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

Comparing AnyEvent-MP/MP.pm (file contents):
Revision 1.35 by root, Thu Aug 6 10:21:48 2009 UTC vs.
Revision 1.77 by elmex, Thu Sep 3 07:57:30 2009 UTC

…		…
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	use AnyEvent::MP;	7	use AnyEvent::MP;
8		8
9	$NODE # contains this node's noderef	9	$NODE # contains this node's node ID
10	NODE # returns this node's noderef	10	NODE # returns this node's node ID
11	NODE $port # returns the noderef of the port
12		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
13	snd $port, type => data...;	20	snd $port, type => data...;
		21	snd $port, @msg;
		22	snd @msg_with_first_element_being_a_port;
14		23
15	$SELF # receiving/own port id in rcv callbacks	24	# creating/using ports, the simple way
		25	my $simple_port = port { my @msg = @_ };
16		26
17	rcv $port, smartmatch => $cb->($port, @msg);	27	# creating/using ports, tagged message matching
18		28	my $port = port;
19	# examples:
20	rcv $port2, ping => sub { snd $_[0], "pong"; 0 };	29	rcv $port, ping => sub { snd $_[0], "pong" };
21	rcv $port1, pong => sub { warn "pong received\n" };	30	rcv $port, pong => sub { warn "pong received\n" };
22	snd $port2, ping => $port1;
23		31
24	# more, smarter, matches (_any_ is exported by this module)	32	# create a port on another node
25	rcv $port, [child_died => $pid] => sub { ...	33	my $port = spawn $node, $initfunc, @initdata;
26	rcv $port, [_any_, _any_, 3] => sub { .. $_[2] is 3
27
28	# linking two ports, so they both crash together
29	lnk $port1, $port2;
30		34
31	# monitoring	35	# monitoring
32	mon $port, $cb->(@msg) # callback is invoked on death	36	mon $port, $cb->(@msg) # callback is invoked on death
33	mon $port, $otherport # kill otherport on abnormal death	37	mon $port, $otherport # kill otherport on abnormal death
34	mon $port, $otherport, @msg # send message on death	38	mon $port, $otherport, @msg # send message on death
35		39
		40	=head1 CURRENT STATUS
		41
		42	bin/aemp - stable.
		43	AnyEvent::MP - stable API, should work.
		44	AnyEvent::MP::Intro - explains most concepts.
		45	AnyEvent::MP::Kernel - mostly stable.
		46	AnyEvent::MP::Global - stable API, protocol not yet final.
		47
		48	stay tuned.
		49
36	=head1 DESCRIPTION	50	=head1 DESCRIPTION
37		51
38	This module (-family) implements a simple message passing framework.	52	This module (-family) implements a simple message passing framework.
39		53
40	Despite its simplicity, you can securely message other processes running	54	Despite its simplicity, you can securely message other processes running
41	on the same or other hosts.	55	on the same or other hosts, and you can supervise entities remotely.
42		56
43	For an introduction to this module family, see the L<AnyEvent::MP::Intro>	57	For an introduction to this module family, see the L<AnyEvent::MP::Intro>
44	manual page.	58	manual page and the examples under F<eg/>.
45
46	At the moment, this module family is severly broken and underdocumented,
47	so do not use. This was uploaded mainly to reserve the CPAN namespace -
48	stay tuned! The basic API should be finished, however.
49		59
50	=head1 CONCEPTS	60	=head1 CONCEPTS
51		61
52	=over 4	62	=over 4
53		63
54	=item port	64	=item port
55		65
56	A port is something you can send messages to (with the C<snd> function).	66	A port is something you can send messages to (with the C<snd> function).
57		67
58	Some ports allow you to register C<rcv> handlers that can match specific	68	Ports allow you to register C<rcv> handlers that can match all or just
59	messages. All C<rcv> handlers will receive messages they match, messages	69	some messages. Messages send to ports will not be queued, regardless of
60	will not be queued.	70	anything was listening for them or not.
61		71
62	=item port id - C<noderef#portname>	72	=item port ID - C<nodeid#portname>
63		73
64	A port id is normaly the concatenation of a noderef, a hash-mark (C<#>) as	74	A port ID is the concatenation of a node ID, a hash-mark (C<#>) as
65	separator, and a port name (a printable string of unspecified format). An	75	separator, and a port name (a printable string of unspecified format).
66	exception is the the node port, whose ID is identical to its node
67	reference.
68		76
69	=item node	77	=item node
70		78
71	A node is a single process containing at least one port - the node	79	A node is a single process containing at least one port - the node port,
72	port. You can send messages to node ports to find existing ports or to	80	which enables nodes to manage each other remotely, and to create new
73	create new ports, among other things.	81	ports.
74		82
75	Nodes are either private (single-process only), slaves (connected to a	83	Nodes are either public (have one or more listening ports) or private
76	master node only) or public nodes (connectable from unrelated nodes).	84	(no listening ports). Private nodes cannot talk to other private nodes
		85	currently.
77		86
78	=item noderef - C<host:port,host:port...>, C<id@noderef>, C<id>	87	=item node ID - C<[a-za-Z0-9_\-.:]+>
79		88
80	A node reference is a string that either simply identifies the node (for	89	A node ID is a string that uniquely identifies the node within a
81	private and slave nodes), or contains a recipe on how to reach a given	90	network. Depending on the configuration used, node IDs can look like a
82	node (for public nodes).	91	hostname, a hostname and a port, or a random string. AnyEvent::MP itself
		92	doesn't interpret node IDs in any way.
83		93
84	This recipe is simply a comma-separated list of C<address:port> pairs (for	94	=item binds - C<ip:port>
85	TCP/IP, other protocols might look different).
86		95
87	Node references come in two flavours: resolved (containing only numerical	96	Nodes can only talk to each other by creating some kind of connection to
88	addresses) or unresolved (where hostnames are used instead of addresses).	97	each other. To do this, nodes should listen on one or more local transport
		98	endpoints - binds. Currently, only standard C<ip:port> specifications can
		99	be used, which specify TCP ports to listen on.
89		100
90	Before using an unresolved node reference in a message you first have to	101	=item seeds - C<host:port>
91	resolve it.	102
		103	When a node starts, it knows nothing about the network. To teach the node
		104	about the network it first has to contact some other node within the
		105	network. This node is called a seed.
		106
		107	Seeds are transport endpoint(s) of as many nodes as one wants. Those nodes
		108	are expected to be long-running, and at least one of those should always
		109	be available. When nodes run out of connections (e.g. due to a network
		110	error), they try to re-establish connections to some seednodes again to
		111	join the network.
		112
		113	Apart from being sued for seeding, seednodes are not special in any way -
		114	every public node can be a seednode.
92		115
93	=back	116	=back
94		117
95	=head1 VARIABLES/FUNCTIONS	118	=head1 VARIABLES/FUNCTIONS
96		119
…		…
98		121
99	=cut	122	=cut
100		123
101	package AnyEvent::MP;	124	package AnyEvent::MP;
102		125
103	use AnyEvent::MP::Base;	126	use AnyEvent::MP::Kernel;
104		127
105	use common::sense;	128	use common::sense;
106		129
107	use Carp ();	130	use Carp ();
108		131
109	use AE ();	132	use AE ();
110		133
111	use base "Exporter";	134	use base "Exporter";
112		135
113	our $VERSION = '0.1';	136	our $VERSION = $AnyEvent::MP::Kernel::VERSION;
		137
114	our @EXPORT = qw(	138	our @EXPORT = qw(
115	NODE $NODE *SELF node_of _any_	139	NODE $NODE *SELF node_of after
116	resolve_node initialise_node	140	configure
117	snd rcv mon kil reg psub	141	snd rcv mon mon_guard kil reg psub spawn
118	port	142	port
119	);	143	);
120		144
121	our $SELF;	145	our $SELF;
122		146
…		…
126	kil $SELF, die => $msg;	150	kil $SELF, die => $msg;
127	}	151	}
128		152
129	=item $thisnode = NODE / $NODE	153	=item $thisnode = NODE / $NODE
130		154
131	The C<NODE> function returns, and the C<$NODE> variable contains	155	The C<NODE> function returns, and the C<$NODE> variable contains, the node
132	the noderef of the local node. The value is initialised by a call	156	ID of the node running in the current process. This value is initialised by
133	to C<become_public> or C<become_slave>, after which all local port	157	a call to C<configure>.
134	identifiers become invalid.
135		158
136	=item $noderef = node_of $port	159	=item $nodeid = node_of $port
137		160
138	Extracts and returns the noderef from a portid or a noderef.	161	Extracts and returns the node ID from a port ID or a node ID.
139		162
140	=item initialise_node $noderef, $seednode, $seednode...	163	=item configure key => value...
141		164
142	=item initialise_node "slave/", $master, $master...
143
144	Before a node can talk to other nodes on the network it has to initialise	165	Before a node can talk to other nodes on the network (i.e. enter
145	itself - the minimum a node needs to know is it's own name, and optionally	166	"distributed mode") it has to configure itself - the minimum a node needs
146	it should know the noderefs of some other nodes in the network.	167	to know is its own name, and optionally it should know the addresses of
		168	some other nodes in the network to discover other nodes.
147		169
148	This function initialises a node - it must be called exactly once (or	170	This function configures a node - it must be called exactly once (or
149	never) before calling other AnyEvent::MP functions.	171	never) before calling other AnyEvent::MP functions.
150		172
151	All arguments are noderefs, which can be either resolved or unresolved.
152
153	There are two types of networked nodes, public nodes and slave nodes:
154
155	=over 4	173	=over 4
156		174
157	=item public nodes	175	=item step 1, gathering configuration from profiles
158		176
159	For public nodes, C<$noderef> must either be a (possibly unresolved)	177	The function first looks up a profile in the aemp configuration (see the
160	noderef, in which case it will be resolved, or C<undef> (or missing), in	178	L<aemp> commandline utility). The profile name can be specified via the
161	which case the noderef will be guessed.	179	named C<profile> parameter. If it is missing, then the nodename (F<uname
		180	-n>) will be used as profile name.
162		181
163	Afterwards, the node will bind itself on all endpoints and try to connect	182	The profile data is then gathered as follows:
164	to all additional C<$seednodes> that are specified. Seednodes are optional
165	and can be used to quickly bootstrap the node into an existing network.
166		183
167	=item slave nodes	184	First, all remaining key => value pairs (all of which are conveniently
		185	undocumented at the moment) will be interpreted as configuration
		186	data. Then they will be overwritten by any values specified in the global
		187	default configuration (see the F<aemp> utility), then the chain of
		188	profiles chosen by the profile name (and any C<parent> attributes).
168		189
169	When the C<$noderef> is the special string C<slave/>, then the node will	190	That means that the values specified in the profile have highest priority
170	become a slave node. Slave nodes cannot be contacted from outside and will	191	and the values specified directly via C<configure> have lowest priority,
171	route most of their traffic to the master node that they attach to.	192	and can only be used to specify defaults.
172		193
173	At least one additional noderef is required: The node will try to connect	194	If the profile specifies a node ID, then this will become the node ID of
174	to all of them and will become a slave attached to the first node it can	195	this process. If not, then the profile name will be used as node ID. The
175	successfully connect to.	196	special node ID of C<anon/> will be replaced by a random node ID.
		197
		198	=item step 2, bind listener sockets
		199
		200	The next step is to look up the binds in the profile, followed by binding
		201	aemp protocol listeners on all binds specified (it is possible and valid
		202	to have no binds, meaning that the node cannot be contacted form the
		203	outside. This means the node cannot talk to other nodes that also have no
		204	binds, but it can still talk to all "normal" nodes).
		205
		206	If the profile does not specify a binds list, then a default of C<*> is
		207	used, meaning the node will bind on a dynamically-assigned port on every
		208	local IP address it finds.
		209
		210	=item step 3, connect to seed nodes
		211
		212	As the last step, the seeds list from the profile is passed to the
		213	L<AnyEvent::MP::Global> module, which will then use it to keep
		214	connectivity with at least one node at any point in time.
176		215
177	=back	216	=back
178		217
179	This function will block until all nodes have been resolved and, for slave	218	Example: become a distributed node using the locla node name as profile.
180	nodes, until it has successfully established a connection to a master	219	This should be the most common form of invocation for "daemon"-type nodes.
181	server.
182		220
183	Example: become a public node listening on the default node.	221	configure
184		222
185	initialise_node;	223	Example: become an anonymous node. This form is often used for commandline
		224	clients.
186		225
187	Example: become a public node, and try to contact some well-known master	226	configure nodeid => "anon/";
188	servers to become part of the network.
189		227
190	initialise_node undef, "master1", "master2";	228	Example: configure a node using a profile called seed, which si suitable
		229	for a seed node as it binds on all local addresses on a fixed port (4040,
		230	customary for aemp).
191		231
192	Example: become a public node listening on port C<4041>.	232	# use the aemp commandline utility
		233	# aemp profile seed nodeid anon/ binds '*:4040'
193		234
194	initialise_node 4041;	235	# then use it
		236	configure profile => "seed";
195		237
196	Example: become a public node, only visible on localhost port 4044.	238	# or simply use aemp from the shell again:
		239	# aemp run profile seed
197		240
198	initialise_node "locahost:4044";	241	# or provide a nicer-to-remember nodeid
199		242	# aemp run profile seed nodeid "$(hostname)"
200	Example: become a slave node to any of the specified master servers.
201
202	initialise_node "slave/", "master1", "192.168.13.17", "mp.example.net";
203
204	=item $cv = resolve_node $noderef
205
206	Takes an unresolved node reference that may contain hostnames and
207	abbreviated IDs, resolves all of them and returns a resolved node
208	reference.
209
210	In addition to C<address:port> pairs allowed in resolved noderefs, the
211	following forms are supported:
212
213	=over 4
214
215	=item the empty string
216
217	An empty-string component gets resolved as if the default port (4040) was
218	specified.
219
220	=item naked port numbers (e.g. C<1234>)
221
222	These are resolved by prepending the local nodename and a colon, to be
223	further resolved.
224
225	=item hostnames (e.g. C<localhost:1234>, C<localhost>)
226
227	These are resolved by using AnyEvent::DNS to resolve them, optionally
228	looking up SRV records for the C<aemp=4040> port, if no port was
229	specified.
230
231	=back
232		243
233	=item $SELF	244	=item $SELF
234		245
235	Contains the current port id while executing C<rcv> callbacks or C<psub>	246	Contains the current port id while executing C<rcv> callbacks or C<psub>
236	blocks.	247	blocks.
237		248
238	=item SELF, %SELF, @SELF...	249	=item *SELF, SELF, %SELF, @SELF...
239		250
240	Due to some quirks in how perl exports variables, it is impossible to	251	Due to some quirks in how perl exports variables, it is impossible to
241	just export C<$SELF>, all the symbols called C<SELF> are exported by this	252	just export C<$SELF>, all the symbols named C<SELF> are exported by this
242	module, but only C<$SELF> is currently used.	253	module, but only C<$SELF> is currently used.
243		254
244	=item snd $port, type => @data	255	=item snd $port, type => @data
245		256
246	=item snd $port, @msg	257	=item snd $port, @msg
247		258
248	Send the given message to the given port ID, which can identify either	259	Send the given message to the given port, which can identify either a
249	a local or a remote port, and can be either a string or soemthignt hat	260	local or a remote port, and must be a port ID.
250	stringifies a sa port ID (such as a port object :).
251		261
252	While the message can be about anything, it is highly recommended to use a	262	While the message can be almost anything, it is highly recommended to
253	string as first element (a portid, or some word that indicates a request	263	use a string as first element (a port ID, or some word that indicates a
254	type etc.).	264	request type etc.) and to consist if only simple perl values (scalars,
		265	arrays, hashes) - if you think you need to pass an object, think again.
255		266
256	The message data effectively becomes read-only after a call to this	267	The message data logically becomes read-only after a call to this
257	function: modifying any argument is not allowed and can cause many	268	function: modifying any argument (or values referenced by them) is
258	problems.	269	forbidden, as there can be considerable time between the call to C<snd>
		270	and the time the message is actually being serialised - in fact, it might
		271	never be copied as within the same process it is simply handed to the
		272	receiving port.
259		273
260	The type of data you can transfer depends on the transport protocol: when	274	The type of data you can transfer depends on the transport protocol: when
261	JSON is used, then only strings, numbers and arrays and hashes consisting	275	JSON is used, then only strings, numbers and arrays and hashes consisting
262	of those are allowed (no objects). When Storable is used, then anything	276	of those are allowed (no objects). When Storable is used, then anything
263	that Storable can serialise and deserialise is allowed, and for the local	277	that Storable can serialise and deserialise is allowed, and for the local
264	node, anything can be passed.	278	node, anything can be passed. Best rely only on the common denominator of
		279	these.
265		280
266	=item $local_port = port	281	=item $local_port = port
267		282
268	Create a new local port object that can be used either as a pattern	283	Create a new local port object and returns its port ID. Initially it has
269	matching port ("full port") or a single-callback port ("miniport"),	284	no callbacks set and will throw an error when it receives messages.
270	depending on how C<rcv> callbacks are bound to the object.
271		285
272	=item $port = port { my @msg = @_; $finished }	286	=item $local_port = port { my @msg = @_ }
273		287
274	Creates a "miniport", that is, a very lightweight port without any pattern	288	Creates a new local port, and returns its ID. Semantically the same as
275	matching behind it, and returns its ID. Semantically the same as creating
276	a port and calling C<rcv $port, $callback> on it.	289	creating a port and calling C<rcv $port, $callback> on it.
277		290
278	The block will be called for every message received on the port. When the	291	The block will be called for every message received on the port, with the
279	callback returns a true value its job is considered "done" and the port	292	global variable C<$SELF> set to the port ID. Runtime errors will cause the
280	will be destroyed. Otherwise it will stay alive.	293	port to be C<kil>ed. The message will be passed as-is, no extra argument
		294	(i.e. no port ID) will be passed to the callback.
281		295
282	The message will be passed as-is, no extra argument (i.e. no port id) will	296	If you want to stop/destroy the port, simply C<kil> it:
283	be passed to the callback.
284		297
285	If you need the local port id in the callback, this works nicely:	298	my $port = port {
286		299	my @msg = @_;
287	my $port; $port = port {	300	...
288	snd $otherport, reply => $port;	301	kil $SELF;
289	};	302	};
290		303
291	=cut	304	=cut
292		305
293	sub rcv($@);	306	sub rcv($@);
		307
		308	sub _kilme {
		309	die "received message on port without callback";
		310	}
294		311
295	sub port(;&) {	312	sub port(;&) {
296	my $id = "$UNIQ." . $ID++;	313	my $id = "$UNIQ." . $ID++;
297	my $port = "$NODE#$id";	314	my $port = "$NODE#$id";
298		315
299	if (@_) {	316	rcv $port, shift \|\| \&_kilme;
300	rcv $port, shift;
301	} else {
302	$PORT{$id} = sub { }; # nop
303	}
304		317
305	$port	318	$port
306	}	319	}
307		320
308	=item reg $port, $name
309
310	Registers the given port under the name C<$name>. If the name already
311	exists it is replaced.
312
313	A port can only be registered under one well known name.
314
315	A port automatically becomes unregistered when it is killed.
316
317	=cut
318
319	sub reg(@) {
320	my ($port, $name) = @_;
321
322	$REG{$name} = $port;
323	}
324
325	=item rcv $port, $callback->(@msg)	321	=item rcv $local_port, $callback->(@msg)
326		322
327	Replaces the callback on the specified miniport (after converting it to	323	Replaces the default callback on the specified port. There is no way to
328	one if required).	324	remove the default callback: use C<sub { }> to disable it, or better
329		325	C<kil> the port when it is no longer needed.
330	=item rcv $port, tagstring => $callback->(@msg), ...
331
332	=item rcv $port, $smartmatch => $callback->(@msg), ...
333
334	=item rcv $port, [$smartmatch...] => $callback->(@msg), ...
335
336	Register callbacks to be called on matching messages on the given full
337	port (after converting it to one if required).
338
339	The callback has to return a true value when its work is done, after
340	which is will be removed, or a false value in which case it will stay
341	registered.
342		326
343	The global C<$SELF> (exported by this module) contains C<$port> while	327	The global C<$SELF> (exported by this module) contains C<$port> while
344	executing the callback.	328	executing the callback. Runtime errors during callback execution will
		329	result in the port being C<kil>ed.
345		330
346	Runtime errors wdurign callback execution will result in the port being	331	The default callback received all messages not matched by a more specific
347	C<kil>ed.	332	C<tag> match.
348		333
349	If the match is an array reference, then it will be matched against the	334	=item rcv $local_port, tag => $callback->(@msg_without_tag), ...
350	first elements of the message, otherwise only the first element is being
351	matched.
352		335
353	Any element in the match that is specified as C<_any_> (a function	336	Register (or replace) callbacks to be called on messages starting with the
354	exported by this module) matches any single element of the message.	337	given tag on the given port (and return the port), or unregister it (when
		338	C<$callback> is C<$undef> or missing). There can only be one callback
		339	registered for each tag.
355		340
356	While not required, it is highly recommended that the first matching	341	The original message will be passed to the callback, after the first
357	element is a string identifying the message. The one-string-only match is	342	element (the tag) has been removed. The callback will use the same
358	also the most efficient match (by far).	343	environment as the default callback (see above).
		344
		345	Example: create a port and bind receivers on it in one go.
		346
		347	my $port = rcv port,
		348	msg1 => sub { ... },
		349	msg2 => sub { ... },
		350	;
		351
		352	Example: create a port, bind receivers and send it in a message elsewhere
		353	in one go:
		354
		355	snd $otherport, reply =>
		356	rcv port,
		357	msg1 => sub { ... },
		358	...
		359	;
		360
		361	Example: temporarily register a rcv callback for a tag matching some port
		362	(e.g. for a rpc reply) and unregister it after a message was received.
		363
		364	rcv $port, $otherport => sub {
		365	my @reply = @_;
		366
		367	rcv $SELF, $otherport;
		368	};
359		369
360	=cut	370	=cut
361		371
362	sub rcv($@) {	372	sub rcv($@) {
363	my $port = shift;	373	my $port = shift;
364	my ($noderef, $portid) = split /#/, $port, 2;	374	my ($nodeid, $portid) = split /#/, $port, 2;
365		375
366	($NODE{$noderef} \|\| add_node $noderef) == $NODE{""}	376	$NODE{$nodeid} == $NODE{""}
367	or Carp::croak "$port: rcv can only be called on local ports, caught";	377	or Carp::croak "$port: rcv can only be called on local ports, caught";
368		378
369	if (@_ == 1) {	379	while (@_) {
		380	if (ref $_[0]) {
		381	if (my $self = $PORT_DATA{$portid}) {
		382	"AnyEvent::MP::Port" eq ref $self
		383	or Carp::croak "$port: rcv can only be called on message matching ports, caught";
		384
		385	$self->[2] = shift;
		386	} else {
370	my $cb = shift;	387	my $cb = shift;
371	delete $PORT_DATA{$portid};
372	$PORT{$portid} = sub {	388	$PORT{$portid} = sub {
373	local $SELF = $port;	389	local $SELF = $port;
374	eval {	390	eval { &$cb }; _self_die if $@;
375	&$cb	391	};
376	and kil $port;
377	};	392	}
378	_self_die if $@;	393	} elsif (defined $_[0]) {
379	};
380	} else {
381	my $self = $PORT_DATA{$portid} \|\|= do {	394	my $self = $PORT_DATA{$portid} \|\|= do {
382	my $self = bless {	395	my $self = bless [$PORT{$port} \|\| sub { }, { }, $port], "AnyEvent::MP::Port";
383	id => $port,
384	}, "AnyEvent::MP::Port";
385		396
386	$PORT{$portid} = sub {	397	$PORT{$portid} = sub {
387	local $SELF = $port;	398	local $SELF = $port;
388		399
389	eval {
390	for (@{ $self->{rc0}{$_[0]} }) {	400	if (my $cb = $self->[1]{$_[0]}) {
391	$_ && &{$_->[0]}	401	shift;
392	&& undef $_;	402	eval { &$cb }; _self_die if $@;
393	}	403	} else {
394
395	for (@{ $self->{rcv}{$_[0]} }) {
396	$_ && [@_[1 .. @{$_->[1]}]] ~~ $_->[1]
397	&& &{$_->[0]}	404	&{ $self->[0] };
398	&& undef $_;
399	}
400
401	for (@{ $self->{any} }) {
402	$_ && [@_[0 .. $#{$_->[1]}]] ~~ $_->[1]
403	&& &{$_->[0]}
404	&& undef $_;
405	}	405	}
406	};	406	};
407	_self_die if $@;	407
		408	$self
408	};	409	};
409		410
410	$self
411	};
412
413	"AnyEvent::MP::Port" eq ref $self	411	"AnyEvent::MP::Port" eq ref $self
414	or Carp::croak "$port: rcv can only be called on message matching ports, caught";	412	or Carp::croak "$port: rcv can only be called on message matching ports, caught";
415		413
416	while (@_) {
417	my ($match, $cb) = splice @_, 0, 2;	414	my ($tag, $cb) = splice @_, 0, 2;
418		415
419	if (!ref $match) {	416	if (defined $cb) {
420	push @{ $self->{rc0}{$match} }, [$cb];	417	$self->[1]{$tag} = $cb;
421	} elsif (("ARRAY" eq ref $match && !ref $match->[0])) {
422	my ($type, @match) = @$match;
423	@match
424	? push @{ $self->{rcv}{$match->[0]} }, [$cb, \@match]
425	: push @{ $self->{rc0}{$match->[0]} }, [$cb];
426	} else {	418	} else {
427	push @{ $self->{any} }, [$cb, $match];	419	delete $self->[1]{$tag};
428	}	420	}
429	}	421	}
430	}	422	}
431		423
432	$port	424	$port
…		…
468	$res	460	$res
469	}	461	}
470	}	462	}
471	}	463	}
472		464
473	=item $guard = mon $port, $cb->(@reason)	465	=item $guard = mon $port, $cb->(@reason) # call $cb when $port dies
474		466
475	=item $guard = mon $port, $otherport	467	=item $guard = mon $port, $rcvport # kill $rcvport when $port dies
476		468
477	=item $guard = mon $port, $otherport, @msg	469	=item $guard = mon $port # kill $SELF when $port dies
478		470
		471	=item $guard = mon $port, $rcvport, @msg # send a message when $port dies
		472
479	Monitor the given port and do something when the port is killed.	473	Monitor the given port and do something when the port is killed or
		474	messages to it were lost, and optionally return a guard that can be used
		475	to stop monitoring again.
480		476
		477	C<mon> effectively guarantees that, in the absence of hardware failures,
		478	after starting the monitor, either all messages sent to the port will
		479	arrive, or the monitoring action will be invoked after possible message
		480	loss has been detected. No messages will be lost "in between" (after
		481	the first lost message no further messages will be received by the
		482	port). After the monitoring action was invoked, further messages might get
		483	delivered again.
		484
		485	Note that monitoring-actions are one-shot: once messages are lost (and a
		486	monitoring alert was raised), they are removed and will not trigger again.
		487
481	In the first form, the callback is simply called with any number	488	In the first form (callback), the callback is simply called with any
482	of C<@reason> elements (no @reason means that the port was deleted	489	number of C<@reason> elements (no @reason means that the port was deleted
483	"normally"). Note also that I<< the callback B<must> never die >>, so use	490	"normally"). Note also that I<< the callback B<must> never die >>, so use
484	C<eval> if unsure.	491	C<eval> if unsure.
485		492
486	In the second form, the other port will be C<kil>'ed with C<@reason>, iff	493	In the second form (another port given), the other port (C<$rcvport>)
487	a @reason was specified, i.e. on "normal" kils nothing happens, while	494	will be C<kil>'ed with C<@reason>, if a @reason was specified, i.e. on
488	under all other conditions, the other port is killed with the same reason.	495	"normal" kils nothing happens, while under all other conditions, the other
		496	port is killed with the same reason.
489		497
		498	The third form (kill self) is the same as the second form, except that
		499	C<$rvport> defaults to C<$SELF>.
		500
490	In the last form, a message of the form C<@msg, @reason> will be C<snd>.	501	In the last form (message), a message of the form C<@msg, @reason> will be
		502	C<snd>.
		503
		504	As a rule of thumb, monitoring requests should always monitor a port from
		505	a local port (or callback). The reason is that kill messages might get
		506	lost, just like any other message. Another less obvious reason is that
		507	even monitoring requests can get lost (for example, when the connection
		508	to the other node goes down permanently). When monitoring a port locally
		509	these problems do not exist.
491		510
492	Example: call a given callback when C<$port> is killed.	511	Example: call a given callback when C<$port> is killed.
493		512
494	mon $port, sub { warn "port died because of <@_>\n" };	513	mon $port, sub { warn "port died because of <@_>\n" };
495		514
496	Example: kill ourselves when C<$port> is killed abnormally.	515	Example: kill ourselves when C<$port> is killed abnormally.
497		516
498	mon $port, $self;	517	mon $port;
499		518
500	Example: send us a restart message another C<$port> is killed.	519	Example: send us a restart message when another C<$port> is killed.
501		520
502	mon $port, $self => "restart";	521	mon $port, $self => "restart";
503		522
504	=cut	523	=cut
505		524
506	sub mon {	525	sub mon {
507	my ($noderef, $port) = split /#/, shift, 2;	526	my ($nodeid, $port) = split /#/, shift, 2;
508		527
509	my $node = $NODE{$noderef} \|\| add_node $noderef;	528	my $node = $NODE{$nodeid} \|\| add_node $nodeid;
510		529
511	my $cb = shift;	530	my $cb = @_ ? shift : $SELF \|\| Carp::croak 'mon: called with one argument only, but $SELF not set,';
512		531
513	unless (ref $cb) {	532	unless (ref $cb) {
514	if (@_) {	533	if (@_) {
515	# send a kill info message	534	# send a kill info message
516	my (@msg) = ($cb, @_);	535	my (@msg) = ($cb, @_);
…		…
534	is killed, the references will be freed.	553	is killed, the references will be freed.
535		554
536	Optionally returns a guard that will stop the monitoring.	555	Optionally returns a guard that will stop the monitoring.
537		556
538	This function is useful when you create e.g. timers or other watchers and	557	This function is useful when you create e.g. timers or other watchers and
539	want to free them when the port gets killed:	558	want to free them when the port gets killed (note the use of C<psub>):
540		559
541	$port->rcv (start => sub {	560	$port->rcv (start => sub {
542	my $timer; $timer = mon_guard $port, AE::timer 1, 1, sub {	561	my $timer; $timer = mon_guard $port, AE::timer 1, 1, psub {
543	undef $timer if 0.9 < rand;	562	undef $timer if 0.9 < rand;
544	});	563	});
545	});	564	});
546		565
547	=cut	566	=cut
548		567
549	sub mon_guard {	568	sub mon_guard {
550	my ($port, @refs) = @_;	569	my ($port, @refs) = @_;
551		570
		571	#TODO: mon-less form?
		572
552	mon $port, sub { 0 && @refs }	573	mon $port, sub { 0 && @refs }
553	}	574	}
554		575
555	=item lnk $port1, $port2
556
557	Link two ports. This is simply a shorthand for:
558
559	mon $port1, $port2;
560	mon $port2, $port1;
561
562	It means that if either one is killed abnormally, the other one gets
563	killed as well.
564
565	=item kil $port[, @reason]	576	=item kil $port[, @reason]
566		577
567	Kill the specified port with the given C<@reason>.	578	Kill the specified port with the given C<@reason>.
568		579
569	If no C<@reason> is specified, then the port is killed "normally" (linked	580	If no C<@reason> is specified, then the port is killed "normally" (ports
570	ports will not be kileld, or even notified).	581	monitoring other ports will not necessarily die because a port dies
		582	"normally").
571		583
572	Otherwise, linked ports get killed with the same reason (second form of	584	Otherwise, linked ports get killed with the same reason (second form of
573	C<mon>, see below).	585	C<mon>, see above).
574		586
575	Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks	587	Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks
576	will be reported as reason C<< die => $@ >>.	588	will be reported as reason C<< die => $@ >>.
577		589
578	Transport/communication errors are reported as C<< transport_error =>	590	Transport/communication errors are reported as C<< transport_error =>
579	$message >>.	591	$message >>.
580		592
581	=back
582
583	=head1 NODE MESSAGES
584
585	Nodes understand the following messages sent to them. Many of them take
586	arguments called C<@reply>, which will simply be used to compose a reply
587	message - C<$reply[0]> is the port to reply to, C<$reply[1]> the type and
588	the remaining arguments are simply the message data.
589
590	While other messages exist, they are not public and subject to change.
591
592	=over 4
593
594	=cut	593	=cut
595		594
596	=item lookup => $name, @reply	595	=item $port = spawn $node, $initfunc[, @initdata]
597		596
598	Replies with the port ID of the specified well-known port, or C<undef>.	597	Creates a port on the node C<$node> (which can also be a port ID, in which
		598	case it's the node where that port resides).
599		599
600	=item devnull => ...	600	The port ID of the newly created port is returned immediately, and it is
		601	possible to immediately start sending messages or to monitor the port.
601		602
602	Generic data sink/CPU heat conversion.	603	After the port has been created, the init function is called on the remote
		604	node, in the same context as a C<rcv> callback. This function must be a
		605	fully-qualified function name (e.g. C<MyApp::Chat::Server::init>). To
		606	specify a function in the main program, use C<::name>.
603		607
604	=item relay => $port, @msg	608	If the function doesn't exist, then the node tries to C<require>
		609	the package, then the package above the package and so on (e.g.
		610	C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function
		611	exists or it runs out of package names.
605		612
606	Simply forwards the message to the given port.	613	The init function is then called with the newly-created port as context
		614	object (C<$SELF>) and the C<@initdata> values as arguments.
607		615
608	=item eval => $string[ @reply]	616	A common idiom is to pass a local port, immediately monitor the spawned
		617	port, and in the remote init function, immediately monitor the passed
		618	local port. This two-way monitoring ensures that both ports get cleaned up
		619	when there is a problem.
609		620
610	Evaluates the given string. If C<@reply> is given, then a message of the	621	Example: spawn a chat server port on C<$othernode>.
611	form C<@reply, $@, @evalres> is sent.
612		622
613	Example: crash another node.	623	# this node, executed from within a port context:
		624	my $server = spawn $othernode, "MyApp::Chat::Server::connect", $SELF;
		625	mon $server;
614		626
615	snd $othernode, eval => "exit";	627	# init function on C<$othernode>
		628	sub connect {
		629	my ($srcport) = @_;
616		630
617	=item time => @reply	631	mon $srcport;
618		632
619	Replies the the current node time to C<@reply>.	633	rcv $SELF, sub {
		634	...
		635	};
		636	}
620		637
621	Example: tell the current node to send the current time to C<$myport> in a	638	=cut
622	C<timereply> message.
623		639
624	snd $NODE, time => $myport, timereply => 1, 2;	640	sub _spawn {
625	# => snd $myport, timereply => 1, 2, <time>	641	my $port = shift;
		642	my $init = shift;
		643
		644	local $SELF = "$NODE#$port";
		645	eval {
		646	&{ load_func $init }
		647	};
		648	_self_die if $@;
		649	}
		650
		651	sub spawn(@) {
		652	my ($nodeid, undef) = split /#/, shift, 2;
		653
		654	my $id = "$RUNIQ." . $ID++;
		655
		656	$_[0] =~ /::/
		657	or Carp::croak "spawn init function must be a fully-qualified name, caught";
		658
		659	snd_to_func $nodeid, "AnyEvent::MP::_spawn" => $id, @_;
		660
		661	"$nodeid#$id"
		662	}
		663
		664	=item after $timeout, @msg
		665
		666	=item after $timeout, $callback
		667
		668	Either sends the given message, or call the given callback, after the
		669	specified number of seconds.
		670
		671	This is simply a utility function that comes in handy at times - the
		672	AnyEvent::MP author is not convinced of the wisdom of having it, though,
		673	so it may go away in the future.
		674
		675	=cut
		676
		677	sub after($@) {
		678	my ($timeout, @action) = @_;
		679
		680	my $t; $t = AE::timer $timeout, 0, sub {
		681	undef $t;
		682	ref $action[0]
		683	? $action[0]()
		684	: snd @action;
		685	};
		686	}
626		687
627	=back	688	=back
628		689
629	=head1 AnyEvent::MP vs. Distributed Erlang	690	=head1 AnyEvent::MP vs. Distributed Erlang
630		691
…		…
640		701
641	Despite the similarities, there are also some important differences:	702	Despite the similarities, there are also some important differences:
642		703
643	=over 4	704	=over 4
644		705
645	=item * Node references contain the recipe on how to contact them.	706	=item * Node IDs are arbitrary strings in AEMP.
646		707
647	Erlang relies on special naming and DNS to work everywhere in the	708	Erlang relies on special naming and DNS to work everywhere in the same
648	same way. AEMP relies on each node knowing it's own address(es), with	709	way. AEMP relies on each node somehow knowing its own address(es) (e.g. by
649	convenience functionality.	710	configuration or DNS), but will otherwise discover other odes itself.
650		711
651	This means that AEMP requires a less tightly controlled environment at the	712	=item * Erlang has a "remote ports are like local ports" philosophy, AEMP
652	cost of longer node references and a slightly higher management overhead.	713	uses "local ports are like remote ports".
		714
		715	The failure modes for local ports are quite different (runtime errors
		716	only) then for remote ports - when a local port dies, you I<know> it dies,
		717	when a connection to another node dies, you know nothing about the other
		718	port.
		719
		720	Erlang pretends remote ports are as reliable as local ports, even when
		721	they are not.
		722
		723	AEMP encourages a "treat remote ports differently" philosophy, with local
		724	ports being the special case/exception, where transport errors cannot
		725	occur.
653		726
654	=item * Erlang uses processes and a mailbox, AEMP does not queue.	727	=item * Erlang uses processes and a mailbox, AEMP does not queue.
655		728
656	Erlang uses processes that selctively receive messages, and therefore	729	Erlang uses processes that selectively receive messages, and therefore
657	needs a queue. AEMP is event based, queuing messages would serve no useful	730	needs a queue. AEMP is event based, queuing messages would serve no
658	purpose.	731	useful purpose. For the same reason the pattern-matching abilities of
		732	AnyEvent::MP are more limited, as there is little need to be able to
		733	filter messages without dequeuing them.
659		734
660	(But see L<Coro::MP> for a more Erlang-like process model on top of AEMP).	735	(But see L<Coro::MP> for a more Erlang-like process model on top of AEMP).
661		736
662	=item * Erlang sends are synchronous, AEMP sends are asynchronous.	737	=item * Erlang sends are synchronous, AEMP sends are asynchronous.
663		738
664	Sending messages in Erlang is synchronous and blocks the process. AEMP	739	Sending messages in Erlang is synchronous and blocks the process (and
665	sends are immediate, connection establishment is handled in the	740	so does not need a queue that can overflow). AEMP sends are immediate,
666	background.	741	connection establishment is handled in the background.
667		742
668	=item * Erlang can silently lose messages, AEMP cannot.	743	=item * Erlang suffers from silent message loss, AEMP does not.
669		744
670	Erlang makes few guarantees on messages delivery - messages can get lost	745	Erlang makes few guarantees on messages delivery - messages can get lost
671	without any of the processes realising it (i.e. you send messages a, b,	746	without any of the processes realising it (i.e. you send messages a, b,
672	and c, and the other side only receives messages a and c).	747	and c, and the other side only receives messages a and c).
673		748
674	AEMP guarantees correct ordering, and the guarantee that there are no	749	AEMP guarantees correct ordering, and the guarantee that after one message
675	holes in the message sequence.	750	is lost, all following ones sent to the same port are lost as well, until
676		751	monitoring raises an error, so there are no silent "holes" in the message
677	=item * In Erlang, processes can be declared dead and later be found to be	752	sequence.
678	alive.
679
680	In Erlang it can happen that a monitored process is declared dead and
681	linked processes get killed, but later it turns out that the process is
682	still alive - and can receive messages.
683
684	In AEMP, when port monitoring detects a port as dead, then that port will
685	eventually be killed - it cannot happen that a node detects a port as dead
686	and then later sends messages to it, finding it is still alive.
687		753
688	=item * Erlang can send messages to the wrong port, AEMP does not.	754	=item * Erlang can send messages to the wrong port, AEMP does not.
689		755
690	In Erlang it is quite possible that a node that restarts reuses a process	756	In Erlang it is quite likely that a node that restarts reuses a process ID
691	ID known to other nodes for a completely different process, causing	757	known to other nodes for a completely different process, causing messages
692	messages destined for that process to end up in an unrelated process.	758	destined for that process to end up in an unrelated process.
693		759
694	AEMP never reuses port IDs, so old messages or old port IDs floating	760	AEMP never reuses port IDs, so old messages or old port IDs floating
695	around in the network will not be sent to an unrelated port.	761	around in the network will not be sent to an unrelated port.
696		762
697	=item * Erlang uses unprotected connections, AEMP uses secure	763	=item * Erlang uses unprotected connections, AEMP uses secure
698	authentication and can use TLS.	764	authentication and can use TLS.
699		765
700	AEMP can use a proven protocol - SSL/TLS - to protect connections and	766	AEMP can use a proven protocol - TLS - to protect connections and
701	securely authenticate nodes.	767	securely authenticate nodes.
702		768
703	=item * The AEMP protocol is optimised for both text-based and binary	769	=item * The AEMP protocol is optimised for both text-based and binary
704	communications.	770	communications.
705		771
706	The AEMP protocol, unlike the Erlang protocol, supports both	772	The AEMP protocol, unlike the Erlang protocol, supports both programming
707	language-independent text-only protocols (good for debugging) and binary,	773	language independent text-only protocols (good for debugging) and binary,
708	language-specific serialisers (e.g. Storable).	774	language-specific serialisers (e.g. Storable). By default, unless TLS is
		775	used, the protocol is actually completely text-based.
709		776
710	It has also been carefully designed to be implementable in other languages	777	It has also been carefully designed to be implementable in other languages
711	with a minimum of work while gracefully degrading fucntionality to make the	778	with a minimum of work while gracefully degrading functionality to make the
712	protocol simple.	779	protocol simple.
713		780
714	=item * AEMP has more flexible monitoring options than Erlang.	781	=item * AEMP has more flexible monitoring options than Erlang.
715		782
716	In Erlang, you can chose to receive I<all> exit signals as messages	783	In Erlang, you can chose to receive I<all> exit signals as messages
717	or I<none>, there is no in-between, so monitoring single processes is	784	or I<none>, there is no in-between, so monitoring single processes is
718	difficult to implement. Monitoring in AEMP is more flexible than in	785	difficult to implement. Monitoring in AEMP is more flexible than in
719	Erlang, as one can choose between automatic kill, exit message or callback	786	Erlang, as one can choose between automatic kill, exit message or callback
720	on a per-process basis.	787	on a per-process basis.
721		788
722	=item * Erlang has different semantics for monitoring and linking, AEMP has the same.	789	=item * Erlang tries to hide remote/local connections, AEMP does not.
723		790
724	Monitoring in Erlang is not an indicator of process death/crashes,	791	Monitoring in Erlang is not an indicator of process death/crashes, in the
725	as linking is (except linking is unreliable in Erlang). In AEMP, the	792	same way as linking is (except linking is unreliable in Erlang).
726	semantics of monitoring and linking are identical, linking is simply	793
727	two-way monitoring with automatic kill.	794	In AEMP, you don't "look up" registered port names or send to named ports
		795	that might or might not be persistent. Instead, you normally spawn a port
		796	on the remote node. The init function monitors you, and you monitor the
		797	remote port. Since both monitors are local to the node, they are much more
		798	reliable (no need for C<spawn_link>).
		799
		800	This also saves round-trips and avoids sending messages to the wrong port
		801	(hard to do in Erlang).
728		802
729	=back	803	=back
730		804
		805	=head1 RATIONALE
		806
		807	=over 4
		808
		809	=item Why strings for port and node IDs, why not objects?
		810
		811	We considered "objects", but found that the actual number of methods
		812	that can be called are quite low. Since port and node IDs travel over
		813	the network frequently, the serialising/deserialising would add lots of
		814	overhead, as well as having to keep a proxy object everywhere.
		815
		816	Strings can easily be printed, easily serialised etc. and need no special
		817	procedures to be "valid".
		818
		819	And as a result, a miniport consists of a single closure stored in a
		820	global hash - it can't become much cheaper.
		821
		822	=item Why favour JSON, why not a real serialising format such as Storable?
		823
		824	In fact, any AnyEvent::MP node will happily accept Storable as framing
		825	format, but currently there is no way to make a node use Storable by
		826	default (although all nodes will accept it).
		827
		828	The default framing protocol is JSON because a) JSON::XS is many times
		829	faster for small messages and b) most importantly, after years of
		830	experience we found that object serialisation is causing more problems
		831	than it solves: Just like function calls, objects simply do not travel
		832	easily over the network, mostly because they will always be a copy, so you
		833	always have to re-think your design.
		834
		835	Keeping your messages simple, concentrating on data structures rather than
		836	objects, will keep your messages clean, tidy and efficient.
		837
		838	=back
		839
731	=head1 SEE ALSO	840	=head1 SEE ALSO
		841
		842	L<AnyEvent::MP::Intro> - a gentle introduction.
		843
		844	L<AnyEvent::MP::Kernel> - more, lower-level, stuff.
		845
		846	L<AnyEvent::MP::Global> - network maintainance and port groups, to find
		847	your applications.
732		848
733	L<AnyEvent>.	849	L<AnyEvent>.
734		850
735	=head1 AUTHOR	851	=head1 AUTHOR
736		852

Diff Legend

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

Comparing AnyEvent-MP/MP.pm (file contents): Revision 1.35 by root, Thu Aug 6 10:21:48 2009 UTC vs. Revision 1.77 by elmex, Thu Sep 3 07:57:30 2009 UTC

Diff Legend

Comparing AnyEvent-MP/MP.pm (file contents):
Revision 1.35 by root, Thu Aug 6 10:21:48 2009 UTC vs.
Revision 1.77 by elmex, Thu Sep 3 07:57:30 2009 UTC