[ViewVC] Diff of: cvs/Coro/Coro.pm

Comparing Coro/Coro.pm (file contents):
Revision 1.5 by root, Tue Jul 10 01:43:21 2001 UTC vs.
Revision 1.247 by root, Mon Dec 15 02:07:11 2008 UTC

1	=head1 NAME	1	=head1 NAME
2		2
3	Coro - create and manage coroutines	3	Coro - the only real threads in perl
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	use Coro;	7	use Coro;
8		8
9	$new = new Coro sub {	9	async {
10	print "in coroutine, switching back\n";	10	# some asynchronous thread of execution
11	$Coro::main->resume;	11	print "2\n";
12	print "in coroutine again, switching back\n";	12	cede; # yield back to main
13	$Coro::main->resume;	13	print "4\n";
14	};	14	};
15		15	print "1\n";
16	print "in main, switching to coroutine\n";	16	cede; # yield to coroutine
17	$new->resume;	17	print "3\n";
18	print "back in main, switch to coroutine again\n";	18	cede; # and again
19	$new->resume;	19
20	print "back in main\n";	20	# use locking
		21	use Coro::Semaphore;
		22	my $lock = new Coro::Semaphore;
		23	my $locked;
		24
		25	$lock->down;
		26	$locked = 1;
		27	$lock->up;
21		28
22	=head1 DESCRIPTION	29	=head1 DESCRIPTION
23		30
24	This module implements coroutines. Coroutines, similar to continuations,	31	For a tutorial-style introduction, please read the L<Coro::Intro>
25	allow you to run more than one "thread of execution" in parallel. Unlike	32	manpage. This manpage mainly contains reference information.
26	threads this, only voluntary switching is used so locking problems are
27	greatly reduced.
28		33
29	Although this is the "main" module of the Coro family it provides only	34	This module collection manages continuations in general, most often
30	low-level functionality. See L<Coro::Process> and related modules for a	35	in the form of cooperative threads (also called coroutines in the
31	more useful process abstraction including scheduling.	36	documentation). They are similar to kernel threads but don't (in general)
		37	run in parallel at the same time even on SMP machines. The specific flavor
		38	of thread offered by this module also guarantees you that it will not
		39	switch between threads unless necessary, at easily-identified points in
		40	your program, so locking and parallel access are rarely an issue, making
		41	thread programming much safer and easier than using other thread models.
		42
		43	Unlike the so-called "Perl threads" (which are not actually real threads
		44	but only the windows process emulation ported to unix), Coro provides a
		45	full shared address space, which makes communication between threads
		46	very easy. And threads are fast, too: disabling the Windows process
		47	emulation code in your perl and using Coro can easily result in a two to
		48	four times speed increase for your programs.
		49
		50	Coro achieves that by supporting multiple running interpreters that share
		51	data, which is especially useful to code pseudo-parallel processes and
		52	for event-based programming, such as multiple HTTP-GET requests running
		53	concurrently. See L<Coro::AnyEvent> to learn more on how to integrate Coro
		54	into an event-based environment.
		55
		56	In this module, a thread is defined as "callchain + lexical variables +
		57	@_ + $_ + $@ + $/ + C stack), that is, a thread has its own callchain,
		58	its own set of lexicals and its own set of perls most important global
		59	variables (see L<Coro::State> for more configuration and background info).
		60
		61	See also the C<SEE ALSO> section at the end of this document - the Coro
		62	module family is quite large.
		63
		64	=cut
		65
		66	package Coro;
		67
		68	use strict qw(vars subs);
		69	no warnings "uninitialized";
		70
		71	use Guard ();
		72
		73	use Coro::State;
		74
		75	use base qw(Coro::State Exporter);
		76
		77	our $idle; # idle handler
		78	our $main; # main coroutine
		79	our $current; # current coroutine
		80
		81	our $VERSION = 5.13;
		82
		83	our @EXPORT = qw(async async_pool cede schedule terminate current unblock_sub);
		84	our %EXPORT_TAGS = (
		85	prio => [qw(PRIO_MAX PRIO_HIGH PRIO_NORMAL PRIO_LOW PRIO_IDLE PRIO_MIN)],
		86	);
		87	our @EXPORT_OK = (@{$EXPORT_TAGS{prio}}, qw(nready));
		88
		89	=head1 GLOBAL VARIABLES
32		90
33	=over 4	91	=over 4
34		92
35	=cut
36
37	package Coro;
38
39	BEGIN {
40	$VERSION = 0.03;
41
42	require XSLoader;
43	XSLoader::load Coro, $VERSION;
44	}
45
46	=item $main	93	=item $Coro::main
47		94
48	This coroutine represents the main program.	95	This variable stores the coroutine object that represents the main
		96	program. While you cna C<ready> it and do most other things you can do to
		97	coroutines, it is mainly useful to compare again C<$Coro::current>, to see
		98	whether you are running in the main program or not.
49		99
		100	=cut
		101
		102	# $main is now being initialised by Coro::State
		103
50	=item $current	104	=item $Coro::current
51		105
52	The current coroutine (the last coroutine switched to). The initial value is C<$main> (of course).	106	The coroutine object representing the current coroutine (the last
		107	coroutine that the Coro scheduler switched to). The initial value is
		108	C<$Coro::main> (of course).
53		109
54	=cut	110	This variable is B<strictly> I<read-only>. You can take copies of the
		111	value stored in it and use it as any other coroutine object, but you must
		112	not otherwise modify the variable itself.
55		113
56	$main = $current = _newprocess {	114	=cut
57	# never being called	115
		116	sub current() { $current } # [DEPRECATED]
		117
		118	=item $Coro::idle
		119
		120	This variable is mainly useful to integrate Coro into event loops. It is
		121	usually better to rely on L<Coro::AnyEvent> or L<Coro::EV>, as this is
		122	pretty low-level functionality.
		123
		124	This variable stores either a coroutine or a callback.
		125
		126	If it is a callback, the it is called whenever the scheduler finds no
		127	ready coroutines to run. The default implementation prints "FATAL:
		128	deadlock detected" and exits, because the program has no other way to
		129	continue.
		130
		131	If it is a coroutine object, then this object will be readied (without
		132	invoking any ready hooks, however) when the scheduler finds no other ready
		133	coroutines to run.
		134
		135	This hook is overwritten by modules such as C<Coro::EV> and
		136	C<Coro::AnyEvent> to wait on an external event that hopefully wake up a
		137	coroutine so the scheduler can run it.
		138
		139	Note that the callback I<must not>, under any circumstances, block
		140	the current coroutine. Normally, this is achieved by having an "idle
		141	coroutine" that calls the event loop and then blocks again, and then
		142	readying that coroutine in the idle handler, or by simply placing the idle
		143	coroutine in this variable.
		144
		145	See L<Coro::Event> or L<Coro::AnyEvent> for examples of using this
		146	technique.
		147
		148	Please note that if your callback recursively invokes perl (e.g. for event
		149	handlers), then it must be prepared to be called recursively itself.
		150
		151	=cut
		152
		153	$idle = sub {
		154	require Carp;
		155	Carp::croak ("FATAL: deadlock detected");
58	};	156	};
59		157
60	=item $error, $error_msg, $error_coro	158	# this coroutine is necessary because a coroutine
		159	# cannot destroy itself.
		160	our @destroy;
		161	our $manager;
61		162
62	This coroutine will be called on fatal errors. C<$error_msg> and	163	$manager = new Coro sub {
63	C<$error_coro> return the error message and the error-causing coroutine,	164	while () {
64	respectively.	165	Coro::_cancel shift @destroy
		166	while @destroy;
65		167
66	=cut	168	&schedule;
67		169	}
68	$error_msg =
69	$error_coro = undef;
70
71	$error = _newprocess {
72	print STDERR "FATAL: $error_msg\nprogram aborted\n";
73	exit 250;
74	};	170	};
		171	$manager->{desc} = "[coro manager]";
		172	$manager->prio (PRIO_MAX);
75		173
76	=item $coro = new $coderef [, @args]	174	=back
77		175
78	Create a new coroutine and return it. The first C<resume> call to this	176	=head1 SIMPLE COROUTINE CREATION
79	coroutine will start execution at the given coderef. If it returns it
80	should return a coroutine to switch to. If, after returning, the coroutine
81	is C<resume>d again it starts execution again at the givne coderef.
82		177
83	=cut	178	=over 4
84		179
		180	=item async { ... } [@args...]
		181
		182	Create a new coroutine and return its coroutine object (usually
		183	unused). The coroutine will be put into the ready queue, so
		184	it will start running automatically on the next scheduler run.
		185
		186	The first argument is a codeblock/closure that should be executed in the
		187	coroutine. When it returns argument returns the coroutine is automatically
		188	terminated.
		189
		190	The remaining arguments are passed as arguments to the closure.
		191
		192	See the C<Coro::State::new> constructor for info about the coroutine
		193	environment in which coroutines are executed.
		194
		195	Calling C<exit> in a coroutine will do the same as calling exit outside
		196	the coroutine. Likewise, when the coroutine dies, the program will exit,
		197	just as it would in the main program.
		198
		199	If you do not want that, you can provide a default C<die> handler, or
		200	simply avoid dieing (by use of C<eval>).
		201
		202	Example: Create a new coroutine that just prints its arguments.
		203
		204	async {
		205	print "@_\n";
		206	} 1,2,3,4;
		207
		208	=cut
		209
		210	sub async(&@) {
		211	my $coro = new Coro @_;
		212	$coro->ready;
		213	$coro
		214	}
		215
		216	=item async_pool { ... } [@args...]
		217
		218	Similar to C<async>, but uses a coroutine pool, so you should not call
		219	terminate or join on it (although you are allowed to), and you get a
		220	coroutine that might have executed other code already (which can be good
		221	or bad :).
		222
		223	On the plus side, this function is about twice as fast as creating (and
		224	destroying) a completely new coroutine, so if you need a lot of generic
		225	coroutines in quick successsion, use C<async_pool>, not C<async>.
		226
		227	The code block is executed in an C<eval> context and a warning will be
		228	issued in case of an exception instead of terminating the program, as
		229	C<async> does. As the coroutine is being reused, stuff like C<on_destroy>
		230	will not work in the expected way, unless you call terminate or cancel,
		231	which somehow defeats the purpose of pooling (but is fine in the
		232	exceptional case).
		233
		234	The priority will be reset to C<0> after each run, tracing will be
		235	disabled, the description will be reset and the default output filehandle
		236	gets restored, so you can change all these. Otherwise the coroutine will
		237	be re-used "as-is": most notably if you change other per-coroutine global
		238	stuff such as C<$/> you I<must needs> revert that change, which is most
		239	simply done by using local as in: C<< local $/ >>.
		240
		241	The idle pool size is limited to C<8> idle coroutines (this can be
		242	adjusted by changing $Coro::POOL_SIZE), but there can be as many non-idle
		243	coros as required.
		244
		245	If you are concerned about pooled coroutines growing a lot because a
		246	single C<async_pool> used a lot of stackspace you can e.g. C<async_pool
		247	{ terminate }> once per second or so to slowly replenish the pool. In
		248	addition to that, when the stacks used by a handler grows larger than 32kb
		249	(adjustable via $Coro::POOL_RSS) it will also be destroyed.
		250
		251	=cut
		252
		253	our $POOL_SIZE = 8;
		254	our $POOL_RSS = 32 * 1024;
		255	our @async_pool;
		256
		257	sub pool_handler {
		258	while () {
		259	eval {
		260	&{&_pool_handler} while 1;
		261	};
		262
		263	warn $@ if $@;
		264	}
		265	}
		266
		267	=back
		268
		269	=head1 STATIC METHODS
		270
		271	Static methods are actually functions that implicitly operate on the
		272	current coroutine.
		273
		274	=over 4
		275
		276	=item schedule
		277
		278	Calls the scheduler. The scheduler will find the next coroutine that is
		279	to be run from the ready queue and switches to it. The next coroutine
		280	to be run is simply the one with the highest priority that is longest
		281	in its ready queue. If there is no coroutine ready, it will clal the
		282	C<$Coro::idle> hook.
		283
		284	Please note that the current coroutine will I<not> be put into the ready
		285	queue, so calling this function usually means you will never be called
		286	again unless something else (e.g. an event handler) calls C<< ->ready >>,
		287	thus waking you up.
		288
		289	This makes C<schedule> I<the> generic method to use to block the current
		290	coroutine and wait for events: first you remember the current coroutine in
		291	a variable, then arrange for some callback of yours to call C<< ->ready
		292	>> on that once some event happens, and last you call C<schedule> to put
		293	yourself to sleep. Note that a lot of things can wake your coroutine up,
		294	so you need to check whether the event indeed happened, e.g. by storing the
		295	status in a variable.
		296
		297	See B<HOW TO WAIT FOR A CALLBACK>, below, for some ways to wait for callbacks.
		298
		299	=item cede
		300
		301	"Cede" to other coroutines. This function puts the current coroutine into
		302	the ready queue and calls C<schedule>, which has the effect of giving
		303	up the current "timeslice" to other coroutines of the same or higher
		304	priority. Once your coroutine gets its turn again it will automatically be
		305	resumed.
		306
		307	This function is often called C<yield> in other languages.
		308
		309	=item Coro::cede_notself
		310
		311	Works like cede, but is not exported by default and will cede to I<any>
		312	coroutine, regardless of priority. This is useful sometimes to ensure
		313	progress is made.
		314
		315	=item terminate [arg...]
		316
		317	Terminates the current coroutine with the given status values (see L<cancel>).
		318
		319	=item Coro::on_enter BLOCK, Coro::on_leave BLOCK
		320
		321	These function install enter and leave winders in the current scope. The
		322	enter block will be executed when on_enter is called and whenever the
		323	current coroutine is re-entered by the scheduler, while the leave block is
		324	executed whenever the current coroutine is blocked by the scheduler, and
		325	also when the containing scope is exited (by whatever means, be it exit,
		326	die, last etc.).
		327
		328	I<Neither invoking the scheduler, nor exceptions, are allowed within those
		329	BLOCKs>. That means: do not even think about calling C<die> without an
		330	eval, and do not even think of entering the scheduler in any way.
		331
		332	Since both BLOCKs are tied to the current scope, they will automatically
		333	be removed when the current scope exits.
		334
		335	These functions implement the same concept as C<dynamic-wind> in scheme
		336	does, and are useful when you want to localise some resource to a specific
		337	coroutine.
		338
		339	They slow down coroutine switching considerably for coroutines that use
		340	them (But coroutine switching is still reasonably fast if the handlers are
		341	fast).
		342
		343	These functions are best understood by an example: The following function
		344	will change the current timezone to "Antarctica/South_Pole", which
		345	requires a call to C<tzset>, but by using C<on_enter> and C<on_leave>,
		346	which remember/change the current timezone and restore the previous
		347	value, respectively, the timezone is only changes for the coroutine that
		348	installed those handlers.
		349
		350	use POSIX qw(tzset);
		351
		352	async {
		353	my $old_tz; # store outside TZ value here
		354
		355	Coro::on_enter {
		356	$old_tz = $ENV{TZ}; # remember the old value
		357
		358	$ENV{TZ} = "Antarctica/South_Pole";
		359	tzset; # enable new value
		360	};
		361
		362	Coro::on_leave {
		363	$ENV{TZ} = $old_tz;
		364	tzset; # restore old value
		365	};
		366
		367	# at this place, the timezone is Antarctica/South_Pole,
		368	# without disturbing the TZ of any other coroutine.
		369	};
		370
		371	This can be used to localise about any resource (locale, uid, current
		372	working directory etc.) to a block, despite the existance of other
		373	coroutines.
		374
		375	=item killall
		376
		377	Kills/terminates/cancels all coroutines except the currently running one.
		378
		379	Note that while this will try to free some of the main interpreter
		380	resources if the calling coroutine isn't the main coroutine, but one
		381	cannot free all of them, so if a coroutine that is not the main coroutine
		382	calls this function, there will be some one-time resource leak.
		383
		384	=cut
		385
		386	sub killall {
		387	for (Coro::State::list) {
		388	$_->cancel
		389	if $_ != $current && UNIVERSAL::isa $_, "Coro";
		390	}
		391	}
		392
		393	=back
		394
		395	=head1 COROUTINE OBJECT METHODS
		396
		397	These are the methods you can call on coroutine objects (or to create
		398	them).
		399
		400	=over 4
		401
		402	=item new Coro \&sub [, @args...]
		403
		404	Create a new coroutine and return it. When the sub returns, the coroutine
		405	automatically terminates as if C<terminate> with the returned values were
		406	called. To make the coroutine run you must first put it into the ready
		407	queue by calling the ready method.
		408
		409	See C<async> and C<Coro::State::new> for additional info about the
		410	coroutine environment.
		411
		412	=cut
		413
		414	sub _coro_run {
		415	terminate &{+shift};
		416	}
		417
		418	=item $success = $coroutine->ready
		419
		420	Put the given coroutine into the end of its ready queue (there is one
		421	queue for each priority) and return true. If the coroutine is already in
		422	the ready queue, do nothing and return false.
		423
		424	This ensures that the scheduler will resume this coroutine automatically
		425	once all the coroutines of higher priority and all coroutines of the same
		426	priority that were put into the ready queue earlier have been resumed.
		427
		428	=item $is_ready = $coroutine->is_ready
		429
		430	Return whether the coroutine is currently the ready queue or not,
		431
		432	=item $coroutine->cancel (arg...)
		433
		434	Terminates the given coroutine and makes it return the given arguments as
		435	status (default: the empty list). Never returns if the coroutine is the
		436	current coroutine.
		437
		438	=cut
		439
85	sub new {	440	sub cancel {
86	my $class = $_[0];	441	my $self = shift;
87	my $proc = $_[1];	442
88	bless _newprocess {	443	if ($current == $self) {
89	do {	444	terminate @_;
90	eval { &$proc->resume };	445	} else {
91	if ($@) {	446	$self->{_status} = [@_];
92	($error_msg, $error_coro) = ($@, $current);	447	$self->_cancel;
93	$error->resume;	448	}
		449	}
		450
		451	=item $coroutine->schedule_to
		452
		453	Puts the current coroutine to sleep (like C<Coro::schedule>), but instead
		454	of continuing with the next coro from the ready queue, always switch to
		455	the given coroutine object (regardless of priority etc.). The readyness
		456	state of that coroutine isn't changed.
		457
		458	This is an advanced method for special cases - I'd love to hear about any
		459	uses for this one.
		460
		461	=item $coroutine->cede_to
		462
		463	Like C<schedule_to>, but puts the current coroutine into the ready
		464	queue. This has the effect of temporarily switching to the given
		465	coroutine, and continuing some time later.
		466
		467	This is an advanced method for special cases - I'd love to hear about any
		468	uses for this one.
		469
		470	=item $coroutine->throw ([$scalar])
		471
		472	If C<$throw> is specified and defined, it will be thrown as an exception
		473	inside the coroutine at the next convenient point in time. Otherwise
		474	clears the exception object.
		475
		476	Coro will check for the exception each time a schedule-like-function
		477	returns, i.e. after each C<schedule>, C<cede>, C<< Coro::Semaphore->down
		478	>>, C<< Coro::Handle->readable >> and so on. Most of these functions
		479	detect this case and return early in case an exception is pending.
		480
		481	The exception object will be thrown "as is" with the specified scalar in
		482	C<$@>, i.e. if it is a string, no line number or newline will be appended
		483	(unlike with C<die>).
		484
		485	This can be used as a softer means than C<cancel> to ask a coroutine to
		486	end itself, although there is no guarantee that the exception will lead to
		487	termination, and if the exception isn't caught it might well end the whole
		488	program.
		489
		490	You might also think of C<throw> as being the moral equivalent of
		491	C<kill>ing a coroutine with a signal (in this case, a scalar).
		492
		493	=item $coroutine->join
		494
		495	Wait until the coroutine terminates and return any values given to the
		496	C<terminate> or C<cancel> functions. C<join> can be called concurrently
		497	from multiple coroutines, and all will be resumed and given the status
		498	return once the C<$coroutine> terminates.
		499
		500	=cut
		501
		502	sub join {
		503	my $self = shift;
		504
		505	unless ($self->{_status}) {
		506	my $current = $current;
		507
		508	push @{$self->{_on_destroy}}, sub {
		509	$current->ready;
		510	undef $current;
		511	};
		512
		513	&schedule while $current;
		514	}
		515
		516	wantarray ? @{$self->{_status}} : $self->{_status}[0];
		517	}
		518
		519	=item $coroutine->on_destroy (\&cb)
		520
		521	Registers a callback that is called when this coroutine gets destroyed,
		522	but before it is joined. The callback gets passed the terminate arguments,
		523	if any, and I<must not> die, under any circumstances.
		524
		525	=cut
		526
		527	sub on_destroy {
		528	my ($self, $cb) = @_;
		529
		530	push @{ $self->{_on_destroy} }, $cb;
		531	}
		532
		533	=item $oldprio = $coroutine->prio ($newprio)
		534
		535	Sets (or gets, if the argument is missing) the priority of the
		536	coroutine. Higher priority coroutines get run before lower priority
		537	coroutines. Priorities are small signed integers (currently -4 .. +3),
		538	that you can refer to using PRIO_xxx constants (use the import tag :prio
		539	to get then):
		540
		541	PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN
		542	3 > 1 > 0 > -1 > -3 > -4
		543
		544	# set priority to HIGH
		545	current->prio(PRIO_HIGH);
		546
		547	The idle coroutine ($Coro::idle) always has a lower priority than any
		548	existing coroutine.
		549
		550	Changing the priority of the current coroutine will take effect immediately,
		551	but changing the priority of coroutines in the ready queue (but not
		552	running) will only take effect after the next schedule (of that
		553	coroutine). This is a bug that will be fixed in some future version.
		554
		555	=item $newprio = $coroutine->nice ($change)
		556
		557	Similar to C<prio>, but subtract the given value from the priority (i.e.
		558	higher values mean lower priority, just as in unix).
		559
		560	=item $olddesc = $coroutine->desc ($newdesc)
		561
		562	Sets (or gets in case the argument is missing) the description for this
		563	coroutine. This is just a free-form string you can associate with a
		564	coroutine.
		565
		566	This method simply sets the C<< $coroutine->{desc} >> member to the given
		567	string. You can modify this member directly if you wish.
		568
		569	=cut
		570
		571	sub desc {
		572	my $old = $_[0]{desc};
		573	$_[0]{desc} = $_[1] if @_ > 1;
		574	$old;
		575	}
		576
		577	sub transfer {
		578	require Carp;
		579	Carp::croak ("You must not call ->transfer on Coro objects. Use Coro::State objects or the ->schedule_to method. Caught");
		580	}
		581
		582	=back
		583
		584	=head1 GLOBAL FUNCTIONS
		585
		586	=over 4
		587
		588	=item Coro::nready
		589
		590	Returns the number of coroutines that are currently in the ready state,
		591	i.e. that can be switched to by calling C<schedule> directory or
		592	indirectly. The value C<0> means that the only runnable coroutine is the
		593	currently running one, so C<cede> would have no effect, and C<schedule>
		594	would cause a deadlock unless there is an idle handler that wakes up some
		595	coroutines.
		596
		597	=item my $guard = Coro::guard { ... }
		598
		599	This function still exists, but is deprecated. Please use the
		600	C<Guard::guard> function instead.
		601
		602	=cut
		603
		604	BEGIN { *guard = \&Guard::guard }
		605
		606	=item unblock_sub { ... }
		607
		608	This utility function takes a BLOCK or code reference and "unblocks" it,
		609	returning a new coderef. Unblocking means that calling the new coderef
		610	will return immediately without blocking, returning nothing, while the
		611	original code ref will be called (with parameters) from within another
		612	coroutine.
		613
		614	The reason this function exists is that many event libraries (such as the
		615	venerable L<Event\|Event> module) are not coroutine-safe (a weaker form
		616	of reentrancy). This means you must not block within event callbacks,
		617	otherwise you might suffer from crashes or worse. The only event library
		618	currently known that is safe to use without C<unblock_sub> is L<EV>.
		619
		620	This function allows your callbacks to block by executing them in another
		621	coroutine where it is safe to block. One example where blocking is handy
		622	is when you use the L<Coro::AIO\|Coro::AIO> functions to save results to
		623	disk, for example.
		624
		625	In short: simply use C<unblock_sub { ... }> instead of C<sub { ... }> when
		626	creating event callbacks that want to block.
		627
		628	If your handler does not plan to block (e.g. simply sends a message to
		629	another coroutine, or puts some other coroutine into the ready queue),
		630	there is no reason to use C<unblock_sub>.
		631
		632	Note that you also need to use C<unblock_sub> for any other callbacks that
		633	are indirectly executed by any C-based event loop. For example, when you
		634	use a module that uses L<AnyEvent> (and you use L<Coro::AnyEvent>) and it
		635	provides callbacks that are the result of some event callback, then you
		636	must not block either, or use C<unblock_sub>.
		637
		638	=cut
		639
		640	our @unblock_queue;
		641
		642	# we create a special coro because we want to cede,
		643	# to reduce pressure on the coro pool (because most callbacks
		644	# return immediately and can be reused) and because we cannot cede
		645	# inside an event callback.
		646	our $unblock_scheduler = new Coro sub {
		647	while () {
		648	while (my $cb = pop @unblock_queue) {
		649	&async_pool (@$cb);
		650
		651	# for short-lived callbacks, this reduces pressure on the coro pool
		652	# as the chance is very high that the async_poll coro will be back
		653	# in the idle state when cede returns
		654	cede;
94	}	655	}
95	} while (1);	656	schedule; # sleep well
96	}, $class;	657	}
97	}	658	};
		659	$unblock_scheduler->{desc} = "[unblock_sub scheduler]";
98		660
99	=item $coro->resume	661	sub unblock_sub(&) {
		662	my $cb = shift;
100		663
101	Resume execution at the given coroutine.	664	sub {
102		665	unshift @unblock_queue, [$cb, @_];
103	=cut	666	$unblock_scheduler->ready;
104		667	}
105	my $prev;
106
107	# I call the _transfer function from a pelr function
108	# because that way perl saves all important things on
109	# the stack.
110	sub resume {
111	$prev = $current; $current = $_[0];
112	_transfer($prev, $current);
113	}	668	}
		669
		670	=item $cb = Coro::rouse_cb
		671
		672	Create and return a "rouse callback". That's a code reference that,
		673	when called, will remember a copy of its arguments and notify the owner
		674	coroutine of the callback.
		675
		676	See the next function.
		677
		678	=item @args = Coro::rouse_wait [$cb]
		679
		680	Wait for the specified rouse callback (or the last one that was created in
		681	this coroutine).
		682
		683	As soon as the callback is invoked (or when the callback was invoked
		684	before C<rouse_wait>), it will return the arguments originally passed to
		685	the rouse callback.
		686
		687	See the section B<HOW TO WAIT FOR A CALLBACK> for an actual usage example.
		688
		689	=back
		690
		691	=cut
114		692
115	1;	693	1;
116		694
		695	=head1 HOW TO WAIT FOR A CALLBACK
		696
		697	It is very common for a coroutine to wait for some callback to be
		698	called. This occurs naturally when you use coroutines in an otherwise
		699	event-based program, or when you use event-based libraries.
		700
		701	These typically register a callback for some event, and call that callback
		702	when the event occured. In a coroutine, however, you typically want to
		703	just wait for the event, simplyifying things.
		704
		705	For example C<< AnyEvent->child >> registers a callback to be called when
		706	a specific child has exited:
		707
		708	my $child_watcher = AnyEvent->child (pid => $pid, cb => sub { ... });
		709
		710	But from withina coroutine, you often just want to write this:
		711
		712	my $status = wait_for_child $pid;
		713
		714	Coro offers two functions specifically designed to make this easy,
		715	C<Coro::rouse_cb> and C<Coro::rouse_wait>.
		716
		717	The first function, C<rouse_cb>, generates and returns a callback that,
		718	when invoked, will save its arguments and notify the coroutine that
		719	created the callback.
		720
		721	The second function, C<rouse_wait>, waits for the callback to be called
		722	(by calling C<schedule> to go to sleep) and returns the arguments
		723	originally passed to the callback.
		724
		725	Using these functions, it becomes easy to write the C<wait_for_child>
		726	function mentioned above:
		727
		728	sub wait_for_child($) {
		729	my ($pid) = @_;
		730
		731	my $watcher = AnyEvent->child (pid => $pid, cb => Coro::rouse_cb);
		732
		733	my ($rpid, $rstatus) = Coro::rouse_wait;
		734	$rstatus
		735	}
		736
		737	In the case where C<rouse_cb> and C<rouse_wait> are not flexible enough,
		738	you can roll your own, using C<schedule>:
		739
		740	sub wait_for_child($) {
		741	my ($pid) = @_;
		742
		743	# store the current coroutine in $current,
		744	# and provide result variables for the closure passed to ->child
		745	my $current = $Coro::current;
		746	my ($done, $rstatus);
		747
		748	# pass a closure to ->child
		749	my $watcher = AnyEvent->child (pid => $pid, cb => sub {
		750	$rstatus = $_[1]; # remember rstatus
		751	$done = 1; # mark $rstatus as valud
		752	});
		753
		754	# wait until the closure has been called
		755	schedule while !$done;
		756
		757	$rstatus
		758	}
		759
		760
		761	=head1 BUGS/LIMITATIONS
		762
		763	=over 4
		764
		765	=item fork with pthread backend
		766
		767	When Coro is compiled using the pthread backend (which isn't recommended
		768	but required on many BSDs as their libcs are completely broken), then
		769	coroutines will not survive a fork. There is no known workaround except to
		770	fix your libc and use a saner backend.
		771
		772	=item perl process emulation ("threads")
		773
		774	This module is not perl-pseudo-thread-safe. You should only ever use this
		775	module from the first thread (this requirement might be removed in the
		776	future to allow per-thread schedulers, but Coro::State does not yet allow
		777	this). I recommend disabling thread support and using processes, as having
		778	the windows process emulation enabled under unix roughly halves perl
		779	performance, even when not used.
		780
		781	=item coroutine switching not signal safe
		782
		783	You must not switch to another coroutine from within a signal handler
		784	(only relevant with %SIG - most event libraries provide safe signals).
		785
		786	That means you I<MUST NOT> call any function that might "block" the
		787	current coroutine - C<cede>, C<schedule> C<< Coro::Semaphore->down >> or
		788	anything that calls those. Everything else, including calling C<ready>,
		789	works.
		790
117	=back	791	=back
118		792
119	=head1 BUGS
120
121	This module has not yet been extensively tested.
122		793
123	=head1 SEE ALSO	794	=head1 SEE ALSO
124		795
125	L<Coro::Process>, L<Coro::Signal>.	796	Event-Loop integration: L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>.
		797
		798	Debugging: L<Coro::Debug>.
		799
		800	Support/Utility: L<Coro::Specific>, L<Coro::Util>.
		801
		802	Locking and IPC: L<Coro::Signal>, L<Coro::Channel>, L<Coro::Semaphore>,
		803	L<Coro::SemaphoreSet>, L<Coro::RWLock>.
		804
		805	I/O and Timers: L<Coro::Timer>, L<Coro::Handle>, L<Coro::Socket>, L<Coro::AIO>.
		806
		807	Compatibility with other modules: L<Coro::LWP> (but see also L<AnyEvent::HTTP> for
		808	a better-working alternative), L<Coro::BDB>, L<Coro::Storable>,
		809	L<Coro::Select>.
		810
		811	XS API: L<Coro::MakeMaker>.
		812
		813	Low level Configuration, Thread Environment, Continuations: L<Coro::State>.
126		814
127	=head1 AUTHOR	815	=head1 AUTHOR
128		816
129	Marc Lehmann <pcg@goof.com>	817	Marc Lehmann <schmorp@schmorp.de>
130	http://www.goof.com/pcg/marc/	818	http://home.schmorp.de/
131		819
132	=cut	820	=cut
133		821

Diff Legend

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

Comparing Coro/Coro.pm (file contents): Revision 1.5 by root, Tue Jul 10 01:43:21 2001 UTC vs. Revision 1.247 by root, Mon Dec 15 02:07:11 2008 UTC

Diff Legend

Comparing Coro/Coro.pm (file contents):
Revision 1.5 by root, Tue Jul 10 01:43:21 2001 UTC vs.
Revision 1.247 by root, Mon Dec 15 02:07:11 2008 UTC