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

Comparing AnyEvent-MP/MP.pm (file contents):
Revision 1.7 by root, Sat Aug 1 15:04:30 2009 UTC vs.
Revision 1.48 by root, Thu Aug 13 02:59:42 2009 UTC

…		…
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 noderef
		10	NODE # returns this node's noderef
		11	NODE $port # returns the noderef of the port
11		12
		13	$SELF # receiving/own port id in rcv callbacks
		14
		15	# initialise the node so it can send/receive messages
		16	initialise_node; # -OR-
		17	initialise_node "localhost:4040"; # -OR-
		18	initialise_node "slave/", "localhost:4040"
		19
		20	# ports are message endpoints
		21
		22	# sending messages
12	snd $port, type => data...;	23	snd $port, type => data...;
		24	snd $port, @msg;
		25	snd @msg_with_first_element_being_a_port;
13		26
		27	# creating/using miniports
		28	my $miniport = port { my @msg = @_; 0 };
		29
		30	# creating/using full ports
		31	my $port = port;
14	rcv $port, smartmatch => $cb->($port, @msg);	32	rcv $port, smartmatch => $cb->(@msg);
15
16	# examples:
17	rcv $port2, ping => sub { snd $_[0], "pong"; 0 };	33	rcv $port, ping => sub { snd $_[0], "pong"; 0 };
18	rcv $port1, pong => sub { warn "pong received\n" };	34	rcv $port, pong => sub { warn "pong received\n"; 0 };
19	snd $port2, ping => $port1;
20		35
21	# more, smarter, matches (_any_ is exported by this module)	36	# more, smarter, matches (_any_ is exported by this module)
22	rcv $port, [child_died => $pid] => sub { ...	37	rcv $port, [child_died => $pid] => sub { ...
23	rcv $port, [_any_, _any_, 3] => sub { .. $_[2] is 3	38	rcv $port, [_any_, _any_, 3] => sub { .. $_[2] is 3
24		39
		40	# create a port on another node
		41	my $port = spawn $node, $initfunc, @initdata;
		42
		43	# monitoring
		44	mon $port, $cb->(@msg) # callback is invoked on death
		45	mon $port, $otherport # kill otherport on abnormal death
		46	mon $port, $otherport, @msg # send message on death
		47
		48	=head1 CURRENT STATUS
		49
		50	AnyEvent::MP - stable API, should work
		51	AnyEvent::MP::Intro - outdated
		52	AnyEvent::MP::Kernel - WIP
		53	AnyEvent::MP::Transport - mostly stable
		54
		55	stay tuned.
		56
25	=head1 DESCRIPTION	57	=head1 DESCRIPTION
26		58
27	This module (-family) implements a simple message passing framework.	59	This module (-family) implements a simple message passing framework.
28		60
29	Despite its simplicity, you can securely message other processes running	61	Despite its simplicity, you can securely message other processes running
30	on the same or other hosts.	62	on the same or other hosts.
31		63
		64	For an introduction to this module family, see the L<AnyEvent::MP::Intro>
		65	manual page.
		66
32	At the moment, this module family is severly brokena nd underdocumented,	67	At the moment, this module family is severly broken and underdocumented,
33	so do not use. This was uploaded mainly to resreve the CPAN namespace -	68	so do not use. This was uploaded mainly to reserve the CPAN namespace -
34	stay tuned!	69	stay tuned!
35		70
36	=head1 CONCEPTS	71	=head1 CONCEPTS
37		72
38	=over 4	73	=over 4
39		74
40	=item port	75	=item port
41		76
42	A port is something you can send messages to with the C<snd> function, and	77	A port is something you can send messages to (with the C<snd> function).
43	you can register C<rcv> handlers with. All C<rcv> handlers will receive	78
44	messages they match, messages will not be queued.	79	Some ports allow you to register C<rcv> handlers that can match specific
		80	messages. All C<rcv> handlers will receive messages they match, messages
		81	will not be queued.
45		82
46	=item port id - C<noderef#portname>	83	=item port id - C<noderef#portname>
47		84
48	A port id is always the noderef, a hash-mark (C<#>) as separator, followed	85	A port id is normaly the concatenation of a noderef, a hash-mark (C<#>) as
49	by a port name (a printable string of unspecified format).	86	separator, and a port name (a printable string of unspecified format). An
		87	exception is the the node port, whose ID is identical to its node
		88	reference.
50		89
51	=item node	90	=item node
52		91
53	A node is a single process containing at least one port - the node	92	A node is a single process containing at least one port - the node
54	port. You can send messages to node ports to let them create new ports,	93	port. You can send messages to node ports to find existing ports or to
55	among other things.	94	create new ports, among other things.
56		95
57	Initially, nodes are either private (single-process only) or hidden	96	Nodes are either private (single-process only), slaves (connected to a
58	(connected to a master node only). Only when they epxlicitly "become	97	master node only) or public nodes (connectable from unrelated nodes).
59	public" can you send them messages from unrelated other nodes.
60		98
61	=item noderef - C<host:port,host:port...>, C<id@noderef>, C<id>	99	=item noderef - C<host:port,host:port...>, C<id@noderef>, C<id>
62		100
63	A noderef is a string that either uniquely identifies a given node (for	101	A node reference is a string that either simply identifies the node (for
64	private and hidden nodes), or contains a recipe on how to reach a given	102	private and slave nodes), or contains a recipe on how to reach a given
65	node (for public nodes).	103	node (for public nodes).
66		104
		105	This recipe is simply a comma-separated list of C<address:port> pairs (for
		106	TCP/IP, other protocols might look different).
		107
		108	Node references come in two flavours: resolved (containing only numerical
		109	addresses) or unresolved (where hostnames are used instead of addresses).
		110
		111	Before using an unresolved node reference in a message you first have to
		112	resolve it.
		113
67	=back	114	=back
68		115
69	=head1 VARIABLES/FUNCTIONS	116	=head1 VARIABLES/FUNCTIONS
70		117
71	=over 4	118	=over 4
72		119
73	=cut	120	=cut
74		121
75	package AnyEvent::MP;	122	package AnyEvent::MP;
76		123
77	use AnyEvent::MP::Util ();
78	use AnyEvent::MP::Node;	124	use AnyEvent::MP::Kernel;
79	use AnyEvent::MP::Transport;
80		125
81	use utf8;
82	use common::sense;	126	use common::sense;
83		127
84	use Carp ();	128	use Carp ();
85		129
86	use AE ();	130	use AE ();
87		131
88	use base "Exporter";	132	use base "Exporter";
89		133
90	our $VERSION = '0.01';	134	our $VERSION = $AnyEvent::MP::Kernel::VERSION;
91	our @EXPORT = qw(NODE $NODE $PORT snd rcv _any_);
92		135
93	our $DEFAULT_SECRET;	136	our @EXPORT = qw(
94	our $DEFAULT_PORT = "4040";	137	NODE $NODE *SELF node_of _any_
		138	resolve_node initialise_node
		139	snd rcv mon kil reg psub spawn
		140	port
		141	);
95		142
96	our $CONNECT_INTERVAL = 5; # new connect every 5s, at least	143	our $SELF;
97	our $CONNECT_TIMEOUT = 30; # includes handshake
98		144
99	sub default_secret {	145	sub _self_die() {
100	unless (defined $DEFAULT_SECRET) {	146	my $msg = $@;
101	if (open my $fh, "<$ENV{HOME}/.aemp-secret") {	147	$msg =~ s/\n+$// unless ref $msg;
102	sysread $fh, $DEFAULT_SECRET, -s $fh;	148	kil $SELF, die => $msg;
103	} else {
104	$DEFAULT_SECRET = AnyEvent::MP::Util::nonce 32;
105	}
106	}
107
108	$DEFAULT_SECRET
109	}	149	}
110		150
111	=item NODE / $NODE	151	=item $thisnode = NODE / $NODE
112		152
113	The C<NODE ()> function and the C<$NODE> variable contain the noderef of	153	The C<NODE> function returns, and the C<$NODE> variable contains
114	the local node. The value is initialised by a call to C<become_public> or	154	the noderef of the local node. The value is initialised by a call
115	C<become_slave>, after which all local port identifiers become invalid.	155	to C<become_public> or C<become_slave>, after which all local port
		156	identifiers become invalid.
116		157
117	=cut	158	=item $noderef = node_of $port
118		159
119	our $UNIQ = sprintf "%x.%x", $$, time; # per-process/node unique cookie	160	Extracts and returns the noderef from a portid or a noderef.
120	our $ID = "a0";
121	our $PUBLIC = 0;
122	our $NODE;
123	our $PORT;
124		161
125	our %NODE; # node id to transport mapping, or "undef", for local node	162	=item initialise_node $noderef, $seednode, $seednode...
126	our %PORT; # local ports
127	our %LISTENER; # local transports
128		163
129	sub NODE() { $NODE }	164	=item initialise_node "slave/", $master, $master...
130		165
131	{	166	Before a node can talk to other nodes on the network it has to initialise
132	use POSIX ();	167	itself - the minimum a node needs to know is it's own name, and optionally
133	my $nodename = (POSIX::uname)[1];	168	it should know the noderefs of some other nodes in the network.
134	$NODE = "$$\@$nodename";
135	}
136		169
137	sub _ANY_() { 1 }	170	This function initialises a node - it must be called exactly once (or
138	sub _any_() { \&_ANY_ }	171	never) before calling other AnyEvent::MP functions.
139		172
140	sub add_node {	173	All arguments are noderefs, which can be either resolved or unresolved.
141	my ($noderef) = @_;
142		174
143	return $NODE{$noderef}	175	There are two types of networked nodes, public nodes and slave nodes:
144	if exists $NODE{$noderef};
145		176
146	for (split /,/, $noderef) {	177	=over 4
147	return $NODE{$noderef} = $NODE{$_}
148	if exists $NODE{$_};
149	}
150		178
151	# for indirect sends, use a different class	179	=item public nodes
152	my $node = new AnyEvent::MP::Node::Direct $noderef;
153		180
154	$NODE{$_} = $node	181	For public nodes, C<$noderef> must either be a (possibly unresolved)
155	for $noderef, split /,/, $noderef;	182	noderef, in which case it will be resolved, or C<undef> (or missing), in
		183	which case the noderef will be guessed.
156		184
157	$node	185	Afterwards, the node will bind itself on all endpoints and try to connect
158	}	186	to all additional C<$seednodes> that are specified. Seednodes are optional
		187	and can be used to quickly bootstrap the node into an existing network.
159		188
		189	=item slave nodes
		190
		191	When the C<$noderef> is the special string C<slave/>, then the node will
		192	become a slave node. Slave nodes cannot be contacted from outside and will
		193	route most of their traffic to the master node that they attach to.
		194
		195	At least one additional noderef is required: The node will try to connect
		196	to all of them and will become a slave attached to the first node it can
		197	successfully connect to.
		198
		199	=back
		200
		201	This function will block until all nodes have been resolved and, for slave
		202	nodes, until it has successfully established a connection to a master
		203	server.
		204
		205	Example: become a public node listening on the default node.
		206
		207	initialise_node;
		208
		209	Example: become a public node, and try to contact some well-known master
		210	servers to become part of the network.
		211
		212	initialise_node undef, "master1", "master2";
		213
		214	Example: become a public node listening on port C<4041>.
		215
		216	initialise_node 4041;
		217
		218	Example: become a public node, only visible on localhost port 4044.
		219
		220	initialise_node "locahost:4044";
		221
		222	Example: become a slave node to any of the specified master servers.
		223
		224	initialise_node "slave/", "master1", "192.168.13.17", "mp.example.net";
		225
		226	=item $cv = resolve_node $noderef
		227
		228	Takes an unresolved node reference that may contain hostnames and
		229	abbreviated IDs, resolves all of them and returns a resolved node
		230	reference.
		231
		232	In addition to C<address:port> pairs allowed in resolved noderefs, the
		233	following forms are supported:
		234
		235	=over 4
		236
		237	=item the empty string
		238
		239	An empty-string component gets resolved as if the default port (4040) was
		240	specified.
		241
		242	=item naked port numbers (e.g. C<1234>)
		243
		244	These are resolved by prepending the local nodename and a colon, to be
		245	further resolved.
		246
		247	=item hostnames (e.g. C<localhost:1234>, C<localhost>)
		248
		249	These are resolved by using AnyEvent::DNS to resolve them, optionally
		250	looking up SRV records for the C<aemp=4040> port, if no port was
		251	specified.
		252
		253	=back
		254
		255	=item $SELF
		256
		257	Contains the current port id while executing C<rcv> callbacks or C<psub>
		258	blocks.
		259
		260	=item SELF, %SELF, @SELF...
		261
		262	Due to some quirks in how perl exports variables, it is impossible to
		263	just export C<$SELF>, all the symbols called C<SELF> are exported by this
		264	module, but only C<$SELF> is currently used.
		265
160	=item snd $portid, type => @data	266	=item snd $port, type => @data
161		267
162	=item snd $portid, @msg	268	=item snd $port, @msg
163		269
164	Send the given message to the given port ID, which can identify either a	270	Send the given message to the given port ID, which can identify either
165	local or a remote port.	271	a local or a remote port, and can be either a string or soemthignt hat
		272	stringifies a sa port ID (such as a port object :).
166		273
167	While the message can be about anything, it is highly recommended to use	274	While the message can be about anything, it is highly recommended to use a
168	a constant string as first element.	275	string as first element (a portid, or some word that indicates a request
		276	type etc.).
169		277
170	The message data effectively becomes read-only after a call to this	278	The message data effectively becomes read-only after a call to this
171	function: modifying any argument is not allowed and can cause many	279	function: modifying any argument is not allowed and can cause many
172	problems.	280	problems.
173		281
…		…
175	JSON is used, then only strings, numbers and arrays and hashes consisting	283	JSON is used, then only strings, numbers and arrays and hashes consisting
176	of those are allowed (no objects). When Storable is used, then anything	284	of those are allowed (no objects). When Storable is used, then anything
177	that Storable can serialise and deserialise is allowed, and for the local	285	that Storable can serialise and deserialise is allowed, and for the local
178	node, anything can be passed.	286	node, anything can be passed.
179		287
180	=cut	288	=item $local_port = port
181		289
182	sub snd(@) {	290	Create a new local port object that can be used either as a pattern
183	my ($noderef, $port) = split /#/, shift, 2;	291	matching port ("full port") or a single-callback port ("miniport"),
		292	depending on how C<rcv> callbacks are bound to the object.
184		293
185	add_node $noderef	294	=item $port = port { my @msg = @_; $finished }
186	unless exists $NODE{$noderef};
187		295
188	$NODE{$noderef}->send (["$port", [@_]]);	296	Creates a "miniport", that is, a very lightweight port without any pattern
		297	matching behind it, and returns its ID. Semantically the same as creating
		298	a port and calling C<rcv $port, $callback> on it.
		299
		300	The block will be called for every message received on the port. When the
		301	callback returns a true value its job is considered "done" and the port
		302	will be destroyed. Otherwise it will stay alive.
		303
		304	The message will be passed as-is, no extra argument (i.e. no port id) will
		305	be passed to the callback.
		306
		307	If you need the local port id in the callback, this works nicely:
		308
		309	my $port; $port = port {
		310	snd $otherport, reply => $port;
		311	};
		312
		313	=cut
		314
		315	sub rcv($@);
		316
		317	sub port(;&) {
		318	my $id = "$UNIQ." . $ID++;
		319	my $port = "$NODE#$id";
		320
		321	if (@_) {
		322	rcv $port, shift;
		323	} else {
		324	$PORT{$id} = sub { }; # nop
		325	}
		326
		327	$port
189	}	328	}
190		329
		330	=item reg $port, $name
		331
		332	=item reg $name
		333
		334	Registers the given port (or C<$SELF><<< if missing) under the name
		335	C<$name>. If the name already exists it is replaced.
		336
		337	A port can only be registered under one well known name.
		338
		339	A port automatically becomes unregistered when it is killed.
		340
		341	=cut
		342
		343	sub reg(@) {
		344	my $port = @_ > 1 ? shift : $SELF \|\| Carp::croak 'reg: called with one argument only, but $SELF not set,';
		345
		346	$REG{$_[0]} = $port;
		347	}
		348
191	=item rcv $portid, type => $callback->(@msg)	349	=item rcv $port, $callback->(@msg)
192		350
		351	Replaces the callback on the specified miniport (after converting it to
		352	one if required).
		353
		354	=item rcv $port, tagstring => $callback->(@msg), ...
		355
193	=item rcv $portid, $smartmatch => $callback->(@msg)	356	=item rcv $port, $smartmatch => $callback->(@msg), ...
194		357
195	=item rcv $portid, [$smartmatch...] => $callback->(@msg)	358	=item rcv $port, [$smartmatch...] => $callback->(@msg), ...
196		359
197	Register a callback on the port identified by C<$portid>, which I<must> be	360	Register callbacks to be called on matching messages on the given full
198	a local port.	361	port (after converting it to one if required) and return the port.
199		362
200	The callback has to return a true value when its work is done, after	363	The callback has to return a true value when its work is done, after
201	which is will be removed, or a false value in which case it will stay	364	which is will be removed, or a false value in which case it will stay
202	registered.	365	registered.
203		366
		367	The global C<$SELF> (exported by this module) contains C<$port> while
		368	executing the callback.
		369
		370	Runtime errors during callback execution will result in the port being
		371	C<kil>ed.
		372
204	If the match is an array reference, then it will be matched against the	373	If the match is an array reference, then it will be matched against the
205	first elements of the message, otherwise only the first element is being	374	first elements of the message, otherwise only the first element is being
206	matched.	375	matched.
207		376
208	Any element in the match that is specified as C<_any_> (a function	377	Any element in the match that is specified as C<_any_> (a function
…		…
210		379
211	While not required, it is highly recommended that the first matching	380	While not required, it is highly recommended that the first matching
212	element is a string identifying the message. The one-string-only match is	381	element is a string identifying the message. The one-string-only match is
213	also the most efficient match (by far).	382	also the most efficient match (by far).
214		383
		384	Example: create a port and bind receivers on it in one go.
		385
		386	my $port = rcv port,
		387	msg1 => sub { ...; 0 },
		388	msg2 => sub { ...; 0 },
		389	;
		390
		391	Example: create a port, bind receivers and send it in a message elsewhere
		392	in one go:
		393
		394	snd $otherport, reply =>
		395	rcv port,
		396	msg1 => sub { ...; 0 },
		397	...
		398	;
		399
215	=cut	400	=cut
216		401
217	sub rcv($@) {	402	sub rcv($@) {
218	my ($port, $match, $cb) = @_;	403	my $port = shift;
219
220	my $port = $PORT{$port}
221	or do {
222	my ($noderef, $lport) = split /#/, $port;	404	my ($noderef, $portid) = split /#/, $port, 2;
223	"AnyEvent::MP::Node::Self" eq ref $NODE{$noderef}	405
		406	($NODE{$noderef} \|\| add_node $noderef) == $NODE{""}
224	or Carp::croak "$port: can only rcv on local ports";	407	or Carp::croak "$port: rcv can only be called on local ports, caught";
225		408
226	$PORT{$lport}	409	if (@_ == 1) {
227	or Carp::croak "$port: port does not exist";	410	my $cb = shift;
		411	delete $PORT_DATA{$portid};
		412	$PORT{$portid} = sub {
		413	local $SELF = $port;
		414	eval {
228		415	&$cb
229	$PORT{$port} = $PORT{$lport} # also return	416	and kil $port;
		417	};
		418	_self_die if $@;
230	};	419	};
231
232	if (!ref $match) {
233	push @{ $port->{rc0}{$match} }, [$cb];
234	} elsif (("ARRAY" eq ref $match && !ref $match->[0])) {
235	my ($type, @match) = @$match;
236	@match
237	? push @{ $port->{rcv}{$match->[0]} }, [$cb, \@match]
238	: push @{ $port->{rc0}{$match->[0]} }, [$cb];
239	} else {	420	} else {
240	push @{ $port->{any} }, [$cb, $match];	421	my $self = $PORT_DATA{$portid} \|\|= do {
241	}	422	my $self = bless {
242	}	423	id => $port,
		424	}, "AnyEvent::MP::Port";
243		425
244	sub _inject {	426	$PORT{$portid} = sub {
245	my ($port, $msg) = @{+shift};	427	local $SELF = $port;
246		428
247	$port = $PORT{$port}	429	eval {
248	or return;	430	for (@{ $self->{rc0}{$_[0]} }) {
249
250	@_ = @$msg;
251
252	for (@{ $port->{rc0}{$msg->[0]} }) {
253	$_ && &{$_->[0]}	431	$_ && &{$_->[0]}
254	&& undef $_;	432	&& undef $_;
255	}
256
257	for (@{ $port->{rcv}{$msg->[0]} }) {
258	$_ && [@_[1..$#{$_->[1]}]] ~~ $_->[1]
259	&& &{$_->[0]}
260	&& undef $_;
261	}
262
263	for (@{ $port->{any} }) {
264	$_ && [@_[0..$#{$_->[1]}]] ~~ $_->[1]
265	&& &{$_->[0]}
266	&& undef $_;
267	}
268	}
269
270	sub normalise_noderef($) {
271	my ($noderef) = @_;
272
273	my $cv = AE::cv;
274	my @res;
275
276	$cv->begin (sub {
277	my %seen;
278	my @refs;
279	for (sort { $a->[0] <=> $b->[0] } @res) {
280	push @refs, $_->[1] unless $seen{$_->[1]}++
281	}
282	shift->send (join ",", @refs);
283	});
284
285	$noderef = $DEFAULT_PORT unless length $noderef;
286
287	my $idx;
288	for my $t (split /,/, $noderef) {
289	my $pri = ++$idx;
290
291	#TODO: this should be outside normalise_noderef and in become_public
292	if ($t =~ /^\d*$/) {
293	my $nodename = (POSIX::uname)[1];
294
295	$cv->begin;
296	AnyEvent::Socket::resolve_sockaddr $nodename, $t \|\| "aemp=$DEFAULT_PORT", "tcp", 0, undef, sub {
297	for (@_) {
298	my ($service, $host) = AnyEvent::Socket::unpack_sockaddr $_->[3];
299	push @res, [
300	$pri += 1e-5,
301	AnyEvent::Socket::format_hostport AnyEvent::Socket::format_address $host, $service
302	];	433	}
		434
		435	for (@{ $self->{rcv}{$_[0]} }) {
		436	$_ && [@_[1 .. @{$_->[1]}]] ~~ $_->[1]
		437	&& &{$_->[0]}
		438	&& undef $_;
		439	}
		440
		441	for (@{ $self->{any} }) {
		442	$_ && [@_[0 .. $#{$_->[1]}]] ~~ $_->[1]
		443	&& &{$_->[0]}
		444	&& undef $_;
		445	}
303	}	446	};
304	$cv->end;	447	_self_die if $@;
305	};	448	};
306		449
307	# my (undef, undef, undef, undef, @ipv4) = gethostbyname $nodename;	450	$self
308	#	451	};
309	# for (@ipv4) {	452
310	# push @res, [	453	"AnyEvent::MP::Port" eq ref $self
311	# $pri,	454	or Carp::croak "$port: rcv can only be called on message matching ports, caught";
312	# AnyEvent::Socket::format_hostport AnyEvent::Socket::format_address $_, $t \|\| $DEFAULT_PORT,	455
313	# ];	456	while (@_) {
314	# }	457	my ($match, $cb) = splice @_, 0, 2;
		458
		459	if (!ref $match) {
		460	push @{ $self->{rc0}{$match} }, [$cb];
		461	} elsif (("ARRAY" eq ref $match && !ref $match->[0])) {
		462	my ($type, @match) = @$match;
		463	@match
		464	? push @{ $self->{rcv}{$match->[0]} }, [$cb, \@match]
		465	: push @{ $self->{rc0}{$match->[0]} }, [$cb];
315	} else {	466	} else {
316	my ($host, $port) = AnyEvent::Socket::parse_hostport $t, "aemp=$DEFAULT_PORT"	467	push @{ $self->{any} }, [$cb, $match];
317	or Carp::croak "$t: unparsable transport descriptor";
318
319	$cv->begin;
320	AnyEvent::Socket::resolve_sockaddr $host, $port, "tcp", 0, undef, sub {
321	for (@_) {
322	my ($service, $host) = AnyEvent::Socket::unpack_sockaddr $_->[3];
323	push @res, [
324	$pri += 1e-5,
325	AnyEvent::Socket::format_hostport AnyEvent::Socket::format_address $host, $service
326	];
327	}
328	$cv->end;
329	}	468	}
330	}	469	}
331	}	470	}
332		471
333	$cv->end;	472	$port
334
335	$cv
336	}	473	}
337		474
338	sub become_public {	475	=item $closure = psub { BLOCK }
339	return if $PUBLIC;
340		476
341	my $noderef = join ",", ref $_[0] ? @{+shift} : shift;	477	Remembers C<$SELF> and creates a closure out of the BLOCK. When the
342	my @args = @_;	478	closure is executed, sets up the environment in the same way as in C<rcv>
		479	callbacks, i.e. runtime errors will cause the port to get C<kil>ed.
343		480
344	$NODE = (normalise_noderef $noderef)->recv;	481	This is useful when you register callbacks from C<rcv> callbacks:
345		482
346	for my $t (split /,/, $NODE) {	483	rcv delayed_reply => sub {
347	$NODE{$t} = $NODE{""};	484	my ($delay, @reply) = @_;
348		485	my $timer = AE::timer $delay, 0, psub {
349	my ($host, $port) = AnyEvent::Socket::parse_hostport $t;	486	snd @reply, $SELF;
350
351	$LISTENER{$t} = AnyEvent::MP::Transport::mp_server $host, $port,
352	@args,
353	on_error => sub {
354	die "on_error<@_>\n";#d#
355	},
356	on_connect => sub {
357	my ($tp) = @_;
358
359	$NODE{$tp->{remote_id}} = $_[0];
360	},
361	sub {
362	my ($tp) = @_;
363
364	$NODE{"$tp->{peerhost}:$tp->{peerport}"} = $tp;
365	},
366	;	487	};
		488	};
		489
		490	=cut
		491
		492	sub psub(&) {
		493	my $cb = shift;
		494
		495	my $port = $SELF
		496	or Carp::croak "psub can only be called from within rcv or psub callbacks, not";
		497
		498	sub {
		499	local $SELF = $port;
		500
		501	if (wantarray) {
		502	my @res = eval { &$cb };
		503	_self_die if $@;
		504	@res
		505	} else {
		506	my $res = eval { &$cb };
		507	_self_die if $@;
		508	$res
		509	}
367	}	510	}
		511	}
368		512
369	$PUBLIC = 1;	513	=item $guard = mon $port, $cb->(@reason)
		514
		515	=item $guard = mon $port, $rcvport
		516
		517	=item $guard = mon $port
		518
		519	=item $guard = mon $port, $rcvport, @msg
		520
		521	Monitor the given port and do something when the port is killed or
		522	messages to it were lost, and optionally return a guard that can be used
		523	to stop monitoring again.
		524
		525	C<mon> effectively guarantees that, in the absence of hardware failures,
		526	that after starting the monitor, either all messages sent to the port
		527	will arrive, or the monitoring action will be invoked after possible
		528	message loss has been detected. No messages will be lost "in between"
		529	(after the first lost message no further messages will be received by the
		530	port). After the monitoring action was invoked, further messages might get
		531	delivered again.
		532
		533	In the first form (callback), the callback is simply called with any
		534	number of C<@reason> elements (no @reason means that the port was deleted
		535	"normally"). Note also that I<< the callback B<must> never die >>, so use
		536	C<eval> if unsure.
		537
		538	In the second form (another port given), the other port (C<$rcvport>)
		539	will be C<kil>'ed with C<@reason>, iff a @reason was specified, i.e. on
		540	"normal" kils nothing happens, while under all other conditions, the other
		541	port is killed with the same reason.
		542
		543	The third form (kill self) is the same as the second form, except that
		544	C<$rvport> defaults to C<$SELF>.
		545
		546	In the last form (message), a message of the form C<@msg, @reason> will be
		547	C<snd>.
		548
		549	As a rule of thumb, monitoring requests should always monitor a port from
		550	a local port (or callback). The reason is that kill messages might get
		551	lost, just like any other message. Another less obvious reason is that
		552	even monitoring requests can get lost (for exmaple, when the connection
		553	to the other node goes down permanently). When monitoring a port locally
		554	these problems do not exist.
		555
		556	Example: call a given callback when C<$port> is killed.
		557
		558	mon $port, sub { warn "port died because of <@_>\n" };
		559
		560	Example: kill ourselves when C<$port> is killed abnormally.
		561
		562	mon $port;
		563
		564	Example: send us a restart message when another C<$port> is killed.
		565
		566	mon $port, $self => "restart";
		567
		568	=cut
		569
		570	sub mon {
		571	my ($noderef, $port) = split /#/, shift, 2;
		572
		573	my $node = $NODE{$noderef} \|\| add_node $noderef;
		574
		575	my $cb = @_ ? shift : $SELF \|\| Carp::croak 'mon: called with one argument only, but $SELF not set,';
		576
		577	unless (ref $cb) {
		578	if (@_) {
		579	# send a kill info message
		580	my (@msg) = ($cb, @_);
		581	$cb = sub { snd @msg, @_ };
		582	} else {
		583	# simply kill other port
		584	my $port = $cb;
		585	$cb = sub { kil $port, @_ if @_ };
		586	}
		587	}
		588
		589	$node->monitor ($port, $cb);
		590
		591	defined wantarray
		592	and AnyEvent::Util::guard { $node->unmonitor ($port, $cb) }
		593	}
		594
		595	=item $guard = mon_guard $port, $ref, $ref...
		596
		597	Monitors the given C<$port> and keeps the passed references. When the port
		598	is killed, the references will be freed.
		599
		600	Optionally returns a guard that will stop the monitoring.
		601
		602	This function is useful when you create e.g. timers or other watchers and
		603	want to free them when the port gets killed:
		604
		605	$port->rcv (start => sub {
		606	my $timer; $timer = mon_guard $port, AE::timer 1, 1, sub {
		607	undef $timer if 0.9 < rand;
		608	});
		609	});
		610
		611	=cut
		612
		613	sub mon_guard {
		614	my ($port, @refs) = @_;
		615
		616	#TODO: mon-less form?
		617
		618	mon $port, sub { 0 && @refs }
		619	}
		620
		621	=item kil $port[, @reason]
		622
		623	Kill the specified port with the given C<@reason>.
		624
		625	If no C<@reason> is specified, then the port is killed "normally" (linked
		626	ports will not be kileld, or even notified).
		627
		628	Otherwise, linked ports get killed with the same reason (second form of
		629	C<mon>, see below).
		630
		631	Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks
		632	will be reported as reason C<< die => $@ >>.
		633
		634	Transport/communication errors are reported as C<< transport_error =>
		635	$message >>.
		636
		637	=cut
		638
		639	=item $port = spawn $node, $initfunc[, @initdata]
		640
		641	Creates a port on the node C<$node> (which can also be a port ID, in which
		642	case it's the node where that port resides).
		643
		644	The port ID of the newly created port is return immediately, and it is
		645	permissible to immediately start sending messages or monitor the port.
		646
		647	After the port has been created, the init function is
		648	called. This function must be a fully-qualified function name
		649	(e.g. C<MyApp::Chat::Server::init>). To specify a function in the main
		650	program, use C<::name>.
		651
		652	If the function doesn't exist, then the node tries to C<require>
		653	the package, then the package above the package and so on (e.g.
		654	C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function
		655	exists or it runs out of package names.
		656
		657	The init function is then called with the newly-created port as context
		658	object (C<$SELF>) and the C<@initdata> values as arguments.
		659
		660	A common idiom is to pass your own port, monitor the spawned port, and
		661	in the init function, monitor the original port. This two-way monitoring
		662	ensures that both ports get cleaned up when there is a problem.
		663
		664	Example: spawn a chat server port on C<$othernode>.
		665
		666	# this node, executed from within a port context:
		667	my $server = spawn $othernode, "MyApp::Chat::Server::connect", $SELF;
		668	mon $server;
		669
		670	# init function on C<$othernode>
		671	sub connect {
		672	my ($srcport) = @_;
		673
		674	mon $srcport;
		675
		676	rcv $SELF, sub {
		677	...
		678	};
		679	}
		680
		681	=cut
		682
		683	sub _spawn {
		684	my $port = shift;
		685	my $init = shift;
		686
		687	local $SELF = "$NODE#$port";
		688	eval {
		689	&{ load_func $init }
		690	};
		691	_self_die if $@;
		692	}
		693
		694	sub spawn(@) {
		695	my ($noderef, undef) = split /#/, shift, 2;
		696
		697	my $id = "$RUNIQ." . $ID++;
		698
		699	$_[0] =~ /::/
		700	or Carp::croak "spawn init function must be a fully-qualified name, caught";
		701
		702	($NODE{$noderef} \|\| add_node $noderef)
		703	->send (["", "AnyEvent::MP::_spawn" => $id, @_]);
		704
		705	"$noderef#$id"
370	}	706	}
371		707
372	=back	708	=back
373		709
374	=head1 NODE MESSAGES	710	=head1 NODE MESSAGES
…		…
376	Nodes understand the following messages sent to them. Many of them take	712	Nodes understand the following messages sent to them. Many of them take
377	arguments called C<@reply>, which will simply be used to compose a reply	713	arguments called C<@reply>, which will simply be used to compose a reply
378	message - C<$reply[0]> is the port to reply to, C<$reply[1]> the type and	714	message - C<$reply[0]> is the port to reply to, C<$reply[1]> the type and
379	the remaining arguments are simply the message data.	715	the remaining arguments are simply the message data.
380		716
		717	While other messages exist, they are not public and subject to change.
		718
381	=over 4	719	=over 4
382		720
383	=cut	721	=cut
384		722
385	#############################################################################	723	=item lookup => $name, @reply
386	# self node code
387		724
388	sub _new_port($) {	725	Replies with the port ID of the specified well-known port, or C<undef>.
389	my ($name) = @_;
390
391	my ($noderef, $portname) = split /#/, $name;
392
393	$PORT{$name} =
394	$PORT{$portname} = {
395	names => [$name, $portname],
396	};
397	}
398
399	$NODE{""} = new AnyEvent::MP::Node::Self noderef => $NODE;
400	_new_port "";
401		726
402	=item devnull => ...	727	=item devnull => ...
403		728
404	Generic data sink/CPU heat conversion.	729	Generic data sink/CPU heat conversion.
405		730
406	=cut
407
408	rcv "", devnull => sub { () };
409
410	=item relay => $port, @msg	731	=item relay => $port, @msg
411		732
412	Simply forwards the message to the given port.	733	Simply forwards the message to the given port.
413
414	=cut
415
416	rcv "", relay => sub { \&snd; () };
417		734
418	=item eval => $string[ @reply]	735	=item eval => $string[ @reply]
419		736
420	Evaluates the given string. If C<@reply> is given, then a message of the	737	Evaluates the given string. If C<@reply> is given, then a message of the
421	form C<@reply, $@, @evalres> is sent.	738	form C<@reply, $@, @evalres> is sent.
422		739
423	Example: crash another node.	740	Example: crash another node.
424		741
425	snd $othernode, eval => "exit";	742	snd $othernode, eval => "exit";
426		743
427	=cut
428
429	rcv "", eval => sub {
430	my (undef, $string, @reply) = @_;
431	my @res = eval $string;
432	snd @reply, "$@", @res if @reply;
433	()
434	};
435
436	=item time => @reply	744	=item time => @reply
437		745
438	Replies the the current node time to C<@reply>.	746	Replies the the current node time to C<@reply>.
439		747
440	Example: tell the current node to send the current time to C<$myport> in a	748	Example: tell the current node to send the current time to C<$myport> in a
441	C<timereply> message.	749	C<timereply> message.
442		750
443	snd $NODE, time => $myport, timereply => 1, 2;	751	snd $NODE, time => $myport, timereply => 1, 2;
444	# => snd $myport, timereply => 1, 2, <time>	752	# => snd $myport, timereply => 1, 2, <time>
445		753
446	=cut	754	=back
447		755
448	rcv "", time => sub { shift; snd @_, AE::time; () };	756	=head1 AnyEvent::MP vs. Distributed Erlang
		757
		758	AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node
		759	== aemp node, Erlang process == aemp port), so many of the documents and
		760	programming techniques employed by Erlang apply to AnyEvent::MP. Here is a
		761	sample:
		762
		763	http://www.Erlang.se/doc/programming_rules.shtml
		764	http://Erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4
		765	http://Erlang.org/download/Erlang-book-part1.pdf # chapters 5 and 6
		766	http://Erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5
		767
		768	Despite the similarities, there are also some important differences:
		769
		770	=over 4
		771
		772	=item * Node references contain the recipe on how to contact them.
		773
		774	Erlang relies on special naming and DNS to work everywhere in the
		775	same way. AEMP relies on each node knowing it's own address(es), with
		776	convenience functionality.
		777
		778	This means that AEMP requires a less tightly controlled environment at the
		779	cost of longer node references and a slightly higher management overhead.
		780
		781	=item * Erlang uses processes and a mailbox, AEMP does not queue.
		782
		783	Erlang uses processes that selctively receive messages, and therefore
		784	needs a queue. AEMP is event based, queuing messages would serve no useful
		785	purpose.
		786
		787	(But see L<Coro::MP> for a more Erlang-like process model on top of AEMP).
		788
		789	=item * Erlang sends are synchronous, AEMP sends are asynchronous.
		790
		791	Sending messages in Erlang is synchronous and blocks the process. AEMP
		792	sends are immediate, connection establishment is handled in the
		793	background.
		794
		795	=item * Erlang can silently lose messages, AEMP cannot.
		796
		797	Erlang makes few guarantees on messages delivery - messages can get lost
		798	without any of the processes realising it (i.e. you send messages a, b,
		799	and c, and the other side only receives messages a and c).
		800
		801	AEMP guarantees correct ordering, and the guarantee that there are no
		802	holes in the message sequence.
		803
		804	=item * In Erlang, processes can be declared dead and later be found to be
		805	alive.
		806
		807	In Erlang it can happen that a monitored process is declared dead and
		808	linked processes get killed, but later it turns out that the process is
		809	still alive - and can receive messages.
		810
		811	In AEMP, when port monitoring detects a port as dead, then that port will
		812	eventually be killed - it cannot happen that a node detects a port as dead
		813	and then later sends messages to it, finding it is still alive.
		814
		815	=item * Erlang can send messages to the wrong port, AEMP does not.
		816
		817	In Erlang it is quite possible that a node that restarts reuses a process
		818	ID known to other nodes for a completely different process, causing
		819	messages destined for that process to end up in an unrelated process.
		820
		821	AEMP never reuses port IDs, so old messages or old port IDs floating
		822	around in the network will not be sent to an unrelated port.
		823
		824	=item * Erlang uses unprotected connections, AEMP uses secure
		825	authentication and can use TLS.
		826
		827	AEMP can use a proven protocol - SSL/TLS - to protect connections and
		828	securely authenticate nodes.
		829
		830	=item * The AEMP protocol is optimised for both text-based and binary
		831	communications.
		832
		833	The AEMP protocol, unlike the Erlang protocol, supports both
		834	language-independent text-only protocols (good for debugging) and binary,
		835	language-specific serialisers (e.g. Storable).
		836
		837	It has also been carefully designed to be implementable in other languages
		838	with a minimum of work while gracefully degrading fucntionality to make the
		839	protocol simple.
		840
		841	=item * AEMP has more flexible monitoring options than Erlang.
		842
		843	In Erlang, you can chose to receive I<all> exit signals as messages
		844	or I<none>, there is no in-between, so monitoring single processes is
		845	difficult to implement. Monitoring in AEMP is more flexible than in
		846	Erlang, as one can choose between automatic kill, exit message or callback
		847	on a per-process basis.
		848
		849	=item * Erlang tries to hide remote/local connections, AEMP does not.
		850
		851	Monitoring in Erlang is not an indicator of process death/crashes,
		852	as linking is (except linking is unreliable in Erlang).
		853
		854	In AEMP, you don't "look up" registered port names or send to named ports
		855	that might or might not be persistent. Instead, you normally spawn a port
		856	on the remote node. The init function monitors the you, and you monitor
		857	the remote port. Since both monitors are local to the node, they are much
		858	more reliable.
		859
		860	This also saves round-trips and avoids sending messages to the wrong port
		861	(hard to do in Erlang).
		862
		863	=back
		864
		865	=head1 RATIONALE
		866
		867	=over 4
		868
		869	=item Why strings for ports and noderefs, why not objects?
		870
		871	We considered "objects", but found that the actual number of methods
		872	thatc an be called are very low. Since port IDs and noderefs travel over
		873	the network frequently, the serialising/deserialising would add lots of
		874	overhead, as well as having to keep a proxy object.
		875
		876	Strings can easily be printed, easily serialised etc. and need no special
		877	procedures to be "valid".
		878
		879	And a a miniport consists of a single closure stored in a global hash - it
		880	can't become much cheaper.
		881
		882	=item Why favour JSON, why not real serialising format such as Storable?
		883
		884	In fact, any AnyEvent::MP node will happily accept Storable as framing
		885	format, but currently there is no way to make a node use Storable by
		886	default.
		887
		888	The default framing protocol is JSON because a) JSON::XS is many times
		889	faster for small messages and b) most importantly, after years of
		890	experience we found that object serialisation is causing more problems
		891	than it gains: Just like function calls, objects simply do not travel
		892	easily over the network, mostly because they will always be a copy, so you
		893	always have to re-think your design.
		894
		895	Keeping your messages simple, concentrating on data structures rather than
		896	objects, will keep your messages clean, tidy and efficient.
449		897
450	=back	898	=back
451		899
452	=head1 SEE ALSO	900	=head1 SEE ALSO
453		901

Diff Legend

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

Comparing AnyEvent-MP/MP.pm (file contents): Revision 1.7 by root, Sat Aug 1 15:04:30 2009 UTC vs. Revision 1.48 by root, Thu Aug 13 02:59:42 2009 UTC

Diff Legend

Comparing AnyEvent-MP/MP.pm (file contents):
Revision 1.7 by root, Sat Aug 1 15:04:30 2009 UTC vs.
Revision 1.48 by root, Thu Aug 13 02:59:42 2009 UTC