[ViewVC] Diff of: cvs/AnyEvent-MP/MP.pm

Comparing AnyEvent-MP/MP.pm (file contents):
Revision 1.20 by root, Mon Aug 3 22:05:55 2009 UTC vs.
Revision 1.123 by root, Thu Mar 1 19:37:59 2012 UTC

…		…
1	=head1 NAME	1	=head1 NAME
2		2
3	AnyEvent::MP - multi-processing/message-passing framework	3	AnyEvent::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
9	NODE # returns this node identifier
10	$NODE # contains this node identifier	9	$NODE # contains this node's node ID
		10	NODE # returns this node's node ID
11		11
		12	$SELF # receiving/own port id in rcv callbacks
		13
		14	# initialise the node so it can send/receive messages
		15	configure;
		16
		17	# ports are message destinations
		18
		19	# sending messages
12	snd $port, type => data...;	20	snd $port, type => data...;
		21	snd $port, @msg;
		22	snd @msg_with_first_element_being_a_port;
13		23
14	rcv $port, smartmatch => $cb->($port, @msg);	24	# creating/using ports, the simple way
		25	my $simple_port = port { my @msg = @_ };
15		26
16	# examples:	27	# creating/using ports, tagged message matching
		28	my $port = port;
17	rcv $port2, ping => sub { snd $_[0], "pong"; 0 };	29	rcv $port, ping => sub { snd $_[0], "pong" };
18	rcv $port1, pong => sub { warn "pong received\n" };	30	rcv $port, pong => sub { warn "pong received\n" };
19	snd $port2, ping => $port1;
20		31
21	# more, smarter, matches (_any_ is exported by this module)	32	# create a port on another node
22	rcv $port, [child_died => $pid] => sub { ...	33	my $port = spawn $node, $initfunc, @initdata;
23	rcv $port, [_any_, _any_, 3] => sub { .. $_[2] is 3	34
		35	# destroy a port again
		36	kil $port; # "normal" kill
		37	kil $port, my_error => "everything is broken"; # error kill
		38
		39	# monitoring
		40	mon $localport, $cb->(@msg) # callback is invoked on death
		41	mon $localport, $otherport # kill otherport on abnormal 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	};
		51
		52	=head1 CURRENT STATUS
		53
		54	bin/aemp - stable.
		55	AnyEvent::MP - stable API, should work.
		56	AnyEvent::MP::Intro - explains most concepts.
		57	AnyEvent::MP::Kernel - mostly stable API.
		58	AnyEvent::MP::Global - stable API.
24		59
25	=head1 DESCRIPTION	60	=head1 DESCRIPTION
26		61
27	This module (-family) implements a simple message passing framework.	62	This module (-family) implements a simple message passing framework.
28		63
29	Despite its simplicity, you can securely message other processes running	64	Despite its simplicity, you can securely message other processes running
30	on the same or other hosts.	65	on the same or other hosts, and you can supervise entities remotely.
31		66
32	At the moment, this module family is severly brokena nd underdocumented,	67	For an introduction to this module family, see the L<AnyEvent::MP::Intro>
33	so do not use. This was uploaded mainly to resreve the CPAN namespace -	68	manual page and the examples under F<eg/>.
34	stay tuned!
35		69
36	=head1 CONCEPTS	70	=head1 CONCEPTS
37		71
38	=over 4	72	=over 4
39		73
40	=item port	74	=item port
41		75
42	A port is something you can send messages to with the C<snd> function, and	76	Not to be confused with a TCP port, a "port" is something you can send
43	you can register C<rcv> handlers with. All C<rcv> handlers will receive	77	messages to (with the C<snd> function).
44	messages they match, messages will not be queued.
45		78
		79	Ports allow you to register C<rcv> handlers that can match all or just
		80	some messages. Messages send to ports will not be queued, regardless of
		81	anything was listening for them or not.
		82
		83	Ports are represented by (printable) strings called "port IDs".
		84
46	=item port id - C<noderef#portname>	85	=item port ID - C<nodeid#portname>
47		86
48	A port id is always the noderef, a hash-mark (C<#>) as separator, followed	87	A port ID is the concatenation of a node ID, a hash-mark (C<#>)
49	by a port name (a printable string of unspecified format).	88	as separator, and a port name (a printable string of unspecified
		89	format created by AnyEvent::MP).
50		90
51	=item node	91	=item node
52		92
53	A node is a single process containing at least one port - the node	93	A node is a single process containing at least one port - the node port,
54	port. You can send messages to node ports to let them create new ports,	94	which enables nodes to manage each other remotely, and to create new
55	among other things.	95	ports.
56		96
57	Initially, nodes are either private (single-process only) or hidden	97	Nodes are either public (have one or more listening ports) or private
58	(connected to a master node only). Only when they epxlicitly "become	98	(no listening ports). Private nodes cannot talk to other private nodes
59	public" can you send them messages from unrelated other nodes.	99	currently, but all nodes can talk to public nodes.
60		100
61	=item noderef - C<host:port,host:port...>, C<id@noderef>, C<id>	101	Nodes is represented by (printable) strings called "node IDs".
62		102
		103	=item node ID - C<[A-Za-z0-9_\-.:]*>
		104
63	A noderef is a string that either uniquely identifies a given node (for	105	A node ID is a string that uniquely identifies the node within a
64	private and hidden nodes), or contains a recipe on how to reach a given	106	network. Depending on the configuration used, node IDs can look like a
65	node (for public nodes).	107	hostname, a hostname and a port, or a random string. AnyEvent::MP itself
		108	doesn't interpret node IDs in any way except to uniquely identify a node.
		109
		110	=item binds - C<ip:port>
		111
		112	Nodes can only talk to each other by creating some kind of connection to
		113	each other. To do this, nodes should listen on one or more local transport
		114	endpoints - binds.
		115
		116	Currently, only standard C<ip:port> specifications can be used, which
		117	specify TCP ports to listen on. So a bind is basically just a tcp socket
		118	in listening mode thta accepts conenctions form other nodes.
		119
		120	=item seed nodes
		121
		122	When a node starts, it knows nothing about the network it is in - it
		123	needs to connect to at least one other node that is already in the
		124	network. These other nodes are called "seed nodes".
		125
		126	Seed nodes themselves are not special - they are seed nodes only because
		127	some other node I<uses> them as such, but any node can be used as seed
		128	node for other nodes, and eahc node cna use a different set of seed nodes.
		129
		130	In addition to discovering the network, seed nodes are also used to
		131	maintain the network - all nodes using the same seed node form are part of
		132	the same network. If a network is split into multiple subnets because e.g.
		133	the network link between the parts goes down, then using the same seed
		134	nodes for all nodes ensures that eventually the subnets get merged again.
		135
		136	Seed nodes are expected to be long-running, and at least one seed node
		137	should always be available. They should also be relatively responsive - a
		138	seed node that blocks for long periods will slow down everybody else.
		139
		140	For small networks, it's best if every node uses the same set of seed
		141	nodes. For large networks, it can be useful to specify "regional" seed
		142	nodes for most nodes in an area, and use all seed nodes as seed nodes for
		143	each other. What's important is that all seed nodes connections form a
		144	complete graph, so that the network cannot split into separate subnets
		145	forever.
		146
		147	Seed nodes are represented by seed IDs.
		148
		149	=item seed IDs - C<host:port>
		150
		151	Seed IDs are transport endpoint(s) (usually a hostname/IP address and a
		152	TCP port) of nodes that should be used as seed nodes.
		153
		154	=item global nodes
		155
		156	An AEMP network needs a discovery service - nodes need to know how to
		157	connect to other nodes they only know by name. In addition, AEMP offers a
		158	distributed "group database", which maps group names to a list of strings
		159	- for example, to register worker ports.
		160
		161	A network needs at least one global node to work, and allows every node to
		162	be a global node.
		163
		164	Any node that loads the L<AnyEvent::MP::Global> module becomes a global
		165	node and tries to keep connections to all other nodes. So while it can
		166	make sense to make every node "global" in small networks, it usually makes
		167	sense to only make seed nodes into global nodes in large networks (nodes
		168	keep connections to seed nodes and global nodes, so makign them the same
		169	reduces overhead).
66		170
67	=back	171	=back
68		172
69	=head1 VARIABLES/FUNCTIONS	173	=head1 VARIABLES/FUNCTIONS
70		174
72		176
73	=cut	177	=cut
74		178
75	package AnyEvent::MP;	179	package AnyEvent::MP;
76		180
		181	use AnyEvent::MP::Config ();
77	use AnyEvent::MP::Base;	182	use AnyEvent::MP::Kernel;
		183	use AnyEvent::MP::Kernel qw(%NODE %PORT %PORT_DATA $UNIQ $RUNIQ $ID);
78		184
79	use common::sense;	185	use common::sense;
80		186
81	use Carp ();	187	use Carp ();
82		188
83	use AE ();	189	use AE ();
84		190
85	use base "Exporter";	191	use base "Exporter";
86		192
87	our $VERSION = '0.02';	193	our $VERSION = $AnyEvent::MP::Config::VERSION;
		194
88	our @EXPORT = qw(	195	our @EXPORT = qw(
89	NODE $NODE $PORT snd rcv mon del _any_	196	NODE $NODE *SELF node_of after
90	create_port create_port_on	197	configure
91	create_miniport	198	snd rcv mon mon_guard kil psub peval spawn cal
92	become_slave become_public	199	port
93	);	200	);
94		201
		202	our $SELF;
		203
		204	sub _self_die() {
		205	my $msg = $@;
		206	$msg =~ s/\n+$// unless ref $msg;
		207	kil $SELF, die => $msg;
		208	}
		209
95	=item NODE / $NODE	210	=item $thisnode = NODE / $NODE
96		211
97	The C<NODE ()> function and the C<$NODE> variable contain the noderef of	212	The C<NODE> function returns, and the C<$NODE> variable contains, the node
98	the local node. The value is initialised by a call to C<become_public> or	213	ID of the node running in the current process. This value is initialised by
99	C<become_slave>, after which all local port identifiers become invalid.	214	a call to C<configure>.
100		215
		216	=item $nodeid = node_of $port
		217
		218	Extracts and returns the node ID from a port ID or a node ID.
		219
		220	=item configure $profile, key => value...
		221
		222	=item configure key => value...
		223
		224	Before a node can talk to other nodes on the network (i.e. enter
		225	"distributed mode") it has to configure itself - the minimum a node needs
		226	to know is its own name, and optionally it should know the addresses of
		227	some other nodes in the network to discover other nodes.
		228
		229	This function configures a node - it must be called exactly once (or
		230	never) before calling other AnyEvent::MP functions.
		231
		232	The key/value pairs are basically the same ones as documented for the
		233	F<aemp> command line utility (sans the set/del prefix), with two additions:
		234
		235	=over 4
		236
		237	=item norc => $boolean (default false)
		238
		239	If true, then the rc file (e.g. F<~/.perl-anyevent-mp>) will I<not>
		240	be consulted - all configuraiton options must be specified in the
		241	C<configure> call.
		242
		243	=item force => $boolean (default false)
		244
		245	IF true, then the values specified in the C<configure> will take
		246	precedence over any values configured via the rc file. The default is for
		247	the rc file to override any options specified in the program.
		248
		249	=back
		250
		251	=over 4
		252
		253	=item step 1, gathering configuration from profiles
		254
		255	The function first looks up a profile in the aemp configuration (see the
		256	L<aemp> commandline utility). The profile name can be specified via the
		257	named C<profile> parameter or can simply be the first parameter). If it is
		258	missing, then the nodename (F<uname -n>) will be used as profile name.
		259
		260	The profile data is then gathered as follows:
		261
		262	First, all remaining key => value pairs (all of which are conveniently
		263	undocumented at the moment) will be interpreted as configuration
		264	data. Then they will be overwritten by any values specified in the global
		265	default configuration (see the F<aemp> utility), then the chain of
		266	profiles chosen by the profile name (and any C<parent> attributes).
		267
		268	That means that the values specified in the profile have highest priority
		269	and the values specified directly via C<configure> have lowest priority,
		270	and can only be used to specify defaults.
		271
		272	If the profile specifies a node ID, then this will become the node ID of
		273	this process. If not, then the profile name will be used as node ID, with
		274	a slash (C</>) attached.
		275
		276	If the node ID (or profile name) ends with a slash (C</>), then a random
		277	string is appended to make it unique.
		278
		279	=item step 2, bind listener sockets
		280
		281	The next step is to look up the binds in the profile, followed by binding
		282	aemp protocol listeners on all binds specified (it is possible and valid
		283	to have no binds, meaning that the node cannot be contacted form the
		284	outside. This means the node cannot talk to other nodes that also have no
		285	binds, but it can still talk to all "normal" nodes).
		286
		287	If the profile does not specify a binds list, then a default of C<*> is
		288	used, meaning the node will bind on a dynamically-assigned port on every
		289	local IP address it finds.
		290
		291	=item step 3, connect to seed nodes
		292
		293	As the last step, the seed ID list from the profile is passed to the
		294	L<AnyEvent::MP::Global> module, which will then use it to keep
		295	connectivity with at least one node at any point in time.
		296
		297	=back
		298
		299	Example: become a distributed node using the local node name as profile.
		300	This should be the most common form of invocation for "daemon"-type nodes.
		301
		302	configure
		303
		304	Example: become an anonymous node. This form is often used for commandline
		305	clients.
		306
		307	configure nodeid => "anon/";
		308
		309	Example: configure a node using a profile called seed, which is suitable
		310	for a seed node as it binds on all local addresses on a fixed port (4040,
		311	customary for aemp).
		312
		313	# use the aemp commandline utility
		314	# aemp profile seed binds '*:4040'
		315
		316	# then use it
		317	configure profile => "seed";
		318
		319	# or simply use aemp from the shell again:
		320	# aemp run profile seed
		321
		322	# or provide a nicer-to-remember nodeid
		323	# aemp run profile seed nodeid "$(hostname)"
		324
		325	=item $SELF
		326
		327	Contains the current port id while executing C<rcv> callbacks or C<psub>
		328	blocks.
		329
		330	=item *SELF, SELF, %SELF, @SELF...
		331
		332	Due to some quirks in how perl exports variables, it is impossible to
		333	just export C<$SELF>, all the symbols named C<SELF> are exported by this
		334	module, but only C<$SELF> is currently used.
		335
101	=item snd $portid, type => @data	336	=item snd $port, type => @data
102		337
103	=item snd $portid, @msg	338	=item snd $port, @msg
104		339
105	Send the given message to the given port ID, which can identify either	340	Send the given message to the given port, which can identify either a
106	a local or a remote port, and can be either a string or soemthignt hat	341	local or a remote port, and must be a port ID.
107	stringifies a sa port ID (such as a port object :).
108		342
109	While the message can be about anything, it is highly recommended to use a	343	While the message can be almost anything, it is highly recommended to
110	string as first element (a portid, or some word that indicates a request	344	use a string as first element (a port ID, or some word that indicates a
111	type etc.).	345	request type etc.) and to consist if only simple perl values (scalars,
		346	arrays, hashes) - if you think you need to pass an object, think again.
112		347
113	The message data effectively becomes read-only after a call to this	348	The message data logically becomes read-only after a call to this
114	function: modifying any argument is not allowed and can cause many	349	function: modifying any argument (or values referenced by them) is
115	problems.	350	forbidden, as there can be considerable time between the call to C<snd>
		351	and the time the message is actually being serialised - in fact, it might
		352	never be copied as within the same process it is simply handed to the
		353	receiving port.
116		354
117	The type of data you can transfer depends on the transport protocol: when	355	The type of data you can transfer depends on the transport protocol: when
118	JSON is used, then only strings, numbers and arrays and hashes consisting	356	JSON is used, then only strings, numbers and arrays and hashes consisting
119	of those are allowed (no objects). When Storable is used, then anything	357	of those are allowed (no objects). When Storable is used, then anything
120	that Storable can serialise and deserialise is allowed, and for the local	358	that Storable can serialise and deserialise is allowed, and for the local
121	node, anything can be passed.	359	node, anything can be passed. Best rely only on the common denominator of
		360	these.
122		361
123	=item $guard = mon $portid, $cb->()	362	=item $local_port = port
124		363
125	Monitor the given port and call the given callback when the port is	364	Create a new local port object and returns its port ID. Initially it has
126	destroyed or connection to it's node is lost.	365	no callbacks set and will throw an error when it receives messages.
127		366
128	#TODO	367	=item $local_port = port { my @msg = @_ }
		368
		369	Creates a new local port, and returns its ID. Semantically the same as
		370	creating a port and calling C<rcv $port, $callback> on it.
		371
		372	The block will be called for every message received on the port, with the
		373	global variable C<$SELF> set to the port ID. Runtime errors will cause the
		374	port to be C<kil>ed. The message will be passed as-is, no extra argument
		375	(i.e. no port ID) will be passed to the callback.
		376
		377	If you want to stop/destroy the port, simply C<kil> it:
		378
		379	my $port = port {
		380	my @msg = @_;
		381	...
		382	kil $SELF;
		383	};
		384
		385	=cut
		386
		387	sub rcv($@);
		388
		389	sub _kilme {
		390	die "received message on port without callback";
		391	}
		392
		393	sub port(;&) {
		394	my $id = $UNIQ . ++$ID;
		395	my $port = "$NODE#$id";
		396
		397	rcv $port, shift \|\| \&_kilme;
		398
		399	$port
		400	}
		401
		402	=item rcv $local_port, $callback->(@msg)
		403
		404	Replaces the default callback on the specified port. There is no way to
		405	remove the default callback: use C<sub { }> to disable it, or better
		406	C<kil> the port when it is no longer needed.
		407
		408	The global C<$SELF> (exported by this module) contains C<$port> while
		409	executing the callback. Runtime errors during callback execution will
		410	result in the port being C<kil>ed.
		411
		412	The default callback received all messages not matched by a more specific
		413	C<tag> match.
		414
		415	=item rcv $local_port, tag => $callback->(@msg_without_tag), ...
		416
		417	Register (or replace) callbacks to be called on messages starting with the
		418	given tag on the given port (and return the port), or unregister it (when
		419	C<$callback> is C<$undef> or missing). There can only be one callback
		420	registered for each tag.
		421
		422	The original message will be passed to the callback, after the first
		423	element (the tag) has been removed. The callback will use the same
		424	environment as the default callback (see above).
		425
		426	Example: create a port and bind receivers on it in one go.
		427
		428	my $port = rcv port,
		429	msg1 => sub { ... },
		430	msg2 => sub { ... },
		431	;
		432
		433	Example: create a port, bind receivers and send it in a message elsewhere
		434	in one go:
		435
		436	snd $otherport, reply =>
		437	rcv port,
		438	msg1 => sub { ... },
		439	...
		440	;
		441
		442	Example: temporarily register a rcv callback for a tag matching some port
		443	(e.g. for an rpc reply) and unregister it after a message was received.
		444
		445	rcv $port, $otherport => sub {
		446	my @reply = @_;
		447
		448	rcv $SELF, $otherport;
		449	};
		450
		451	=cut
		452
		453	sub rcv($@) {
		454	my $port = shift;
		455	my ($nodeid, $portid) = split /#/, $port, 2;
		456
		457	$NODE{$nodeid} == $NODE{""}
		458	or Carp::croak "$port: rcv can only be called on local ports, caught";
		459
		460	while (@_) {
		461	if (ref $_[0]) {
		462	if (my $self = $PORT_DATA{$portid}) {
		463	"AnyEvent::MP::Port" eq ref $self
		464	or Carp::croak "$port: rcv can only be called on message matching ports, caught";
		465
		466	$self->[0] = shift;
		467	} else {
		468	my $cb = shift;
		469	$PORT{$portid} = sub {
		470	local $SELF = $port;
		471	eval { &$cb }; _self_die if $@;
		472	};
		473	}
		474	} elsif (defined $_[0]) {
		475	my $self = $PORT_DATA{$portid} \|\|= do {
		476	my $self = bless [$PORT{$portid} \|\| sub { }, { }, $port], "AnyEvent::MP::Port";
		477
		478	$PORT{$portid} = sub {
		479	local $SELF = $port;
		480
		481	if (my $cb = $self->[1]{$_[0]}) {
		482	shift;
		483	eval { &$cb }; _self_die if $@;
		484	} else {
		485	&{ $self->[0] };
		486	}
		487	};
		488
		489	$self
		490	};
		491
		492	"AnyEvent::MP::Port" eq ref $self
		493	or Carp::croak "$port: rcv can only be called on message matching ports, caught";
		494
		495	my ($tag, $cb) = splice @_, 0, 2;
		496
		497	if (defined $cb) {
		498	$self->[1]{$tag} = $cb;
		499	} else {
		500	delete $self->[1]{$tag};
		501	}
		502	}
		503	}
		504
		505	$port
		506	}
		507
		508	=item peval $port, $coderef[, @args]
		509
		510	Evaluates the given C<$codref> within the contetx of C<$port>, that is,
		511	when the code throews an exception the C<$port> will be killed.
		512
		513	Any remaining args will be passed to the callback. Any return values will
		514	be returned to the caller.
		515
		516	This is useful when you temporarily want to execute code in the context of
		517	a port.
		518
		519	Example: create a port and run some initialisation code in it's context.
		520
		521	my $port = port { ... };
		522
		523	peval $port, sub {
		524	init
		525	or die "unable to init";
		526	};
		527
		528	=cut
		529
		530	sub peval($$) {
		531	local $SELF = shift;
		532	my $cb = shift;
		533
		534	if (wantarray) {
		535	my @res = eval { &$cb };
		536	_self_die if $@;
		537	@res
		538	} else {
		539	my $res = eval { &$cb };
		540	_self_die if $@;
		541	$res
		542	}
		543	}
		544
		545	=item $closure = psub { BLOCK }
		546
		547	Remembers C<$SELF> and creates a closure out of the BLOCK. When the
		548	closure is executed, sets up the environment in the same way as in C<rcv>
		549	callbacks, i.e. runtime errors will cause the port to get C<kil>ed.
		550
		551	The effect is basically as if it returned C<< sub { peval $SELF, sub {
		552	BLOCK }, @_ } >>.
		553
		554	This is useful when you register callbacks from C<rcv> callbacks:
		555
		556	rcv delayed_reply => sub {
		557	my ($delay, @reply) = @_;
		558	my $timer = AE::timer $delay, 0, psub {
		559	snd @reply, $SELF;
		560	};
		561	};
		562
		563	=cut
		564
		565	sub psub(&) {
		566	my $cb = shift;
		567
		568	my $port = $SELF
		569	or Carp::croak "psub can only be called from within rcv or psub callbacks, not";
		570
		571	sub {
		572	local $SELF = $port;
		573
		574	if (wantarray) {
		575	my @res = eval { &$cb };
		576	_self_die if $@;
		577	@res
		578	} else {
		579	my $res = eval { &$cb };
		580	_self_die if $@;
		581	$res
		582	}
		583	}
		584	}
		585
		586	=item $guard = mon $port, $cb->(@reason) # call $cb when $port dies
		587
		588	=item $guard = mon $port, $rcvport # kill $rcvport when $port dies
		589
		590	=item $guard = mon $port # kill $SELF when $port dies
		591
		592	=item $guard = mon $port, $rcvport, @msg # send a message when $port dies
		593
		594	Monitor the given port and do something when the port is killed or
		595	messages to it were lost, and optionally return a guard that can be used
		596	to stop monitoring again.
		597
		598	In the first form (callback), the callback is simply called with any
		599	number of C<@reason> elements (no @reason means that the port was deleted
		600	"normally"). Note also that I<< the callback B<must> never die >>, so use
		601	C<eval> if unsure.
		602
		603	In the second form (another port given), the other port (C<$rcvport>)
		604	will be C<kil>'ed with C<@reason>, if a @reason was specified, i.e. on
		605	"normal" kils nothing happens, while under all other conditions, the other
		606	port is killed with the same reason.
		607
		608	The third form (kill self) is the same as the second form, except that
		609	C<$rvport> defaults to C<$SELF>.
		610
		611	In the last form (message), a message of the form C<@msg, @reason> will be
		612	C<snd>.
		613
		614	Monitoring-actions are one-shot: once messages are lost (and a monitoring
		615	alert was raised), they are removed and will not trigger again.
		616
		617	As a rule of thumb, monitoring requests should always monitor a port from
		618	a local port (or callback). The reason is that kill messages might get
		619	lost, just like any other message. Another less obvious reason is that
		620	even monitoring requests can get lost (for example, when the connection
		621	to the other node goes down permanently). When monitoring a port locally
		622	these problems do not exist.
		623
		624	C<mon> effectively guarantees that, in the absence of hardware failures,
		625	after starting the monitor, either all messages sent to the port will
		626	arrive, or the monitoring action will be invoked after possible message
		627	loss has been detected. No messages will be lost "in between" (after
		628	the first lost message no further messages will be received by the
		629	port). After the monitoring action was invoked, further messages might get
		630	delivered again.
		631
		632	Inter-host-connection timeouts and monitoring depend on the transport
		633	used. The only transport currently implemented is TCP, and AnyEvent::MP
		634	relies on TCP to detect node-downs (this can take 10-15 minutes on a
		635	non-idle connection, and usually around two hours for idle connections).
		636
		637	This means that monitoring is good for program errors and cleaning up
		638	stuff eventually, but they are no replacement for a timeout when you need
		639	to ensure some maximum latency.
		640
		641	Example: call a given callback when C<$port> is killed.
		642
		643	mon $port, sub { warn "port died because of <@_>\n" };
		644
		645	Example: kill ourselves when C<$port> is killed abnormally.
		646
		647	mon $port;
		648
		649	Example: send us a restart message when another C<$port> is killed.
		650
		651	mon $port, $self => "restart";
129		652
130	=cut	653	=cut
131		654
132	sub mon {	655	sub mon {
133	my ($noderef, $port) = split /#/, shift, 2;	656	my ($nodeid, $port) = split /#/, shift, 2;
134		657
135	my $node = AnyEvent::MP::Base::add_node $noderef;	658	my $node = $NODE{$nodeid} \|\| add_node $nodeid;
136		659
137	my $cb = shift;	660	my $cb = @_ ? shift : $SELF \|\| Carp::croak 'mon: called with one argument only, but $SELF not set,';
		661
		662	unless (ref $cb) {
		663	if (@_) {
		664	# send a kill info message
		665	my (@msg) = ($cb, @_);
		666	$cb = sub { snd @msg, @_ };
		667	} else {
		668	# simply kill other port
		669	my $port = $cb;
		670	$cb = sub { kil $port, @_ if @_ };
		671	}
		672	}
138		673
139	$node->monitor ($port, $cb);	674	$node->monitor ($port, $cb);
140		675
141	defined wantarray	676	defined wantarray
142	and AnyEvent::Util::guard { $node->unmonitor ($port, $cb) }	677	and ($cb += 0, AnyEvent::Util::guard { $node->unmonitor ($port, $cb) })
143	}	678	}
144		679
145	=item $local_port = create_port	680	=item $guard = mon_guard $port, $ref, $ref...
146		681
147	Create a new local port object. See the next section for allowed methods.	682	Monitors the given C<$port> and keeps the passed references. When the port
		683	is killed, the references will be freed.
148		684
149	=cut	685	Optionally returns a guard that will stop the monitoring.
150		686
151	sub create_port {	687	This function is useful when you create e.g. timers or other watchers and
152	my $id = "$AnyEvent::MP::Base::UNIQ." . $AnyEvent::MP::Base::ID++;	688	want to free them when the port gets killed (note the use of C<psub>):
153		689
154	my $self = bless {	690	$port->rcv (start => sub {
155	id => "$NODE#$id",	691	my $timer; $timer = mon_guard $port, AE::timer 1, 1, psub {
156	names => [$id],	692	undef $timer if 0.9 < rand;
157	}, "AnyEvent::MP::Port";	693	});
		694	});
158		695
159	$AnyEvent::MP::Base::PORT{$id} = sub {	696	=cut
160	unshift @_, $self;
161		697
162	for (@{ $self->{rc0}{$_[1]} }) {	698	sub mon_guard {
163	$_ && &{$_->[0]}	699	my ($port, @refs) = @_;
164	&& undef $_;	700
		701	#TODO: mon-less form?
		702
		703	mon $port, sub { 0 && @refs }
		704	}
		705
		706	=item kil $port[, @reason]
		707
		708	Kill the specified port with the given C<@reason>.
		709
		710	If no C<@reason> is specified, then the port is killed "normally" -
		711	monitor callback will be invoked, but the kil will not cause linked ports
		712	(C<mon $mport, $lport> form) to get killed.
		713
		714	If a C<@reason> is specified, then linked ports (C<mon $mport, $lport>
		715	form) get killed with the same reason.
		716
		717	Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks
		718	will be reported as reason C<< die => $@ >>.
		719
		720	Transport/communication errors are reported as C<< transport_error =>
		721	$message >>.
		722
		723	=cut
		724
		725	=item $port = spawn $node, $initfunc[, @initdata]
		726
		727	Creates a port on the node C<$node> (which can also be a port ID, in which
		728	case it's the node where that port resides).
		729
		730	The port ID of the newly created port is returned immediately, and it is
		731	possible to immediately start sending messages or to monitor the port.
		732
		733	After the port has been created, the init function is called on the remote
		734	node, in the same context as a C<rcv> callback. This function must be a
		735	fully-qualified function name (e.g. C<MyApp::Chat::Server::init>). To
		736	specify a function in the main program, use C<::name>.
		737
		738	If the function doesn't exist, then the node tries to C<require>
		739	the package, then the package above the package and so on (e.g.
		740	C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function
		741	exists or it runs out of package names.
		742
		743	The init function is then called with the newly-created port as context
		744	object (C<$SELF>) and the C<@initdata> values as arguments. It I<must>
		745	call one of the C<rcv> functions to set callbacks on C<$SELF>, otherwise
		746	the port might not get created.
		747
		748	A common idiom is to pass a local port, immediately monitor the spawned
		749	port, and in the remote init function, immediately monitor the passed
		750	local port. This two-way monitoring ensures that both ports get cleaned up
		751	when there is a problem.
		752
		753	C<spawn> guarantees that the C<$initfunc> has no visible effects on the
		754	caller before C<spawn> returns (by delaying invocation when spawn is
		755	called for the local node).
		756
		757	Example: spawn a chat server port on C<$othernode>.
		758
		759	# this node, executed from within a port context:
		760	my $server = spawn $othernode, "MyApp::Chat::Server::connect", $SELF;
		761	mon $server;
		762
		763	# init function on C<$othernode>
		764	sub connect {
		765	my ($srcport) = @_;
		766
		767	mon $srcport;
		768
		769	rcv $SELF, sub {
		770	...
165	}	771	};
		772	}
166		773
167	for (@{ $self->{rcv}{$_[1]} }) {	774	=cut
168	$_ && [@_[1 .. @{$_->[1]}]] ~~ $_->[1]
169	&& &{$_->[0]}
170	&& undef $_;
171	}
172		775
173	for (@{ $self->{any} }) {	776	sub _spawn {
174	$_ && [@_[0 .. $#{$_->[1]}]] ~~ $_->[1]	777	my $port = shift;
175	&& &{$_->[0]}	778	my $init = shift;
176	&& undef $_;	779
177	}	780	# rcv will create the actual port
		781	local $SELF = "$NODE#$port";
		782	eval {
		783	&{ load_func $init }
178	};	784	};
179		785	_self_die if $@;
180	$self
181	}	786	}
182		787
183	=item $portid = miniport { my @msg = @_; $finished }	788	sub spawn(@) {
		789	my ($nodeid, undef) = split /#/, shift, 2;
184		790
185	Creates a "mini port", that is, a very lightweight port without any	791	my $id = $RUNIQ . ++$ID;
186	pattern matching behind it, and returns its ID.
187		792
188	The block will be called for every message received on the port. When the	793	$_[0] =~ /::/
189	callback returns a true value its job is considered "done" and the port	794	or Carp::croak "spawn init function must be a fully-qualified name, caught";
190	will be destroyed. Otherwise it will stay alive.
191		795
192	The message will be passed as-is, no extra argument (i.e. no port id) will	796	snd_to_func $nodeid, "AnyEvent::MP::_spawn" => $id, @_;
193	be passed to the callback.
194		797
195	If you need the local port id in the callback, this works nicely:	798	"$nodeid#$id"
		799	}
196		800
197	my $port; $port = miniport {	801
198	snd $otherport, reply => $port;	802	=item after $timeout, @msg
		803
		804	=item after $timeout, $callback
		805
		806	Either sends the given message, or call the given callback, after the
		807	specified number of seconds.
		808
		809	This is simply a utility function that comes in handy at times - the
		810	AnyEvent::MP author is not convinced of the wisdom of having it, though,
		811	so it may go away in the future.
		812
		813	=cut
		814
		815	sub after($@) {
		816	my ($timeout, @action) = @_;
		817
		818	my $t; $t = AE::timer $timeout, 0, sub {
		819	undef $t;
		820	ref $action[0]
		821	? $action[0]()
		822	: snd @action;
199	};	823	};
		824	}
200		825
201	=cut	826	=item cal $port, @msg, $callback[, $timeout]
202		827
203	sub miniport(&) {	828	A simple form of RPC - sends a message to the given C<$port> with the
		829	given contents (C<@msg>), but adds a reply port to the message.
		830
		831	The reply port is created temporarily just for the purpose of receiving
		832	the reply, and will be C<kil>ed when no longer needed.
		833
		834	A reply message sent to the port is passed to the C<$callback> as-is.
		835
		836	If an optional time-out (in seconds) is given and it is not C<undef>,
		837	then the callback will be called without any arguments after the time-out
		838	elapsed and the port is C<kil>ed.
		839
		840	If no time-out is given (or it is C<undef>), then the local port will
		841	monitor the remote port instead, so it eventually gets cleaned-up.
		842
		843	Currently this function returns the temporary port, but this "feature"
		844	might go in future versions unless you can make a convincing case that
		845	this is indeed useful for something.
		846
		847	=cut
		848
		849	sub cal(@) {
		850	my $timeout = ref $_[-1] ? undef : pop;
204	my $cb = shift;	851	my $cb = pop;
205	my $id = "$AnyEvent::MP::Base::UNIQ." . $AnyEvent::MP::Base::ID++;
206		852
207	$AnyEvent::MP::Base::PORT{$id} = sub {	853	my $port = port {
		854	undef $timeout;
		855	kil $SELF;
208	&$cb	856	&$cb;
209	and del $id;
210	};	857	};
211		858
212	"$NODE#$id"	859	if (defined $timeout) {
213	}	860	$timeout = AE::timer $timeout, 0, sub {
		861	undef $timeout;
		862	kil $port;
		863	$cb->();
		864	};
		865	} else {
		866	mon $_[0], sub {
		867	kil $port;
		868	$cb->();
		869	};
		870	}
214		871
215	package AnyEvent::MP::Port;	872	push @_, $port;
		873	&snd;
		874
		875	$port
		876	}
216		877
217	=back	878	=back
218		879
219	=head1 METHODS FOR PORT OBJECTS	880	=head1 AnyEvent::MP vs. Distributed Erlang
		881
		882	AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node
		883	== aemp node, Erlang process == aemp port), so many of the documents and
		884	programming techniques employed by Erlang apply to AnyEvent::MP. Here is a
		885	sample:
		886
		887	http://www.erlang.se/doc/programming_rules.shtml
		888	http://erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4
		889	http://erlang.org/download/erlang-book-part1.pdf # chapters 5 and 6
		890	http://erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5
		891
		892	Despite the similarities, there are also some important differences:
220		893
221	=over 4	894	=over 4
222		895
223	=item "$port"	896	=item * Node IDs are arbitrary strings in AEMP.
224		897
225	A port object stringifies to its port ID, so can be used directly for	898	Erlang relies on special naming and DNS to work everywhere in the same
226	C<snd> operations.	899	way. AEMP relies on each node somehow knowing its own address(es) (e.g. by
		900	configuration or DNS), and possibly the addresses of some seed nodes, but
		901	will otherwise discover other nodes (and their IDs) itself.
227		902
228	=cut	903	=item * Erlang has a "remote ports are like local ports" philosophy, AEMP
		904	uses "local ports are like remote ports".
229		905
230	use overload	906	The failure modes for local ports are quite different (runtime errors
231	'""' => sub { $_[0]{id} },	907	only) then for remote ports - when a local port dies, you I<know> it dies,
232	fallback => 1;	908	when a connection to another node dies, you know nothing about the other
		909	port.
233		910
234	sub TO_JSON { $_[0]{id} }	911	Erlang pretends remote ports are as reliable as local ports, even when
		912	they are not.
235		913
236	=item $port->rcv (type => $callback->($port, @msg))	914	AEMP encourages a "treat remote ports differently" philosophy, with local
		915	ports being the special case/exception, where transport errors cannot
		916	occur.
237		917
238	=item $port->rcv ($smartmatch => $callback->($port, @msg))	918	=item * Erlang uses processes and a mailbox, AEMP does not queue.
239		919
240	=item $port->rcv ([$smartmatch...] => $callback->($port, @msg))	920	Erlang uses processes that selectively receive messages out of order, and
		921	therefore needs a queue. AEMP is event based, queuing messages would serve
		922	no useful purpose. For the same reason the pattern-matching abilities
		923	of AnyEvent::MP are more limited, as there is little need to be able to
		924	filter messages without dequeuing them.
241		925
242	Register a callback on the given port.	926	This is not a philosophical difference, but simply stems from AnyEvent::MP
		927	being event-based, while Erlang is process-based.
243		928
244	The callback has to return a true value when its work is done, after	929	You cna have a look at L<Coro::MP> for a more Erlang-like process model on
245	which is will be removed, or a false value in which case it will stay	930	top of AEMP and Coro threads.
246	registered.
247		931
248	If the match is an array reference, then it will be matched against the	932	=item * Erlang sends are synchronous, AEMP sends are asynchronous.
249	first elements of the message, otherwise only the first element is being
250	matched.
251		933
252	Any element in the match that is specified as C<_any_> (a function	934	Sending messages in Erlang is synchronous and blocks the process until
253	exported by this module) matches any single element of the message.	935	a conenction has been established and the message sent (and so does not
		936	need a queue that can overflow). AEMP sends return immediately, connection
		937	establishment is handled in the background.
254		938
255	While not required, it is highly recommended that the first matching	939	=item * Erlang suffers from silent message loss, AEMP does not.
256	element is a string identifying the message. The one-string-only match is
257	also the most efficient match (by far).
258		940
259	=cut	941	Erlang implements few guarantees on messages delivery - messages can get
		942	lost without any of the processes realising it (i.e. you send messages a,
		943	b, and c, and the other side only receives messages a and c).
260		944
261	sub rcv($@) {	945	AEMP guarantees (modulo hardware errors) correct ordering, and the
262	my ($self, $match, $cb) = @_;	946	guarantee that after one message is lost, all following ones sent to the
		947	same port are lost as well, until monitoring raises an error, so there are
		948	no silent "holes" in the message sequence.
263		949
264	if (!ref $match) {	950	If you want your software to be very reliable, you have to cope with
265	push @{ $self->{rc0}{$match} }, [$cb];	951	corrupted and even out-of-order messages in both Erlang and AEMP. AEMP
266	} elsif (("ARRAY" eq ref $match && !ref $match->[0])) {	952	simply tries to work better in common error cases, such as when a network
267	my ($type, @match) = @$match;	953	link goes down.
268	@match
269	? push @{ $self->{rcv}{$match->[0]} }, [$cb, \@match]
270	: push @{ $self->{rc0}{$match->[0]} }, [$cb];
271	} else {
272	push @{ $self->{any} }, [$cb, $match];
273	}
274	}
275		954
276	=item $port->register ($name)	955	=item * Erlang can send messages to the wrong port, AEMP does not.
277		956
278	Registers the given port under the well known name C<$name>. If the name	957	In Erlang it is quite likely that a node that restarts reuses an Erlang
279	already exists it is replaced.	958	process ID known to other nodes for a completely different process,
		959	causing messages destined for that process to end up in an unrelated
		960	process.
280		961
281	A port can only be registered under one well known name.	962	AEMP does not reuse port IDs, so old messages or old port IDs floating
		963	around in the network will not be sent to an unrelated port.
282		964
283	=cut	965	=item * Erlang uses unprotected connections, AEMP uses secure
		966	authentication and can use TLS.
284		967
285	sub register {	968	AEMP can use a proven protocol - TLS - to protect connections and
286	my ($self, $name) = @_;	969	securely authenticate nodes.
287		970
288	$self->{wkname} = $name;	971	=item * The AEMP protocol is optimised for both text-based and binary
289	$AnyEvent::MP::Base::WKP{$name} = "$self";	972	communications.
290	}
291		973
292	=item $port->destroy	974	The AEMP protocol, unlike the Erlang protocol, supports both programming
		975	language independent text-only protocols (good for debugging), and binary,
		976	language-specific serialisers (e.g. Storable). By default, unless TLS is
		977	used, the protocol is actually completely text-based.
293		978
294	Explicitly destroy/remove/nuke/vaporise the port.	979	It has also been carefully designed to be implementable in other languages
		980	with a minimum of work while gracefully degrading functionality to make the
		981	protocol simple.
295		982
296	Ports are normally kept alive by there mere existance alone, and need to	983	=item * AEMP has more flexible monitoring options than Erlang.
297	be destroyed explicitly.
298		984
299	=cut	985	In Erlang, you can chose to receive I<all> exit signals as messages or
		986	I<none>, there is no in-between, so monitoring single Erlang processes is
		987	difficult to implement.
300		988
301	sub destroy {	989	Monitoring in AEMP is more flexible than in Erlang, as one can choose
302	my ($self) = @_;	990	between automatic kill, exit message or callback on a per-port basis.
303		991
304	AnyEvent::MP::Base::del $self->{id};	992	=item * Erlang tries to hide remote/local connections, AEMP does not.
305		993
306	delete $AnyEvent::MP::Base::WKP{ $self->{wkname} };	994	Monitoring in Erlang is not an indicator of process death/crashes, in the
		995	same way as linking is (except linking is unreliable in Erlang).
307		996
308	delete $AnyEvent::MP::Base::PORT{$_}	997	In AEMP, you don't "look up" registered port names or send to named ports
309	for @{ $self->{names} };	998	that might or might not be persistent. Instead, you normally spawn a port
310	}	999	on the remote node. The init function monitors you, and you monitor the
		1000	remote port. Since both monitors are local to the node, they are much more
		1001	reliable (no need for C<spawn_link>).
		1002
		1003	This also saves round-trips and avoids sending messages to the wrong port
		1004	(hard to do in Erlang).
311		1005
312	=back	1006	=back
313		1007
314	=head1 FUNCTIONS FOR NODES	1008	=head1 RATIONALE
315		1009
316	=over 4	1010	=over 4
317		1011
318	=item mon $noderef, $callback->($noderef, $status, $)	1012	=item Why strings for port and node IDs, why not objects?
319		1013
320	Monitors the given noderef.	1014	We considered "objects", but found that the actual number of methods
		1015	that can be called are quite low. Since port and node IDs travel over
		1016	the network frequently, the serialising/deserialising would add lots of
		1017	overhead, as well as having to keep a proxy object everywhere.
321		1018
322	=item become_public endpoint...	1019	Strings can easily be printed, easily serialised etc. and need no special
		1020	procedures to be "valid".
323		1021
324	Tells the node to become a public node, i.e. reachable from other nodes.	1022	And as a result, a port with just a default receiver consists of a single
		1023	code reference stored in a global hash - it can't become much cheaper.
325		1024
326	If no arguments are given, or the first argument is C<undef>, then	1025	=item Why favour JSON, why not a real serialising format such as Storable?
327	AnyEvent::MP tries to bind on port C<4040> on all IP addresses that the
328	local nodename resolves to.
329		1026
330	Otherwise the first argument must be an array-reference with transport	1027	In fact, any AnyEvent::MP node will happily accept Storable as framing
331	endpoints ("ip:port", "hostname:port") or port numbers (in which case the	1028	format, but currently there is no way to make a node use Storable by
332	local nodename is used as hostname). The endpoints are all resolved and	1029	default (although all nodes will accept it).
333	will become the node reference.
334		1030
335	=cut	1031	The default framing protocol is JSON because a) JSON::XS is many times
		1032	faster for small messages and b) most importantly, after years of
		1033	experience we found that object serialisation is causing more problems
		1034	than it solves: Just like function calls, objects simply do not travel
		1035	easily over the network, mostly because they will always be a copy, so you
		1036	always have to re-think your design.
		1037
		1038	Keeping your messages simple, concentrating on data structures rather than
		1039	objects, will keep your messages clean, tidy and efficient.
336		1040
337	=back	1041	=back
338		1042
339	=head1 NODE MESSAGES
340
341	Nodes understand the following messages sent to them. Many of them take
342	arguments called C<@reply>, which will simply be used to compose a reply
343	message - C<$reply[0]> is the port to reply to, C<$reply[1]> the type and
344	the remaining arguments are simply the message data.
345
346	=over 4
347
348	=cut
349
350	=item wkp => $name, @reply
351
352	Replies with the port ID of the specified well-known port, or C<undef>.
353
354	=item devnull => ...
355
356	Generic data sink/CPU heat conversion.
357
358	=item relay => $port, @msg
359
360	Simply forwards the message to the given port.
361
362	=item eval => $string[ @reply]
363
364	Evaluates the given string. If C<@reply> is given, then a message of the
365	form C<@reply, $@, @evalres> is sent.
366
367	Example: crash another node.
368
369	snd $othernode, eval => "exit";
370
371	=item time => @reply
372
373	Replies the the current node time to C<@reply>.
374
375	Example: tell the current node to send the current time to C<$myport> in a
376	C<timereply> message.
377
378	snd $NODE, time => $myport, timereply => 1, 2;
379	# => snd $myport, timereply => 1, 2, <time>
380
381	=back
382
383	=head1 SEE ALSO	1043	=head1 SEE ALSO
		1044
		1045	L<AnyEvent::MP::Intro> - a gentle introduction.
		1046
		1047	L<AnyEvent::MP::Kernel> - more, lower-level, stuff.
		1048
		1049	L<AnyEvent::MP::Global> - network maintenance and port groups, to find
		1050	your applications.
		1051
		1052	L<AnyEvent::MP::DataConn> - establish data connections between nodes.
		1053
		1054	L<AnyEvent::MP::LogCatcher> - simple service to display log messages from
		1055	all nodes.
384		1056
385	L<AnyEvent>.	1057	L<AnyEvent>.
386		1058
387	=head1 AUTHOR	1059	=head1 AUTHOR
388		1060

Diff Legend

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

Comparing AnyEvent-MP/MP.pm (file contents): Revision 1.20 by root, Mon Aug 3 22:05:55 2009 UTC vs. Revision 1.123 by root, Thu Mar 1 19:37:59 2012 UTC

Diff Legend

Comparing AnyEvent-MP/MP.pm (file contents):
Revision 1.20 by root, Mon Aug 3 22:05:55 2009 UTC vs.
Revision 1.123 by root, Thu Mar 1 19:37:59 2012 UTC