | |
|
|
268 | The event loop is running and has just invoked one of your callback |
|
|
269 | at time=500 (assume no other callbacks delay processing). In your |
|
|
270 | callback, you wait a second by executing "sleep 1" (blocking the |
|
|
271 | process for a second) and then (at time=501) you create a relative |
|
|
272 | timer that fires after three seconds. |
|
|
273 | |
|
|
274 | With Event::Lib, "AnyEvent->time" and "AnyEvent->now" will both |
|
|
275 | return 501, because that is the current time, and the timer will be |
|
|
276 | scheduled to fire at time=504 (501 + 3). |
|
|
277 | |
|
|
278 | With EV, "AnyEvent->time" returns 501 (as that is the current time), |
|
|
279 | but "AnyEvent->now" returns 500, as that is the time the last event |
|
|
280 | processing phase started. With EV, your timer gets scheduled to run |
|
|
281 | at time=503 (500 + 3). |
|
|
282 | |
|
|
283 | In one sense, Event::Lib is more exact, as it uses the current time |
|
|
284 | regardless of any delays introduced by event processing. However, |
|
|
285 | most callbacks do not expect large delays in processing, so this |
|
|
286 | causes a higher drift (and a lot more system calls to get the |
|
|
287 | current time). |
|
|
288 | |
|
|
289 | In another sense, EV is more exact, as your timer will be scheduled |
|
|
290 | at the same time, regardless of how long event processing actually |
|
|
291 | took. |
|
|
292 | |
|
|
293 | In either case, if you care (and in most cases, you don't), then you |
|
|
294 | can get whatever behaviour you want with any event loop, by taking |
|
|
295 | the difference between "AnyEvent->time" and "AnyEvent->now" into |
|
|
296 | account. |
149 | |
297 | |
150 | SIGNAL WATCHERS |
298 | SIGNAL WATCHERS |
151 | You can listen for signals using a signal watcher, "signal" is the |
299 | You can watch for signals using a signal watcher, "signal" is the signal |
152 | signal *name* without any "SIG" prefix. Multiple signals events can be |
300 | *name* in uppercase and without any "SIG" prefix, "cb" is the Perl |
153 | clumped together into one callback invocation, and callback invocation |
301 | callback to be invoked whenever a signal occurs. |
154 | might or might not be asynchronous. |
|
|
155 | |
302 | |
|
|
303 | Although the callback might get passed parameters, their value and |
|
|
304 | presence is undefined and you cannot rely on them. Portable AnyEvent |
|
|
305 | callbacks cannot use arguments passed to signal watcher callbacks. |
|
|
306 | |
|
|
307 | Multiple signal occurrences can be clumped together into one callback |
|
|
308 | invocation, and callback invocation will be synchronous. Synchronous |
|
|
309 | means that it might take a while until the signal gets handled by the |
|
|
310 | process, but it is guaranteed not to interrupt any other callbacks. |
|
|
311 | |
|
|
312 | The main advantage of using these watchers is that you can share a |
|
|
313 | signal between multiple watchers. |
|
|
314 | |
156 | These watchers might use %SIG, so programs overwriting those signals |
315 | This watcher might use %SIG, so programs overwriting those signals |
157 | directly will likely not work correctly. |
316 | directly will likely not work correctly. |
158 | |
317 | |
159 | Example: exit on SIGINT |
318 | Example: exit on SIGINT |
160 | |
319 | |
161 | my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 }); |
320 | my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 }); |
162 | |
321 | |
163 | CHILD PROCESS WATCHERS |
322 | CHILD PROCESS WATCHERS |
164 | You can also listen for the status of a child process specified by the |
323 | You can also watch on a child process exit and catch its exit status. |
165 | "pid" argument. The watcher will only trigger once. This works by |
|
|
166 | installing a signal handler for "SIGCHLD". |
|
|
167 | |
324 | |
|
|
325 | The child process is specified by the "pid" argument (if set to 0, it |
|
|
326 | watches for any child process exit). The watcher will trigger as often |
|
|
327 | as status change for the child are received. This works by installing a |
|
|
328 | signal handler for "SIGCHLD". The callback will be called with the pid |
|
|
329 | and exit status (as returned by waitpid), so unlike other watcher types, |
|
|
330 | you *can* rely on child watcher callback arguments. |
|
|
331 | |
|
|
332 | There is a slight catch to child watchers, however: you usually start |
|
|
333 | them *after* the child process was created, and this means the process |
|
|
334 | could have exited already (and no SIGCHLD will be sent anymore). |
|
|
335 | |
|
|
336 | Not all event models handle this correctly (POE doesn't), but even for |
|
|
337 | event models that *do* handle this correctly, they usually need to be |
|
|
338 | loaded before the process exits (i.e. before you fork in the first |
|
|
339 | place). |
|
|
340 | |
|
|
341 | This means you cannot create a child watcher as the very first thing in |
|
|
342 | an AnyEvent program, you *have* to create at least one watcher before |
|
|
343 | you "fork" the child (alternatively, you can call "AnyEvent::detect"). |
|
|
344 | |
|
|
345 | Example: fork a process and wait for it |
|
|
346 | |
|
|
347 | my $done = AnyEvent->condvar; |
|
|
348 | |
|
|
349 | my $pid = fork or exit 5; |
|
|
350 | |
|
|
351 | my $w = AnyEvent->child ( |
|
|
352 | pid => $pid, |
|
|
353 | cb => sub { |
|
|
354 | my ($pid, $status) = @_; |
|
|
355 | warn "pid $pid exited with status $status"; |
|
|
356 | $done->send; |
|
|
357 | }, |
|
|
358 | ); |
|
|
359 | |
|
|
360 | # do something else, then wait for process exit |
|
|
361 | $done->recv; |
|
|
362 | |
|
|
363 | CONDITION VARIABLES |
|
|
364 | If you are familiar with some event loops you will know that all of them |
|
|
365 | require you to run some blocking "loop", "run" or similar function that |
|
|
366 | will actively watch for new events and call your callbacks. |
|
|
367 | |
|
|
368 | AnyEvent is different, it expects somebody else to run the event loop |
|
|
369 | and will only block when necessary (usually when told by the user). |
|
|
370 | |
|
|
371 | The instrument to do that is called a "condition variable", so called |
|
|
372 | because they represent a condition that must become true. |
|
|
373 | |
|
|
374 | Condition variables can be created by calling the "AnyEvent->condvar" |
|
|
375 | method, usually without arguments. The only argument pair allowed is |
|
|
376 | "cb", which specifies a callback to be called when the condition |
|
|
377 | variable becomes true. |
|
|
378 | |
|
|
379 | After creation, the condition variable is "false" until it becomes |
|
|
380 | "true" by calling the "send" method (or calling the condition variable |
|
|
381 | as if it were a callback, read about the caveats in the description for |
|
|
382 | the "->send" method). |
|
|
383 | |
|
|
384 | Condition variables are similar to callbacks, except that you can |
|
|
385 | optionally wait for them. They can also be called merge points - points |
|
|
386 | in time where multiple outstanding events have been processed. And yet |
|
|
387 | another way to call them is transactions - each condition variable can |
|
|
388 | be used to represent a transaction, which finishes at some point and |
|
|
389 | delivers a result. |
|
|
390 | |
|
|
391 | Condition variables are very useful to signal that something has |
|
|
392 | finished, for example, if you write a module that does asynchronous http |
|
|
393 | requests, then a condition variable would be the ideal candidate to |
|
|
394 | signal the availability of results. The user can either act when the |
|
|
395 | callback is called or can synchronously "->recv" for the results. |
|
|
396 | |
|
|
397 | You can also use them to simulate traditional event loops - for example, |
|
|
398 | you can block your main program until an event occurs - for example, you |
|
|
399 | could "->recv" in your main program until the user clicks the Quit |
|
|
400 | button of your app, which would "->send" the "quit" event. |
|
|
401 | |
|
|
402 | Note that condition variables recurse into the event loop - if you have |
|
|
403 | two pieces of code that call "->recv" in a round-robin fashion, you |
|
|
404 | lose. Therefore, condition variables are good to export to your caller, |
|
|
405 | but you should avoid making a blocking wait yourself, at least in |
|
|
406 | callbacks, as this asks for trouble. |
|
|
407 | |
|
|
408 | Condition variables are represented by hash refs in perl, and the keys |
|
|
409 | used by AnyEvent itself are all named "_ae_XXX" to make subclassing easy |
|
|
410 | (it is often useful to build your own transaction class on top of |
|
|
411 | AnyEvent). To subclass, use "AnyEvent::CondVar" as base class and call |
|
|
412 | it's "new" method in your own "new" method. |
|
|
413 | |
|
|
414 | There are two "sides" to a condition variable - the "producer side" |
|
|
415 | which eventually calls "-> send", and the "consumer side", which waits |
|
|
416 | for the send to occur. |
|
|
417 | |
168 | Example: wait for pid 1333 |
418 | Example: wait for a timer. |
169 | |
419 | |
170 | my $w = AnyEvent->child (pid => 1333, cb => sub { warn "exit status $?" }); |
420 | # wait till the result is ready |
|
|
421 | my $result_ready = AnyEvent->condvar; |
171 | |
422 | |
172 | GLOBALS |
423 | # do something such as adding a timer |
|
|
424 | # or socket watcher the calls $result_ready->send |
|
|
425 | # when the "result" is ready. |
|
|
426 | # in this case, we simply use a timer: |
|
|
427 | my $w = AnyEvent->timer ( |
|
|
428 | after => 1, |
|
|
429 | cb => sub { $result_ready->send }, |
|
|
430 | ); |
|
|
431 | |
|
|
432 | # this "blocks" (while handling events) till the callback |
|
|
433 | # calls send |
|
|
434 | $result_ready->recv; |
|
|
435 | |
|
|
436 | Example: wait for a timer, but take advantage of the fact that condition |
|
|
437 | variables are also code references. |
|
|
438 | |
|
|
439 | my $done = AnyEvent->condvar; |
|
|
440 | my $delay = AnyEvent->timer (after => 5, cb => $done); |
|
|
441 | $done->recv; |
|
|
442 | |
|
|
443 | METHODS FOR PRODUCERS |
|
|
444 | These methods should only be used by the producing side, i.e. the |
|
|
445 | code/module that eventually sends the signal. Note that it is also the |
|
|
446 | producer side which creates the condvar in most cases, but it isn't |
|
|
447 | uncommon for the consumer to create it as well. |
|
|
448 | |
|
|
449 | $cv->send (...) |
|
|
450 | Flag the condition as ready - a running "->recv" and all further |
|
|
451 | calls to "recv" will (eventually) return after this method has been |
|
|
452 | called. If nobody is waiting the send will be remembered. |
|
|
453 | |
|
|
454 | If a callback has been set on the condition variable, it is called |
|
|
455 | immediately from within send. |
|
|
456 | |
|
|
457 | Any arguments passed to the "send" call will be returned by all |
|
|
458 | future "->recv" calls. |
|
|
459 | |
|
|
460 | Condition variables are overloaded so one can call them directly (as |
|
|
461 | a code reference). Calling them directly is the same as calling |
|
|
462 | "send". Note, however, that many C-based event loops do not handle |
|
|
463 | overloading, so as tempting as it may be, passing a condition |
|
|
464 | variable instead of a callback does not work. Both the pure perl and |
|
|
465 | EV loops support overloading, however, as well as all functions that |
|
|
466 | use perl to invoke a callback (as in AnyEvent::Socket and |
|
|
467 | AnyEvent::DNS for example). |
|
|
468 | |
|
|
469 | $cv->croak ($error) |
|
|
470 | Similar to send, but causes all call's to "->recv" to invoke |
|
|
471 | "Carp::croak" with the given error message/object/scalar. |
|
|
472 | |
|
|
473 | This can be used to signal any errors to the condition variable |
|
|
474 | user/consumer. |
|
|
475 | |
|
|
476 | $cv->begin ([group callback]) |
|
|
477 | $cv->end |
|
|
478 | These two methods are EXPERIMENTAL and MIGHT CHANGE. |
|
|
479 | |
|
|
480 | These two methods can be used to combine many transactions/events |
|
|
481 | into one. For example, a function that pings many hosts in parallel |
|
|
482 | might want to use a condition variable for the whole process. |
|
|
483 | |
|
|
484 | Every call to "->begin" will increment a counter, and every call to |
|
|
485 | "->end" will decrement it. If the counter reaches 0 in "->end", the |
|
|
486 | (last) callback passed to "begin" will be executed. That callback is |
|
|
487 | *supposed* to call "->send", but that is not required. If no |
|
|
488 | callback was set, "send" will be called without any arguments. |
|
|
489 | |
|
|
490 | Let's clarify this with the ping example: |
|
|
491 | |
|
|
492 | my $cv = AnyEvent->condvar; |
|
|
493 | |
|
|
494 | my %result; |
|
|
495 | $cv->begin (sub { $cv->send (\%result) }); |
|
|
496 | |
|
|
497 | for my $host (@list_of_hosts) { |
|
|
498 | $cv->begin; |
|
|
499 | ping_host_then_call_callback $host, sub { |
|
|
500 | $result{$host} = ...; |
|
|
501 | $cv->end; |
|
|
502 | }; |
|
|
503 | } |
|
|
504 | |
|
|
505 | $cv->end; |
|
|
506 | |
|
|
507 | This code fragment supposedly pings a number of hosts and calls |
|
|
508 | "send" after results for all then have have been gathered - in any |
|
|
509 | order. To achieve this, the code issues a call to "begin" when it |
|
|
510 | starts each ping request and calls "end" when it has received some |
|
|
511 | result for it. Since "begin" and "end" only maintain a counter, the |
|
|
512 | order in which results arrive is not relevant. |
|
|
513 | |
|
|
514 | There is an additional bracketing call to "begin" and "end" outside |
|
|
515 | the loop, which serves two important purposes: first, it sets the |
|
|
516 | callback to be called once the counter reaches 0, and second, it |
|
|
517 | ensures that "send" is called even when "no" hosts are being pinged |
|
|
518 | (the loop doesn't execute once). |
|
|
519 | |
|
|
520 | This is the general pattern when you "fan out" into multiple |
|
|
521 | subrequests: use an outer "begin"/"end" pair to set the callback and |
|
|
522 | ensure "end" is called at least once, and then, for each subrequest |
|
|
523 | you start, call "begin" and for each subrequest you finish, call |
|
|
524 | "end". |
|
|
525 | |
|
|
526 | METHODS FOR CONSUMERS |
|
|
527 | These methods should only be used by the consuming side, i.e. the code |
|
|
528 | awaits the condition. |
|
|
529 | |
|
|
530 | $cv->recv |
|
|
531 | Wait (blocking if necessary) until the "->send" or "->croak" methods |
|
|
532 | have been called on c<$cv>, while servicing other watchers normally. |
|
|
533 | |
|
|
534 | You can only wait once on a condition - additional calls are valid |
|
|
535 | but will return immediately. |
|
|
536 | |
|
|
537 | If an error condition has been set by calling "->croak", then this |
|
|
538 | function will call "croak". |
|
|
539 | |
|
|
540 | In list context, all parameters passed to "send" will be returned, |
|
|
541 | in scalar context only the first one will be returned. |
|
|
542 | |
|
|
543 | Not all event models support a blocking wait - some die in that case |
|
|
544 | (programs might want to do that to stay interactive), so *if you are |
|
|
545 | using this from a module, never require a blocking wait*, but let |
|
|
546 | the caller decide whether the call will block or not (for example, |
|
|
547 | by coupling condition variables with some kind of request results |
|
|
548 | and supporting callbacks so the caller knows that getting the result |
|
|
549 | will not block, while still supporting blocking waits if the caller |
|
|
550 | so desires). |
|
|
551 | |
|
|
552 | Another reason *never* to "->recv" in a module is that you cannot |
|
|
553 | sensibly have two "->recv"'s in parallel, as that would require |
|
|
554 | multiple interpreters or coroutines/threads, none of which |
|
|
555 | "AnyEvent" can supply. |
|
|
556 | |
|
|
557 | The Coro module, however, *can* and *does* supply coroutines and, in |
|
|
558 | fact, Coro::AnyEvent replaces AnyEvent's condvars by coroutine-safe |
|
|
559 | versions and also integrates coroutines into AnyEvent, making |
|
|
560 | blocking "->recv" calls perfectly safe as long as they are done from |
|
|
561 | another coroutine (one that doesn't run the event loop). |
|
|
562 | |
|
|
563 | You can ensure that "-recv" never blocks by setting a callback and |
|
|
564 | only calling "->recv" from within that callback (or at a later |
|
|
565 | time). This will work even when the event loop does not support |
|
|
566 | blocking waits otherwise. |
|
|
567 | |
|
|
568 | $bool = $cv->ready |
|
|
569 | Returns true when the condition is "true", i.e. whether "send" or |
|
|
570 | "croak" have been called. |
|
|
571 | |
|
|
572 | $cb = $cv->cb ([new callback]) |
|
|
573 | This is a mutator function that returns the callback set and |
|
|
574 | optionally replaces it before doing so. |
|
|
575 | |
|
|
576 | The callback will be called when the condition becomes "true", i.e. |
|
|
577 | when "send" or "croak" are called, with the only argument being the |
|
|
578 | condition variable itself. Calling "recv" inside the callback or at |
|
|
579 | any later time is guaranteed not to block. |
|
|
580 | |
|
|
581 | GLOBAL VARIABLES AND FUNCTIONS |
173 | $AnyEvent::MODEL |
582 | $AnyEvent::MODEL |
174 | Contains "undef" until the first watcher is being created. Then it |
583 | Contains "undef" until the first watcher is being created. Then it |
175 | contains the event model that is being used, which is the name of |
584 | contains the event model that is being used, which is the name of |
176 | the Perl class implementing the model. This class is usually one of |
585 | the Perl class implementing the model. This class is usually one of |
177 | the "AnyEvent::Impl:xxx" modules, but can be any other class in the |
586 | the "AnyEvent::Impl:xxx" modules, but can be any other class in the |
178 | case AnyEvent has been extended at runtime (e.g. in *rxvt-unicode*). |
587 | case AnyEvent has been extended at runtime (e.g. in *rxvt-unicode*). |
179 | |
588 | |
180 | The known classes so far are: |
589 | The known classes so far are: |
181 | |
590 | |
182 | AnyEvent::Impl::Coro based on Coro::Event, best choise. |
591 | AnyEvent::Impl::EV based on EV (an interface to libev, best choice). |
183 | AnyEvent::Impl::Event based on Event, also best choice :) |
592 | AnyEvent::Impl::Event based on Event, second best choice. |
|
|
593 | AnyEvent::Impl::Perl pure-perl implementation, fast and portable. |
184 | AnyEvent::Impl::Glib based on Glib, second-best choice. |
594 | AnyEvent::Impl::Glib based on Glib, third-best choice. |
185 | AnyEvent::Impl::Tk based on Tk, very bad choice. |
595 | AnyEvent::Impl::Tk based on Tk, very bad choice. |
186 | AnyEvent::Impl::Perl pure-perl implementation, inefficient. |
596 | AnyEvent::Impl::Qt based on Qt, cannot be autoprobed (see its docs). |
|
|
597 | AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse. |
|
|
598 | AnyEvent::Impl::POE based on POE, not generic enough for full support. |
|
|
599 | |
|
|
600 | There is no support for WxWidgets, as WxWidgets has no support for |
|
|
601 | watching file handles. However, you can use WxWidgets through the |
|
|
602 | POE Adaptor, as POE has a Wx backend that simply polls 20 times per |
|
|
603 | second, which was considered to be too horrible to even consider for |
|
|
604 | AnyEvent. Likewise, other POE backends can be used by AnyEvent by |
|
|
605 | using it's adaptor. |
|
|
606 | |
|
|
607 | AnyEvent knows about Prima and Wx and will try to use POE when |
|
|
608 | autodetecting them. |
187 | |
609 | |
188 | AnyEvent::detect |
610 | AnyEvent::detect |
189 | Returns $AnyEvent::MODEL, forcing autodetection of the event model |
611 | Returns $AnyEvent::MODEL, forcing autodetection of the event model |
190 | if necessary. You should only call this function right before you |
612 | if necessary. You should only call this function right before you |
191 | would have created an AnyEvent watcher anyway, that is, very late at |
613 | would have created an AnyEvent watcher anyway, that is, as late as |
192 | runtime. |
614 | possible at runtime. |
|
|
615 | |
|
|
616 | $guard = AnyEvent::post_detect { BLOCK } |
|
|
617 | Arranges for the code block to be executed as soon as the event |
|
|
618 | model is autodetected (or immediately if this has already happened). |
|
|
619 | |
|
|
620 | If called in scalar or list context, then it creates and returns an |
|
|
621 | object that automatically removes the callback again when it is |
|
|
622 | destroyed. See Coro::BDB for a case where this is useful. |
|
|
623 | |
|
|
624 | @AnyEvent::post_detect |
|
|
625 | If there are any code references in this array (you can "push" to it |
|
|
626 | before or after loading AnyEvent), then they will called directly |
|
|
627 | after the event loop has been chosen. |
|
|
628 | |
|
|
629 | You should check $AnyEvent::MODEL before adding to this array, |
|
|
630 | though: if it contains a true value then the event loop has already |
|
|
631 | been detected, and the array will be ignored. |
|
|
632 | |
|
|
633 | Best use "AnyEvent::post_detect { BLOCK }" instead. |
193 | |
634 | |
194 | WHAT TO DO IN A MODULE |
635 | WHAT TO DO IN A MODULE |
195 | As a module author, you should "use AnyEvent" and call AnyEvent methods |
636 | As a module author, you should "use AnyEvent" and call AnyEvent methods |
196 | freely, but you should not load a specific event module or rely on it. |
637 | freely, but you should not load a specific event module or rely on it. |
197 | |
638 | |
198 | Be careful when you create watchers in the module body - Anyevent will |
639 | Be careful when you create watchers in the module body - AnyEvent will |
199 | decide which event module to use as soon as the first method is called, |
640 | decide which event module to use as soon as the first method is called, |
200 | so by calling AnyEvent in your module body you force the user of your |
641 | so by calling AnyEvent in your module body you force the user of your |
201 | module to load the event module first. |
642 | module to load the event module first. |
202 | |
643 | |
|
|
644 | Never call "->recv" on a condition variable unless you *know* that the |
|
|
645 | "->send" method has been called on it already. This is because it will |
|
|
646 | stall the whole program, and the whole point of using events is to stay |
|
|
647 | interactive. |
|
|
648 | |
|
|
649 | It is fine, however, to call "->recv" when the user of your module |
|
|
650 | requests it (i.e. if you create a http request object ad have a method |
|
|
651 | called "results" that returns the results, it should call "->recv" |
|
|
652 | freely, as the user of your module knows what she is doing. always). |
|
|
653 | |
203 | WHAT TO DO IN THE MAIN PROGRAM |
654 | WHAT TO DO IN THE MAIN PROGRAM |
204 | There will always be a single main program - the only place that should |
655 | There will always be a single main program - the only place that should |
205 | dictate which event model to use. |
656 | dictate which event model to use. |
206 | |
657 | |
207 | If it doesn't care, it can just "use AnyEvent" and use it itself, or not |
658 | If it doesn't care, it can just "use AnyEvent" and use it itself, or not |
208 | do anything special and let AnyEvent decide which implementation to |
659 | do anything special (it does not need to be event-based) and let |
209 | chose. |
660 | AnyEvent decide which implementation to chose if some module relies on |
|
|
661 | it. |
210 | |
662 | |
211 | If the main program relies on a specific event model (for example, in |
663 | If the main program relies on a specific event model - for example, in |
212 | Gtk2 programs you have to rely on either Glib or Glib::Event), you |
664 | Gtk2 programs you have to rely on the Glib module - you should load the |
213 | should load it before loading AnyEvent or any module that uses it, |
665 | event module before loading AnyEvent or any module that uses it: |
214 | generally, as early as possible. The reason is that modules might create |
666 | generally speaking, you should load it as early as possible. The reason |
215 | watchers when they are loaded, and AnyEvent will decide on the event |
667 | is that modules might create watchers when they are loaded, and AnyEvent |
216 | model to use as soon as it creates watchers, and it might chose the |
668 | will decide on the event model to use as soon as it creates watchers, |
217 | wrong one unless you load the correct one yourself. |
669 | and it might chose the wrong one unless you load the correct one |
|
|
670 | yourself. |
218 | |
671 | |
219 | You can chose to use a rather inefficient pure-perl implementation by |
672 | You can chose to use a pure-perl implementation by loading the |
220 | loading the "AnyEvent::Impl::Perl" module, but letting AnyEvent chose is |
673 | "AnyEvent::Impl::Perl" module, which gives you similar behaviour |
221 | generally better. |
674 | everywhere, but letting AnyEvent chose the model is generally better. |
|
|
675 | |
|
|
676 | MAINLOOP EMULATION |
|
|
677 | Sometimes (often for short test scripts, or even standalone programs who |
|
|
678 | only want to use AnyEvent), you do not want to run a specific event |
|
|
679 | loop. |
|
|
680 | |
|
|
681 | In that case, you can use a condition variable like this: |
|
|
682 | |
|
|
683 | AnyEvent->condvar->recv; |
|
|
684 | |
|
|
685 | This has the effect of entering the event loop and looping forever. |
|
|
686 | |
|
|
687 | Note that usually your program has some exit condition, in which case it |
|
|
688 | is better to use the "traditional" approach of storing a condition |
|
|
689 | variable somewhere, waiting for it, and sending it when the program |
|
|
690 | should exit cleanly. |
|
|
691 | |
|
|
692 | OTHER MODULES |
|
|
693 | The following is a non-exhaustive list of additional modules that use |
|
|
694 | AnyEvent and can therefore be mixed easily with other AnyEvent modules |
|
|
695 | in the same program. Some of the modules come with AnyEvent, some are |
|
|
696 | available via CPAN. |
|
|
697 | |
|
|
698 | AnyEvent::Util |
|
|
699 | Contains various utility functions that replace often-used but |
|
|
700 | blocking functions such as "inet_aton" by event-/callback-based |
|
|
701 | versions. |
|
|
702 | |
|
|
703 | AnyEvent::Socket |
|
|
704 | Provides various utility functions for (internet protocol) sockets, |
|
|
705 | addresses and name resolution. Also functions to create non-blocking |
|
|
706 | tcp connections or tcp servers, with IPv6 and SRV record support and |
|
|
707 | more. |
|
|
708 | |
|
|
709 | AnyEvent::Handle |
|
|
710 | Provide read and write buffers, manages watchers for reads and |
|
|
711 | writes, supports raw and formatted I/O, I/O queued and fully |
|
|
712 | transparent and non-blocking SSL/TLS. |
|
|
713 | |
|
|
714 | AnyEvent::DNS |
|
|
715 | Provides rich asynchronous DNS resolver capabilities. |
|
|
716 | |
|
|
717 | AnyEvent::HTTP |
|
|
718 | A simple-to-use HTTP library that is capable of making a lot of |
|
|
719 | concurrent HTTP requests. |
|
|
720 | |
|
|
721 | AnyEvent::HTTPD |
|
|
722 | Provides a simple web application server framework. |
|
|
723 | |
|
|
724 | AnyEvent::FastPing |
|
|
725 | The fastest ping in the west. |
|
|
726 | |
|
|
727 | AnyEvent::DBI |
|
|
728 | Executes DBI requests asynchronously in a proxy process. |
|
|
729 | |
|
|
730 | AnyEvent::AIO |
|
|
731 | Truly asynchronous I/O, should be in the toolbox of every event |
|
|
732 | programmer. AnyEvent::AIO transparently fuses IO::AIO and AnyEvent |
|
|
733 | together. |
|
|
734 | |
|
|
735 | AnyEvent::BDB |
|
|
736 | Truly asynchronous Berkeley DB access. AnyEvent::BDB transparently |
|
|
737 | fuses BDB and AnyEvent together. |
|
|
738 | |
|
|
739 | AnyEvent::GPSD |
|
|
740 | A non-blocking interface to gpsd, a daemon delivering GPS |
|
|
741 | information. |
|
|
742 | |
|
|
743 | AnyEvent::IGS |
|
|
744 | A non-blocking interface to the Internet Go Server protocol (used by |
|
|
745 | App::IGS). |
|
|
746 | |
|
|
747 | Net::IRC3 |
|
|
748 | AnyEvent based IRC client module family. |
|
|
749 | |
|
|
750 | Net::XMPP2 |
|
|
751 | AnyEvent based XMPP (Jabber protocol) module family. |
|
|
752 | |
|
|
753 | Net::FCP |
|
|
754 | AnyEvent-based implementation of the Freenet Client Protocol, |
|
|
755 | birthplace of AnyEvent. |
|
|
756 | |
|
|
757 | Event::ExecFlow |
|
|
758 | High level API for event-based execution flow control. |
|
|
759 | |
|
|
760 | Coro |
|
|
761 | Has special support for AnyEvent via Coro::AnyEvent. |
|
|
762 | |
|
|
763 | IO::Lambda |
|
|
764 | The lambda approach to I/O - don't ask, look there. Can use |
|
|
765 | AnyEvent. |
222 | |
766 | |
223 | SUPPLYING YOUR OWN EVENT MODEL INTERFACE |
767 | SUPPLYING YOUR OWN EVENT MODEL INTERFACE |
|
|
768 | This is an advanced topic that you do not normally need to use AnyEvent |
|
|
769 | in a module. This section is only of use to event loop authors who want |
|
|
770 | to provide AnyEvent compatibility. |
|
|
771 | |
224 | If you need to support another event library which isn't directly |
772 | If you need to support another event library which isn't directly |
225 | supported by AnyEvent, you can supply your own interface to it by |
773 | supported by AnyEvent, you can supply your own interface to it by |
226 | pushing, before the first watcher gets created, the package name of the |
774 | pushing, before the first watcher gets created, the package name of the |
227 | event module and the package name of the interface to use onto |
775 | event module and the package name of the interface to use onto |
228 | @AnyEvent::REGISTRY. You can do that before and even without loading |
776 | @AnyEvent::REGISTRY. You can do that before and even without loading |
229 | AnyEvent. |
777 | AnyEvent, so it is reasonably cheap. |
230 | |
778 | |
231 | Example: |
779 | Example: |
232 | |
780 | |
233 | push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::]; |
781 | push @AnyEvent::REGISTRY, [urxvt => urxvt::anyevent::]; |
234 | |
782 | |
235 | This tells AnyEvent to (literally) use the "urxvt::anyevent::" |
783 | This tells AnyEvent to (literally) use the "urxvt::anyevent::" |
236 | package/class when it finds the "urxvt" package/module is loaded. When |
784 | package/class when it finds the "urxvt" package/module is already |
|
|
785 | loaded. |
|
|
786 | |
237 | AnyEvent is loaded and asked to find a suitable event model, it will |
787 | When AnyEvent is loaded and asked to find a suitable event model, it |
238 | first check for the presence of urxvt. |
788 | will first check for the presence of urxvt by trying to "use" the |
|
|
789 | "urxvt::anyevent" module. |
239 | |
790 | |
240 | The class should provide implementations for all watcher types (see |
791 | The class should provide implementations for all watcher types. See |
241 | AnyEvent::Impl::Event (source code), AnyEvent::Impl::Glib (Source code) |
792 | AnyEvent::Impl::EV (source code), AnyEvent::Impl::Glib (Source code) and |
242 | and so on for actual examples, use "perldoc -m AnyEvent::Impl::Glib" to |
793 | so on for actual examples. Use "perldoc -m AnyEvent::Impl::Glib" to see |
243 | see the sources). |
794 | the sources. |
244 | |
795 | |
|
|
796 | If you don't provide "signal" and "child" watchers than AnyEvent will |
|
|
797 | provide suitable (hopefully) replacements. |
|
|
798 | |
245 | The above isn't fictitious, the *rxvt-unicode* (a.k.a. urxvt) uses the |
799 | The above example isn't fictitious, the *rxvt-unicode* (a.k.a. urxvt) |
246 | above line as-is. An interface isn't included in AnyEvent because it |
800 | terminal emulator uses the above line as-is. An interface isn't included |
247 | doesn't make sense outside the embedded interpreter inside |
801 | in AnyEvent because it doesn't make sense outside the embedded |
248 | *rxvt-unicode*, and it is updated and maintained as part of the |
802 | interpreter inside *rxvt-unicode*, and it is updated and maintained as |
249 | *rxvt-unicode* distribution. |
803 | part of the *rxvt-unicode* distribution. |
250 | |
804 | |
251 | *rxvt-unicode* also cheats a bit by not providing blocking access to |
805 | *rxvt-unicode* also cheats a bit by not providing blocking access to |
252 | condition variables: code blocking while waiting for a condition will |
806 | condition variables: code blocking while waiting for a condition will |
253 | "die". This still works with most modules/usages, and blocking calls |
807 | "die". This still works with most modules/usages, and blocking calls |
254 | must not be in an interactive application, so it makes sense. |
808 | must not be done in an interactive application, so it makes sense. |
255 | |
809 | |
256 | ENVIRONMENT VARIABLES |
810 | ENVIRONMENT VARIABLES |
257 | The following environment variables are used by this module: |
811 | The following environment variables are used by this module: |
258 | |
812 | |
259 | "PERL_ANYEVENT_VERBOSE" when set to 2 or higher, reports which event |
813 | "PERL_ANYEVENT_VERBOSE" |
260 | model gets used. |
814 | By default, AnyEvent will be completely silent except in fatal |
|
|
815 | conditions. You can set this environment variable to make AnyEvent |
|
|
816 | more talkative. |
261 | |
817 | |
262 | EXAMPLE |
818 | When set to 1 or higher, causes AnyEvent to warn about unexpected |
|
|
819 | conditions, such as not being able to load the event model specified |
|
|
820 | by "PERL_ANYEVENT_MODEL". |
|
|
821 | |
|
|
822 | When set to 2 or higher, cause AnyEvent to report to STDERR which |
|
|
823 | event model it chooses. |
|
|
824 | |
|
|
825 | "PERL_ANYEVENT_STRICT" |
|
|
826 | AnyEvent does not do much argument checking by default, as thorough |
|
|
827 | argument checking is very costly. Setting this variable to a true |
|
|
828 | value will cause AnyEvent to load "AnyEvent::Strict" and then to |
|
|
829 | thoroughly check the arguments passed to most method calls. If it |
|
|
830 | finds any problems it will croak. |
|
|
831 | |
|
|
832 | In other words, enables "strict" mode. |
|
|
833 | |
|
|
834 | Unlike "use strict" it is definitely recommended ot keep it off in |
|
|
835 | production. |
|
|
836 | |
|
|
837 | "PERL_ANYEVENT_MODEL" |
|
|
838 | This can be used to specify the event model to be used by AnyEvent, |
|
|
839 | before auto detection and -probing kicks in. It must be a string |
|
|
840 | consisting entirely of ASCII letters. The string "AnyEvent::Impl::" |
|
|
841 | gets prepended and the resulting module name is loaded and if the |
|
|
842 | load was successful, used as event model. If it fails to load |
|
|
843 | AnyEvent will proceed with auto detection and -probing. |
|
|
844 | |
|
|
845 | This functionality might change in future versions. |
|
|
846 | |
|
|
847 | For example, to force the pure perl model (AnyEvent::Impl::Perl) you |
|
|
848 | could start your program like this: |
|
|
849 | |
|
|
850 | PERL_ANYEVENT_MODEL=Perl perl ... |
|
|
851 | |
|
|
852 | "PERL_ANYEVENT_PROTOCOLS" |
|
|
853 | Used by both AnyEvent::DNS and AnyEvent::Socket to determine |
|
|
854 | preferences for IPv4 or IPv6. The default is unspecified (and might |
|
|
855 | change, or be the result of auto probing). |
|
|
856 | |
|
|
857 | Must be set to a comma-separated list of protocols or address |
|
|
858 | families, current supported: "ipv4" and "ipv6". Only protocols |
|
|
859 | mentioned will be used, and preference will be given to protocols |
|
|
860 | mentioned earlier in the list. |
|
|
861 | |
|
|
862 | This variable can effectively be used for denial-of-service attacks |
|
|
863 | against local programs (e.g. when setuid), although the impact is |
|
|
864 | likely small, as the program has to handle connection errors |
|
|
865 | already- |
|
|
866 | |
|
|
867 | Examples: "PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6" - prefer IPv4 over |
|
|
868 | IPv6, but support both and try to use both. |
|
|
869 | "PERL_ANYEVENT_PROTOCOLS=ipv4" - only support IPv4, never try to |
|
|
870 | resolve or contact IPv6 addresses. |
|
|
871 | "PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4" support either IPv4 or IPv6, but |
|
|
872 | prefer IPv6 over IPv4. |
|
|
873 | |
|
|
874 | "PERL_ANYEVENT_EDNS0" |
|
|
875 | Used by AnyEvent::DNS to decide whether to use the EDNS0 extension |
|
|
876 | for DNS. This extension is generally useful to reduce DNS traffic, |
|
|
877 | but some (broken) firewalls drop such DNS packets, which is why it |
|
|
878 | is off by default. |
|
|
879 | |
|
|
880 | Setting this variable to 1 will cause AnyEvent::DNS to announce |
|
|
881 | EDNS0 in its DNS requests. |
|
|
882 | |
|
|
883 | "PERL_ANYEVENT_MAX_FORKS" |
|
|
884 | The maximum number of child processes that |
|
|
885 | "AnyEvent::Util::fork_call" will create in parallel. |
|
|
886 | |
|
|
887 | EXAMPLE PROGRAM |
263 | The following program uses an io watcher to read data from stdin, a |
888 | The following program uses an I/O watcher to read data from STDIN, a |
264 | timer to display a message once per second, and a condvar to exit the |
889 | timer to display a message once per second, and a condition variable to |
265 | program when the user enters quit: |
890 | quit the program when the user enters quit: |
266 | |
891 | |
267 | use AnyEvent; |
892 | use AnyEvent; |
268 | |
893 | |
269 | my $cv = AnyEvent->condvar; |
894 | my $cv = AnyEvent->condvar; |
270 | |
895 | |
271 | my $io_watcher = AnyEvent->io (fh => \*STDIN, poll => 'r', cb => sub { |
896 | my $io_watcher = AnyEvent->io ( |
|
|
897 | fh => \*STDIN, |
|
|
898 | poll => 'r', |
|
|
899 | cb => sub { |
272 | warn "io event <$_[0]>\n"; # will always output <r> |
900 | warn "io event <$_[0]>\n"; # will always output <r> |
273 | chomp (my $input = <STDIN>); # read a line |
901 | chomp (my $input = <STDIN>); # read a line |
274 | warn "read: $input\n"; # output what has been read |
902 | warn "read: $input\n"; # output what has been read |
275 | $cv->broadcast if $input =~ /^q/i; # quit program if /^q/i |
903 | $cv->send if $input =~ /^q/i; # quit program if /^q/i |
|
|
904 | }, |
276 | }); |
905 | ); |
277 | |
906 | |
278 | my $time_watcher; # can only be used once |
907 | my $time_watcher; # can only be used once |
279 | |
908 | |
280 | sub new_timer { |
909 | sub new_timer { |
281 | $timer = AnyEvent->timer (after => 1, cb => sub { |
910 | $timer = AnyEvent->timer (after => 1, cb => sub { |
… | |
… | |
284 | }); |
913 | }); |
285 | } |
914 | } |
286 | |
915 | |
287 | new_timer; # create first timer |
916 | new_timer; # create first timer |
288 | |
917 | |
289 | $cv->wait; # wait until user enters /^q/i |
918 | $cv->recv; # wait until user enters /^q/i |
290 | |
919 | |
291 | REAL-WORLD EXAMPLE |
920 | REAL-WORLD EXAMPLE |
292 | Consider the Net::FCP module. It features (among others) the following |
921 | Consider the Net::FCP module. It features (among others) the following |
293 | API calls, which are to freenet what HTTP GET requests are to http: |
922 | API calls, which are to freenet what HTTP GET requests are to http: |
294 | |
923 | |
… | |
… | |
343 | syswrite $txn->{fh}, $txn->{request} |
972 | syswrite $txn->{fh}, $txn->{request} |
344 | or die "connection or write error"; |
973 | or die "connection or write error"; |
345 | $txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r }); |
974 | $txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r }); |
346 | |
975 | |
347 | Again, "fh_ready_r" waits till all data has arrived, and then stores the |
976 | Again, "fh_ready_r" waits till all data has arrived, and then stores the |
348 | result and signals any possible waiters that the request ahs finished: |
977 | result and signals any possible waiters that the request has finished: |
349 | |
978 | |
350 | sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf}; |
979 | sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf}; |
351 | |
980 | |
352 | if (end-of-file or data complete) { |
981 | if (end-of-file or data complete) { |
353 | $txn->{result} = $txn->{buf}; |
982 | $txn->{result} = $txn->{buf}; |
354 | $txn->{finished}->broadcast; |
983 | $txn->{finished}->send; |
355 | $txb->{cb}->($txn) of $txn->{cb}; # also call callback |
984 | $txb->{cb}->($txn) of $txn->{cb}; # also call callback |
356 | } |
985 | } |
357 | |
986 | |
358 | The "result" method, finally, just waits for the finished signal (if the |
987 | The "result" method, finally, just waits for the finished signal (if the |
359 | request was already finished, it doesn't wait, of course, and returns |
988 | request was already finished, it doesn't wait, of course, and returns |
360 | the data: |
989 | the data: |
361 | |
990 | |
362 | $txn->{finished}->wait; |
991 | $txn->{finished}->recv; |
363 | return $txn->{result}; |
992 | return $txn->{result}; |
364 | |
993 | |
365 | The actual code goes further and collects all errors ("die"s, |
994 | The actual code goes further and collects all errors ("die"s, |
366 | exceptions) that occured during request processing. The "result" method |
995 | exceptions) that occurred during request processing. The "result" method |
367 | detects wether an exception as thrown (it is stored inside the $txn |
996 | detects whether an exception as thrown (it is stored inside the $txn |
368 | object) and just throws the exception, which means connection errors and |
997 | object) and just throws the exception, which means connection errors and |
369 | other problems get reported tot he code that tries to use the result, |
998 | other problems get reported tot he code that tries to use the result, |
370 | not in a random callback. |
999 | not in a random callback. |
371 | |
1000 | |
372 | All of this enables the following usage styles: |
1001 | All of this enables the following usage styles: |
373 | |
1002 | |
374 | 1. Blocking: |
1003 | 1. Blocking: |
375 | |
1004 | |
376 | my $data = $fcp->client_get ($url); |
1005 | my $data = $fcp->client_get ($url); |
377 | |
1006 | |
378 | 2. Blocking, but parallelizing: |
1007 | 2. Blocking, but running in parallel: |
379 | |
1008 | |
380 | my @datas = map $_->result, |
1009 | my @datas = map $_->result, |
381 | map $fcp->txn_client_get ($_), |
1010 | map $fcp->txn_client_get ($_), |
382 | @urls; |
1011 | @urls; |
383 | |
1012 | |
384 | Both blocking examples work without the module user having to know |
1013 | Both blocking examples work without the module user having to know |
385 | anything about events. |
1014 | anything about events. |
386 | |
1015 | |
387 | 3a. Event-based in a main program, using any support Event module: |
1016 | 3a. Event-based in a main program, using any supported event module: |
388 | |
1017 | |
389 | use Event; |
1018 | use EV; |
390 | |
1019 | |
391 | $fcp->txn_client_get ($url)->cb (sub { |
1020 | $fcp->txn_client_get ($url)->cb (sub { |
392 | my $txn = shift; |
1021 | my $txn = shift; |
393 | my $data = $txn->result; |
1022 | my $data = $txn->result; |
394 | ... |
1023 | ... |
395 | }); |
1024 | }); |
396 | |
1025 | |
397 | Event::loop; |
1026 | EV::loop; |
398 | |
1027 | |
399 | 3b. The module user could use AnyEvent, too: |
1028 | 3b. The module user could use AnyEvent, too: |
400 | |
1029 | |
401 | use AnyEvent; |
1030 | use AnyEvent; |
402 | |
1031 | |
403 | my $quit = AnyEvent->condvar; |
1032 | my $quit = AnyEvent->condvar; |
404 | |
1033 | |
405 | $fcp->txn_client_get ($url)->cb (sub { |
1034 | $fcp->txn_client_get ($url)->cb (sub { |
406 | ... |
1035 | ... |
407 | $quit->broadcast; |
1036 | $quit->send; |
408 | }); |
1037 | }); |
409 | |
1038 | |
410 | $quit->wait; |
1039 | $quit->recv; |
|
|
1040 | |
|
|
1041 | BENCHMARKS |
|
|
1042 | To give you an idea of the performance and overheads that AnyEvent adds |
|
|
1043 | over the event loops themselves and to give you an impression of the |
|
|
1044 | speed of various event loops I prepared some benchmarks. |
|
|
1045 | |
|
|
1046 | BENCHMARKING ANYEVENT OVERHEAD |
|
|
1047 | Here is a benchmark of various supported event models used natively and |
|
|
1048 | through AnyEvent. The benchmark creates a lot of timers (with a zero |
|
|
1049 | timeout) and I/O watchers (watching STDOUT, a pty, to become writable, |
|
|
1050 | which it is), lets them fire exactly once and destroys them again. |
|
|
1051 | |
|
|
1052 | Source code for this benchmark is found as eg/bench in the AnyEvent |
|
|
1053 | distribution. |
|
|
1054 | |
|
|
1055 | Explanation of the columns |
|
|
1056 | *watcher* is the number of event watchers created/destroyed. Since |
|
|
1057 | different event models feature vastly different performances, each event |
|
|
1058 | loop was given a number of watchers so that overall runtime is |
|
|
1059 | acceptable and similar between tested event loop (and keep them from |
|
|
1060 | crashing): Glib would probably take thousands of years if asked to |
|
|
1061 | process the same number of watchers as EV in this benchmark. |
|
|
1062 | |
|
|
1063 | *bytes* is the number of bytes (as measured by the resident set size, |
|
|
1064 | RSS) consumed by each watcher. This method of measuring captures both C |
|
|
1065 | and Perl-based overheads. |
|
|
1066 | |
|
|
1067 | *create* is the time, in microseconds (millionths of seconds), that it |
|
|
1068 | takes to create a single watcher. The callback is a closure shared |
|
|
1069 | between all watchers, to avoid adding memory overhead. That means |
|
|
1070 | closure creation and memory usage is not included in the figures. |
|
|
1071 | |
|
|
1072 | *invoke* is the time, in microseconds, used to invoke a simple callback. |
|
|
1073 | The callback simply counts down a Perl variable and after it was invoked |
|
|
1074 | "watcher" times, it would "->send" a condvar once to signal the end of |
|
|
1075 | this phase. |
|
|
1076 | |
|
|
1077 | *destroy* is the time, in microseconds, that it takes to destroy a |
|
|
1078 | single watcher. |
|
|
1079 | |
|
|
1080 | Results |
|
|
1081 | name watchers bytes create invoke destroy comment |
|
|
1082 | EV/EV 400000 244 0.56 0.46 0.31 EV native interface |
|
|
1083 | EV/Any 100000 244 2.50 0.46 0.29 EV + AnyEvent watchers |
|
|
1084 | CoroEV/Any 100000 244 2.49 0.44 0.29 coroutines + Coro::Signal |
|
|
1085 | Perl/Any 100000 513 4.92 0.87 1.12 pure perl implementation |
|
|
1086 | Event/Event 16000 516 31.88 31.30 0.85 Event native interface |
|
|
1087 | Event/Any 16000 590 35.75 31.42 1.08 Event + AnyEvent watchers |
|
|
1088 | Glib/Any 16000 1357 98.22 12.41 54.00 quadratic behaviour |
|
|
1089 | Tk/Any 2000 1860 26.97 67.98 14.00 SEGV with >> 2000 watchers |
|
|
1090 | POE/Event 2000 6644 108.64 736.02 14.73 via POE::Loop::Event |
|
|
1091 | POE/Select 2000 6343 94.13 809.12 565.96 via POE::Loop::Select |
|
|
1092 | |
|
|
1093 | Discussion |
|
|
1094 | The benchmark does *not* measure scalability of the event loop very |
|
|
1095 | well. For example, a select-based event loop (such as the pure perl one) |
|
|
1096 | can never compete with an event loop that uses epoll when the number of |
|
|
1097 | file descriptors grows high. In this benchmark, all events become ready |
|
|
1098 | at the same time, so select/poll-based implementations get an unnatural |
|
|
1099 | speed boost. |
|
|
1100 | |
|
|
1101 | Also, note that the number of watchers usually has a nonlinear effect on |
|
|
1102 | overall speed, that is, creating twice as many watchers doesn't take |
|
|
1103 | twice the time - usually it takes longer. This puts event loops tested |
|
|
1104 | with a higher number of watchers at a disadvantage. |
|
|
1105 | |
|
|
1106 | To put the range of results into perspective, consider that on the |
|
|
1107 | benchmark machine, handling an event takes roughly 1600 CPU cycles with |
|
|
1108 | EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 |
|
|
1109 | CPU cycles with POE. |
|
|
1110 | |
|
|
1111 | "EV" is the sole leader regarding speed and memory use, which are both |
|
|
1112 | maximal/minimal, respectively. Even when going through AnyEvent, it uses |
|
|
1113 | far less memory than any other event loop and is still faster than Event |
|
|
1114 | natively. |
|
|
1115 | |
|
|
1116 | The pure perl implementation is hit in a few sweet spots (both the |
|
|
1117 | constant timeout and the use of a single fd hit optimisations in the |
|
|
1118 | perl interpreter and the backend itself). Nevertheless this shows that |
|
|
1119 | it adds very little overhead in itself. Like any select-based backend |
|
|
1120 | its performance becomes really bad with lots of file descriptors (and |
|
|
1121 | few of them active), of course, but this was not subject of this |
|
|
1122 | benchmark. |
|
|
1123 | |
|
|
1124 | The "Event" module has a relatively high setup and callback invocation |
|
|
1125 | cost, but overall scores in on the third place. |
|
|
1126 | |
|
|
1127 | "Glib"'s memory usage is quite a bit higher, but it features a faster |
|
|
1128 | callback invocation and overall ends up in the same class as "Event". |
|
|
1129 | However, Glib scales extremely badly, doubling the number of watchers |
|
|
1130 | increases the processing time by more than a factor of four, making it |
|
|
1131 | completely unusable when using larger numbers of watchers (note that |
|
|
1132 | only a single file descriptor was used in the benchmark, so |
|
|
1133 | inefficiencies of "poll" do not account for this). |
|
|
1134 | |
|
|
1135 | The "Tk" adaptor works relatively well. The fact that it crashes with |
|
|
1136 | more than 2000 watchers is a big setback, however, as correctness takes |
|
|
1137 | precedence over speed. Nevertheless, its performance is surprising, as |
|
|
1138 | the file descriptor is dup()ed for each watcher. This shows that the |
|
|
1139 | dup() employed by some adaptors is not a big performance issue (it does |
|
|
1140 | incur a hidden memory cost inside the kernel which is not reflected in |
|
|
1141 | the figures above). |
|
|
1142 | |
|
|
1143 | "POE", regardless of underlying event loop (whether using its pure perl |
|
|
1144 | select-based backend or the Event module, the POE-EV backend couldn't be |
|
|
1145 | tested because it wasn't working) shows abysmal performance and memory |
|
|
1146 | usage with AnyEvent: Watchers use almost 30 times as much memory as EV |
|
|
1147 | watchers, and 10 times as much memory as Event (the high memory |
|
|
1148 | requirements are caused by requiring a session for each watcher). |
|
|
1149 | Watcher invocation speed is almost 900 times slower than with AnyEvent's |
|
|
1150 | pure perl implementation. |
|
|
1151 | |
|
|
1152 | The design of the POE adaptor class in AnyEvent can not really account |
|
|
1153 | for the performance issues, though, as session creation overhead is |
|
|
1154 | small compared to execution of the state machine, which is coded pretty |
|
|
1155 | optimally within AnyEvent::Impl::POE (and while everybody agrees that |
|
|
1156 | using multiple sessions is not a good approach, especially regarding |
|
|
1157 | memory usage, even the author of POE could not come up with a faster |
|
|
1158 | design). |
|
|
1159 | |
|
|
1160 | Summary |
|
|
1161 | * Using EV through AnyEvent is faster than any other event loop (even |
|
|
1162 | when used without AnyEvent), but most event loops have acceptable |
|
|
1163 | performance with or without AnyEvent. |
|
|
1164 | |
|
|
1165 | * The overhead AnyEvent adds is usually much smaller than the overhead |
|
|
1166 | of the actual event loop, only with extremely fast event loops such |
|
|
1167 | as EV adds AnyEvent significant overhead. |
|
|
1168 | |
|
|
1169 | * You should avoid POE like the plague if you want performance or |
|
|
1170 | reasonable memory usage. |
|
|
1171 | |
|
|
1172 | BENCHMARKING THE LARGE SERVER CASE |
|
|
1173 | This benchmark actually benchmarks the event loop itself. It works by |
|
|
1174 | creating a number of "servers": each server consists of a socket pair, a |
|
|
1175 | timeout watcher that gets reset on activity (but never fires), and an |
|
|
1176 | I/O watcher waiting for input on one side of the socket. Each time the |
|
|
1177 | socket watcher reads a byte it will write that byte to a random other |
|
|
1178 | "server". |
|
|
1179 | |
|
|
1180 | The effect is that there will be a lot of I/O watchers, only part of |
|
|
1181 | which are active at any one point (so there is a constant number of |
|
|
1182 | active fds for each loop iteration, but which fds these are is random). |
|
|
1183 | The timeout is reset each time something is read because that reflects |
|
|
1184 | how most timeouts work (and puts extra pressure on the event loops). |
|
|
1185 | |
|
|
1186 | In this benchmark, we use 10000 socket pairs (20000 sockets), of which |
|
|
1187 | 100 (1%) are active. This mirrors the activity of large servers with |
|
|
1188 | many connections, most of which are idle at any one point in time. |
|
|
1189 | |
|
|
1190 | Source code for this benchmark is found as eg/bench2 in the AnyEvent |
|
|
1191 | distribution. |
|
|
1192 | |
|
|
1193 | Explanation of the columns |
|
|
1194 | *sockets* is the number of sockets, and twice the number of "servers" |
|
|
1195 | (as each server has a read and write socket end). |
|
|
1196 | |
|
|
1197 | *create* is the time it takes to create a socket pair (which is |
|
|
1198 | nontrivial) and two watchers: an I/O watcher and a timeout watcher. |
|
|
1199 | |
|
|
1200 | *request*, the most important value, is the time it takes to handle a |
|
|
1201 | single "request", that is, reading the token from the pipe and |
|
|
1202 | forwarding it to another server. This includes deleting the old timeout |
|
|
1203 | and creating a new one that moves the timeout into the future. |
|
|
1204 | |
|
|
1205 | Results |
|
|
1206 | name sockets create request |
|
|
1207 | EV 20000 69.01 11.16 |
|
|
1208 | Perl 20000 73.32 35.87 |
|
|
1209 | Event 20000 212.62 257.32 |
|
|
1210 | Glib 20000 651.16 1896.30 |
|
|
1211 | POE 20000 349.67 12317.24 uses POE::Loop::Event |
|
|
1212 | |
|
|
1213 | Discussion |
|
|
1214 | This benchmark *does* measure scalability and overall performance of the |
|
|
1215 | particular event loop. |
|
|
1216 | |
|
|
1217 | EV is again fastest. Since it is using epoll on my system, the setup |
|
|
1218 | time is relatively high, though. |
|
|
1219 | |
|
|
1220 | Perl surprisingly comes second. It is much faster than the C-based event |
|
|
1221 | loops Event and Glib. |
|
|
1222 | |
|
|
1223 | Event suffers from high setup time as well (look at its code and you |
|
|
1224 | will understand why). Callback invocation also has a high overhead |
|
|
1225 | compared to the "$_->() for .."-style loop that the Perl event loop |
|
|
1226 | uses. Event uses select or poll in basically all documented |
|
|
1227 | configurations. |
|
|
1228 | |
|
|
1229 | Glib is hit hard by its quadratic behaviour w.r.t. many watchers. It |
|
|
1230 | clearly fails to perform with many filehandles or in busy servers. |
|
|
1231 | |
|
|
1232 | POE is still completely out of the picture, taking over 1000 times as |
|
|
1233 | long as EV, and over 100 times as long as the Perl implementation, even |
|
|
1234 | though it uses a C-based event loop in this case. |
|
|
1235 | |
|
|
1236 | Summary |
|
|
1237 | * The pure perl implementation performs extremely well. |
|
|
1238 | |
|
|
1239 | * Avoid Glib or POE in large projects where performance matters. |
|
|
1240 | |
|
|
1241 | BENCHMARKING SMALL SERVERS |
|
|
1242 | While event loops should scale (and select-based ones do not...) even to |
|
|
1243 | large servers, most programs we (or I :) actually write have only a few |
|
|
1244 | I/O watchers. |
|
|
1245 | |
|
|
1246 | In this benchmark, I use the same benchmark program as in the large |
|
|
1247 | server case, but it uses only eight "servers", of which three are active |
|
|
1248 | at any one time. This should reflect performance for a small server |
|
|
1249 | relatively well. |
|
|
1250 | |
|
|
1251 | The columns are identical to the previous table. |
|
|
1252 | |
|
|
1253 | Results |
|
|
1254 | name sockets create request |
|
|
1255 | EV 16 20.00 6.54 |
|
|
1256 | Perl 16 25.75 12.62 |
|
|
1257 | Event 16 81.27 35.86 |
|
|
1258 | Glib 16 32.63 15.48 |
|
|
1259 | POE 16 261.87 276.28 uses POE::Loop::Event |
|
|
1260 | |
|
|
1261 | Discussion |
|
|
1262 | The benchmark tries to test the performance of a typical small server. |
|
|
1263 | While knowing how various event loops perform is interesting, keep in |
|
|
1264 | mind that their overhead in this case is usually not as important, due |
|
|
1265 | to the small absolute number of watchers (that is, you need efficiency |
|
|
1266 | and speed most when you have lots of watchers, not when you only have a |
|
|
1267 | few of them). |
|
|
1268 | |
|
|
1269 | EV is again fastest. |
|
|
1270 | |
|
|
1271 | Perl again comes second. It is noticeably faster than the C-based event |
|
|
1272 | loops Event and Glib, although the difference is too small to really |
|
|
1273 | matter. |
|
|
1274 | |
|
|
1275 | POE also performs much better in this case, but is is still far behind |
|
|
1276 | the others. |
|
|
1277 | |
|
|
1278 | Summary |
|
|
1279 | * C-based event loops perform very well with small number of watchers, |
|
|
1280 | as the management overhead dominates. |
|
|
1281 | |
|
|
1282 | FORK |
|
|
1283 | Most event libraries are not fork-safe. The ones who are usually are |
|
|
1284 | because they rely on inefficient but fork-safe "select" or "poll" calls. |
|
|
1285 | Only EV is fully fork-aware. |
|
|
1286 | |
|
|
1287 | If you have to fork, you must either do so *before* creating your first |
|
|
1288 | watcher OR you must not use AnyEvent at all in the child. |
|
|
1289 | |
|
|
1290 | SECURITY CONSIDERATIONS |
|
|
1291 | AnyEvent can be forced to load any event model via |
|
|
1292 | $ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used |
|
|
1293 | to execute arbitrary code or directly gain access, it can easily be used |
|
|
1294 | to make the program hang or malfunction in subtle ways, as AnyEvent |
|
|
1295 | watchers will not be active when the program uses a different event |
|
|
1296 | model than specified in the variable. |
|
|
1297 | |
|
|
1298 | You can make AnyEvent completely ignore this variable by deleting it |
|
|
1299 | before the first watcher gets created, e.g. with a "BEGIN" block: |
|
|
1300 | |
|
|
1301 | BEGIN { delete $ENV{PERL_ANYEVENT_MODEL} } |
|
|
1302 | |
|
|
1303 | use AnyEvent; |
|
|
1304 | |
|
|
1305 | Similar considerations apply to $ENV{PERL_ANYEVENT_VERBOSE}, as that can |
|
|
1306 | be used to probe what backend is used and gain other information (which |
|
|
1307 | is probably even less useful to an attacker than PERL_ANYEVENT_MODEL), |
|
|
1308 | and $ENV{PERL_ANYEGENT_STRICT}. |
|
|
1309 | |
|
|
1310 | BUGS |
|
|
1311 | Perl 5.8 has numerous memleaks that sometimes hit this module and are |
|
|
1312 | hard to work around. If you suffer from memleaks, first upgrade to Perl |
|
|
1313 | 5.10 and check wether the leaks still show up. (Perl 5.10.0 has other |
|
|
1314 | annoying mamleaks, such as leaking on "map" and "grep" but it is usually |
|
|
1315 | not as pronounced). |
411 | |
1316 | |
412 | SEE ALSO |
1317 | SEE ALSO |
413 | Event modules: Coro::Event, Coro, Event, Glib::Event, Glib. |
1318 | Utility functions: AnyEvent::Util. |
414 | |
1319 | |
|
|
1320 | Event modules: EV, EV::Glib, Glib::EV, Event, Glib::Event, Glib, Tk, |
|
|
1321 | Event::Lib, Qt, POE. |
|
|
1322 | |
415 | Implementations: AnyEvent::Impl::Coro, AnyEvent::Impl::Event, |
1323 | Implementations: AnyEvent::Impl::EV, AnyEvent::Impl::Event, |
416 | AnyEvent::Impl::Glib, AnyEvent::Impl::Tk. |
1324 | AnyEvent::Impl::Glib, AnyEvent::Impl::Tk, AnyEvent::Impl::Perl, |
|
|
1325 | AnyEvent::Impl::EventLib, AnyEvent::Impl::Qt, AnyEvent::Impl::POE. |
417 | |
1326 | |
418 | Nontrivial usage example: Net::FCP. |
1327 | Non-blocking file handles, sockets, TCP clients and servers: |
|
|
1328 | AnyEvent::Handle, AnyEvent::Socket. |
419 | |
1329 | |
|
|
1330 | Asynchronous DNS: AnyEvent::DNS. |
420 | |
1331 | |
|
|
1332 | Coroutine support: Coro, Coro::AnyEvent, Coro::EV, Coro::Event, |
|
|
1333 | |
|
|
1334 | Nontrivial usage examples: Net::FCP, Net::XMPP2, AnyEvent::DNS. |
|
|
1335 | |
|
|
1336 | AUTHOR |
|
|
1337 | Marc Lehmann <schmorp@schmorp.de> |
|
|
1338 | http://home.schmorp.de/ |
|
|
1339 | |