ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent-MP/MP.pm
(Generate patch)

Comparing AnyEvent-MP/MP.pm (file contents):
Revision 1.81 by root, Mon Sep 7 18:33:44 2009 UTC vs.
Revision 1.105 by root, Sun Nov 8 23:58:02 2009 UTC

1=head1 NAME 1=head1 NAME
2 2
3AnyEvent::MP - multi-processing/message-passing framework 3AnyEvent::MP - erlang-style multi-processing/message-passing framework
4 4
5=head1 SYNOPSIS 5=head1 SYNOPSIS
6 6
7 use AnyEvent::MP; 7 use AnyEvent::MP;
8 8
30 rcv $port, pong => sub { warn "pong received\n" }; 30 rcv $port, pong => sub { warn "pong received\n" };
31 31
32 # create a port on another node 32 # create a port on another node
33 my $port = spawn $node, $initfunc, @initdata; 33 my $port = spawn $node, $initfunc, @initdata;
34 34
35 # destroy a prot again
36 kil $port; # "normal" kill
37 kil $port, my_error => "everything is broken"; # error kill
38
35 # monitoring 39 # monitoring
36 mon $port, $cb->(@msg) # callback is invoked on death 40 mon $localport, $cb->(@msg) # callback is invoked on death
37 mon $port, $otherport # kill otherport on abnormal death 41 mon $localport, $otherport # kill otherport on abnormal death
38 mon $port, $otherport, @msg # send message on death 42 mon $localport, $otherport, @msg # send message on death
43
44 # temporarily execute code in port context
45 peval $port, sub { die "kill the port!" };
46
47 # execute callbacks in $SELF port context
48 my $timer = AE::timer 1, 0, psub {
49 die "kill the port, delayed";
50 };
39 51
40=head1 CURRENT STATUS 52=head1 CURRENT STATUS
41 53
42 bin/aemp - stable. 54 bin/aemp - stable.
43 AnyEvent::MP - stable API, should work. 55 AnyEvent::MP - stable API, should work.
44 AnyEvent::MP::Intro - explains most concepts. 56 AnyEvent::MP::Intro - explains most concepts.
45 AnyEvent::MP::Kernel - mostly stable. 57 AnyEvent::MP::Kernel - mostly stable API.
46 AnyEvent::MP::Global - stable but incomplete, protocol not yet final. 58 AnyEvent::MP::Global - stable API.
47
48stay tuned.
49 59
50=head1 DESCRIPTION 60=head1 DESCRIPTION
51 61
52This module (-family) implements a simple message passing framework. 62This module (-family) implements a simple message passing framework.
53 63
83 93
84Nodes are either public (have one or more listening ports) or private 94Nodes are either public (have one or more listening ports) or private
85(no listening ports). Private nodes cannot talk to other private nodes 95(no listening ports). Private nodes cannot talk to other private nodes
86currently. 96currently.
87 97
88=item node ID - C<[a-za-Z0-9_\-.:]+> 98=item node ID - C<[A-Z_][a-zA-Z0-9_\-.:]*>
89 99
90A node ID is a string that uniquely identifies the node within a 100A node ID is a string that uniquely identifies the node within a
91network. Depending on the configuration used, node IDs can look like a 101network. Depending on the configuration used, node IDs can look like a
92hostname, a hostname and a port, or a random string. AnyEvent::MP itself 102hostname, a hostname and a port, or a random string. AnyEvent::MP itself
93doesn't interpret node IDs in any way. 103doesn't interpret node IDs in any way.
97Nodes can only talk to each other by creating some kind of connection to 107Nodes can only talk to each other by creating some kind of connection to
98each other. To do this, nodes should listen on one or more local transport 108each other. To do this, nodes should listen on one or more local transport
99endpoints - binds. Currently, only standard C<ip:port> specifications can 109endpoints - binds. Currently, only standard C<ip:port> specifications can
100be used, which specify TCP ports to listen on. 110be used, which specify TCP ports to listen on.
101 111
102=item seeds - C<host:port> 112=item seed nodes
103 113
104When a node starts, it knows nothing about the network. To teach the node 114When a node starts, it knows nothing about the network. To teach the node
105about the network it first has to contact some other node within the 115about the network it first has to contact some other node within the
106network. This node is called a seed. 116network. This node is called a seed.
107 117
108Seeds are transport endpoint(s) of as many nodes as one wants. Those nodes 118Apart from the fact that other nodes know them as seed nodes and they have
119to have fixed listening addresses, seed nodes are perfectly normal nodes -
120any node can function as a seed node for others.
121
122In addition to discovering the network, seed nodes are also used to
123maintain the network and to connect nodes that otherwise would have
124trouble connecting. They form the backbone of an AnyEvent::MP network.
125
109are expected to be long-running, and at least one of those should always 126Seed nodes are expected to be long-running, and at least one seed node
110be available. When nodes run out of connections (e.g. due to a network 127should always be available. They should also be relatively responsive - a
111error), they try to re-establish connections to some seednodes again to 128seed node that blocks for long periods will slow down everybody else.
112join the network.
113 129
114Apart from being sued for seeding, seednodes are not special in any way - 130=item seeds - C<host:port>
115every public node can be a seednode. 131
132Seeds are transport endpoint(s) (usually a hostname/IP address and a
133TCP port) of nodes that should be used as seed nodes.
134
135The nodes listening on those endpoints are expected to be long-running,
136and at least one of those should always be available. When nodes run out
137of connections (e.g. due to a network error), they try to re-establish
138connections to some seednodes again to join the network.
116 139
117=back 140=back
118 141
119=head1 VARIABLES/FUNCTIONS 142=head1 VARIABLES/FUNCTIONS
120 143
132 155
133use AE (); 156use AE ();
134 157
135use base "Exporter"; 158use base "Exporter";
136 159
137our $VERSION = $AnyEvent::MP::Kernel::VERSION; 160our $VERSION = 1.23;
138 161
139our @EXPORT = qw( 162our @EXPORT = qw(
140 NODE $NODE *SELF node_of after 163 NODE $NODE *SELF node_of after
141 configure 164 configure
142 snd rcv mon mon_guard kil reg psub spawn 165 snd rcv mon mon_guard kil psub peval spawn cal
143 port 166 port
144); 167);
145 168
146our $SELF; 169our $SELF;
147 170
216L<AnyEvent::MP::Global> module, which will then use it to keep 239L<AnyEvent::MP::Global> module, which will then use it to keep
217connectivity with at least one node at any point in time. 240connectivity with at least one node at any point in time.
218 241
219=back 242=back
220 243
221Example: become a distributed node using the locla node name as profile. 244Example: become a distributed node using the local node name as profile.
222This should be the most common form of invocation for "daemon"-type nodes. 245This should be the most common form of invocation for "daemon"-type nodes.
223 246
224 configure 247 configure
225 248
226Example: become an anonymous node. This form is often used for commandline 249Example: become an anonymous node. This form is often used for commandline
360 msg1 => sub { ... }, 383 msg1 => sub { ... },
361 ... 384 ...
362 ; 385 ;
363 386
364Example: temporarily register a rcv callback for a tag matching some port 387Example: temporarily register a rcv callback for a tag matching some port
365(e.g. for a rpc reply) and unregister it after a message was received. 388(e.g. for an rpc reply) and unregister it after a message was received.
366 389
367 rcv $port, $otherport => sub { 390 rcv $port, $otherport => sub {
368 my @reply = @_; 391 my @reply = @_;
369 392
370 rcv $SELF, $otherport; 393 rcv $SELF, $otherport;
383 if (ref $_[0]) { 406 if (ref $_[0]) {
384 if (my $self = $PORT_DATA{$portid}) { 407 if (my $self = $PORT_DATA{$portid}) {
385 "AnyEvent::MP::Port" eq ref $self 408 "AnyEvent::MP::Port" eq ref $self
386 or Carp::croak "$port: rcv can only be called on message matching ports, caught"; 409 or Carp::croak "$port: rcv can only be called on message matching ports, caught";
387 410
388 $self->[2] = shift; 411 $self->[0] = shift;
389 } else { 412 } else {
390 my $cb = shift; 413 my $cb = shift;
391 $PORT{$portid} = sub { 414 $PORT{$portid} = sub {
392 local $SELF = $port; 415 local $SELF = $port;
393 eval { &$cb }; _self_die if $@; 416 eval { &$cb }; _self_die if $@;
394 }; 417 };
395 } 418 }
396 } elsif (defined $_[0]) { 419 } elsif (defined $_[0]) {
397 my $self = $PORT_DATA{$portid} ||= do { 420 my $self = $PORT_DATA{$portid} ||= do {
398 my $self = bless [$PORT{$port} || sub { }, { }, $port], "AnyEvent::MP::Port"; 421 my $self = bless [$PORT{$portid} || sub { }, { }, $port], "AnyEvent::MP::Port";
399 422
400 $PORT{$portid} = sub { 423 $PORT{$portid} = sub {
401 local $SELF = $port; 424 local $SELF = $port;
402 425
403 if (my $cb = $self->[1]{$_[0]}) { 426 if (my $cb = $self->[1]{$_[0]}) {
425 } 448 }
426 449
427 $port 450 $port
428} 451}
429 452
453=item peval $port, $coderef[, @args]
454
455Evaluates the given C<$codref> within the contetx of C<$port>, that is,
456when the code throews an exception the C<$port> will be killed.
457
458Any remaining args will be passed to the callback. Any return values will
459be returned to the caller.
460
461This is useful when you temporarily want to execute code in the context of
462a port.
463
464Example: create a port and run some initialisation code in it's context.
465
466 my $port = port { ... };
467
468 peval $port, sub {
469 init
470 or die "unable to init";
471 };
472
473=cut
474
475sub peval($$) {
476 local $SELF = shift;
477 my $cb = shift;
478
479 if (wantarray) {
480 my @res = eval { &$cb };
481 _self_die if $@;
482 @res
483 } else {
484 my $res = eval { &$cb };
485 _self_die if $@;
486 $res
487 }
488}
489
430=item $closure = psub { BLOCK } 490=item $closure = psub { BLOCK }
431 491
432Remembers C<$SELF> and creates a closure out of the BLOCK. When the 492Remembers C<$SELF> and creates a closure out of the BLOCK. When the
433closure is executed, sets up the environment in the same way as in C<rcv> 493closure is executed, sets up the environment in the same way as in C<rcv>
434callbacks, i.e. runtime errors will cause the port to get C<kil>ed. 494callbacks, i.e. runtime errors will cause the port to get C<kil>ed.
495
496The effect is basically as if it returned C<< sub { peval $SELF, sub {
497BLOCK } } >>.
435 498
436This is useful when you register callbacks from C<rcv> callbacks: 499This is useful when you register callbacks from C<rcv> callbacks:
437 500
438 rcv delayed_reply => sub { 501 rcv delayed_reply => sub {
439 my ($delay, @reply) = @_; 502 my ($delay, @reply) = @_;
512delivered again. 575delivered again.
513 576
514Inter-host-connection timeouts and monitoring depend on the transport 577Inter-host-connection timeouts and monitoring depend on the transport
515used. The only transport currently implemented is TCP, and AnyEvent::MP 578used. The only transport currently implemented is TCP, and AnyEvent::MP
516relies on TCP to detect node-downs (this can take 10-15 minutes on a 579relies on TCP to detect node-downs (this can take 10-15 minutes on a
517non-idle connection, and usually around two hours for idle conenctions). 580non-idle connection, and usually around two hours for idle connections).
518 581
519This means that monitoring is good for program errors and cleaning up 582This means that monitoring is good for program errors and cleaning up
520stuff eventually, but they are no replacement for a timeout when you need 583stuff eventually, but they are no replacement for a timeout when you need
521to ensure some maximum latency. 584to ensure some maximum latency.
522 585
554 } 617 }
555 618
556 $node->monitor ($port, $cb); 619 $node->monitor ($port, $cb);
557 620
558 defined wantarray 621 defined wantarray
559 and AnyEvent::Util::guard { $node->unmonitor ($port, $cb) } 622 and ($cb += 0, AnyEvent::Util::guard { $node->unmonitor ($port, $cb) })
560} 623}
561 624
562=item $guard = mon_guard $port, $ref, $ref... 625=item $guard = mon_guard $port, $ref, $ref...
563 626
564Monitors the given C<$port> and keeps the passed references. When the port 627Monitors the given C<$port> and keeps the passed references. When the port
621the package, then the package above the package and so on (e.g. 684the package, then the package above the package and so on (e.g.
622C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function 685C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function
623exists or it runs out of package names. 686exists or it runs out of package names.
624 687
625The init function is then called with the newly-created port as context 688The init function is then called with the newly-created port as context
626object (C<$SELF>) and the C<@initdata> values as arguments. 689object (C<$SELF>) and the C<@initdata> values as arguments. It I<must>
690call one of the C<rcv> functions to set callbacks on C<$SELF>, otherwise
691the port might not get created.
627 692
628A common idiom is to pass a local port, immediately monitor the spawned 693A common idiom is to pass a local port, immediately monitor the spawned
629port, and in the remote init function, immediately monitor the passed 694port, and in the remote init function, immediately monitor the passed
630local port. This two-way monitoring ensures that both ports get cleaned up 695local port. This two-way monitoring ensures that both ports get cleaned up
631when there is a problem. 696when there is a problem.
655 720
656sub _spawn { 721sub _spawn {
657 my $port = shift; 722 my $port = shift;
658 my $init = shift; 723 my $init = shift;
659 724
725 # rcv will create the actual port
660 local $SELF = "$NODE#$port"; 726 local $SELF = "$NODE#$port";
661 eval { 727 eval {
662 &{ load_func $init } 728 &{ load_func $init }
663 }; 729 };
664 _self_die if $@; 730 _self_die if $@;
699 ? $action[0]() 765 ? $action[0]()
700 : snd @action; 766 : snd @action;
701 }; 767 };
702} 768}
703 769
770=item cal $port, @msg, $callback[, $timeout]
771
772A simple form of RPC - sends a message to the given C<$port> with the
773given contents (C<@msg>), but adds a reply port to the message.
774
775The reply port is created temporarily just for the purpose of receiving
776the reply, and will be C<kil>ed when no longer needed.
777
778A reply message sent to the port is passed to the C<$callback> as-is.
779
780If an optional time-out (in seconds) is given and it is not C<undef>,
781then the callback will be called without any arguments after the time-out
782elapsed and the port is C<kil>ed.
783
784If no time-out is given (or it is C<undef>), then the local port will
785monitor the remote port instead, so it eventually gets cleaned-up.
786
787Currently this function returns the temporary port, but this "feature"
788might go in future versions unless you can make a convincing case that
789this is indeed useful for something.
790
791=cut
792
793sub cal(@) {
794 my $timeout = ref $_[-1] ? undef : pop;
795 my $cb = pop;
796
797 my $port = port {
798 undef $timeout;
799 kil $SELF;
800 &$cb;
801 };
802
803 if (defined $timeout) {
804 $timeout = AE::timer $timeout, 0, sub {
805 undef $timeout;
806 kil $port;
807 $cb->();
808 };
809 } else {
810 mon $_[0], sub {
811 kil $port;
812 $cb->();
813 };
814 }
815
816 push @_, $port;
817 &snd;
818
819 $port
820}
821
704=back 822=back
705 823
706=head1 AnyEvent::MP vs. Distributed Erlang 824=head1 AnyEvent::MP vs. Distributed Erlang
707 825
708AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node 826AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node
709== aemp node, Erlang process == aemp port), so many of the documents and 827== aemp node, Erlang process == aemp port), so many of the documents and
710programming techniques employed by Erlang apply to AnyEvent::MP. Here is a 828programming techniques employed by Erlang apply to AnyEvent::MP. Here is a
711sample: 829sample:
712 830
713 http://www.Erlang.se/doc/programming_rules.shtml 831 http://www.erlang.se/doc/programming_rules.shtml
714 http://Erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4 832 http://erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4
715 http://Erlang.org/download/Erlang-book-part1.pdf # chapters 5 and 6 833 http://erlang.org/download/erlang-book-part1.pdf # chapters 5 and 6
716 http://Erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5 834 http://erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5
717 835
718Despite the similarities, there are also some important differences: 836Despite the similarities, there are also some important differences:
719 837
720=over 4 838=over 4
721 839
722=item * Node IDs are arbitrary strings in AEMP. 840=item * Node IDs are arbitrary strings in AEMP.
723 841
724Erlang relies on special naming and DNS to work everywhere in the same 842Erlang relies on special naming and DNS to work everywhere in the same
725way. AEMP relies on each node somehow knowing its own address(es) (e.g. by 843way. AEMP relies on each node somehow knowing its own address(es) (e.g. by
726configuration or DNS), but will otherwise discover other odes itself. 844configuration or DNS), and possibly the addresses of some seed nodes, but
845will otherwise discover other nodes (and their IDs) itself.
727 846
728=item * Erlang has a "remote ports are like local ports" philosophy, AEMP 847=item * Erlang has a "remote ports are like local ports" philosophy, AEMP
729uses "local ports are like remote ports". 848uses "local ports are like remote ports".
730 849
731The failure modes for local ports are quite different (runtime errors 850The failure modes for local ports are quite different (runtime errors
756so does not need a queue that can overflow). AEMP sends are immediate, 875so does not need a queue that can overflow). AEMP sends are immediate,
757connection establishment is handled in the background. 876connection establishment is handled in the background.
758 877
759=item * Erlang suffers from silent message loss, AEMP does not. 878=item * Erlang suffers from silent message loss, AEMP does not.
760 879
761Erlang makes few guarantees on messages delivery - messages can get lost 880Erlang implements few guarantees on messages delivery - messages can get
762without any of the processes realising it (i.e. you send messages a, b, 881lost without any of the processes realising it (i.e. you send messages a,
763and c, and the other side only receives messages a and c). 882b, and c, and the other side only receives messages a and c).
764 883
765AEMP guarantees correct ordering, and the guarantee that after one message 884AEMP guarantees correct ordering, and the guarantee that after one message
766is lost, all following ones sent to the same port are lost as well, until 885is lost, all following ones sent to the same port are lost as well, until
767monitoring raises an error, so there are no silent "holes" in the message 886monitoring raises an error, so there are no silent "holes" in the message
768sequence. 887sequence.
860L<AnyEvent::MP::Kernel> - more, lower-level, stuff. 979L<AnyEvent::MP::Kernel> - more, lower-level, stuff.
861 980
862L<AnyEvent::MP::Global> - network maintainance and port groups, to find 981L<AnyEvent::MP::Global> - network maintainance and port groups, to find
863your applications. 982your applications.
864 983
984L<AnyEvent::MP::DataConn> - establish data connections between nodes.
985
865L<AnyEvent::MP::LogCatcher> - simple service to display log messages from 986L<AnyEvent::MP::LogCatcher> - simple service to display log messages from
866all nodes. 987all nodes.
867 988
868L<AnyEvent>. 989L<AnyEvent>.
869 990

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines