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

Comparing cvsroot/Coro/Coro.pm (file contents):
Revision 1.23 by root, Mon Jul 23 04:23:32 2001 UTC vs.
Revision 1.245 by root, Sun Dec 14 19:52:58 2008 UTC

1	=head1 NAME	1	=head1 NAME
2		2
3	Coro - coroutine process abstraction	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	async {	9	async {
10	# some asynchronous thread of execution	10	# some asynchronous thread of execution
		11	print "2\n";
		12	cede; # yield back to main
		13	print "4\n";
11	};	14	};
12		15	print "1\n";
13	# alternatively create an async process like this:	16	cede; # yield to coroutine
14		17	print "3\n";
15	sub some_func : Coro {	18	cede; # and again
16	# some more async code	19
17	}	20	# use locking
18		21	use Coro::Semaphore;
19	cede;	22	my $lock = new Coro::Semaphore;
		23	my $locked;
		24
		25	$lock->down;
		26	$locked = 1;
		27	$lock->up;
20		28
21	=head1 DESCRIPTION	29	=head1 DESCRIPTION
22		30
23	This module collection manages coroutines. Coroutines are similar to	31	For a tutorial-style introduction, please read the L<Coro::Intro>
24	Threads but don't run in parallel.	32	manpage. This manpage mainly contains reference information.
25		33
26	This module is still experimental, see the BUGS section below.	34	This module collection manages continuations in general, most often
		35	in the form of cooperative threads (also called coroutines in the
		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.
27		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
28	In this module, coroutines are defined as "callchain + lexical variables	56	In this module, a thread is defined as "callchain + lexical variables +
29	+ @_ + $_ + $@ + $^W + C stack), that is, a coroutine has it's own	57	@_ + $_ + $@ + $/ + C stack), that is, a thread has its own callchain,
30	callchain, it's own set of lexicals and it's own set of perl's most	58	its own set of lexicals and its own set of perls most important global
31	important global variables.	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.
32		63
33	=cut	64	=cut
34		65
35	package Coro;	66	package Coro;
36		67
		68	use strict qw(vars subs);
		69	no warnings "uninitialized";
		70
37	use Coro::State;	71	use Coro::State;
38		72
39	use base Exporter;	73	use base qw(Coro::State Exporter);
40		74
		75	our $idle; # idle handler
		76	our $main; # main coroutine
		77	our $current; # current coroutine
		78
41	$VERSION = 0.10;	79	our $VERSION = 5.13;
42		80
43	@EXPORT = qw(async cede schedule terminate current);	81	our @EXPORT = qw(async async_pool cede schedule terminate current unblock_sub);
44	@EXPORT_OK = qw($current);	82	our %EXPORT_TAGS = (
		83	prio => [qw(PRIO_MAX PRIO_HIGH PRIO_NORMAL PRIO_LOW PRIO_IDLE PRIO_MIN)],
		84	);
		85	our @EXPORT_OK = (@{$EXPORT_TAGS{prio}}, qw(nready));
45		86
46	{	87	=head1 GLOBAL VARIABLES
47	my @async;
48		88
49	# this way of handling attributes simply is NOT scalable ;()	89	=over 4
50	sub import {
51	Coro->export_to_level(1, @_);
52	my $old = *{(caller)[0]."::MODIFY_CODE_ATTRIBUTES"}{CODE};
53	*{(caller)[0]."::MODIFY_CODE_ATTRIBUTES"} = sub {
54	my ($package, $ref) = (shift, shift);
55	my @attrs;
56	for (@_) {
57	if ($_ eq "Coro") {
58	push @async, $ref;
59	} else {
60	push @attrs, $_;
61	}
62	}
63	return $old ? $old->($package, $ref, @attrs) : @attrs;
64	};
65	}
66		90
67	sub INIT {
68	&async(pop @async) while @async;
69	}
70	}
71
72	=item $main	91	=item $Coro::main
73		92
74	This coroutine represents the main program.	93	This variable stores the coroutine object that represents the main
		94	program. While you cna C<ready> it and do most other things you can do to
		95	coroutines, it is mainly useful to compare again C<$Coro::current>, to see
		96	whether you are running in the main program or not.
75		97
76	=cut	98	=cut
77		99
78	our $main = new Coro;	100	# $main is now being initialised by Coro::State
79		101
80	=item $current (or as function: current)	102	=item $Coro::current
81		103
82	The current coroutine (the last coroutine switched to). The initial value is C<$main> (of course).	104	The coroutine object representing the current coroutine (the last
		105	coroutine that the Coro scheduler switched to). The initial value is
		106	C<$Coro::main> (of course).
83		107
84	=cut	108	This variable is B<strictly> I<read-only>. You can take copies of the
		109	value stored in it and use it as any other coroutine object, but you must
		110	not otherwise modify the variable itself.
85		111
86	# maybe some other module used Coro::Specific before...	112	=cut
87	if ($current) {
88	$main->{specific} = $current->{specific};
89	}
90		113
91	our $current = $main;
92
93	sub current() { $current }	114	sub current() { $current } # [DEPRECATED]
94		115
95	=item $idle	116	=item $Coro::idle
96		117
97	The coroutine to switch to when no other coroutine is running. The default	118	This variable is mainly useful to integrate Coro into event loops. It is
98	implementation prints "FATAL: deadlock detected" and exits.	119	usually better to rely on L<Coro::AnyEvent> or L<Coro::EV>, as this is
		120	pretty low-level functionality.
99		121
100	=cut	122	This variable stores either a coroutine or a callback.
101		123
102	# should be done using priorities :(	124	If it is a callback, the it is called whenever the scheduler finds no
103	our $idle = new Coro sub {	125	ready coroutines to run. The default implementation prints "FATAL:
104	print STDERR "FATAL: deadlock detected\n";	126	deadlock detected" and exits, because the program has no other way to
105	exit(51);	127	continue.
		128
		129	If it is a coroutine object, then this object will be readied (without
		130	invoking any ready hooks, however) when the scheduler finds no other ready
		131	coroutines to run.
		132
		133	This hook is overwritten by modules such as C<Coro::EV> and
		134	C<Coro::AnyEvent> to wait on an external event that hopefully wake up a
		135	coroutine so the scheduler can run it.
		136
		137	Note that the callback I<must not>, under any circumstances, block
		138	the current coroutine. Normally, this is achieved by having an "idle
		139	coroutine" that calls the event loop and then blocks again, and then
		140	readying that coroutine in the idle handler, or by simply placing the idle
		141	coroutine in this variable.
		142
		143	See L<Coro::Event> or L<Coro::AnyEvent> for examples of using this
		144	technique.
		145
		146	Please note that if your callback recursively invokes perl (e.g. for event
		147	handlers), then it must be prepared to be called recursively itself.
		148
		149	=cut
		150
		151	$idle = sub {
		152	require Carp;
		153	Carp::croak ("FATAL: deadlock detected");
106	};	154	};
107		155
108	# we really need priorities...	156	# this coroutine is necessary because a coroutine
109	my @ready; # the ready queue. hehe, rather broken ;)	157	# cannot destroy itself.
		158	our @destroy;
		159	our $manager;
110		160
111	# static methods. not really.	161	$manager = new Coro sub {
		162	while () {
		163	Coro::_cancel shift @destroy
		164	while @destroy;
112		165
113	=head2 STATIC METHODS	166	&schedule;
		167	}
		168	};
		169	$manager->{desc} = "[coro manager]";
		170	$manager->prio (PRIO_MAX);
114		171
115	Static methods are actually functions that operate on the current process only.	172	=back
		173
		174	=head1 SIMPLE COROUTINE CREATION
116		175
117	=over 4	176	=over 4
118		177
119	=item async { ... } [@args...]	178	=item async { ... } [@args...]
120		179
121	Create a new asynchronous process and return it's process object	180	Create a new coroutine and return its coroutine object (usually
122	(usually unused). When the sub returns the new process is automatically	181	unused). The coroutine will be put into the ready queue, so
		182	it will start running automatically on the next scheduler run.
		183
		184	The first argument is a codeblock/closure that should be executed in the
		185	coroutine. When it returns argument returns the coroutine is automatically
123	terminated.	186	terminated.
124		187
		188	The remaining arguments are passed as arguments to the closure.
		189
		190	See the C<Coro::State::new> constructor for info about the coroutine
		191	environment in which coroutines are executed.
		192
		193	Calling C<exit> in a coroutine will do the same as calling exit outside
		194	the coroutine. Likewise, when the coroutine dies, the program will exit,
		195	just as it would in the main program.
		196
		197	If you do not want that, you can provide a default C<die> handler, or
		198	simply avoid dieing (by use of C<eval>).
		199
125	# create a new coroutine that just prints its arguments	200	Example: Create a new coroutine that just prints its arguments.
		201
126	async {	202	async {
127	print "@_\n";	203	print "@_\n";
128	} 1,2,3,4;	204	} 1,2,3,4;
129		205
130	The coderef you submit MUST NOT be a closure that refers to variables
131	in an outer scope. This does NOT work. Pass arguments into it instead.
132
133	=cut	206	=cut
134		207
135	sub async(&@) {	208	sub async(&@) {
136	my $pid = new Coro @_;	209	my $coro = new Coro @_;
137	$pid->ready;	210	$coro->ready;
138	$pid;	211	$coro
139	}	212	}
		213
		214	=item async_pool { ... } [@args...]
		215
		216	Similar to C<async>, but uses a coroutine pool, so you should not call
		217	terminate or join on it (although you are allowed to), and you get a
		218	coroutine that might have executed other code already (which can be good
		219	or bad :).
		220
		221	On the plus side, this function is about twice as fast as creating (and
		222	destroying) a completely new coroutine, so if you need a lot of generic
		223	coroutines in quick successsion, use C<async_pool>, not C<async>.
		224
		225	The code block is executed in an C<eval> context and a warning will be
		226	issued in case of an exception instead of terminating the program, as
		227	C<async> does. As the coroutine is being reused, stuff like C<on_destroy>
		228	will not work in the expected way, unless you call terminate or cancel,
		229	which somehow defeats the purpose of pooling (but is fine in the
		230	exceptional case).
		231
		232	The priority will be reset to C<0> after each run, tracing will be
		233	disabled, the description will be reset and the default output filehandle
		234	gets restored, so you can change all these. Otherwise the coroutine will
		235	be re-used "as-is": most notably if you change other per-coroutine global
		236	stuff such as C<$/> you I<must needs> revert that change, which is most
		237	simply done by using local as in: C<< local $/ >>.
		238
		239	The idle pool size is limited to C<8> idle coroutines (this can be
		240	adjusted by changing $Coro::POOL_SIZE), but there can be as many non-idle
		241	coros as required.
		242
		243	If you are concerned about pooled coroutines growing a lot because a
		244	single C<async_pool> used a lot of stackspace you can e.g. C<async_pool
		245	{ terminate }> once per second or so to slowly replenish the pool. In
		246	addition to that, when the stacks used by a handler grows larger than 32kb
		247	(adjustable via $Coro::POOL_RSS) it will also be destroyed.
		248
		249	=cut
		250
		251	our $POOL_SIZE = 8;
		252	our $POOL_RSS = 32 * 1024;
		253	our @async_pool;
		254
		255	sub pool_handler {
		256	while () {
		257	eval {
		258	&{&_pool_handler} while 1;
		259	};
		260
		261	warn $@ if $@;
		262	}
		263	}
		264
		265	=back
		266
		267	=head1 STATIC METHODS
		268
		269	Static methods are actually functions that implicitly operate on the
		270	current coroutine.
		271
		272	=over 4
140		273
141	=item schedule	274	=item schedule
142		275
143	Calls the scheduler. Please note that the current process will not be put	276	Calls the scheduler. The scheduler will find the next coroutine that is
		277	to be run from the ready queue and switches to it. The next coroutine
		278	to be run is simply the one with the highest priority that is longest
		279	in its ready queue. If there is no coroutine ready, it will clal the
		280	C<$Coro::idle> hook.
		281
		282	Please note that the current coroutine will I<not> be put into the ready
144	into the ready queue, so calling this function usually means you will	283	queue, so calling this function usually means you will never be called
145	never be called again.	284	again unless something else (e.g. an event handler) calls C<< ->ready >>,
		285	thus waking you up.
146		286
147	=cut	287	This makes C<schedule> I<the> generic method to use to block the current
		288	coroutine and wait for events: first you remember the current coroutine in
		289	a variable, then arrange for some callback of yours to call C<< ->ready
		290	>> on that once some event happens, and last you call C<schedule> to put
		291	yourself to sleep. Note that a lot of things can wake your coroutine up,
		292	so you need to check whether the event indeed happened, e.g. by storing the
		293	status in a variable.
148		294
149	my $prev;	295	See B<HOW TO WAIT FOR A CALLBACK>, below, for some ways to wait for callbacks.
150
151	sub schedule {
152	# should be done using priorities :(
153	($prev, $current) = ($current, shift @ready \|\| $idle);
154	Coro::State::transfer($prev, $current);
155	}
156		296
157	=item cede	297	=item cede
158		298
159	"Cede" to other processes. This function puts the current process into the	299	"Cede" to other coroutines. This function puts the current coroutine into
160	ready queue and calls C<schedule>, which has the effect of giving up the	300	the ready queue and calls C<schedule>, which has the effect of giving
161	current "timeslice" to other coroutines of the same or higher priority.	301	up the current "timeslice" to other coroutines of the same or higher
		302	priority. Once your coroutine gets its turn again it will automatically be
		303	resumed.
162		304
163	=cut	305	This function is often called C<yield> in other languages.
164		306
165	sub cede {	307	=item Coro::cede_notself
		308
		309	Works like cede, but is not exported by default and will cede to I<any>
		310	coroutine, regardless of priority. This is useful sometimes to ensure
		311	progress is made.
		312
		313	=item terminate [arg...]
		314
		315	Terminates the current coroutine with the given status values (see L<cancel>).
		316
		317	=item killall
		318
		319	Kills/terminates/cancels all coroutines except the currently running
		320	one. This can be useful after a fork, either in the child or the parent,
		321	as usually only one of them should inherit the running coroutines.
		322
		323	Note that in the implementation, destructors run as normal, making this
		324	function not so useful after a fork. Future versions of this function
		325	might try to free resources without running any code.
		326
		327	Note that while this will try to free some of the main programs resources,
		328	you cannot free all of them, so if a coroutine that is not the main
		329	program calls this function, there will be some one-time resource leak.
		330
		331	=cut
		332
		333	sub killall {
		334	for (Coro::State::list) {
		335	$_->cancel
		336	if $_ != $current && UNIVERSAL::isa $_, "Coro";
		337	}
		338	}
		339
		340	=back
		341
		342	=head1 COROUTINE OBJECT METHODS
		343
		344	These are the methods you can call on coroutine objects (or to create
		345	them).
		346
		347	=over 4
		348
		349	=item new Coro \&sub [, @args...]
		350
		351	Create a new coroutine and return it. When the sub returns, the coroutine
		352	automatically terminates as if C<terminate> with the returned values were
		353	called. To make the coroutine run you must first put it into the ready
		354	queue by calling the ready method.
		355
		356	See C<async> and C<Coro::State::new> for additional info about the
		357	coroutine environment.
		358
		359	=cut
		360
		361	sub _coro_run {
		362	terminate &{+shift};
		363	}
		364
		365	=item $success = $coroutine->ready
		366
		367	Put the given coroutine into the end of its ready queue (there is one
		368	queue for each priority) and return true. If the coroutine is already in
		369	the ready queue, do nothing and return false.
		370
		371	This ensures that the scheduler will resume this coroutine automatically
		372	once all the coroutines of higher priority and all coroutines of the same
		373	priority that were put into the ready queue earlier have been resumed.
		374
		375	=item $is_ready = $coroutine->is_ready
		376
		377	Return whether the coroutine is currently the ready queue or not,
		378
		379	=item $coroutine->cancel (arg...)
		380
		381	Terminates the given coroutine and makes it return the given arguments as
		382	status (default: the empty list). Never returns if the coroutine is the
		383	current coroutine.
		384
		385	=cut
		386
		387	sub cancel {
		388	my $self = shift;
		389
		390	if ($current == $self) {
		391	terminate @_;
		392	} else {
		393	$self->{_status} = [@_];
		394	$self->_cancel;
		395	}
		396	}
		397
		398	=item $coroutine->schedule_to
		399
		400	Puts the current coroutine to sleep (like C<Coro::schedule>), but instead
		401	of continuing with the next coro from the ready queue, always switch to
		402	the given coroutine object (regardless of priority etc.). The readyness
		403	state of that coroutine isn't changed.
		404
		405	This is an advanced method for special cases - I'd love to hear about any
		406	uses for this one.
		407
		408	=item $coroutine->cede_to
		409
		410	Like C<schedule_to>, but puts the current coroutine into the ready
		411	queue. This has the effect of temporarily switching to the given
		412	coroutine, and continuing some time later.
		413
		414	This is an advanced method for special cases - I'd love to hear about any
		415	uses for this one.
		416
		417	=item $coroutine->throw ([$scalar])
		418
		419	If C<$throw> is specified and defined, it will be thrown as an exception
		420	inside the coroutine at the next convenient point in time. Otherwise
		421	clears the exception object.
		422
		423	Coro will check for the exception each time a schedule-like-function
		424	returns, i.e. after each C<schedule>, C<cede>, C<< Coro::Semaphore->down
		425	>>, C<< Coro::Handle->readable >> and so on. Most of these functions
		426	detect this case and return early in case an exception is pending.
		427
		428	The exception object will be thrown "as is" with the specified scalar in
		429	C<$@>, i.e. if it is a string, no line number or newline will be appended
		430	(unlike with C<die>).
		431
		432	This can be used as a softer means than C<cancel> to ask a coroutine to
		433	end itself, although there is no guarantee that the exception will lead to
		434	termination, and if the exception isn't caught it might well end the whole
		435	program.
		436
		437	You might also think of C<throw> as being the moral equivalent of
		438	C<kill>ing a coroutine with a signal (in this case, a scalar).
		439
		440	=item $coroutine->join
		441
		442	Wait until the coroutine terminates and return any values given to the
		443	C<terminate> or C<cancel> functions. C<join> can be called concurrently
		444	from multiple coroutines, and all will be resumed and given the status
		445	return once the C<$coroutine> terminates.
		446
		447	=cut
		448
		449	sub join {
		450	my $self = shift;
		451
		452	unless ($self->{_status}) {
		453	my $current = $current;
		454
		455	push @{$self->{_on_destroy}}, sub {
166	$current->ready;	456	$current->ready;
167	&schedule;	457	undef $current;
168	}	458	};
169		459
170	=item terminate	460	&schedule while $current;
		461	}
171		462
172	Terminates the current process.	463	wantarray ? @{$self->{_status}} : $self->{_status}[0];
		464	}
173		465
174	Future versions of this function will allow result arguments.	466	=item $coroutine->on_destroy (\&cb)
175		467
176	=cut	468	Registers a callback that is called when this coroutine gets destroyed,
		469	but before it is joined. The callback gets passed the terminate arguments,
		470	if any, and I<must not> die, under any circumstances.
177		471
178	# this coroutine is necessary because a coroutine	472	=cut
179	# cannot destroy itself.	473
180	my @destroy;	474	sub on_destroy {
181	my $terminate = new Coro sub {	475	my ($self, $cb) = @_;
		476
		477	push @{ $self->{_on_destroy} }, $cb;
		478	}
		479
		480	=item $oldprio = $coroutine->prio ($newprio)
		481
		482	Sets (or gets, if the argument is missing) the priority of the
		483	coroutine. Higher priority coroutines get run before lower priority
		484	coroutines. Priorities are small signed integers (currently -4 .. +3),
		485	that you can refer to using PRIO_xxx constants (use the import tag :prio
		486	to get then):
		487
		488	PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN
		489	3 > 1 > 0 > -1 > -3 > -4
		490
		491	# set priority to HIGH
		492	current->prio(PRIO_HIGH);
		493
		494	The idle coroutine ($Coro::idle) always has a lower priority than any
		495	existing coroutine.
		496
		497	Changing the priority of the current coroutine will take effect immediately,
		498	but changing the priority of coroutines in the ready queue (but not
		499	running) will only take effect after the next schedule (of that
		500	coroutine). This is a bug that will be fixed in some future version.
		501
		502	=item $newprio = $coroutine->nice ($change)
		503
		504	Similar to C<prio>, but subtract the given value from the priority (i.e.
		505	higher values mean lower priority, just as in unix).
		506
		507	=item $olddesc = $coroutine->desc ($newdesc)
		508
		509	Sets (or gets in case the argument is missing) the description for this
		510	coroutine. This is just a free-form string you can associate with a
		511	coroutine.
		512
		513	This method simply sets the C<< $coroutine->{desc} >> member to the given
		514	string. You can modify this member directly if you wish.
		515
		516	=cut
		517
		518	sub desc {
		519	my $old = $_[0]{desc};
		520	$_[0]{desc} = $_[1] if @_ > 1;
		521	$old;
		522	}
		523
		524	sub transfer {
		525	require Carp;
		526	Carp::croak ("You must not call ->transfer on Coro objects. Use Coro::State objects or the ->schedule_to method. Caught");
		527	}
		528
		529	=back
		530
		531	=head1 GLOBAL FUNCTIONS
		532
		533	=over 4
		534
		535	=item Coro::nready
		536
		537	Returns the number of coroutines that are currently in the ready state,
		538	i.e. that can be switched to by calling C<schedule> directory or
		539	indirectly. The value C<0> means that the only runnable coroutine is the
		540	currently running one, so C<cede> would have no effect, and C<schedule>
		541	would cause a deadlock unless there is an idle handler that wakes up some
		542	coroutines.
		543
		544	=item my $guard = Coro::guard { ... }
		545
		546	This function still exists, but is deprecated. Please use the
		547	C<Guard::guard> function instead.
		548
		549	=cut
		550
		551	BEGIN { *guard = \&Guard::guard }
		552
		553	=item unblock_sub { ... }
		554
		555	This utility function takes a BLOCK or code reference and "unblocks" it,
		556	returning a new coderef. Unblocking means that calling the new coderef
		557	will return immediately without blocking, returning nothing, while the
		558	original code ref will be called (with parameters) from within another
		559	coroutine.
		560
		561	The reason this function exists is that many event libraries (such as the
		562	venerable L<Event\|Event> module) are not coroutine-safe (a weaker form
		563	of reentrancy). This means you must not block within event callbacks,
		564	otherwise you might suffer from crashes or worse. The only event library
		565	currently known that is safe to use without C<unblock_sub> is L<EV>.
		566
		567	This function allows your callbacks to block by executing them in another
		568	coroutine where it is safe to block. One example where blocking is handy
		569	is when you use the L<Coro::AIO\|Coro::AIO> functions to save results to
		570	disk, for example.
		571
		572	In short: simply use C<unblock_sub { ... }> instead of C<sub { ... }> when
		573	creating event callbacks that want to block.
		574
		575	If your handler does not plan to block (e.g. simply sends a message to
		576	another coroutine, or puts some other coroutine into the ready queue),
		577	there is no reason to use C<unblock_sub>.
		578
		579	Note that you also need to use C<unblock_sub> for any other callbacks that
		580	are indirectly executed by any C-based event loop. For example, when you
		581	use a module that uses L<AnyEvent> (and you use L<Coro::AnyEvent>) and it
		582	provides callbacks that are the result of some event callback, then you
		583	must not block either, or use C<unblock_sub>.
		584
		585	=cut
		586
		587	our @unblock_queue;
		588
		589	# we create a special coro because we want to cede,
		590	# to reduce pressure on the coro pool (because most callbacks
		591	# return immediately and can be reused) and because we cannot cede
		592	# inside an event callback.
		593	our $unblock_scheduler = new Coro sub {
182	while() {	594	while () {
183	delete ((pop @destroy)->{_coro_state}) while @destroy;	595	while (my $cb = pop @unblock_queue) {
184	&schedule;	596	&async_pool (@$cb);
		597
		598	# for short-lived callbacks, this reduces pressure on the coro pool
		599	# as the chance is very high that the async_poll coro will be back
		600	# in the idle state when cede returns
		601	cede;
		602	}
		603	schedule; # sleep well
185	}	604	}
186	};	605	};
		606	$unblock_scheduler->{desc} = "[unblock_sub scheduler]";
187		607
188	sub terminate {	608	sub unblock_sub(&) {
189	push @destroy, $current;	609	my $cb = shift;
190	$terminate->ready;	610
191	&schedule;	611	sub {
192	# NORETURN	612	unshift @unblock_queue, [$cb, @_];
		613	$unblock_scheduler->ready;
		614	}
193	}	615	}
		616
		617	=item $cb = Coro::rouse_cb
		618
		619	Create and return a "rouse callback". That's a code reference that,
		620	when called, will remember a copy of its arguments and notify the owner
		621	coroutine of the callback.
		622
		623	See the next function.
		624
		625	=item @args = Coro::rouse_wait [$cb]
		626
		627	Wait for the specified rouse callback (or the last one that was created in
		628	this coroutine).
		629
		630	As soon as the callback is invoked (or when the callback was invoked
		631	before C<rouse_wait>), it will return the arguments originally passed to
		632	the rouse callback.
		633
		634	See the section B<HOW TO WAIT FOR A CALLBACK> for an actual usage example.
194		635
195	=back	636	=back
196		637
197	# dynamic methods	638	=cut
198		639
199	=head2 PROCESS METHODS	640	1;
200		641
201	These are the methods you can call on process objects.	642	=head1 HOW TO WAIT FOR A CALLBACK
		643
		644	It is very common for a coroutine to wait for some callback to be
		645	called. This occurs naturally when you use coroutines in an otherwise
		646	event-based program, or when you use event-based libraries.
		647
		648	These typically register a callback for some event, and call that callback
		649	when the event occured. In a coroutine, however, you typically want to
		650	just wait for the event, simplyifying things.
		651
		652	For example C<< AnyEvent->child >> registers a callback to be called when
		653	a specific child has exited:
		654
		655	my $child_watcher = AnyEvent->child (pid => $pid, cb => sub { ... });
		656
		657	But from withina coroutine, you often just want to write this:
		658
		659	my $status = wait_for_child $pid;
		660
		661	Coro offers two functions specifically designed to make this easy,
		662	C<Coro::rouse_cb> and C<Coro::rouse_wait>.
		663
		664	The first function, C<rouse_cb>, generates and returns a callback that,
		665	when invoked, will save its arguments and notify the coroutine that
		666	created the callback.
		667
		668	The second function, C<rouse_wait>, waits for the callback to be called
		669	(by calling C<schedule> to go to sleep) and returns the arguments
		670	originally passed to the callback.
		671
		672	Using these functions, it becomes easy to write the C<wait_for_child>
		673	function mentioned above:
		674
		675	sub wait_for_child($) {
		676	my ($pid) = @_;
		677
		678	my $watcher = AnyEvent->child (pid => $pid, cb => Coro::rouse_cb);
		679
		680	my ($rpid, $rstatus) = Coro::rouse_wait;
		681	$rstatus
		682	}
		683
		684	In the case where C<rouse_cb> and C<rouse_wait> are not flexible enough,
		685	you can roll your own, using C<schedule>:
		686
		687	sub wait_for_child($) {
		688	my ($pid) = @_;
		689
		690	# store the current coroutine in $current,
		691	# and provide result variables for the closure passed to ->child
		692	my $current = $Coro::current;
		693	my ($done, $rstatus);
		694
		695	# pass a closure to ->child
		696	my $watcher = AnyEvent->child (pid => $pid, cb => sub {
		697	$rstatus = $_[1]; # remember rstatus
		698	$done = 1; # mark $rstatus as valud
		699	});
		700
		701	# wait until the closure has been called
		702	schedule while !$done;
		703
		704	$rstatus
		705	}
		706
		707
		708	=head1 BUGS/LIMITATIONS
202		709
203	=over 4	710	=over 4
204		711
205	=item new Coro \&sub [, @args...]	712	=item fork with pthread backend
206		713
207	Create a new process and return it. When the sub returns the process	714	When Coro is compiled using the pthread backend (which isn't recommended
208	automatically terminates. To start the process you must first put it into	715	but required on many BSDs as their libcs are completely broken), then
209	the ready queue by calling the ready method.	716	coroutines will not survive a fork. There is no known workaround except to
		717	fix your libc and use a saner backend.
210		718
211	The coderef you submit MUST NOT be a closure that refers to variables	719	=item perl process emulation ("threads")
212	in an outer scope. This does NOT work. Pass arguments into it instead.
213		720
214	=cut	721	This module is not perl-pseudo-thread-safe. You should only ever use this
		722	module from the first thread (this requirement might be removed in the
		723	future to allow per-thread schedulers, but Coro::State does not yet allow
		724	this). I recommend disabling thread support and using processes, as having
		725	the windows process emulation enabled under unix roughly halves perl
		726	performance, even when not used.
215		727
216	sub _newcoro {	728	=item coroutine switching not signal safe
217	terminate &{+shift};
218	}
219		729
220	sub new {	730	You must not switch to another coroutine from within a signal handler
221	my $class = shift;	731	(only relevant with %SIG - most event libraries provide safe signals).
222	bless {
223	_coro_state => (new Coro::State $_[0] && \&_newcoro, @_),
224	}, $class;
225	}
226		732
227	=item $process->ready	733	That means you I<MUST NOT> call any function that might "block" the
228		734	current coroutine - C<cede>, C<schedule> C<< Coro::Semaphore->down >> or
229	Put the current process into the ready queue.	735	anything that calls those. Everything else, including calling C<ready>,
230		736	works.
231	=cut
232
233	sub ready {
234	push @ready, $_[0];
235	}
236		737
237	=back	738	=back
238		739
239	=cut
240
241	1;
242
243	=head1 BUGS/LIMITATIONS
244
245	- could be faster, especially when the core would introduce special
246	support for coroutines (like it does for threads).
247	- there is still a memleak on coroutine termination that I could not
248	identify. Could be as small as a single SV.
249	- this module is not well-tested.
250	- if variables or arguments "disappear" (become undef) or become
251	corrupted please contact the author so he cen iron out the
252	remaining bugs.
253	- this module is not thread-safe. You must only ever use this module from
254	the same thread (this requirement might be loosened in the future to
255	allow per-thread schedulers, but Coro::State does not yet allow this).
256		740
257	=head1 SEE ALSO	741	=head1 SEE ALSO
258		742
259	L<Coro::Channel>, L<Coro::Cont>, L<Coro::Specific>, L<Coro::Semaphore>,	743	Event-Loop integration: L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>.
260	L<Coro::Signal>, L<Coro::State>, L<Coro::Event>.	744
		745	Debugging: L<Coro::Debug>.
		746
		747	Support/Utility: L<Coro::Specific>, L<Coro::Util>.
		748
		749	Locking and IPC: L<Coro::Signal>, L<Coro::Channel>, L<Coro::Semaphore>,
		750	L<Coro::SemaphoreSet>, L<Coro::RWLock>.
		751
		752	I/O and Timers: L<Coro::Timer>, L<Coro::Handle>, L<Coro::Socket>, L<Coro::AIO>.
		753
		754	Compatibility with other modules: L<Coro::LWP> (but see also L<AnyEvent::HTTP> for
		755	a better-working alternative), L<Coro::BDB>, L<Coro::Storable>,
		756	L<Coro::Select>.
		757
		758	XS API: L<Coro::MakeMaker>.
		759
		760	Low level Configuration, Thread Environment, Continuations: L<Coro::State>.
261		761
262	=head1 AUTHOR	762	=head1 AUTHOR
263		763
264	Marc Lehmann <pcg@goof.com>	764	Marc Lehmann <schmorp@schmorp.de>
265	http://www.goof.com/pcg/marc/	765	http://home.schmorp.de/
266		766
267	=cut	767	=cut
268		768

Diff Legend

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

Comparing cvsroot/Coro/Coro.pm (file contents): Revision 1.23 by root, Mon Jul 23 04:23:32 2001 UTC vs. Revision 1.245 by root, Sun Dec 14 19:52:58 2008 UTC

Diff Legend

Comparing cvsroot/Coro/Coro.pm (file contents):
Revision 1.23 by root, Mon Jul 23 04:23:32 2001 UTC vs.
Revision 1.245 by root, Sun Dec 14 19:52:58 2008 UTC