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

Comparing AnyEvent-MP/MP.pm (file contents):
Revision 1.1 by root, Thu Jul 30 08:38:50 2009 UTC vs.
Revision 1.67 by root, Fri Aug 28 22:21:53 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
		10	NODE # returns this node's noderef
		11	NODE $port # returns the noderef of the port
		12
		13	$SELF # receiving/own port id in rcv callbacks
		14
		15	# initialise the node so it can send/receive messages
		16	initialise_node;
		17
		18	# ports are message endpoints
		19
		20	# sending messages
		21	snd $port, type => data...;
		22	snd $port, @msg;
		23	snd @msg_with_first_element_being_a_port;
		24
		25	# creating/using ports, the simple way
		26	my $simple_port = port { my @msg = @_; 0 };
		27
		28	# creating/using ports, tagged message matching
		29	my $port = port;
		30	rcv $port, ping => sub { snd $_[0], "pong"; 0 };
		31	rcv $port, pong => sub { warn "pong received\n"; 0 };
		32
		33	# create a port on another node
		34	my $port = spawn $node, $initfunc, @initdata;
		35
		36	# monitoring
		37	mon $port, $cb->(@msg) # callback is invoked on death
		38	mon $port, $otherport # kill otherport on abnormal death
		39	mon $port, $otherport, @msg # send message on death
		40
		41	=head1 CURRENT STATUS
		42
		43	AnyEvent::MP - stable API, should work
		44	AnyEvent::MP::Intro - outdated
		45	AnyEvent::MP::Kernel - mostly stable
		46	AnyEvent::MP::Global - mostly stable
		47	AnyEvent::MP::Node - mostly stable, but internal anyways
		48	AnyEvent::MP::Transport - mostly stable, but internal anyways
		49
		50	stay tuned.
		51
9	=head1 DESCRIPTION	52	=head1 DESCRIPTION
10		53
		54	This module (-family) implements a simple message passing framework.
		55
		56	Despite its simplicity, you can securely message other processes running
		57	on the same or other hosts, and you can supervise entities remotely.
		58
		59	For an introduction to this module family, see the L<AnyEvent::MP::Intro>
		60	manual page and the examples under F<eg/>.
		61
		62	At the moment, this module family is a bit underdocumented.
		63
		64	=head1 CONCEPTS
		65
		66	=over 4
		67
		68	=item port
		69
		70	A port is something you can send messages to (with the C<snd> function).
		71
		72	Ports allow you to register C<rcv> handlers that can match all or just
		73	some messages. Messages send to ports will not be queued, regardless of
		74	anything was listening for them or not.
		75
		76	=item port ID - C<nodeid#portname>
		77
		78	A port ID is the concatenation of a node ID, a hash-mark (C<#>) as
		79	separator, and a port name (a printable string of unspecified format).
		80
		81	=item node
		82
		83	A node is a single process containing at least one port - the node port,
		84	which enables nodes to manage each other remotely, and to create new
		85	ports.
		86
		87	Nodes are either public (have one or more listening ports) or private
		88	(no listening ports). Private nodes cannot talk to other private nodes
		89	currently.
		90
		91	=item node ID - C<[a-za-Z0-9_\-.:]+>
		92
		93	A node ID is a string that uniquely identifies the node within a
		94	network. Depending on the configuration used, node IDs can look like a
		95	hostname, a hostname and a port, or a random string. AnyEvent::MP itself
		96	doesn't interpret node IDs in any way.
		97
		98	=item binds - C<ip:port>
		99
		100	Nodes can only talk to each other by creating some kind of connection to
		101	each other. To do this, nodes should listen on one or more local transport
		102	endpoints - binds. Currently, only standard C<ip:port> specifications can
		103	be used, which specify TCP ports to listen on.
		104
		105	=item seeds - C<host:port>
		106
		107	When a node starts, it knows nothing about the network. To teach the node
		108	about the network it first has to contact some other node within the
		109	network. This node is called a seed.
		110
		111	Seeds are transport endpoint(s) of as many nodes as one wants. Those nodes
		112	are expected to be long-running, and at least one of those should always
		113	be available. When nodes run out of connections (e.g. due to a network
		114	error), they try to re-establish connections to some seednodes again to
		115	join the network.
		116
		117	Apart from being sued for seeding, seednodes are not special in any way -
		118	every public node can be a seednode.
		119
		120	=back
		121
		122	=head1 VARIABLES/FUNCTIONS
		123
		124	=over 4
		125
11	=cut	126	=cut
12		127
13	package AnyEvent::MP;	128	package AnyEvent::MP;
14		129
		130	use AnyEvent::MP::Kernel;
		131
15	use common::sense;	132	use common::sense;
16		133
		134	use Carp ();
		135
17	use AE ();	136	use AE ();
18		137
19	our $VERSION = '0.0';	138	use base "Exporter";
20		139
21	sub nonce($) {	140	our $VERSION = $AnyEvent::MP::Kernel::VERSION;
22	my $nonce;
23		141
24	if (open my $fh, "</dev/urandom") {	142	our @EXPORT = qw(
25	sysread $fh, $nonce, $_[0];	143	NODE $NODE *SELF node_of after
26	} else {	144	initialise_node
		145	snd rcv mon mon_guard kil reg psub spawn
		146	port
		147	);
		148
		149	our $SELF;
		150
		151	sub _self_die() {
		152	my $msg = $@;
		153	$msg =~ s/\n+$// unless ref $msg;
		154	kil $SELF, die => $msg;
		155	}
		156
		157	=item $thisnode = NODE / $NODE
		158
		159	The C<NODE> function returns, and the C<$NODE> variable contains, the node
		160	ID of the node running in the current process. This value is initialised by
		161	a call to C<initialise_node>.
		162
		163	=item $nodeid = node_of $port
		164
		165	Extracts and returns the node ID from a port ID or a node ID.
		166
		167	=item initialise_node $profile_name
		168
		169	Before a node can talk to other nodes on the network (i.e. enter
		170	"distributed mode") it has to initialise itself - the minimum a node needs
		171	to know is its own name, and optionally it should know the addresses of
		172	some other nodes in the network to discover other nodes.
		173
		174	This function initialises a node - it must be called exactly once (or
		175	never) before calling other AnyEvent::MP functions.
		176
		177	The first argument is a profile name. If it is C<undef> or missing, then
		178	the current nodename will be used instead (i.e. F<uname -n>).
		179
		180	The function then looks up the profile in the aemp configuration (see the
		181	L<aemp> commandline utility).
		182
		183	If the profile specifies a node ID, then this will become the node ID of
		184	this process. If not, then the profile name will be used as node ID. The
		185	special node ID of C<anon/> will be replaced by a random node ID.
		186
		187	The next step is to look up the binds in the profile, followed by binding
		188	aemp protocol listeners on all binds specified (it is possible and valid
		189	to have no binds, meaning that the node cannot be contacted form the
		190	outside. This means the node cannot talk to other nodes that also have no
		191	binds, but it can still talk to all "normal" nodes).
		192
		193	If the profile does not specify a binds list, then the node ID will be
		194	treated as if it were of the form C<host:port>, which will be resolved and
		195	used as binds list.
		196
		197	Lastly, the seeds list from the profile is passed to the
		198	L<AnyEvent::MP::Global> module, which will then use it to keep
		199	connectivity with at least on of those seed nodes at any point in time.
		200
		201	Example: become a distributed node listening on the guessed noderef, or
		202	the one specified via C<aemp> for the current node. This should be the
		203	most common form of invocation for "daemon"-type nodes.
		204
		205	initialise_node;
		206
		207	Example: become an anonymous node. This form is often used for commandline
		208	clients.
		209
		210	initialise_node "anon/";
		211
		212	Example: become a distributed node. If there is no profile of the given
		213	name, or no binds list was specified, resolve C<localhost:4044> and bind
		214	on the resulting addresses.
		215
		216	initialise_node "localhost:4044";
		217
		218	=item $SELF
		219
		220	Contains the current port id while executing C<rcv> callbacks or C<psub>
		221	blocks.
		222
		223	=item *SELF, SELF, %SELF, @SELF...
		224
		225	Due to some quirks in how perl exports variables, it is impossible to
		226	just export C<$SELF>, all the symbols named C<SELF> are exported by this
		227	module, but only C<$SELF> is currently used.
		228
		229	=item snd $port, type => @data
		230
		231	=item snd $port, @msg
		232
		233	Send the given message to the given port, which can identify either a
		234	local or a remote port, and must be a port ID.
		235
		236	While the message can be almost anything, it is highly recommended to
		237	use a string as first element (a port ID, or some word that indicates a
		238	request type etc.) and to consist if only simple perl values (scalars,
		239	arrays, hashes) - if you think you need to pass an object, think again.
		240
		241	The message data logically becomes read-only after a call to this
		242	function: modifying any argument (or values referenced by them) is
		243	forbidden, as there can be considerable time between the call to C<snd>
		244	and the time the message is actually being serialised - in fact, it might
		245	never be copied as within the same process it is simply handed to the
		246	receiving port.
		247
		248	The type of data you can transfer depends on the transport protocol: when
		249	JSON is used, then only strings, numbers and arrays and hashes consisting
		250	of those are allowed (no objects). When Storable is used, then anything
		251	that Storable can serialise and deserialise is allowed, and for the local
		252	node, anything can be passed. Best rely only on the common denominator of
		253	these.
		254
		255	=item $local_port = port
		256
		257	Create a new local port object and returns its port ID. Initially it has
		258	no callbacks set and will throw an error when it receives messages.
		259
		260	=item $local_port = port { my @msg = @_ }
		261
		262	Creates a new local port, and returns its ID. Semantically the same as
		263	creating a port and calling C<rcv $port, $callback> on it.
		264
		265	The block will be called for every message received on the port, with the
		266	global variable C<$SELF> set to the port ID. Runtime errors will cause the
		267	port to be C<kil>ed. The message will be passed as-is, no extra argument
		268	(i.e. no port ID) will be passed to the callback.
		269
		270	If you want to stop/destroy the port, simply C<kil> it:
		271
		272	my $port = port {
		273	my @msg = @_;
27	# shit...	274	...
28	our $nonce_init;	275	kil $SELF;
29	unless ($nonce_init++) {
30	srand time ^ $$ ^ unpack "%L*", qx"ps -edalf" . qx"ipconfig /all";
31	}
32
33	$nonce = join "", map +(chr rand 256), 1 .. $_[0]
34	}	276	};
35		277
36	$nonce	278	=cut
37	}
38		279
39	our $DEFAULT_SECRET;	280	sub rcv($@);
40		281
41	sub default_secret {	282	sub _kilme {
42	unless (defined $DEFAULT_SECRET) {	283	die "received message on port without callback";
43	if (open my $fh, "<$ENV{HOME}/.aemp-secret") {	284	}
44	sysread $fh, $DEFAULT_SECRET, -s $fh;	285
		286	sub port(;&) {
		287	my $id = "$UNIQ." . $ID++;
		288	my $port = "$NODE#$id";
		289
		290	rcv $port, shift \|\| \&_kilme;
		291
		292	$port
		293	}
		294
		295	=item rcv $local_port, $callback->(@msg)
		296
		297	Replaces the default callback on the specified port. There is no way to
		298	remove the default callback: use C<sub { }> to disable it, or better
		299	C<kil> the port when it is no longer needed.
		300
		301	The global C<$SELF> (exported by this module) contains C<$port> while
		302	executing the callback. Runtime errors during callback execution will
		303	result in the port being C<kil>ed.
		304
		305	The default callback received all messages not matched by a more specific
		306	C<tag> match.
		307
		308	=item rcv $local_port, tag => $callback->(@msg_without_tag), ...
		309
		310	Register (or replace) callbacks to be called on messages starting with the
		311	given tag on the given port (and return the port), or unregister it (when
		312	C<$callback> is C<$undef> or missing). There can only be one callback
		313	registered for each tag.
		314
		315	The original message will be passed to the callback, after the first
		316	element (the tag) has been removed. The callback will use the same
		317	environment as the default callback (see above).
		318
		319	Example: create a port and bind receivers on it in one go.
		320
		321	my $port = rcv port,
		322	msg1 => sub { ... },
		323	msg2 => sub { ... },
		324	;
		325
		326	Example: create a port, bind receivers and send it in a message elsewhere
		327	in one go:
		328
		329	snd $otherport, reply =>
		330	rcv port,
		331	msg1 => sub { ... },
		332	...
		333	;
		334
		335	Example: temporarily register a rcv callback for a tag matching some port
		336	(e.g. for a rpc reply) and unregister it after a message was received.
		337
		338	rcv $port, $otherport => sub {
		339	my @reply = @_;
		340
		341	rcv $SELF, $otherport;
		342	};
		343
		344	=cut
		345
		346	sub rcv($@) {
		347	my $port = shift;
		348	my ($noderef, $portid) = split /#/, $port, 2;
		349
		350	$NODE{$noderef} == $NODE{""}
		351	or Carp::croak "$port: rcv can only be called on local ports, caught";
		352
		353	while (@_) {
		354	if (ref $_[0]) {
		355	if (my $self = $PORT_DATA{$portid}) {
		356	"AnyEvent::MP::Port" eq ref $self
		357	or Carp::croak "$port: rcv can only be called on message matching ports, caught";
		358
		359	$self->[2] = shift;
45	} else {	360	} else {
46	$DEFAULT_SECRET = nonce 32;	361	my $cb = shift;
		362	$PORT{$portid} = sub {
		363	local $SELF = $port;
		364	eval { &$cb }; _self_die if $@;
		365	};
		366	}
		367	} elsif (defined $_[0]) {
		368	my $self = $PORT_DATA{$portid} \|\|= do {
		369	my $self = bless [$PORT{$port} \|\| sub { }, { }, $port], "AnyEvent::MP::Port";
		370
		371	$PORT{$portid} = sub {
		372	local $SELF = $port;
		373
		374	if (my $cb = $self->[1]{$_[0]}) {
		375	shift;
		376	eval { &$cb }; _self_die if $@;
		377	} else {
		378	&{ $self->[0] };
		379	}
		380	};
		381
		382	$self
		383	};
		384
		385	"AnyEvent::MP::Port" eq ref $self
		386	or Carp::croak "$port: rcv can only be called on message matching ports, caught";
		387
		388	my ($tag, $cb) = splice @_, 0, 2;
		389
		390	if (defined $cb) {
		391	$self->[1]{$tag} = $cb;
		392	} else {
		393	delete $self->[1]{$tag};
		394	}
47	}	395	}
48	}	396	}
49		397
50	$DEFAULT_SECRET	398	$port
51	}	399	}
		400
		401	=item $closure = psub { BLOCK }
		402
		403	Remembers C<$SELF> and creates a closure out of the BLOCK. When the
		404	closure is executed, sets up the environment in the same way as in C<rcv>
		405	callbacks, i.e. runtime errors will cause the port to get C<kil>ed.
		406
		407	This is useful when you register callbacks from C<rcv> callbacks:
		408
		409	rcv delayed_reply => sub {
		410	my ($delay, @reply) = @_;
		411	my $timer = AE::timer $delay, 0, psub {
		412	snd @reply, $SELF;
		413	};
		414	};
		415
		416	=cut
		417
		418	sub psub(&) {
		419	my $cb = shift;
		420
		421	my $port = $SELF
		422	or Carp::croak "psub can only be called from within rcv or psub callbacks, not";
		423
		424	sub {
		425	local $SELF = $port;
		426
		427	if (wantarray) {
		428	my @res = eval { &$cb };
		429	_self_die if $@;
		430	@res
		431	} else {
		432	my $res = eval { &$cb };
		433	_self_die if $@;
		434	$res
		435	}
		436	}
		437	}
		438
		439	=item $guard = mon $port, $cb->(@reason) # call $cb when $port dies
		440
		441	=item $guard = mon $port, $rcvport # kill $rcvport when $port dies
		442
		443	=item $guard = mon $port # kill $SELF when $port dies
		444
		445	=item $guard = mon $port, $rcvport, @msg # send a message when $port dies
		446
		447	Monitor the given port and do something when the port is killed or
		448	messages to it were lost, and optionally return a guard that can be used
		449	to stop monitoring again.
		450
		451	C<mon> effectively guarantees that, in the absence of hardware failures,
		452	after starting the monitor, either all messages sent to the port will
		453	arrive, or the monitoring action will be invoked after possible message
		454	loss has been detected. No messages will be lost "in between" (after
		455	the first lost message no further messages will be received by the
		456	port). After the monitoring action was invoked, further messages might get
		457	delivered again.
		458
		459	Note that monitoring-actions are one-shot: once messages are lost (and a
		460	monitoring alert was raised), they are removed and will not trigger again.
		461
		462	In the first form (callback), the callback is simply called with any
		463	number of C<@reason> elements (no @reason means that the port was deleted
		464	"normally"). Note also that I<< the callback B<must> never die >>, so use
		465	C<eval> if unsure.
		466
		467	In the second form (another port given), the other port (C<$rcvport>)
		468	will be C<kil>'ed with C<@reason>, iff a @reason was specified, i.e. on
		469	"normal" kils nothing happens, while under all other conditions, the other
		470	port is killed with the same reason.
		471
		472	The third form (kill self) is the same as the second form, except that
		473	C<$rvport> defaults to C<$SELF>.
		474
		475	In the last form (message), a message of the form C<@msg, @reason> will be
		476	C<snd>.
		477
		478	As a rule of thumb, monitoring requests should always monitor a port from
		479	a local port (or callback). The reason is that kill messages might get
		480	lost, just like any other message. Another less obvious reason is that
		481	even monitoring requests can get lost (for exmaple, when the connection
		482	to the other node goes down permanently). When monitoring a port locally
		483	these problems do not exist.
		484
		485	Example: call a given callback when C<$port> is killed.
		486
		487	mon $port, sub { warn "port died because of <@_>\n" };
		488
		489	Example: kill ourselves when C<$port> is killed abnormally.
		490
		491	mon $port;
		492
		493	Example: send us a restart message when another C<$port> is killed.
		494
		495	mon $port, $self => "restart";
		496
		497	=cut
		498
		499	sub mon {
		500	my ($noderef, $port) = split /#/, shift, 2;
		501
		502	my $node = $NODE{$noderef} \|\| add_node $noderef;
		503
		504	my $cb = @_ ? shift : $SELF \|\| Carp::croak 'mon: called with one argument only, but $SELF not set,';
		505
		506	unless (ref $cb) {
		507	if (@_) {
		508	# send a kill info message
		509	my (@msg) = ($cb, @_);
		510	$cb = sub { snd @msg, @_ };
		511	} else {
		512	# simply kill other port
		513	my $port = $cb;
		514	$cb = sub { kil $port, @_ if @_ };
		515	}
		516	}
		517
		518	$node->monitor ($port, $cb);
		519
		520	defined wantarray
		521	and AnyEvent::Util::guard { $node->unmonitor ($port, $cb) }
		522	}
		523
		524	=item $guard = mon_guard $port, $ref, $ref...
		525
		526	Monitors the given C<$port> and keeps the passed references. When the port
		527	is killed, the references will be freed.
		528
		529	Optionally returns a guard that will stop the monitoring.
		530
		531	This function is useful when you create e.g. timers or other watchers and
		532	want to free them when the port gets killed (note the use of C<psub>):
		533
		534	$port->rcv (start => sub {
		535	my $timer; $timer = mon_guard $port, AE::timer 1, 1, psub {
		536	undef $timer if 0.9 < rand;
		537	});
		538	});
		539
		540	=cut
		541
		542	sub mon_guard {
		543	my ($port, @refs) = @_;
		544
		545	#TODO: mon-less form?
		546
		547	mon $port, sub { 0 && @refs }
		548	}
		549
		550	=item kil $port[, @reason]
		551
		552	Kill the specified port with the given C<@reason>.
		553
		554	If no C<@reason> is specified, then the port is killed "normally" (ports
		555	monitoring other ports will not necessarily die because a port dies
		556	"normally").
		557
		558	Otherwise, linked ports get killed with the same reason (second form of
		559	C<mon>, see above).
		560
		561	Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks
		562	will be reported as reason C<< die => $@ >>.
		563
		564	Transport/communication errors are reported as C<< transport_error =>
		565	$message >>.
		566
		567	=cut
		568
		569	=item $port = spawn $node, $initfunc[, @initdata]
		570
		571	Creates a port on the node C<$node> (which can also be a port ID, in which
		572	case it's the node where that port resides).
		573
		574	The port ID of the newly created port is returned immediately, and it is
		575	possible to immediately start sending messages or to monitor the port.
		576
		577	After the port has been created, the init function is called on the remote
		578	node, in the same context as a C<rcv> callback. This function must be a
		579	fully-qualified function name (e.g. C<MyApp::Chat::Server::init>). To
		580	specify a function in the main program, use C<::name>.
		581
		582	If the function doesn't exist, then the node tries to C<require>
		583	the package, then the package above the package and so on (e.g.
		584	C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function
		585	exists or it runs out of package names.
		586
		587	The init function is then called with the newly-created port as context
		588	object (C<$SELF>) and the C<@initdata> values as arguments.
		589
		590	A common idiom is to pass a local port, immediately monitor the spawned
		591	port, and in the remote init function, immediately monitor the passed
		592	local port. This two-way monitoring ensures that both ports get cleaned up
		593	when there is a problem.
		594
		595	Example: spawn a chat server port on C<$othernode>.
		596
		597	# this node, executed from within a port context:
		598	my $server = spawn $othernode, "MyApp::Chat::Server::connect", $SELF;
		599	mon $server;
		600
		601	# init function on C<$othernode>
		602	sub connect {
		603	my ($srcport) = @_;
		604
		605	mon $srcport;
		606
		607	rcv $SELF, sub {
		608	...
		609	};
		610	}
		611
		612	=cut
		613
		614	sub _spawn {
		615	my $port = shift;
		616	my $init = shift;
		617
		618	local $SELF = "$NODE#$port";
		619	eval {
		620	&{ load_func $init }
		621	};
		622	_self_die if $@;
		623	}
		624
		625	sub spawn(@) {
		626	my ($noderef, undef) = split /#/, shift, 2;
		627
		628	my $id = "$RUNIQ." . $ID++;
		629
		630	$_[0] =~ /::/
		631	or Carp::croak "spawn init function must be a fully-qualified name, caught";
		632
		633	snd_to_func $noderef, "AnyEvent::MP::_spawn" => $id, @_;
		634
		635	"$noderef#$id"
		636	}
		637
		638	=item after $timeout, @msg
		639
		640	=item after $timeout, $callback
		641
		642	Either sends the given message, or call the given callback, after the
		643	specified number of seconds.
		644
		645	This is simply a utility function that comes in handy at times - the
		646	AnyEvent::MP author is not convinced of the wisdom of having it, though,
		647	so it may go away in the future.
		648
		649	=cut
		650
		651	sub after($@) {
		652	my ($timeout, @action) = @_;
		653
		654	my $t; $t = AE::timer $timeout, 0, sub {
		655	undef $t;
		656	ref $action[0]
		657	? $action[0]()
		658	: snd @action;
		659	};
		660	}
		661
		662	=back
		663
		664	=head1 AnyEvent::MP vs. Distributed Erlang
		665
		666	AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node
		667	== aemp node, Erlang process == aemp port), so many of the documents and
		668	programming techniques employed by Erlang apply to AnyEvent::MP. Here is a
		669	sample:
		670
		671	http://www.Erlang.se/doc/programming_rules.shtml
		672	http://Erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4
		673	http://Erlang.org/download/Erlang-book-part1.pdf # chapters 5 and 6
		674	http://Erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5
		675
		676	Despite the similarities, there are also some important differences:
		677
		678	=over 4
		679
		680	=item * Node IDs are arbitrary strings in AEMP.
		681
		682	Erlang relies on special naming and DNS to work everywhere in the same
		683	way. AEMP relies on each node somehow knowing its own address(es) (e.g. by
		684	configuraiton or DNS), but will otherwise discover other odes itself.
		685
		686	=item * Erlang has a "remote ports are like local ports" philosophy, AEMP
		687	uses "local ports are like remote ports".
		688
		689	The failure modes for local ports are quite different (runtime errors
		690	only) then for remote ports - when a local port dies, you I<know> it dies,
		691	when a connection to another node dies, you know nothing about the other
		692	port.
		693
		694	Erlang pretends remote ports are as reliable as local ports, even when
		695	they are not.
		696
		697	AEMP encourages a "treat remote ports differently" philosophy, with local
		698	ports being the special case/exception, where transport errors cannot
		699	occur.
		700
		701	=item * Erlang uses processes and a mailbox, AEMP does not queue.
		702
		703	Erlang uses processes that selectively receive messages, and therefore
		704	needs a queue. AEMP is event based, queuing messages would serve no
		705	useful purpose. For the same reason the pattern-matching abilities of
		706	AnyEvent::MP are more limited, as there is little need to be able to
		707	filter messages without dequeing them.
		708
		709	(But see L<Coro::MP> for a more Erlang-like process model on top of AEMP).
		710
		711	=item * Erlang sends are synchronous, AEMP sends are asynchronous.
		712
		713	Sending messages in Erlang is synchronous and blocks the process (and
		714	so does not need a queue that can overflow). AEMP sends are immediate,
		715	connection establishment is handled in the background.
		716
		717	=item * Erlang suffers from silent message loss, AEMP does not.
		718
		719	Erlang makes few guarantees on messages delivery - messages can get lost
		720	without any of the processes realising it (i.e. you send messages a, b,
		721	and c, and the other side only receives messages a and c).
		722
		723	AEMP guarantees correct ordering, and the guarantee that after one message
		724	is lost, all following ones sent to the same port are lost as well, until
		725	monitoring raises an error, so there are no silent "holes" in the message
		726	sequence.
		727
		728	=item * Erlang can send messages to the wrong port, AEMP does not.
		729
		730	In Erlang it is quite likely that a node that restarts reuses a process ID
		731	known to other nodes for a completely different process, causing messages
		732	destined for that process to end up in an unrelated process.
		733
		734	AEMP never reuses port IDs, so old messages or old port IDs floating
		735	around in the network will not be sent to an unrelated port.
		736
		737	=item * Erlang uses unprotected connections, AEMP uses secure
		738	authentication and can use TLS.
		739
		740	AEMP can use a proven protocol - TLS - to protect connections and
		741	securely authenticate nodes.
		742
		743	=item * The AEMP protocol is optimised for both text-based and binary
		744	communications.
		745
		746	The AEMP protocol, unlike the Erlang protocol, supports both programming
		747	language independent text-only protocols (good for debugging) and binary,
		748	language-specific serialisers (e.g. Storable). By default, unless TLS is
		749	used, the protocol is actually completely text-based.
		750
		751	It has also been carefully designed to be implementable in other languages
		752	with a minimum of work while gracefully degrading functionality to make the
		753	protocol simple.
		754
		755	=item * AEMP has more flexible monitoring options than Erlang.
		756
		757	In Erlang, you can chose to receive I<all> exit signals as messages
		758	or I<none>, there is no in-between, so monitoring single processes is
		759	difficult to implement. Monitoring in AEMP is more flexible than in
		760	Erlang, as one can choose between automatic kill, exit message or callback
		761	on a per-process basis.
		762
		763	=item * Erlang tries to hide remote/local connections, AEMP does not.
		764
		765	Monitoring in Erlang is not an indicator of process death/crashes, in the
		766	same way as linking is (except linking is unreliable in Erlang).
		767
		768	In AEMP, you don't "look up" registered port names or send to named ports
		769	that might or might not be persistent. Instead, you normally spawn a port
		770	on the remote node. The init function monitors you, and you monitor the
		771	remote port. Since both monitors are local to the node, they are much more
		772	reliable (no need for C<spawn_link>).
		773
		774	This also saves round-trips and avoids sending messages to the wrong port
		775	(hard to do in Erlang).
		776
		777	=back
		778
		779	=head1 RATIONALE
		780
		781	=over 4
		782
		783	=item Why strings for port and node IDs, why not objects?
		784
		785	We considered "objects", but found that the actual number of methods
		786	that can be called are quite low. Since port and node IDs travel over
		787	the network frequently, the serialising/deserialising would add lots of
		788	overhead, as well as having to keep a proxy object everywhere.
		789
		790	Strings can easily be printed, easily serialised etc. and need no special
		791	procedures to be "valid".
		792
		793	And as a result, a miniport consists of a single closure stored in a
		794	global hash - it can't become much cheaper.
		795
		796	=item Why favour JSON, why not a real serialising format such as Storable?
		797
		798	In fact, any AnyEvent::MP node will happily accept Storable as framing
		799	format, but currently there is no way to make a node use Storable by
		800	default (although all nodes will accept it).
		801
		802	The default framing protocol is JSON because a) JSON::XS is many times
		803	faster for small messages and b) most importantly, after years of
		804	experience we found that object serialisation is causing more problems
		805	than it solves: Just like function calls, objects simply do not travel
		806	easily over the network, mostly because they will always be a copy, so you
		807	always have to re-think your design.
		808
		809	Keeping your messages simple, concentrating on data structures rather than
		810	objects, will keep your messages clean, tidy and efficient.
		811
		812	=back
52		813
53	=head1 SEE ALSO	814	=head1 SEE ALSO
54		815
55	L<AnyEvent>.	816	L<AnyEvent>.
56		817

Diff Legend

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

Comparing AnyEvent-MP/MP.pm (file contents): Revision 1.1 by root, Thu Jul 30 08:38:50 2009 UTC vs. Revision 1.67 by root, Fri Aug 28 22:21:53 2009 UTC

Diff Legend

Comparing AnyEvent-MP/MP.pm (file contents):
Revision 1.1 by root, Thu Jul 30 08:38:50 2009 UTC vs.
Revision 1.67 by root, Fri Aug 28 22:21:53 2009 UTC