… | |
… | |
354 | time, which might affect timers and time-outs. |
354 | time, which might affect timers and time-outs. |
355 | |
355 | |
356 | When this is the case, you can call this method, which will update |
356 | When this is the case, you can call this method, which will update |
357 | the event loop's idea of "current time". |
357 | the event loop's idea of "current time". |
358 | |
358 | |
|
|
359 | A typical example would be a script in a web server (e.g. |
|
|
360 | "mod_perl") - when mod_perl executes the script, then the event loop |
|
|
361 | will have the wrong idea about the "current time" (being potentially |
|
|
362 | far in the past, when the script ran the last time). In that case |
|
|
363 | you should arrange a call to "AnyEvent->now_update" each time the |
|
|
364 | web server process wakes up again (e.g. at the start of your script, |
|
|
365 | or in a handler). |
|
|
366 | |
359 | Note that updating the time *might* cause some events to be handled. |
367 | Note that updating the time *might* cause some events to be handled. |
360 | |
368 | |
361 | SIGNAL WATCHERS |
369 | SIGNAL WATCHERS |
362 | $w = AnyEvent->signal (signal => <uppercase_signal_name>, cb => <callback>); |
370 | $w = AnyEvent->signal (signal => <uppercase_signal_name>, cb => <callback>); |
363 | |
371 | |
… | |
… | |
383 | correctly. |
391 | correctly. |
384 | |
392 | |
385 | Example: exit on SIGINT |
393 | Example: exit on SIGINT |
386 | |
394 | |
387 | my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 }); |
395 | my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 }); |
|
|
396 | |
|
|
397 | Restart Behaviour |
|
|
398 | While restart behaviour is up to the event loop implementation, most |
|
|
399 | will not restart syscalls (that includes Async::Interrupt and AnyEvent's |
|
|
400 | pure perl implementation). |
|
|
401 | |
|
|
402 | Safe/Unsafe Signals |
|
|
403 | Perl signals can be either "safe" (synchronous to opcode handling) or |
|
|
404 | "unsafe" (asynchronous) - the former might get delayed indefinitely, the |
|
|
405 | latter might corrupt your memory. |
|
|
406 | |
|
|
407 | AnyEvent signal handlers are, in addition, synchronous to the event |
|
|
408 | loop, i.e. they will not interrupt your running perl program but will |
|
|
409 | only be called as part of the normal event handling (just like timer, |
|
|
410 | I/O etc. callbacks, too). |
388 | |
411 | |
389 | Signal Races, Delays and Workarounds |
412 | Signal Races, Delays and Workarounds |
390 | Many event loops (e.g. Glib, Tk, Qt, IO::Async) do not support attaching |
413 | Many event loops (e.g. Glib, Tk, Qt, IO::Async) do not support attaching |
391 | callbacks to signals in a generic way, which is a pity, as you cannot do |
414 | callbacks to signals in a generic way, which is a pity, as you cannot do |
392 | race-free signal handling in perl, requiring C libraries for this. |
415 | race-free signal handling in perl, requiring C libraries for this. |
… | |
… | |
463 | $done->recv; |
486 | $done->recv; |
464 | |
487 | |
465 | IDLE WATCHERS |
488 | IDLE WATCHERS |
466 | $w = AnyEvent->idle (cb => <callback>); |
489 | $w = AnyEvent->idle (cb => <callback>); |
467 | |
490 | |
468 | Sometimes there is a need to do something, but it is not so important to |
491 | Repeatedly invoke the callback after the process becomes idle, until |
469 | do it instantly, but only when there is nothing better to do. This |
492 | either the watcher is destroyed or new events have been detected. |
470 | "nothing better to do" is usually defined to be "no other events need |
|
|
471 | attention by the event loop". |
|
|
472 | |
493 | |
473 | Idle watchers ideally get invoked when the event loop has nothing better |
494 | Idle watchers are useful when there is a need to do something, but it is |
474 | to do, just before it would block the process to wait for new events. |
495 | not so important (or wise) to do it instantly. The callback will be |
475 | Instead of blocking, the idle watcher is invoked. |
496 | invoked only when there is "nothing better to do", which is usually |
|
|
497 | defined as "all outstanding events have been handled and no new events |
|
|
498 | have been detected". That means that idle watchers ideally get invoked |
|
|
499 | when the event loop has just polled for new events but none have been |
|
|
500 | detected. Instead of blocking to wait for more events, the idle watchers |
|
|
501 | will be invoked. |
476 | |
502 | |
477 | Most event loops unfortunately do not really support idle watchers (only |
503 | Unfortunately, most event loops do not really support idle watchers |
478 | EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent |
504 | (only EV, Event and Glib do it in a usable fashion) - for the rest, |
479 | will simply call the callback "from time to time". |
505 | AnyEvent will simply call the callback "from time to time". |
480 | |
506 | |
481 | Example: read lines from STDIN, but only process them when the program |
507 | Example: read lines from STDIN, but only process them when the program |
482 | is otherwise idle: |
508 | is otherwise idle: |
483 | |
509 | |
484 | my @lines; # read data |
510 | my @lines; # read data |
… | |
… | |
576 | after => 1, |
602 | after => 1, |
577 | cb => sub { $result_ready->send }, |
603 | cb => sub { $result_ready->send }, |
578 | ); |
604 | ); |
579 | |
605 | |
580 | # this "blocks" (while handling events) till the callback |
606 | # this "blocks" (while handling events) till the callback |
581 | # calls -<send |
607 | # calls ->send |
582 | $result_ready->recv; |
608 | $result_ready->recv; |
583 | |
609 | |
584 | Example: wait for a timer, but take advantage of the fact that condition |
610 | Example: wait for a timer, but take advantage of the fact that condition |
585 | variables are also callable directly. |
611 | variables are also callable directly. |
586 | |
612 | |
… | |
… | |
643 | into one. For example, a function that pings many hosts in parallel |
669 | into one. For example, a function that pings many hosts in parallel |
644 | might want to use a condition variable for the whole process. |
670 | might want to use a condition variable for the whole process. |
645 | |
671 | |
646 | Every call to "->begin" will increment a counter, and every call to |
672 | Every call to "->begin" will increment a counter, and every call to |
647 | "->end" will decrement it. If the counter reaches 0 in "->end", the |
673 | "->end" will decrement it. If the counter reaches 0 in "->end", the |
648 | (last) callback passed to "begin" will be executed. That callback is |
674 | (last) callback passed to "begin" will be executed, passing the |
649 | *supposed* to call "->send", but that is not required. If no |
675 | condvar as first argument. That callback is *supposed* to call |
|
|
676 | "->send", but that is not required. If no group callback was set, |
650 | callback was set, "send" will be called without any arguments. |
677 | "send" will be called without any arguments. |
651 | |
678 | |
652 | You can think of "$cv->send" giving you an OR condition (one call |
679 | You can think of "$cv->send" giving you an OR condition (one call |
653 | sends), while "$cv->begin" and "$cv->end" giving you an AND |
680 | sends), while "$cv->begin" and "$cv->end" giving you an AND |
654 | condition (all "begin" calls must be "end"'ed before the condvar |
681 | condition (all "begin" calls must be "end"'ed before the condvar |
655 | sends). |
682 | sends). |
… | |
… | |
683 | that are begung can potentially be zero: |
710 | that are begung can potentially be zero: |
684 | |
711 | |
685 | my $cv = AnyEvent->condvar; |
712 | my $cv = AnyEvent->condvar; |
686 | |
713 | |
687 | my %result; |
714 | my %result; |
688 | $cv->begin (sub { $cv->send (\%result) }); |
715 | $cv->begin (sub { shift->send (\%result) }); |
689 | |
716 | |
690 | for my $host (@list_of_hosts) { |
717 | for my $host (@list_of_hosts) { |
691 | $cv->begin; |
718 | $cv->begin; |
692 | ping_host_then_call_callback $host, sub { |
719 | ping_host_then_call_callback $host, sub { |
693 | $result{$host} = ...; |
720 | $result{$host} = ...; |
… | |
… | |
771 | SUPPORTED EVENT LOOPS/BACKENDS |
798 | SUPPORTED EVENT LOOPS/BACKENDS |
772 | The available backend classes are (every class has its own manpage): |
799 | The available backend classes are (every class has its own manpage): |
773 | |
800 | |
774 | Backends that are autoprobed when no other event loop can be found. |
801 | Backends that are autoprobed when no other event loop can be found. |
775 | EV is the preferred backend when no other event loop seems to be in |
802 | EV is the preferred backend when no other event loop seems to be in |
776 | use. If EV is not installed, then AnyEvent will try Event, and, |
803 | use. If EV is not installed, then AnyEvent will fall back to its own |
777 | failing that, will fall back to its own pure-perl implementation, |
804 | pure-perl implementation, which is available everywhere as it comes |
778 | which is available everywhere as it comes with AnyEvent itself. |
805 | with AnyEvent itself. |
779 | |
806 | |
780 | AnyEvent::Impl::EV based on EV (interface to libev, best choice). |
807 | AnyEvent::Impl::EV based on EV (interface to libev, best choice). |
781 | AnyEvent::Impl::Event based on Event, very stable, few glitches. |
|
|
782 | AnyEvent::Impl::Perl pure-perl implementation, fast and portable. |
808 | AnyEvent::Impl::Perl pure-perl implementation, fast and portable. |
783 | |
809 | |
784 | Backends that are transparently being picked up when they are used. |
810 | Backends that are transparently being picked up when they are used. |
785 | These will be used when they are currently loaded when the first |
811 | These will be used when they are currently loaded when the first |
786 | watcher is created, in which case it is assumed that the application |
812 | watcher is created, in which case it is assumed that the application |
787 | is using them. This means that AnyEvent will automatically pick the |
813 | is using them. This means that AnyEvent will automatically pick the |
788 | right backend when the main program loads an event module before |
814 | right backend when the main program loads an event module before |
789 | anything starts to create watchers. Nothing special needs to be done |
815 | anything starts to create watchers. Nothing special needs to be done |
790 | by the main program. |
816 | by the main program. |
791 | |
817 | |
|
|
818 | AnyEvent::Impl::Event based on Event, very stable, few glitches. |
792 | AnyEvent::Impl::Glib based on Glib, slow but very stable. |
819 | AnyEvent::Impl::Glib based on Glib, slow but very stable. |
793 | AnyEvent::Impl::Tk based on Tk, very broken. |
820 | AnyEvent::Impl::Tk based on Tk, very broken. |
794 | AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse. |
821 | AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse. |
795 | AnyEvent::Impl::POE based on POE, very slow, some limitations. |
822 | AnyEvent::Impl::POE based on POE, very slow, some limitations. |
796 | AnyEvent::Impl::Irssi used when running within irssi. |
823 | AnyEvent::Impl::Irssi used when running within irssi. |
… | |
… | |
895 | You should check $AnyEvent::MODEL before adding to this array, |
922 | You should check $AnyEvent::MODEL before adding to this array, |
896 | though: if it is defined then the event loop has already been |
923 | though: if it is defined then the event loop has already been |
897 | detected, and the array will be ignored. |
924 | detected, and the array will be ignored. |
898 | |
925 | |
899 | Best use "AnyEvent::post_detect { BLOCK }" when your application |
926 | Best use "AnyEvent::post_detect { BLOCK }" when your application |
900 | allows it,as it takes care of these details. |
927 | allows it, as it takes care of these details. |
901 | |
928 | |
902 | This variable is mainly useful for modules that can do something |
929 | This variable is mainly useful for modules that can do something |
903 | useful when AnyEvent is used and thus want to know when it is |
930 | useful when AnyEvent is used and thus want to know when it is |
904 | initialised, but do not need to even load it by default. This array |
931 | initialised, but do not need to even load it by default. This array |
905 | provides the means to hook into AnyEvent passively, without loading |
932 | provides the means to hook into AnyEvent passively, without loading |
906 | it. |
933 | it. |
|
|
934 | |
|
|
935 | Example: To load Coro::AnyEvent whenever Coro and AnyEvent are used |
|
|
936 | together, you could put this into Coro (this is the actual code used |
|
|
937 | by Coro to accomplish this): |
|
|
938 | |
|
|
939 | if (defined $AnyEvent::MODEL) { |
|
|
940 | # AnyEvent already initialised, so load Coro::AnyEvent |
|
|
941 | require Coro::AnyEvent; |
|
|
942 | } else { |
|
|
943 | # AnyEvent not yet initialised, so make sure to load Coro::AnyEvent |
|
|
944 | # as soon as it is |
|
|
945 | push @AnyEvent::post_detect, sub { require Coro::AnyEvent }; |
|
|
946 | } |
907 | |
947 | |
908 | WHAT TO DO IN A MODULE |
948 | WHAT TO DO IN A MODULE |
909 | As a module author, you should "use AnyEvent" and call AnyEvent methods |
949 | As a module author, you should "use AnyEvent" and call AnyEvent methods |
910 | freely, but you should not load a specific event module or rely on it. |
950 | freely, but you should not load a specific event module or rely on it. |
911 | |
951 | |
… | |
… | |
1032 | Event::ExecFlow |
1072 | Event::ExecFlow |
1033 | High level API for event-based execution flow control. |
1073 | High level API for event-based execution flow control. |
1034 | |
1074 | |
1035 | Coro |
1075 | Coro |
1036 | Has special support for AnyEvent via Coro::AnyEvent. |
1076 | Has special support for AnyEvent via Coro::AnyEvent. |
|
|
1077 | |
|
|
1078 | SIMPLIFIED AE API |
|
|
1079 | Starting with version 5.0, AnyEvent officially supports a second, much |
|
|
1080 | simpler, API that is designed to reduce the calling, typing and memory |
|
|
1081 | overhead. |
|
|
1082 | |
|
|
1083 | See the AE manpage for details. |
1037 | |
1084 | |
1038 | ERROR AND EXCEPTION HANDLING |
1085 | ERROR AND EXCEPTION HANDLING |
1039 | In general, AnyEvent does not do any error handling - it relies on the |
1086 | In general, AnyEvent does not do any error handling - it relies on the |
1040 | caller to do that if required. The AnyEvent::Strict module (see also the |
1087 | caller to do that if required. The AnyEvent::Strict module (see also the |
1041 | "PERL_ANYEVENT_STRICT" environment variable, below) provides strict |
1088 | "PERL_ANYEVENT_STRICT" environment variable, below) provides strict |
… | |
… | |
1220 | warn "read: $input\n"; # output what has been read |
1267 | warn "read: $input\n"; # output what has been read |
1221 | $cv->send if $input =~ /^q/i; # quit program if /^q/i |
1268 | $cv->send if $input =~ /^q/i; # quit program if /^q/i |
1222 | }, |
1269 | }, |
1223 | ); |
1270 | ); |
1224 | |
1271 | |
1225 | my $time_watcher; # can only be used once |
|
|
1226 | |
|
|
1227 | sub new_timer { |
|
|
1228 | $timer = AnyEvent->timer (after => 1, cb => sub { |
1272 | my $time_watcher = AnyEvent->timer (after => 1, interval => 1, cb => sub { |
1229 | warn "timeout\n"; # print 'timeout' about every second |
1273 | warn "timeout\n"; # print 'timeout' at most every second |
1230 | &new_timer; # and restart the time |
|
|
1231 | }); |
|
|
1232 | } |
1274 | }); |
1233 | |
|
|
1234 | new_timer; # create first timer |
|
|
1235 | |
1275 | |
1236 | $cv->recv; # wait until user enters /^q/i |
1276 | $cv->recv; # wait until user enters /^q/i |
1237 | |
1277 | |
1238 | REAL-WORLD EXAMPLE |
1278 | REAL-WORLD EXAMPLE |
1239 | Consider the Net::FCP module. It features (among others) the following |
1279 | Consider the Net::FCP module. It features (among others) the following |
… | |
… | |
1366 | through AnyEvent. The benchmark creates a lot of timers (with a zero |
1406 | through AnyEvent. The benchmark creates a lot of timers (with a zero |
1367 | timeout) and I/O watchers (watching STDOUT, a pty, to become writable, |
1407 | timeout) and I/O watchers (watching STDOUT, a pty, to become writable, |
1368 | which it is), lets them fire exactly once and destroys them again. |
1408 | which it is), lets them fire exactly once and destroys them again. |
1369 | |
1409 | |
1370 | Source code for this benchmark is found as eg/bench in the AnyEvent |
1410 | Source code for this benchmark is found as eg/bench in the AnyEvent |
1371 | distribution. |
1411 | distribution. It uses the AE interface, which makes a real difference |
|
|
1412 | for the EV and Perl backends only. |
1372 | |
1413 | |
1373 | Explanation of the columns |
1414 | Explanation of the columns |
1374 | *watcher* is the number of event watchers created/destroyed. Since |
1415 | *watcher* is the number of event watchers created/destroyed. Since |
1375 | different event models feature vastly different performances, each event |
1416 | different event models feature vastly different performances, each event |
1376 | loop was given a number of watchers so that overall runtime is |
1417 | loop was given a number of watchers so that overall runtime is |
… | |
… | |
1395 | *destroy* is the time, in microseconds, that it takes to destroy a |
1436 | *destroy* is the time, in microseconds, that it takes to destroy a |
1396 | single watcher. |
1437 | single watcher. |
1397 | |
1438 | |
1398 | Results |
1439 | Results |
1399 | name watchers bytes create invoke destroy comment |
1440 | name watchers bytes create invoke destroy comment |
1400 | EV/EV 400000 224 0.47 0.35 0.27 EV native interface |
1441 | EV/EV 100000 223 0.47 0.43 0.27 EV native interface |
1401 | EV/Any 100000 224 2.88 0.34 0.27 EV + AnyEvent watchers |
1442 | EV/Any 100000 223 0.48 0.42 0.26 EV + AnyEvent watchers |
1402 | CoroEV/Any 100000 224 2.85 0.35 0.28 coroutines + Coro::Signal |
1443 | Coro::EV/Any 100000 223 0.47 0.42 0.26 coroutines + Coro::Signal |
1403 | Perl/Any 100000 452 4.13 0.73 0.95 pure perl implementation |
1444 | Perl/Any 100000 431 2.70 0.74 0.92 pure perl implementation |
1404 | Event/Event 16000 517 32.20 31.80 0.81 Event native interface |
1445 | Event/Event 16000 516 31.16 31.84 0.82 Event native interface |
1405 | Event/Any 16000 590 35.85 31.55 1.06 Event + AnyEvent watchers |
1446 | Event/Any 16000 1203 42.61 34.79 1.80 Event + AnyEvent watchers |
1406 | IOAsync/Any 16000 989 38.10 32.77 11.13 via IO::Async::Loop::IO_Poll |
1447 | IOAsync/Any 16000 1911 41.92 27.45 16.81 via IO::Async::Loop::IO_Poll |
1407 | IOAsync/Any 16000 990 37.59 29.50 10.61 via IO::Async::Loop::Epoll |
1448 | IOAsync/Any 16000 1726 40.69 26.37 15.25 via IO::Async::Loop::Epoll |
1408 | Glib/Any 16000 1357 102.33 12.31 51.00 quadratic behaviour |
1449 | Glib/Any 16000 1118 89.00 12.57 51.17 quadratic behaviour |
1409 | Tk/Any 2000 1860 27.20 66.31 14.00 SEGV with >> 2000 watchers |
1450 | Tk/Any 2000 1346 20.96 10.75 8.00 SEGV with >> 2000 watchers |
1410 | POE/Event 2000 6328 109.99 751.67 14.02 via POE::Loop::Event |
1451 | POE/Any 2000 6951 108.97 795.32 14.24 via POE::Loop::Event |
1411 | POE/Select 2000 6027 94.54 809.13 579.80 via POE::Loop::Select |
1452 | POE/Any 2000 6648 94.79 774.40 575.51 via POE::Loop::Select |
1412 | |
1453 | |
1413 | Discussion |
1454 | Discussion |
1414 | The benchmark does *not* measure scalability of the event loop very |
1455 | The benchmark does *not* measure scalability of the event loop very |
1415 | well. For example, a select-based event loop (such as the pure perl one) |
1456 | well. For example, a select-based event loop (such as the pure perl one) |
1416 | can never compete with an event loop that uses epoll when the number of |
1457 | can never compete with an event loop that uses epoll when the number of |
… | |
… | |
1427 | benchmark machine, handling an event takes roughly 1600 CPU cycles with |
1468 | benchmark machine, handling an event takes roughly 1600 CPU cycles with |
1428 | EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 |
1469 | EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 |
1429 | CPU cycles with POE. |
1470 | CPU cycles with POE. |
1430 | |
1471 | |
1431 | "EV" is the sole leader regarding speed and memory use, which are both |
1472 | "EV" is the sole leader regarding speed and memory use, which are both |
1432 | maximal/minimal, respectively. Even when going through AnyEvent, it uses |
1473 | maximal/minimal, respectively. When using the AE API there is zero |
|
|
1474 | overhead (when going through the AnyEvent API create is about 5-6 times |
|
|
1475 | slower, with other times being equal, so still uses far less memory than |
1433 | far less memory than any other event loop and is still faster than Event |
1476 | any other event loop and is still faster than Event natively). |
1434 | natively. |
|
|
1435 | |
1477 | |
1436 | The pure perl implementation is hit in a few sweet spots (both the |
1478 | The pure perl implementation is hit in a few sweet spots (both the |
1437 | constant timeout and the use of a single fd hit optimisations in the |
1479 | constant timeout and the use of a single fd hit optimisations in the |
1438 | perl interpreter and the backend itself). Nevertheless this shows that |
1480 | perl interpreter and the backend itself). Nevertheless this shows that |
1439 | it adds very little overhead in itself. Like any select-based backend |
1481 | it adds very little overhead in itself. Like any select-based backend |
… | |
… | |
1509 | In this benchmark, we use 10000 socket pairs (20000 sockets), of which |
1551 | In this benchmark, we use 10000 socket pairs (20000 sockets), of which |
1510 | 100 (1%) are active. This mirrors the activity of large servers with |
1552 | 100 (1%) are active. This mirrors the activity of large servers with |
1511 | many connections, most of which are idle at any one point in time. |
1553 | many connections, most of which are idle at any one point in time. |
1512 | |
1554 | |
1513 | Source code for this benchmark is found as eg/bench2 in the AnyEvent |
1555 | Source code for this benchmark is found as eg/bench2 in the AnyEvent |
1514 | distribution. |
1556 | distribution. It uses the AE interface, which makes a real difference |
|
|
1557 | for the EV and Perl backends only. |
1515 | |
1558 | |
1516 | Explanation of the columns |
1559 | Explanation of the columns |
1517 | *sockets* is the number of sockets, and twice the number of "servers" |
1560 | *sockets* is the number of sockets, and twice the number of "servers" |
1518 | (as each server has a read and write socket end). |
1561 | (as each server has a read and write socket end). |
1519 | |
1562 | |
… | |
… | |
1525 | forwarding it to another server. This includes deleting the old timeout |
1568 | forwarding it to another server. This includes deleting the old timeout |
1526 | and creating a new one that moves the timeout into the future. |
1569 | and creating a new one that moves the timeout into the future. |
1527 | |
1570 | |
1528 | Results |
1571 | Results |
1529 | name sockets create request |
1572 | name sockets create request |
1530 | EV 20000 69.01 11.16 |
1573 | EV 20000 62.66 7.99 |
1531 | Perl 20000 73.32 35.87 |
1574 | Perl 20000 68.32 32.64 |
1532 | IOAsync 20000 157.00 98.14 epoll |
1575 | IOAsync 20000 174.06 101.15 epoll |
1533 | IOAsync 20000 159.31 616.06 poll |
1576 | IOAsync 20000 174.67 610.84 poll |
1534 | Event 20000 212.62 257.32 |
1577 | Event 20000 202.69 242.91 |
1535 | Glib 20000 651.16 1896.30 |
1578 | Glib 20000 557.01 1689.52 |
1536 | POE 20000 349.67 12317.24 uses POE::Loop::Event |
1579 | POE 20000 341.54 12086.32 uses POE::Loop::Event |
1537 | |
1580 | |
1538 | Discussion |
1581 | Discussion |
1539 | This benchmark *does* measure scalability and overall performance of the |
1582 | This benchmark *does* measure scalability and overall performance of the |
1540 | particular event loop. |
1583 | particular event loop. |
1541 | |
1584 | |
… | |
… | |
1654 | As you can see, the AnyEvent + EV combination even beats the |
1697 | As you can see, the AnyEvent + EV combination even beats the |
1655 | hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl |
1698 | hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl |
1656 | backend easily beats IO::Lambda and POE. |
1699 | backend easily beats IO::Lambda and POE. |
1657 | |
1700 | |
1658 | And even the 100% non-blocking version written using the high-level (and |
1701 | And even the 100% non-blocking version written using the high-level (and |
1659 | slow :) AnyEvent::Handle abstraction beats both POE and IO::Lambda by a |
1702 | slow :) AnyEvent::Handle abstraction beats both POE and IO::Lambda |
1660 | large margin, even though it does all of DNS, tcp-connect and socket I/O |
1703 | higher level ("unoptimised") abstractions by a large margin, even though |
1661 | in a non-blocking way. |
1704 | it does all of DNS, tcp-connect and socket I/O in a non-blocking way. |
1662 | |
1705 | |
1663 | The two AnyEvent benchmarks programs can be found as eg/ae0.pl and |
1706 | The two AnyEvent benchmarks programs can be found as eg/ae0.pl and |
1664 | eg/ae2.pl in the AnyEvent distribution, the remaining benchmarks are |
1707 | eg/ae2.pl in the AnyEvent distribution, the remaining benchmarks are |
1665 | part of the IO::lambda distribution and were used without any changes. |
1708 | part of the IO::Lambda distribution and were used without any changes. |
1666 | |
1709 | |
1667 | SIGNALS |
1710 | SIGNALS |
1668 | AnyEvent currently installs handlers for these signals: |
1711 | AnyEvent currently installs handlers for these signals: |
1669 | |
1712 | |
1670 | SIGCHLD |
1713 | SIGCHLD |
… | |
… | |
1697 | it's built-in modules) are required to use it. |
1740 | it's built-in modules) are required to use it. |
1698 | |
1741 | |
1699 | That does not mean that AnyEvent won't take advantage of some additional |
1742 | That does not mean that AnyEvent won't take advantage of some additional |
1700 | modules if they are installed. |
1743 | modules if they are installed. |
1701 | |
1744 | |
1702 | This section epxlains which additional modules will be used, and how |
1745 | This section explains which additional modules will be used, and how |
1703 | they affect AnyEvent's operetion. |
1746 | they affect AnyEvent's operation. |
1704 | |
1747 | |
1705 | Async::Interrupt |
1748 | Async::Interrupt |
1706 | This slightly arcane module is used to implement fast signal |
1749 | This slightly arcane module is used to implement fast signal |
1707 | handling: To my knowledge, there is no way to do completely |
1750 | handling: To my knowledge, there is no way to do completely |
1708 | race-free and quick signal handling in pure perl. To ensure that |
1751 | race-free and quick signal handling in pure perl. To ensure that |
… | |
… | |
1711 | 10 seconds, look for $AnyEvent::MAX_SIGNAL_LATENCY). |
1754 | 10 seconds, look for $AnyEvent::MAX_SIGNAL_LATENCY). |
1712 | |
1755 | |
1713 | If this module is available, then it will be used to implement |
1756 | If this module is available, then it will be used to implement |
1714 | signal catching, which means that signals will not be delayed, and |
1757 | signal catching, which means that signals will not be delayed, and |
1715 | the event loop will not be interrupted regularly, which is more |
1758 | the event loop will not be interrupted regularly, which is more |
1716 | efficient (And good for battery life on laptops). |
1759 | efficient (and good for battery life on laptops). |
1717 | |
1760 | |
1718 | This affects not just the pure-perl event loop, but also other event |
1761 | This affects not just the pure-perl event loop, but also other event |
1719 | loops that have no signal handling on their own (e.g. Glib, Tk, Qt). |
1762 | loops that have no signal handling on their own (e.g. Glib, Tk, Qt). |
1720 | |
1763 | |
1721 | Some event loops (POE, Event, Event::Lib) offer signal watchers |
1764 | Some event loops (POE, Event, Event::Lib) offer signal watchers |
… | |
… | |
1738 | "AnyEvent::Util::guard". This speeds up guards considerably (and |
1781 | "AnyEvent::Util::guard". This speeds up guards considerably (and |
1739 | uses a lot less memory), but otherwise doesn't affect guard |
1782 | uses a lot less memory), but otherwise doesn't affect guard |
1740 | operation much. It is purely used for performance. |
1783 | operation much. It is purely used for performance. |
1741 | |
1784 | |
1742 | JSON and JSON::XS |
1785 | JSON and JSON::XS |
1743 | This module is required when you want to read or write JSON data via |
1786 | One of these modules is required when you want to read or write JSON |
1744 | AnyEvent::Handle. It is also written in pure-perl, but can take |
1787 | data via AnyEvent::Handle. It is also written in pure-perl, but can |
1745 | advantage of the ultra-high-speed JSON::XS module when it is |
1788 | take advantage of the ultra-high-speed JSON::XS module when it is |
1746 | installed. |
1789 | installed. |
1747 | |
1790 | |
1748 | In fact, AnyEvent::Handle will use JSON::XS by default if it is |
1791 | In fact, AnyEvent::Handle will use JSON::XS by default if it is |
1749 | installed. |
1792 | installed. |
1750 | |
1793 | |
… | |
… | |
1760 | additionally use it to try to use a monotonic clock for timing |
1803 | additionally use it to try to use a monotonic clock for timing |
1761 | stability. |
1804 | stability. |
1762 | |
1805 | |
1763 | FORK |
1806 | FORK |
1764 | Most event libraries are not fork-safe. The ones who are usually are |
1807 | Most event libraries are not fork-safe. The ones who are usually are |
1765 | because they rely on inefficient but fork-safe "select" or "poll" calls. |
1808 | because they rely on inefficient but fork-safe "select" or "poll" calls |
1766 | Only EV is fully fork-aware. |
1809 | - higher performance APIs such as BSD's kqueue or the dreaded Linux |
|
|
1810 | epoll are usually badly thought-out hacks that are incompatible with |
|
|
1811 | fork in one way or another. Only EV is fully fork-aware and ensures that |
|
|
1812 | you continue event-processing in both parent and child (or both, if you |
|
|
1813 | know what you are doing). |
|
|
1814 | |
|
|
1815 | This means that, in general, you cannot fork and do event processing in |
|
|
1816 | the child if the event library was initialised before the fork (which |
|
|
1817 | usually happens when the first AnyEvent watcher is created, or the |
|
|
1818 | library is loaded). |
1767 | |
1819 | |
1768 | If you have to fork, you must either do so *before* creating your first |
1820 | If you have to fork, you must either do so *before* creating your first |
1769 | watcher OR you must not use AnyEvent at all in the child OR you must do |
1821 | watcher OR you must not use AnyEvent at all in the child OR you must do |
1770 | something completely out of the scope of AnyEvent. |
1822 | something completely out of the scope of AnyEvent. |
|
|
1823 | |
|
|
1824 | The problem of doing event processing in the parent *and* the child is |
|
|
1825 | much more complicated: even for backends that *are* fork-aware or |
|
|
1826 | fork-safe, their behaviour is not usually what you want: fork clones all |
|
|
1827 | watchers, that means all timers, I/O watchers etc. are active in both |
|
|
1828 | parent and child, which is almost never what you want. USing "exec" to |
|
|
1829 | start worker children from some kind of manage rprocess is usually |
|
|
1830 | preferred, because it is much easier and cleaner, at the expense of |
|
|
1831 | having to have another binary. |
1771 | |
1832 | |
1772 | SECURITY CONSIDERATIONS |
1833 | SECURITY CONSIDERATIONS |
1773 | AnyEvent can be forced to load any event model via |
1834 | AnyEvent can be forced to load any event model via |
1774 | $ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used |
1835 | $ENV{PERL_ANYEVENT_MODEL}. While this cannot (to my knowledge) be used |
1775 | to execute arbitrary code or directly gain access, it can easily be used |
1836 | to execute arbitrary code or directly gain access, it can easily be used |