ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/AnyEvent-MP/MP.pm

Comparing AnyEvent-MP/MP.pm (file contents):
Revision 1.25 by root, Tue Aug 4 21:07:37 2009 UTC vs.
Revision 1.134 by root, Mon Mar 12 14:47:23 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 # 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	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 $port, $cb->(@msg) # callback is invoked on death
		41	mon $port, $localport # kill localport on abnormal death
		42	mon $port, $localport, @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.
27		59
28	=head1 DESCRIPTION	60	=head1 DESCRIPTION
29		61
30	This module (-family) implements a simple message passing framework.	62	This module (-family) implements a simple message passing framework.
31		63
32	Despite its simplicity, you can securely message other processes running	64	Despite its simplicity, you can securely message other processes running
33	on the same or other hosts.	65	on the same or other hosts, and you can supervise entities remotely.
34		66
35	For an introduction to this module family, see the L<AnyEvent::MP::Intro>	67	For an introduction to this module family, see the L<AnyEvent::MP::Intro>
36	manual page.	68	manual page and the examples under F<eg/>.
37
38	At the moment, this module family is severly broken and underdocumented,
39	so do not use. This was uploaded mainly to reserve the CPAN namespace -
40	stay tuned! The basic API should be finished, however.
41		69
42	=head1 CONCEPTS	70	=head1 CONCEPTS
43		71
44	=over 4	72	=over 4
45		73
46	=item port	74	=item port
47		75
48	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
49	you can register C<rcv> handlers with. All C<rcv> handlers will receive	77	messages to (with the C<snd> function).
50	messages they match, messages will not be queued.
51		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
52	=item port id - C<noderef#portname>	85	=item port ID - C<nodeid#portname>
53		86
54	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<#>)
55	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).
56		90
57	=item node	91	=item node
58		92
59	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,
60	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
61	among other things.	95	ports.
62		96
63	Initially, nodes are either private (single-process only) or hidden	97	Nodes are either public (have one or more listening ports) or private
64	(connected to a master node only). Only when they epxlicitly "become	98	(no listening ports). Private nodes cannot talk to other private nodes
65	public" can you send them messages from unrelated other nodes.	99	currently, but all nodes can talk to public nodes.
66		100
67	=item noderef - C<host:port,host:port...>, C<id@noderef>, C<id>	101	Nodes is represented by (printable) strings called "node IDs".
68		102
		103	=item node ID - C<[A-Za-z0-9_\-.:]*>
		104
69	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
70	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
71	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).
72		170
73	=back	171	=back
74		172
75	=head1 VARIABLES/FUNCTIONS	173	=head1 VARIABLES/FUNCTIONS
76		174
…		…
78		176
79	=cut	177	=cut
80		178
81	package AnyEvent::MP;	179	package AnyEvent::MP;
82		180
		181	use AnyEvent::MP::Config ();
83	use AnyEvent::MP::Base;	182	use AnyEvent::MP::Kernel;
		183	use AnyEvent::MP::Kernel qw(%NODE %PORT %PORT_DATA $UNIQ $RUNIQ $ID);
84		184
85	use common::sense;	185	use common::sense;
86		186
87	use Carp ();	187	use Carp ();
88		188
89	use AE ();	189	use AE ();
		190	use Guard ();
90		191
91	use base "Exporter";	192	use base "Exporter";
92		193
93	our $VERSION = '0.1';	194	our $VERSION = $AnyEvent::MP::Config::VERSION;
		195
94	our @EXPORT = qw(	196	our @EXPORT = qw(
95	NODE $NODE *SELF node_of _any_	197	NODE $NODE *SELF node_of after
96	become_slave become_public	198	configure
97	snd rcv mon kil reg psub	199	snd rcv mon mon_guard kil psub peval spawn cal
98	port	200	port
		201	db_set db_del db_reg
		202	db_mon db_family db_keys db_values
99	);	203	);
100		204
101	our $SELF;	205	our $SELF;
102		206
103	sub _self_die() {	207	sub _self_die() {
…		…
106	kil $SELF, die => $msg;	210	kil $SELF, die => $msg;
107	}	211	}
108		212
109	=item $thisnode = NODE / $NODE	213	=item $thisnode = NODE / $NODE
110		214
111	The C<NODE> function returns, and the C<$NODE> variable contains	215	The C<NODE> function returns, and the C<$NODE> variable contains, the node
112	the noderef of the local node. The value is initialised by a call	216	ID of the node running in the current process. This value is initialised by
113	to C<become_public> or C<become_slave>, after which all local port	217	a call to C<configure>.
114	identifiers become invalid.
115		218
116	=item $noderef = node_of $portid	219	=item $nodeid = node_of $port
117		220
118	Extracts and returns the noderef from a portid or a noderef.	221	Extracts and returns the node ID from a port ID or a node ID.
		222
		223	=item configure $profile, key => value...
		224
		225	=item configure key => value...
		226
		227	Before a node can talk to other nodes on the network (i.e. enter
		228	"distributed mode") it has to configure itself - the minimum a node needs
		229	to know is its own name, and optionally it should know the addresses of
		230	some other nodes in the network to discover other nodes.
		231
		232	This function configures a node - it must be called exactly once (or
		233	never) before calling other AnyEvent::MP functions.
		234
		235	The key/value pairs are basically the same ones as documented for the
		236	F<aemp> command line utility (sans the set/del prefix), with these additions:
		237
		238	=over 4
		239
		240	=item norc => $boolean (default false)
		241
		242	If true, then the rc file (e.g. F<~/.perl-anyevent-mp>) will I<not>
		243	be consulted - all configuraiton options must be specified in the
		244	C<configure> call.
		245
		246	=item force => $boolean (default false)
		247
		248	IF true, then the values specified in the C<configure> will take
		249	precedence over any values configured via the rc file. The default is for
		250	the rc file to override any options specified in the program.
		251
		252	=item secure => $pass->($nodeid)
		253
		254	In addition to specifying a boolean, you can specify a code reference that
		255	is called for every remote execution attempt - the execution request is
		256	granted iff the callback returns a true value.
		257
		258	See F<semp setsecure> for more info.
		259
		260	=back
		261
		262	=over 4
		263
		264	=item step 1, gathering configuration from profiles
		265
		266	The function first looks up a profile in the aemp configuration (see the
		267	L<aemp> commandline utility). The profile name can be specified via the
		268	named C<profile> parameter or can simply be the first parameter). If it is
		269	missing, then the nodename (F<uname -n>) will be used as profile name.
		270
		271	The profile data is then gathered as follows:
		272
		273	First, all remaining key => value pairs (all of which are conveniently
		274	undocumented at the moment) will be interpreted as configuration
		275	data. Then they will be overwritten by any values specified in the global
		276	default configuration (see the F<aemp> utility), then the chain of
		277	profiles chosen by the profile name (and any C<parent> attributes).
		278
		279	That means that the values specified in the profile have highest priority
		280	and the values specified directly via C<configure> have lowest priority,
		281	and can only be used to specify defaults.
		282
		283	If the profile specifies a node ID, then this will become the node ID of
		284	this process. If not, then the profile name will be used as node ID, with
		285	a unique randoms tring (C</%u>) appended.
		286
		287	The node ID can contain some C<%> sequences that are expanded: C<%n>
		288	is expanded to the local nodename, C<%u> is replaced by a random
		289	strign to make the node unique. For example, the F<aemp> commandline
		290	utility uses C<aemp/%n/%u> as nodename, which might expand to
		291	C<aemp/cerebro/ZQDGSIkRhEZQDGSIkRhE>.
		292
		293	=item step 2, bind listener sockets
		294
		295	The next step is to look up the binds in the profile, followed by binding
		296	aemp protocol listeners on all binds specified (it is possible and valid
		297	to have no binds, meaning that the node cannot be contacted form the
		298	outside. This means the node cannot talk to other nodes that also have no
		299	binds, but it can still talk to all "normal" nodes).
		300
		301	If the profile does not specify a binds list, then a default of C<*> is
		302	used, meaning the node will bind on a dynamically-assigned port on every
		303	local IP address it finds.
		304
		305	=item step 3, connect to seed nodes
		306
		307	As the last step, the seed ID list from the profile is passed to the
		308	L<AnyEvent::MP::Global> module, which will then use it to keep
		309	connectivity with at least one node at any point in time.
		310
		311	=back
		312
		313	Example: become a distributed node using the local node name as profile.
		314	This should be the most common form of invocation for "daemon"-type nodes.
		315
		316	configure
		317
		318	Example: become a semi-anonymous node. This form is often used for
		319	commandline clients.
		320
		321	configure nodeid => "myscript/%n/%u";
		322
		323	Example: configure a node using a profile called seed, which is suitable
		324	for a seed node as it binds on all local addresses on a fixed port (4040,
		325	customary for aemp).
		326
		327	# use the aemp commandline utility
		328	# aemp profile seed binds '*:4040'
		329
		330	# then use it
		331	configure profile => "seed";
		332
		333	# or simply use aemp from the shell again:
		334	# aemp run profile seed
		335
		336	# or provide a nicer-to-remember nodeid
		337	# aemp run profile seed nodeid "$(hostname)"
119		338
120	=item $SELF	339	=item $SELF
121		340
122	Contains the current port id while executing C<rcv> callbacks or C<psub>	341	Contains the current port id while executing C<rcv> callbacks or C<psub>
123	blocks.	342	blocks.
124		343
125	=item SELF, %SELF, @SELF...	344	=item *SELF, SELF, %SELF, @SELF...
126		345
127	Due to some quirks in how perl exports variables, it is impossible to	346	Due to some quirks in how perl exports variables, it is impossible to
128	just export C<$SELF>, all the symbols called C<SELF> are exported by this	347	just export C<$SELF>, all the symbols named C<SELF> are exported by this
129	module, but only C<$SELF> is currently used.	348	module, but only C<$SELF> is currently used.
130		349
131	=item snd $portid, type => @data	350	=item snd $port, type => @data
132		351
133	=item snd $portid, @msg	352	=item snd $port, @msg
134		353
135	Send the given message to the given port ID, which can identify either	354	Send the given message to the given port, which can identify either a
136	a local or a remote port, and can be either a string or soemthignt hat	355	local or a remote port, and must be a port ID.
137	stringifies a sa port ID (such as a port object :).
138		356
139	While the message can be about anything, it is highly recommended to use a	357	While the message can be almost anything, it is highly recommended to
140	string as first element (a portid, or some word that indicates a request	358	use a string as first element (a port ID, or some word that indicates a
141	type etc.).	359	request type etc.) and to consist if only simple perl values (scalars,
		360	arrays, hashes) - if you think you need to pass an object, think again.
142		361
143	The message data effectively becomes read-only after a call to this	362	The message data logically becomes read-only after a call to this
144	function: modifying any argument is not allowed and can cause many	363	function: modifying any argument (or values referenced by them) is
145	problems.	364	forbidden, as there can be considerable time between the call to C<snd>
		365	and the time the message is actually being serialised - in fact, it might
		366	never be copied as within the same process it is simply handed to the
		367	receiving port.
146		368
147	The type of data you can transfer depends on the transport protocol: when	369	The type of data you can transfer depends on the transport protocol: when
148	JSON is used, then only strings, numbers and arrays and hashes consisting	370	JSON is used, then only strings, numbers and arrays and hashes consisting
149	of those are allowed (no objects). When Storable is used, then anything	371	of those are allowed (no objects). When Storable is used, then anything
150	that Storable can serialise and deserialise is allowed, and for the local	372	that Storable can serialise and deserialise is allowed, and for the local
151	node, anything can be passed.	373	node, anything can be passed. Best rely only on the common denominator of
		374	these.
152		375
153	=item kil $portid[, @reason]	376	=item $local_port = port
154		377
155	Kill the specified port with the given C<@reason>.	378	Create a new local port object and returns its port ID. Initially it has
		379	no callbacks set and will throw an error when it receives messages.
156		380
157	If no C<@reason> is specified, then the port is killed "normally" (linked	381	=item $local_port = port { my @msg = @_ }
158	ports will not be kileld, or even notified).
159		382
160	Otherwise, linked ports get killed with the same reason (second form of	383	Creates a new local port, and returns its ID. Semantically the same as
161	C<mon>, see below).	384	creating a port and calling C<rcv $port, $callback> on it.
162		385
163	Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks	386	The block will be called for every message received on the port, with the
164	will be reported as reason C<< die => $@ >>.	387	global variable C<$SELF> set to the port ID. Runtime errors will cause the
		388	port to be C<kil>ed. The message will be passed as-is, no extra argument
		389	(i.e. no port ID) will be passed to the callback.
165		390
166	Transport/communication errors are reported as C<< transport_error =>	391	If you want to stop/destroy the port, simply C<kil> it:
167	$message >>.
168		392
169	=item $guard = mon $portid, $cb->(@reason)	393	my $port = port {
170		394	my @msg = @_;
171	=item $guard = mon $portid, $otherport	395	...
172		396	kil $SELF;
173	=item $guard = mon $portid, $otherport, @msg	397	};
174
175	Monitor the given port and do something when the port is killed.
176
177	In the first form, the callback is simply called with any number
178	of C<@reason> elements (no @reason means that the port was deleted
179	"normally"). Note also that I<< the callback B<must> never die >>, so use
180	C<eval> if unsure.
181
182	In the second form, the other port will be C<kil>'ed with C<@reason>, iff
183	a @reason was specified, i.e. on "normal" kils nothing happens, while
184	under all other conditions, the other port is killed with the same reason.
185
186	In the last form, a message of the form C<@msg, @reason> will be C<snd>.
187
188	Example: call a given callback when C<$port> is killed.
189
190	mon $port, sub { warn "port died because of <@_>\n" };
191
192	Example: kill ourselves when C<$port> is killed abnormally.
193
194	mon $port, $self;
195
196	Example: send us a restart message another C<$port> is killed.
197
198	mon $port, $self => "restart";
199		398
200	=cut	399	=cut
201		400
202	sub mon {	401	sub rcv($@);
203	my ($noderef, $port, $cb) = ((split /#/, shift, 2), shift);
204		402
205	my $node = $NODE{$noderef} || add_node $noderef;	403	my $KILME = sub {
		404	(my $tag = substr $_[0], 0, 30) =~ s/([\x20-\x7e])/./g;
		405	kil $SELF, unhandled_message => "missing (tag or fallback) callback for message '$tag'";
		406	};
206		407
207	#TODO: ports must not be references	408	sub port(;&) {
208	if (!ref $cb or "AnyEvent::MP::Port" eq ref $cb) {	409	my $id = $UNIQ . ++$ID;
		410	my $port = "$NODE#$id";
		411
		412	rcv $port, shift || $KILME;
		413
		414	$port
		415	}
		416
		417	=item rcv $local_port, $callback->(@msg)
		418
		419	Replaces the default callback on the specified port. There is no way to
		420	remove the default callback: use C<sub { }> to disable it, or better
		421	C<kil> the port when it is no longer needed.
		422
		423	The global C<$SELF> (exported by this module) contains C<$port> while
		424	executing the callback. Runtime errors during callback execution will
		425	result in the port being C<kil>ed.
		426
		427	The default callback receives all messages not matched by a more specific
		428	C<tag> match.
		429
		430	=item rcv $local_port, tag => $callback->(@msg_without_tag), ...
		431
		432	Register (or replace) callbacks to be called on messages starting with the
		433	given tag on the given port (and return the port), or unregister it (when
		434	C<$callback> is C<$undef> or missing). There can only be one callback
		435	registered for each tag.
		436
		437	The original message will be passed to the callback, after the first
		438	element (the tag) has been removed. The callback will use the same
		439	environment as the default callback (see above).
		440
		441	Example: create a port and bind receivers on it in one go.
		442
		443	my $port = rcv port,
		444	msg1 => sub { ... },
		445	msg2 => sub { ... },
		446	;
		447
		448	Example: create a port, bind receivers and send it in a message elsewhere
		449	in one go:
		450
		451	snd $otherport, reply =>
		452	rcv port,
		453	msg1 => sub { ... },
		454	...
		455	;
		456
		457	Example: temporarily register a rcv callback for a tag matching some port
		458	(e.g. for an rpc reply) and unregister it after a message was received.
		459
		460	rcv $port, $otherport => sub {
		461	my @reply = @_;
		462
		463	rcv $SELF, $otherport;
		464	};
		465
		466	=cut
		467
		468	sub rcv($@) {
		469	my $port = shift;
		470	my ($nodeid, $portid) = split /#/, $port, 2;
		471
		472	$NODE{$nodeid} == $NODE{""}
		473	or Carp::croak "$port: rcv can only be called on local ports, caught";
		474
		475	while (@_) {
209	if (@_) {	476	if (ref $_[0]) {
210	# send a kill info message	477	if (my $self = $PORT_DATA{$portid}) {
211	my (@msg) = ($cb, @_);	478	"AnyEvent::MP::Port" eq ref $self
212	$cb = sub { snd @msg, @_ };	479	or Carp::croak "$port: rcv can only be called on message matching ports, caught";
		480
		481	$self->[0] = shift;
213	} else {	482	} else {
214	# simply kill other port	483	my $cb = shift;
215	my $port = $cb;	484	$PORT{$portid} = sub {
216	$cb = sub { kil $port, @_ if @_ };	485	local $SELF = $port;
		486	eval { &$cb }; _self_die if $@;
		487	};
		488	}
		489	} elsif (defined $_[0]) {
		490	my $self = $PORT_DATA{$portid} ||= do {
		491	my $self = bless [$PORT{$portid} || sub { }, { }, $port], "AnyEvent::MP::Port";
		492
		493	$PORT{$portid} = sub {
		494	local $SELF = $port;
		495
		496	if (my $cb = $self->[1]{$_[0]}) {
		497	shift;
		498	eval { &$cb }; _self_die if $@;
		499	} else {
		500	&{ $self->[0] };
		501	}
		502	};
		503
		504	$self
		505	};
		506
		507	"AnyEvent::MP::Port" eq ref $self
		508	or Carp::croak "$port: rcv can only be called on message matching ports, caught";
		509
		510	my ($tag, $cb) = splice @_, 0, 2;
		511
		512	if (defined $cb) {
		513	$self->[1]{$tag} = $cb;
		514	} else {
		515	delete $self->[1]{$tag};
		516	}
217	}	517	}
218	}	518	}
219		519
220	$node->monitor ($port, $cb);
221
222	defined wantarray
223	and AnyEvent::Util::guard { $node->unmonitor ($port, $cb) }
224	}
225
226	=item $guard = mon_guard $port, $ref, $ref...
227
228	Monitors the given C<$port> and keeps the passed references. When the port
229	is killed, the references will be freed.
230
231	Optionally returns a guard that will stop the monitoring.
232
233	This function is useful when you create e.g. timers or other watchers and
234	want to free them when the port gets killed:
235
236	$port->rcv (start => sub {
237	my $timer; $timer = mon_guard $port, AE::timer 1, 1, sub {
238	undef $timer if 0.9 < rand;
239	});
240	});
241
242	=cut
243
244	sub mon_guard {
245	my ($port, @refs) = @_;
246
247	mon $port, sub { 0 && @refs }
248	}
249
250	=item lnk $port1, $port2
251
252	Link two ports. This is simply a shorthand for:
253
254	mon $port1, $port2;
255	mon $port2, $port1;
256
257	It means that if either one is killed abnormally, the other one gets
258	killed as well.
259
260	=item $local_port = port
261
262	Create a new local port object that supports message matching.
263
264	=item $portid = port { my @msg = @_; $finished }
265
266	Creates a "mini port", that is, a very lightweight port without any
267	pattern matching behind it, and returns its ID.
268
269	The block will be called for every message received on the port. When the
270	callback returns a true value its job is considered "done" and the port
271	will be destroyed. Otherwise it will stay alive.
272
273	The message will be passed as-is, no extra argument (i.e. no port id) will
274	be passed to the callback.
275
276	If you need the local port id in the callback, this works nicely:
277
278	my $port; $port = miniport {
279	snd $otherport, reply => $port;
280	};
281
282	=cut
283
284	sub port(;&) {
285	my $id = "$UNIQ." . $ID++;
286	my $port = "$NODE#$id";
287
288	if (@_) {
289	my $cb = shift;
290	$PORT{$id} = sub {
291	local $SELF = $port;
292	eval {
293	&$cb
294	and kil $id;
295	};
296	_self_die if $@;
297	};
298	} else {
299	my $self = bless {
300	id => "$NODE#$id",
301	}, "AnyEvent::MP::Port";
302
303	$PORT_DATA{$id} = $self;
304	$PORT{$id} = sub {
305	local $SELF = $port;
306
307	eval {
308	for (@{ $self->{rc0}{$_[0]} }) {
309	$_ && &{$_->[0]}
310	&& undef $_;
311	}
312
313	for (@{ $self->{rcv}{$_[0]} }) {
314	$_ && [@_[1 .. @{$_->[1]}]] ~~ $_->[1]
315	&& &{$_->[0]}
316	&& undef $_;
317	}
318
319	for (@{ $self->{any} }) {
320	$_ && [@_[0 .. $#{$_->[1]}]] ~~ $_->[1]
321	&& &{$_->[0]}
322	&& undef $_;
323	}
324	};
325	_self_die if $@;
326	};
327	}
328
329	$port	520	$port
330	}	521	}
331		522
332	=item reg $portid, $name	523	=item peval $port, $coderef[, @args]
333		524
334	Registers the given port under the name C<$name>. If the name already	525	Evaluates the given C<$codref> within the contetx of C<$port>, that is,
335	exists it is replaced.	526	when the code throews an exception the C<$port> will be killed.
336		527
337	A port can only be registered under one well known name.	528	Any remaining args will be passed to the callback. Any return values will
		529	be returned to the caller.
338		530
339	A port automatically becomes unregistered when it is killed.	531	This is useful when you temporarily want to execute code in the context of
		532	a port.
		533
		534	Example: create a port and run some initialisation code in it's context.
		535
		536	my $port = port { ... };
		537
		538	peval $port, sub {
		539	init
		540	or die "unable to init";
		541	};
340		542
341	=cut	543	=cut
342		544
343	sub reg(@) {	545	sub peval($$) {
344	my ($portid, $name) = @_;	546	local $SELF = shift;
		547	my $cb = shift;
345		548
346	$REG{$name} = $portid;	549	if (wantarray) {
347	}	550	my @res = eval { &$cb };
348		551	_self_die if $@;
349	=item rcv $portid, tagstring => $callback->(@msg), ...	552	@res
350
351	=item rcv $portid, $smartmatch => $callback->(@msg), ...
352
353	=item rcv $portid, [$smartmatch...] => $callback->(@msg), ...
354
355	Register callbacks to be called on matching messages on the given port.
356
357	The callback has to return a true value when its work is done, after
358	which is will be removed, or a false value in which case it will stay
359	registered.
360
361	The global C<$SELF> (exported by this module) contains C<$portid> while
362	executing the callback.
363
364	Runtime errors wdurign callback execution will result in the port being
365	C<kil>ed.
366
367	If the match is an array reference, then it will be matched against the
368	first elements of the message, otherwise only the first element is being
369	matched.
370
371	Any element in the match that is specified as C<_any_> (a function
372	exported by this module) matches any single element of the message.
373
374	While not required, it is highly recommended that the first matching
375	element is a string identifying the message. The one-string-only match is
376	also the most efficient match (by far).
377
378	=cut
379
380	sub rcv($@) {
381	my ($noderef, $port) = split /#/, shift, 2;
382
383	($NODE{$noderef} || add_node $noderef) == $NODE{""}
384	or Carp::croak "$noderef#$port: rcv can only be called on local ports, caught";
385
386	my $self = $PORT_DATA{$port}
387	or Carp::croak "$noderef#$port: rcv can only be called on message matching ports, caught";
388
389	"AnyEvent::MP::Port" eq ref $self
390	or Carp::croak "$noderef#$port: rcv can only be called on message matching ports, caught";
391
392	while (@_) {
393	my ($match, $cb) = splice @_, 0, 2;
394
395	if (!ref $match) {
396	push @{ $self->{rc0}{$match} }, [$cb];
397	} elsif (("ARRAY" eq ref $match && !ref $match->[0])) {
398	my ($type, @match) = @$match;
399	@match
400	? push @{ $self->{rcv}{$match->[0]} }, [$cb, \@match]
401	: push @{ $self->{rc0}{$match->[0]} }, [$cb];
402	} else {	553	} else {
403	push @{ $self->{any} }, [$cb, $match];	554	my $res = eval { &$cb };
404	}	555	_self_die if $@;
		556	$res
405	}	557	}
406	}	558	}
407		559
408	=item $closure = psub { BLOCK }	560	=item $closure = psub { BLOCK }
409		561
410	Remembers C<$SELF> and creates a closure out of the BLOCK. When the	562	Remembers C<$SELF> and creates a closure out of the BLOCK. When the
411	closure is executed, sets up the environment in the same way as in C<rcv>	563	closure is executed, sets up the environment in the same way as in C<rcv>
412	callbacks, i.e. runtime errors will cause the port to get C<kil>ed.	564	callbacks, i.e. runtime errors will cause the port to get C<kil>ed.
		565
		566	The effect is basically as if it returned C<< sub { peval $SELF, sub {
		567	BLOCK }, @_ } >>.
413		568
414	This is useful when you register callbacks from C<rcv> callbacks:	569	This is useful when you register callbacks from C<rcv> callbacks:
415		570
416	rcv delayed_reply => sub {	571	rcv delayed_reply => sub {
417	my ($delay, @reply) = @_;	572	my ($delay, @reply) = @_;
…		…
441	$res	596	$res
442	}	597	}
443	}	598	}
444	}	599	}
445		600
		601	=item $guard = mon $port, $cb->(@reason) # call $cb when $port dies
		602
		603	=item $guard = mon $port, $rcvport # kill $rcvport when $port dies
		604
		605	=item $guard = mon $port # kill $SELF when $port dies
		606
		607	=item $guard = mon $port, $rcvport, @msg # send a message when $port dies
		608
		609	Monitor the given port and do something when the port is killed or
		610	messages to it were lost, and optionally return a guard that can be used
		611	to stop monitoring again.
		612
		613	In the first form (callback), the callback is simply called with any
		614	number of C<@reason> elements (no @reason means that the port was deleted
		615	"normally"). Note also that I<< the callback B<must> never die >>, so use
		616	C<eval> if unsure.
		617
		618	In the second form (another port given), the other port (C<$rcvport>)
		619	will be C<kil>'ed with C<@reason>, if a @reason was specified, i.e. on
		620	"normal" kils nothing happens, while under all other conditions, the other
		621	port is killed with the same reason.
		622
		623	The third form (kill self) is the same as the second form, except that
		624	C<$rvport> defaults to C<$SELF>.
		625
		626	In the last form (message), a message of the form C<@msg, @reason> will be
		627	C<snd>.
		628
		629	Monitoring-actions are one-shot: once messages are lost (and a monitoring
		630	alert was raised), they are removed and will not trigger again.
		631
		632	As a rule of thumb, monitoring requests should always monitor a port from
		633	a local port (or callback). The reason is that kill messages might get
		634	lost, just like any other message. Another less obvious reason is that
		635	even monitoring requests can get lost (for example, when the connection
		636	to the other node goes down permanently). When monitoring a port locally
		637	these problems do not exist.
		638
		639	C<mon> effectively guarantees that, in the absence of hardware failures,
		640	after starting the monitor, either all messages sent to the port will
		641	arrive, or the monitoring action will be invoked after possible message
		642	loss has been detected. No messages will be lost "in between" (after
		643	the first lost message no further messages will be received by the
		644	port). After the monitoring action was invoked, further messages might get
		645	delivered again.
		646
		647	Inter-host-connection timeouts and monitoring depend on the transport
		648	used. The only transport currently implemented is TCP, and AnyEvent::MP
		649	relies on TCP to detect node-downs (this can take 10-15 minutes on a
		650	non-idle connection, and usually around two hours for idle connections).
		651
		652	This means that monitoring is good for program errors and cleaning up
		653	stuff eventually, but they are no replacement for a timeout when you need
		654	to ensure some maximum latency.
		655
		656	Example: call a given callback when C<$port> is killed.
		657
		658	mon $port, sub { warn "port died because of <@_>\n" };
		659
		660	Example: kill ourselves when C<$port> is killed abnormally.
		661
		662	mon $port;
		663
		664	Example: send us a restart message when another C<$port> is killed.
		665
		666	mon $port, $self => "restart";
		667
		668	=cut
		669
		670	sub mon {
		671	my ($nodeid, $port) = split /#/, shift, 2;
		672
		673	my $node = $NODE{$nodeid} || add_node $nodeid;
		674
		675	my $cb = @_ ? shift : $SELF || Carp::croak 'mon: called with one argument only, but $SELF not set,';
		676
		677	unless (ref $cb) {
		678	if (@_) {
		679	# send a kill info message
		680	my (@msg) = ($cb, @_);
		681	$cb = sub { snd @msg, @_ };
		682	} else {
		683	# simply kill other port
		684	my $port = $cb;
		685	$cb = sub { kil $port, @_ if @_ };
		686	}
		687	}
		688
		689	$node->monitor ($port, $cb);
		690
		691	defined wantarray
		692	and ($cb += 0, Guard::guard { $node->unmonitor ($port, $cb) })
		693	}
		694
		695	=item $guard = mon_guard $port, $ref, $ref...
		696
		697	Monitors the given C<$port> and keeps the passed references. When the port
		698	is killed, the references will be freed.
		699
		700	Optionally returns a guard that will stop the monitoring.
		701
		702	This function is useful when you create e.g. timers or other watchers and
		703	want to free them when the port gets killed (note the use of C<psub>):
		704
		705	$port->rcv (start => sub {
		706	my $timer; $timer = mon_guard $port, AE::timer 1, 1, psub {
		707	undef $timer if 0.9 < rand;
		708	});
		709	});
		710
		711	=cut
		712
		713	sub mon_guard {
		714	my ($port, @refs) = @_;
		715
		716	#TODO: mon-less form?
		717
		718	mon $port, sub { 0 && @refs }
		719	}
		720
		721	=item kil $port[, @reason]
		722
		723	Kill the specified port with the given C<@reason>.
		724
		725	If no C<@reason> is specified, then the port is killed "normally" -
		726	monitor callback will be invoked, but the kil will not cause linked ports
		727	(C<mon $mport, $lport> form) to get killed.
		728
		729	If a C<@reason> is specified, then linked ports (C<mon $mport, $lport>
		730	form) get killed with the same reason.
		731
		732	Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks
		733	will be reported as reason C<< die => $@ >>.
		734
		735	Transport/communication errors are reported as C<< transport_error =>
		736	$message >>.
		737
		738	Common idioms:
		739
		740	# silently remove yourself, do not kill linked ports
		741	kil $SELF;
		742
		743	# report a failure in some detail
		744	kil $SELF, failure_mode_1 => "it failed with too high temperature";
		745
		746	# do not waste much time with killing, just die when something goes wrong
		747	open my $fh, "<file"
		748	or die "file: $!";
		749
		750	=item $port = spawn $node, $initfunc[, @initdata]
		751
		752	Creates a port on the node C<$node> (which can also be a port ID, in which
		753	case it's the node where that port resides).
		754
		755	The port ID of the newly created port is returned immediately, and it is
		756	possible to immediately start sending messages or to monitor the port.
		757
		758	After the port has been created, the init function is called on the remote
		759	node, in the same context as a C<rcv> callback. This function must be a
		760	fully-qualified function name (e.g. C<MyApp::Chat::Server::init>). To
		761	specify a function in the main program, use C<::name>.
		762
		763	If the function doesn't exist, then the node tries to C<require>
		764	the package, then the package above the package and so on (e.g.
		765	C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function
		766	exists or it runs out of package names.
		767
		768	The init function is then called with the newly-created port as context
		769	object (C<$SELF>) and the C<@initdata> values as arguments. It I<must>
		770	call one of the C<rcv> functions to set callbacks on C<$SELF>, otherwise
		771	the port might not get created.
		772
		773	A common idiom is to pass a local port, immediately monitor the spawned
		774	port, and in the remote init function, immediately monitor the passed
		775	local port. This two-way monitoring ensures that both ports get cleaned up
		776	when there is a problem.
		777
		778	C<spawn> guarantees that the C<$initfunc> has no visible effects on the
		779	caller before C<spawn> returns (by delaying invocation when spawn is
		780	called for the local node).
		781
		782	Example: spawn a chat server port on C<$othernode>.
		783
		784	# this node, executed from within a port context:
		785	my $server = spawn $othernode, "MyApp::Chat::Server::connect", $SELF;
		786	mon $server;
		787
		788	# init function on C<$othernode>
		789	sub connect {
		790	my ($srcport) = @_;
		791
		792	mon $srcport;
		793
		794	rcv $SELF, sub {
		795	...
		796	};
		797	}
		798
		799	=cut
		800
		801	sub _spawn {
		802	my $port = shift;
		803	my $init = shift;
		804
		805	# rcv will create the actual port
		806	local $SELF = "$NODE#$port";
		807	eval {
		808	&{ load_func $init }
		809	};
		810	_self_die if $@;
		811	}
		812
		813	sub spawn(@) {
		814	my ($nodeid, undef) = split /#/, shift, 2;
		815
		816	my $id = $RUNIQ . ++$ID;
		817
		818	$_[0] =~ /::/
		819	or Carp::croak "spawn init function must be a fully-qualified name, caught";
		820
		821	snd_to_func $nodeid, "AnyEvent::MP::_spawn" => $id, @_;
		822
		823	"$nodeid#$id"
		824	}
		825
		826
		827	=item after $timeout, @msg
		828
		829	=item after $timeout, $callback
		830
		831	Either sends the given message, or call the given callback, after the
		832	specified number of seconds.
		833
		834	This is simply a utility function that comes in handy at times - the
		835	AnyEvent::MP author is not convinced of the wisdom of having it, though,
		836	so it may go away in the future.
		837
		838	=cut
		839
		840	sub after($@) {
		841	my ($timeout, @action) = @_;
		842
		843	my $t; $t = AE::timer $timeout, 0, sub {
		844	undef $t;
		845	ref $action[0]
		846	? $action[0]()
		847	: snd @action;
		848	};
		849	}
		850
		851	#=item $cb2 = timeout $seconds, $cb[, @args]
		852
		853	=item cal $port, @msg, $callback[, $timeout]
		854
		855	A simple form of RPC - sends a message to the given C<$port> with the
		856	given contents (C<@msg>), but adds a reply port to the message.
		857
		858	The reply port is created temporarily just for the purpose of receiving
		859	the reply, and will be C<kil>ed when no longer needed.
		860
		861	A reply message sent to the port is passed to the C<$callback> as-is.
		862
		863	If an optional time-out (in seconds) is given and it is not C<undef>,
		864	then the callback will be called without any arguments after the time-out
		865	elapsed and the port is C<kil>ed.
		866
		867	If no time-out is given (or it is C<undef>), then the local port will
		868	monitor the remote port instead, so it eventually gets cleaned-up.
		869
		870	Currently this function returns the temporary port, but this "feature"
		871	might go in future versions unless you can make a convincing case that
		872	this is indeed useful for something.
		873
		874	=cut
		875
		876	sub cal(@) {
		877	my $timeout = ref $_[-1] ? undef : pop;
		878	my $cb = pop;
		879
		880	my $port = port {
		881	undef $timeout;
		882	kil $SELF;
		883	&$cb;
		884	};
		885
		886	if (defined $timeout) {
		887	$timeout = AE::timer $timeout, 0, sub {
		888	undef $timeout;
		889	kil $port;
		890	$cb->();
		891	};
		892	} else {
		893	mon $_[0], sub {
		894	kil $port;
		895	$cb->();
		896	};
		897	}
		898
		899	push @_, $port;
		900	&snd;
		901
		902	$port
		903	}
		904
446	=back	905	=back
447		906
448	=head1 FUNCTIONS FOR NODES	907	=head1 DISTRIBUTED DATABASE
		908
		909	AnyEvent::MP comes with a simple distributed database. The database will
		910	be mirrored asynchronously on all global nodes. Other nodes bind to one
		911	of the global nodes for their needs. Every node has a "local database"
		912	which contains all the values that are set locally. All local databases
		913	are merged together to form the global database, which can be queried.
		914
		915	The database structure is that of a two-level hash - the database hash
		916	contains hashes which contain values, similarly to a perl hash of hashes,
		917	i.e.:
		918
		919	$DATABASE{$family}{$subkey} = $value
		920
		921	The top level hash key is called "family", and the second-level hash key
		922	is called "subkey" or simply "key".
		923
		924	The family must be alphanumeric, i.e. start with a letter and consist
		925	of letters, digits, underscores and colons (C<[A-Za-z][A-Za-z0-9_:]*>,
		926	pretty much like Perl module names.
		927
		928	As the family namespace is global, it is recommended to prefix family names
		929	with the name of the application or module using it.
		930
		931	The subkeys must be non-empty strings, with no further restrictions.
		932
		933	The values should preferably be strings, but other perl scalars should
		934	work as well (such as C<undef>, arrays and hashes).
		935
		936	Every database entry is owned by one node - adding the same family/subkey
		937	combination on multiple nodes will not cause discomfort for AnyEvent::MP,
		938	but the result might be nondeterministic, i.e. the key might have
		939	different values on different nodes.
		940
		941	Different subkeys in the same family can be owned by different nodes
		942	without problems, and in fact, this is the common method to create worker
		943	pools. For example, a worker port for image scaling might do this:
		944
		945	db_set my_image_scalers => $port;
		946
		947	And clients looking for an image scaler will want to get the
		948	C<my_image_scalers> keys from time to time:
		949
		950	db_keys my_image_scalers => sub {
		951	@ports = @{ $_[0] };
		952	};
		953
		954	Or better yet, they want to monitor the database family, so they always
		955	have a reasonable up-to-date copy:
		956
		957	db_mon my_image_scalers => sub {
		958	@ports = keys %{ $_[0] };
		959	};
		960
		961	In general, you can set or delete single subkeys, but query and monitor
		962	whole families only.
		963
		964	If you feel the need to monitor or query a single subkey, try giving it
		965	it's own family.
		966
		967	=over
		968
		969	=item db_set $family => $subkey [=> $value]
		970
		971	Sets (or replaces) a key to the database - if C<$value> is omitted,
		972	C<undef> is used instead.
		973
		974	=item db_del $family => $subkey...
		975
		976	Deletes one or more subkeys from the database family.
		977
		978	=item $guard = db_reg $family => $subkey [=> $value]
		979
		980	Sets the key on the database and returns a guard. When the guard is
		981	destroyed, the key is deleted from the database. If C<$value> is missing,
		982	then C<undef> is used.
		983
		984	=item db_family $family => $cb->(\%familyhash)
		985
		986	Queries the named database C<$family> and call the callback with the
		987	family represented as a hash. You can keep and freely modify the hash.
		988
		989	=item db_keys $family => $cb->(\@keys)
		990
		991	Same as C<db_family>, except it only queries the family I<subkeys> and passes
		992	them as array reference to the callback.
		993
		994	=item db_values $family => $cb->(\@values)
		995
		996	Same as C<db_family>, except it only queries the family I<values> and passes them
		997	as array reference to the callback.
		998
		999	=item $guard = db_mon $family => $cb->($familyhash, \@added, \@changed, \@deleted)
		1000
		1001	Creates a monitor on the given database family. Each time a key is set
		1002	or or is deleted the callback is called with a hash containing the
		1003	database family and three lists of added, changed and deleted subkeys,
		1004	respectively. If no keys have changed then the array reference might be
		1005	C<undef> or even missing.
		1006
		1007	If not called in void context, a guard object is returned that, when
		1008	destroyed, stops the monitor.
		1009
		1010	The family hash reference and the key arrays belong to AnyEvent::MP and
		1011	B<must not be modified or stored> by the callback. When in doubt, make a
		1012	copy.
		1013
		1014	As soon as possible after the monitoring starts, the callback will be
		1015	called with the intiial contents of the family, even if it is empty,
		1016	i.e. there will always be a timely call to the callback with the current
		1017	contents.
		1018
		1019	It is possible that the callback is called with a change event even though
		1020	the subkey is already present and the value has not changed.
		1021
		1022	The monitoring stops when the guard object is destroyed.
		1023
		1024	Example: on every change to the family "mygroup", print out all keys.
		1025
		1026	my $guard = db_mon mygroup => sub {
		1027	my ($family, $a, $c, $d) = @_;
		1028	print "mygroup members: ", (join " ", keys %$family), "\n";
		1029	};
		1030
		1031	Exmaple: wait until the family "My::Module::workers" is non-empty.
		1032
		1033	my $guard; $guard = db_mon My::Module::workers => sub {
		1034	my ($family, $a, $c, $d) = @_;
		1035	return unless %$family;
		1036	undef $guard;
		1037	print "My::Module::workers now nonempty\n";
		1038	};
		1039
		1040	Example: print all changes to the family "AnyRvent::Fantasy::Module".
		1041
		1042	my $guard = db_mon AnyRvent::Fantasy::Module => sub {
		1043	my ($family, $a, $c, $d) = @_;
		1044
		1045	print "+$_=$family->{$_}\n" for @$a;
		1046	print "*$_=$family->{$_}\n" for @$c;
		1047	print "-$_=$family->{$_}\n" for @$d;
		1048	};
		1049
		1050	=cut
		1051
		1052	=back
		1053
		1054	=head1 AnyEvent::MP vs. Distributed Erlang
		1055
		1056	AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node
		1057	== aemp node, Erlang process == aemp port), so many of the documents and
		1058	programming techniques employed by Erlang apply to AnyEvent::MP. Here is a
		1059	sample:
		1060
		1061	http://www.erlang.se/doc/programming_rules.shtml
		1062	http://erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4
		1063	http://erlang.org/download/erlang-book-part1.pdf # chapters 5 and 6
		1064	http://erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5
		1065
		1066	Despite the similarities, there are also some important differences:
449		1067
450	=over 4	1068	=over 4
451		1069
452	=item become_public endpoint...	1070	=item * Node IDs are arbitrary strings in AEMP.
453		1071
454	Tells the node to become a public node, i.e. reachable from other nodes.	1072	Erlang relies on special naming and DNS to work everywhere in the same
		1073	way. AEMP relies on each node somehow knowing its own address(es) (e.g. by
		1074	configuration or DNS), and possibly the addresses of some seed nodes, but
		1075	will otherwise discover other nodes (and their IDs) itself.
455		1076
456	If no arguments are given, or the first argument is C<undef>, then	1077	=item * Erlang has a "remote ports are like local ports" philosophy, AEMP
457	AnyEvent::MP tries to bind on port C<4040> on all IP addresses that the	1078	uses "local ports are like remote ports".
458	local nodename resolves to.
459		1079
460	Otherwise the first argument must be an array-reference with transport	1080	The failure modes for local ports are quite different (runtime errors
461	endpoints ("ip:port", "hostname:port") or port numbers (in which case the	1081	only) then for remote ports - when a local port dies, you I<know> it dies,
462	local nodename is used as hostname). The endpoints are all resolved and	1082	when a connection to another node dies, you know nothing about the other
463	will become the node reference.	1083	port.
464		1084
465	=cut	1085	Erlang pretends remote ports are as reliable as local ports, even when
		1086	they are not.
		1087
		1088	AEMP encourages a "treat remote ports differently" philosophy, with local
		1089	ports being the special case/exception, where transport errors cannot
		1090	occur.
		1091
		1092	=item * Erlang uses processes and a mailbox, AEMP does not queue.
		1093
		1094	Erlang uses processes that selectively receive messages out of order, and
		1095	therefore needs a queue. AEMP is event based, queuing messages would serve
		1096	no useful purpose. For the same reason the pattern-matching abilities
		1097	of AnyEvent::MP are more limited, as there is little need to be able to
		1098	filter messages without dequeuing them.
		1099
		1100	This is not a philosophical difference, but simply stems from AnyEvent::MP
		1101	being event-based, while Erlang is process-based.
		1102
		1103	You cna have a look at L<Coro::MP> for a more Erlang-like process model on
		1104	top of AEMP and Coro threads.
		1105
		1106	=item * Erlang sends are synchronous, AEMP sends are asynchronous.
		1107
		1108	Sending messages in Erlang is synchronous and blocks the process until
		1109	a conenction has been established and the message sent (and so does not
		1110	need a queue that can overflow). AEMP sends return immediately, connection
		1111	establishment is handled in the background.
		1112
		1113	=item * Erlang suffers from silent message loss, AEMP does not.
		1114
		1115	Erlang implements few guarantees on messages delivery - messages can get
		1116	lost without any of the processes realising it (i.e. you send messages a,
		1117	b, and c, and the other side only receives messages a and c).
		1118
		1119	AEMP guarantees (modulo hardware errors) correct ordering, and the
		1120	guarantee that after one message is lost, all following ones sent to the
		1121	same port are lost as well, until monitoring raises an error, so there are
		1122	no silent "holes" in the message sequence.
		1123
		1124	If you want your software to be very reliable, you have to cope with
		1125	corrupted and even out-of-order messages in both Erlang and AEMP. AEMP
		1126	simply tries to work better in common error cases, such as when a network
		1127	link goes down.
		1128
		1129	=item * Erlang can send messages to the wrong port, AEMP does not.
		1130
		1131	In Erlang it is quite likely that a node that restarts reuses an Erlang
		1132	process ID known to other nodes for a completely different process,
		1133	causing messages destined for that process to end up in an unrelated
		1134	process.
		1135
		1136	AEMP does not reuse port IDs, so old messages or old port IDs floating
		1137	around in the network will not be sent to an unrelated port.
		1138
		1139	=item * Erlang uses unprotected connections, AEMP uses secure
		1140	authentication and can use TLS.
		1141
		1142	AEMP can use a proven protocol - TLS - to protect connections and
		1143	securely authenticate nodes.
		1144
		1145	=item * The AEMP protocol is optimised for both text-based and binary
		1146	communications.
		1147
		1148	The AEMP protocol, unlike the Erlang protocol, supports both programming
		1149	language independent text-only protocols (good for debugging), and binary,
		1150	language-specific serialisers (e.g. Storable). By default, unless TLS is
		1151	used, the protocol is actually completely text-based.
		1152
		1153	It has also been carefully designed to be implementable in other languages
		1154	with a minimum of work while gracefully degrading functionality to make the
		1155	protocol simple.
		1156
		1157	=item * AEMP has more flexible monitoring options than Erlang.
		1158
		1159	In Erlang, you can chose to receive I<all> exit signals as messages or
		1160	I<none>, there is no in-between, so monitoring single Erlang processes is
		1161	difficult to implement.
		1162
		1163	Monitoring in AEMP is more flexible than in Erlang, as one can choose
		1164	between automatic kill, exit message or callback on a per-port basis.
		1165
		1166	=item * Erlang tries to hide remote/local connections, AEMP does not.
		1167
		1168	Monitoring in Erlang is not an indicator of process death/crashes, in the
		1169	same way as linking is (except linking is unreliable in Erlang).
		1170
		1171	In AEMP, you don't "look up" registered port names or send to named ports
		1172	that might or might not be persistent. Instead, you normally spawn a port
		1173	on the remote node. The init function monitors you, and you monitor the
		1174	remote port. Since both monitors are local to the node, they are much more
		1175	reliable (no need for C<spawn_link>).
		1176
		1177	This also saves round-trips and avoids sending messages to the wrong port
		1178	(hard to do in Erlang).
466		1179
467	=back	1180	=back
468		1181
469	=head1 NODE MESSAGES	1182	=head1 RATIONALE
470
471	Nodes understand the following messages sent to them. Many of them take
472	arguments called C<@reply>, which will simply be used to compose a reply
473	message - C<$reply[0]> is the port to reply to, C<$reply[1]> the type and
474	the remaining arguments are simply the message data.
475		1183
476	=over 4	1184	=over 4
477		1185
478	=cut	1186	=item Why strings for port and node IDs, why not objects?
479		1187
480	=item lookup => $name, @reply	1188	We considered "objects", but found that the actual number of methods
		1189	that can be called are quite low. Since port and node IDs travel over
		1190	the network frequently, the serialising/deserialising would add lots of
		1191	overhead, as well as having to keep a proxy object everywhere.
481		1192
482	Replies with the port ID of the specified well-known port, or C<undef>.	1193	Strings can easily be printed, easily serialised etc. and need no special
		1194	procedures to be "valid".
483		1195
484	=item devnull => ...	1196	And as a result, a port with just a default receiver consists of a single
		1197	code reference stored in a global hash - it can't become much cheaper.
485		1198
486	Generic data sink/CPU heat conversion.	1199	=item Why favour JSON, why not a real serialising format such as Storable?
487		1200
488	=item relay => $port, @msg	1201	In fact, any AnyEvent::MP node will happily accept Storable as framing
		1202	format, but currently there is no way to make a node use Storable by
		1203	default (although all nodes will accept it).
489		1204
490	Simply forwards the message to the given port.	1205	The default framing protocol is JSON because a) JSON::XS is many times
		1206	faster for small messages and b) most importantly, after years of
		1207	experience we found that object serialisation is causing more problems
		1208	than it solves: Just like function calls, objects simply do not travel
		1209	easily over the network, mostly because they will always be a copy, so you
		1210	always have to re-think your design.
491		1211
492	=item eval => $string[ @reply]	1212	Keeping your messages simple, concentrating on data structures rather than
493		1213	objects, will keep your messages clean, tidy and efficient.
494	Evaluates the given string. If C<@reply> is given, then a message of the
495	form C<@reply, $@, @evalres> is sent.
496
497	Example: crash another node.
498
499	snd $othernode, eval => "exit";
500
501	=item time => @reply
502
503	Replies the the current node time to C<@reply>.
504
505	Example: tell the current node to send the current time to C<$myport> in a
506	C<timereply> message.
507
508	snd $NODE, time => $myport, timereply => 1, 2;
509	# => snd $myport, timereply => 1, 2, <time>
510		1214
511	=back	1215	=back
512		1216
513	=head1 SEE ALSO	1217	=head1 SEE ALSO
		1218
		1219	L<AnyEvent::MP::Intro> - a gentle introduction.
		1220
		1221	L<AnyEvent::MP::Kernel> - more, lower-level, stuff.
		1222
		1223	L<AnyEvent::MP::Global> - network maintenance and port groups, to find
		1224	your applications.
		1225
		1226	L<AnyEvent::MP::DataConn> - establish data connections between nodes.
		1227
		1228	L<AnyEvent::MP::LogCatcher> - simple service to display log messages from
		1229	all nodes.
514		1230
515	L<AnyEvent>.	1231	L<AnyEvent>.
516		1232
517	=head1 AUTHOR	1233	=head1 AUTHOR
518		1234

Diff Legend

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