ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/Coro/Coro.pm

Comparing Coro/Coro.pm (file contents):
Revision 1.214 by root, Tue Nov 11 03:30:21 2008 UTC vs.
Revision 1.321 by root, Sun Feb 2 03:26:06 2014 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
…		…
11	print "2\n";	11	print "2\n";
12	cede; # yield back to main	12	cede; # yield back to main
13	print "4\n";	13	print "4\n";
14	};	14	};
15	print "1\n";	15	print "1\n";
16	cede; # yield to coroutine	16	cede; # yield to coro
17	print "3\n";	17	print "3\n";
18	cede; # and again	18	cede; # and again
19		19
20	# use locking	20	# use locking
21	use Coro::Semaphore;
22	my $lock = new Coro::Semaphore;	21	my $lock = new Coro::Semaphore;
23	my $locked;	22	my $locked;
24		23
25	$lock->down;	24	$lock->down;
26	$locked = 1;	25	$locked = 1;
27	$lock->up;	26	$lock->up;
28		27
29	=head1 DESCRIPTION	28	=head1 DESCRIPTION
30		29
31	This module collection manages coroutines. Coroutines are similar to	30	For a tutorial-style introduction, please read the L<Coro::Intro>
32	threads but don't (in general) run in parallel at the same time even	31	manpage. This manpage mainly contains reference information.
33	on SMP machines. The specific flavor of coroutine used in this module
34	also guarantees you that it will not switch between coroutines unless
35	necessary, at easily-identified points in your program, so locking and
36	parallel access are rarely an issue, making coroutine programming much
37	safer and easier than threads programming.
38		32
39	Unlike a normal perl program, however, coroutines allow you to have	33	This module collection manages continuations in general, most often in
40	multiple running interpreters that share data, which is especially useful	34	the form of cooperative threads (also called coros, or simply "coro"
41	to code pseudo-parallel processes and for event-based programming, such as	35	in the documentation). They are similar to kernel threads but don't (in
42	multiple HTTP-GET requests running concurrently. See L<Coro::AnyEvent> to	36	general) run in parallel at the same time even on SMP machines. The
43	learn more.	37	specific flavor of thread offered by this module also guarantees you that
		38	it will not switch between threads unless necessary, at easily-identified
		39	points in your program, so locking and parallel access are rarely an
		40	issue, making thread programming much safer and easier than using other
		41	thread models.
44		42
45	Coroutines are also useful because Perl has no support for threads (the so	43	Unlike the so-called "Perl threads" (which are not actually real threads
46	called "threads" that perl offers are nothing more than the (bad) process	44	but only the windows process emulation (see section of same name for
47	emulation coming from the Windows platform: On standard operating systems	45	more details) ported to UNIX, and as such act as processes), Coro
48	they serve no purpose whatsoever, except by making your programs slow and	46	provides a full shared address space, which makes communication between
49	making them use a lot of memory. Best disable them when building perl, or	47	threads very easy. And coro threads are fast, too: disabling the Windows
50	aks your software vendor/distributor to do it for you).	48	process emulation code in your perl and using Coro can easily result in
		49	a two to four times speed increase for your programs. A parallel matrix
		50	multiplication benchmark (very communication-intensive) runs over 300
		51	times faster on a single core than perls pseudo-threads on a quad core
		52	using all four cores.
51		53
		54	Coro achieves that by supporting multiple running interpreters that share
		55	data, which is especially useful to code pseudo-parallel processes and
		56	for event-based programming, such as multiple HTTP-GET requests running
		57	concurrently. See L<Coro::AnyEvent> to learn more on how to integrate Coro
		58	into an event-based environment.
		59
52	In this module, coroutines are defined as "callchain + lexical variables +	60	In this module, a thread is defined as "callchain + lexical variables +
53	@_ + $_ + $@ + $/ + C stack), that is, a coroutine has its own callchain,	61	some package variables + C stack), that is, a thread has its own callchain,
54	its own set of lexicals and its own set of perls most important global	62	its own set of lexicals and its own set of perls most important global
55	variables (see L<Coro::State> for more configuration).	63	variables (see L<Coro::State> for more configuration and background info).
		64
		65	See also the C<SEE ALSO> section at the end of this document - the Coro
		66	module family is quite large.
		67
		68	=head1 CORO THREAD LIFE CYCLE
		69
		70	During the long and exciting (or not) life of a coro thread, it goes
		71	through a number of states:
		72
		73	=over 4
		74
		75	=item 1. Creation
		76
		77	The first thing in the life of a coro thread is it's creation -
		78	obviously. The typical way to create a thread is to call the C<async
		79	BLOCK> function:
		80
		81	async {
		82	# thread code goes here
		83	};
		84
		85	You can also pass arguments, which are put in C<@_>:
		86
		87	async {
		88	print $_[1]; # prints 2
		89	} 1, 2, 3;
		90
		91	This creates a new coro thread and puts it into the ready queue, meaning
		92	it will run as soon as the CPU is free for it.
		93
		94	C<async> will return a Coro object - you can store this for future
		95	reference or ignore it - a thread that is running, ready to run or waiting
		96	for some event is alive on it's own.
		97
		98	Another way to create a thread is to call the C<new> constructor with a
		99	code-reference:
		100
		101	new Coro sub {
		102	# thread code goes here
		103	}, @optional_arguments;
		104
		105	This is quite similar to calling C<async>, but the important difference is
		106	that the new thread is not put into the ready queue, so the thread will
		107	not run until somebody puts it there. C<async> is, therefore, identical to
		108	this sequence:
		109
		110	my $coro = new Coro sub {
		111	# thread code goes here
		112	};
		113	$coro->ready;
		114	return $coro;
		115
		116	=item 2. Startup
		117
		118	When a new coro thread is created, only a copy of the code reference
		119	and the arguments are stored, no extra memory for stacks and so on is
		120	allocated, keeping the coro thread in a low-memory state.
		121
		122	Only when it actually starts executing will all the resources be finally
		123	allocated.
		124
		125	The optional arguments specified at coro creation are available in C<@_>,
		126	similar to function calls.
		127
		128	=item 3. Running / Blocking
		129
		130	A lot can happen after the coro thread has started running. Quite usually,
		131	it will not run to the end in one go (because you could use a function
		132	instead), but it will give up the CPU regularly because it waits for
		133	external events.
		134
		135	As long as a coro thread runs, its Coro object is available in the global
		136	variable C<$Coro::current>.
		137
		138	The low-level way to give up the CPU is to call the scheduler, which
		139	selects a new coro thread to run:
		140
		141	Coro::schedule;
		142
		143	Since running threads are not in the ready queue, calling the scheduler
		144	without doing anything else will block the coro thread forever - you need
		145	to arrange either for the coro to put woken up (readied) by some other
		146	event or some other thread, or you can put it into the ready queue before
		147	scheduling:
		148
		149	# this is exactly what Coro::cede does
		150	$Coro::current->ready;
		151	Coro::schedule;
		152
		153	All the higher-level synchronisation methods (Coro::Semaphore,
		154	Coro::rouse_*...) are actually implemented via C<< ->ready >> and C<<
		155	Coro::schedule >>.
		156
		157	While the coro thread is running it also might get assigned a C-level
		158	thread, or the C-level thread might be unassigned from it, as the Coro
		159	runtime wishes. A C-level thread needs to be assigned when your perl
		160	thread calls into some C-level function and that function in turn calls
		161	perl and perl then wants to switch coroutines. This happens most often
		162	when you run an event loop and block in the callback, or when perl
		163	itself calls some function such as C<AUTOLOAD> or methods via the C<tie>
		164	mechanism.
		165
		166	=item 4. Termination
		167
		168	Many threads actually terminate after some time. There are a number of
		169	ways to terminate a coro thread, the simplest is returning from the
		170	top-level code reference:
		171
		172	async {
		173	# after returning from here, the coro thread is terminated
		174	};
		175
		176	async {
		177	return if 0.5 < rand; # terminate a little earlier, maybe
		178	print "got a chance to print this\n";
		179	# or here
		180	};
		181
		182	Any values returned from the coroutine can be recovered using C<< ->join
		183	>>:
		184
		185	my $coro = async {
		186	"hello, world\n" # return a string
		187	};
		188
		189	my $hello_world = $coro->join;
		190
		191	print $hello_world;
		192
		193	Another way to terminate is to call C<< Coro::terminate >>, which at any
		194	subroutine call nesting level:
		195
		196	async {
		197	Coro::terminate "return value 1", "return value 2";
		198	};
		199
		200	Yet another way is to C<< ->cancel >> (or C<< ->safe_cancel >>) the coro
		201	thread from another thread:
		202
		203	my $coro = async {
		204	exit 1;
		205	};
		206
		207	$coro->cancel; # also accepts values for ->join to retrieve
		208
		209	Cancellation I<can> be dangerous - it's a bit like calling C<exit> without
		210	actually exiting, and might leave C libraries and XS modules in a weird
		211	state. Unlike other thread implementations, however, Coro is exceptionally
		212	safe with regards to cancellation, as perl will always be in a consistent
		213	state, and for those cases where you want to do truly marvellous things
		214	with your coro while it is being cancelled - that is, make sure all
		215	cleanup code is executed from the thread being cancelled - there is even a
		216	C<< ->safe_cancel >> method.
		217
		218	So, cancelling a thread that runs in an XS event loop might not be the
		219	best idea, but any other combination that deals with perl only (cancelling
		220	when a thread is in a C<tie> method or an C<AUTOLOAD> for example) is
		221	safe.
		222
		223	Last not least, a coro thread object that isn't referenced is C<<
		224	->cancel >>'ed automatically - just like other objects in Perl. This
		225	is not such a common case, however - a running thread is referencedy by
		226	C<$Coro::current>, a thread ready to run is referenced by the ready queue,
		227	a thread waiting on a lock or semaphore is referenced by being in some
		228	wait list and so on. But a thread that isn't in any of those queues gets
		229	cancelled:
		230
		231	async {
		232	schedule; # cede to other coros, don't go into the ready queue
		233	};
		234
		235	cede;
		236	# now the async above is destroyed, as it is not referenced by anything.
		237
		238	A slightly embellished example might make it clearer:
		239
		240	async {
		241	my $guard = Guard::guard { print "destroyed\n" };
		242	schedule while 1;
		243	};
		244
		245	cede;
		246
		247	Superficially one might not expect any output - since the C<async>
		248	implements an endless loop, the C<$guard> will not be cleaned up. However,
		249	since the thread object returned by C<async> is not stored anywhere, the
		250	thread is initially referenced because it is in the ready queue, when it
		251	runs it is referenced by C<$Coro::current>, but when it calls C<schedule>,
		252	it gets C<cancel>ed causing the guard object to be destroyed (see the next
		253	section), and printing it's message.
		254
		255	If this seems a bit drastic, remember that this only happens when nothing
		256	references the thread anymore, which means there is no way to further
		257	execute it, ever. The only options at this point are leaking the thread,
		258	or cleaning it up, which brings us to...
		259
		260	=item 5. Cleanup
		261
		262	Threads will allocate various resources. Most but not all will be returned
		263	when a thread terminates, during clean-up.
		264
		265	Cleanup is quite similar to throwing an uncaught exception: perl will
		266	work it's way up through all subroutine calls and blocks. On it's way, it
		267	will release all C<my> variables, undo all C<local>'s and free any other
		268	resources truly local to the thread.
		269
		270	So, a common way to free resources is to keep them referenced only by my
		271	variables:
		272
		273	async {
		274	my $big_cache = new Cache ...;
		275	};
		276
		277	If there are no other references, then the C<$big_cache> object will be
		278	freed when the thread terminates, regardless of how it does so.
		279
		280	What it does C<NOT> do is unlock any Coro::Semaphores or similar
		281	resources, but that's where the C<guard> methods come in handy:
		282
		283	my $sem = new Coro::Semaphore;
		284
		285	async {
		286	my $lock_guard = $sem->guard;
		287	# if we return, or die or get cancelled, here,
		288	# then the semaphore will be "up"ed.
		289	};
		290
		291	The C<Guard::guard> function comes in handy for any custom cleanup you
		292	might want to do (but you cannot switch to other coroutines from those
		293	code blocks):
		294
		295	async {
		296	my $window = new Gtk2::Window "toplevel";
		297	# The window will not be cleaned up automatically, even when $window
		298	# gets freed, so use a guard to ensure it's destruction
		299	# in case of an error:
		300	my $window_guard = Guard::guard { $window->destroy };
		301
		302	# we are safe here
		303	};
		304
		305	Last not least, C<local> can often be handy, too, e.g. when temporarily
		306	replacing the coro thread description:
		307
		308	sub myfunction {
		309	local $Coro::current->{desc} = "inside myfunction(@_)";
		310
		311	# if we return or die here, the description will be restored
		312	}
		313
		314	=item 6. Viva La Zombie Muerte
		315
		316	Even after a thread has terminated and cleaned up its resources, the Coro
		317	object still is there and stores the return values of the thread.
		318
		319	When there are no other references, it will simply be cleaned up and
		320	freed.
		321
		322	If there areany references, the Coro object will stay around, and you
		323	can call C<< ->join >> as many times as you wish to retrieve the result
		324	values:
		325
		326	async {
		327	print "hi\n";
		328	1
		329	};
		330
		331	# run the async above, and free everything before returning
		332	# from Coro::cede:
		333	Coro::cede;
		334
		335	{
		336	my $coro = async {
		337	print "hi\n";
		338	1
		339	};
		340
		341	# run the async above, and clean up, but do not free the coro
		342	# object:
		343	Coro::cede;
		344
		345	# optionally retrieve the result values
		346	my @results = $coro->join;
		347
		348	# now $coro goes out of scope, and presumably gets freed
		349	};
		350
		351	=back
56		352
57	=cut	353	=cut
58		354
59	package Coro;	355	package Coro;
60		356
61	use strict qw(vars subs);	357	use common::sense;
62	no warnings "uninitialized";	358
		359	use Carp ();
		360
		361	use Guard ();
63		362
64	use Coro::State;	363	use Coro::State;
65		364
66	use base qw(Coro::State Exporter);	365	use base qw(Coro::State Exporter);
67		366
68	our $idle; # idle handler	367	our $idle; # idle handler
69	our $main; # main coroutine	368	our $main; # main coro
70	our $current; # current coroutine	369	our $current; # current coro
71		370
72	our $VERSION = 4.911;	371	our $VERSION = 6.33;
73		372
74	our @EXPORT = qw(async async_pool cede schedule terminate current unblock_sub);	373	our @EXPORT = qw(async async_pool cede schedule terminate current unblock_sub rouse_cb rouse_wait);
75	our %EXPORT_TAGS = (	374	our %EXPORT_TAGS = (
76	prio => [qw(PRIO_MAX PRIO_HIGH PRIO_NORMAL PRIO_LOW PRIO_IDLE PRIO_MIN)],	375	prio => [qw(PRIO_MAX PRIO_HIGH PRIO_NORMAL PRIO_LOW PRIO_IDLE PRIO_MIN)],
77	);	376	);
78	our @EXPORT_OK = (@{$EXPORT_TAGS{prio}}, qw(nready));	377	our @EXPORT_OK = (@{$EXPORT_TAGS{prio}}, qw(nready));
79		378
		379	=head1 GLOBAL VARIABLES
		380
80	=over 4	381	=over 4
81		382
82	=item $Coro::main	383	=item $Coro::main
83		384
84	This variable stores the coroutine object that represents the main	385	This variable stores the Coro object that represents the main
85	program. While you cna C<ready> it and do most other things you can do to	386	program. While you can C<ready> it and do most other things you can do to
86	coroutines, it is mainly useful to compare again C<$Coro::current>, to see	387	coro, it is mainly useful to compare again C<$Coro::current>, to see
87	whether you are running in the main program or not.	388	whether you are running in the main program or not.
88		389
89	=cut	390	=cut
90		391
91	$main = new Coro;	392	# $main is now being initialised by Coro::State
92		393
93	=item $Coro::current	394	=item $Coro::current
94		395
95	The coroutine object representing the current coroutine (the last	396	The Coro object representing the current coro (the last
96	coroutine that the Coro scheduler switched to). The initial value is	397	coro that the Coro scheduler switched to). The initial value is
97	C<$main> (of course).	398	C<$Coro::main> (of course).
98		399
99	This variable is B<strictly> I<read-only>. You can take copies of the	400	This variable is B<strictly> I<read-only>. You can take copies of the
100	value stored in it and use it as any other coroutine object, but you must	401	value stored in it and use it as any other Coro object, but you must
101	not otherwise modify the variable itself.	402	not otherwise modify the variable itself.
102		403
103	=cut	404	=cut
104		405
105	$main->{desc} = "[main::]";
106
107	# maybe some other module used Coro::Specific before...
108	$main->{_specific} = $current->{_specific}
109	if $current;
110
111	_set_current $main;
112
113	sub current() { $current } # [DEPRECATED]	406	sub current() { $current } # [DEPRECATED]
114		407
115	=item $Coro::idle	408	=item $Coro::idle
116		409
117	This variable is mainly useful to integrate Coro into event loops. It is	410	This variable is mainly useful to integrate Coro into event loops. It is
118	usually better to rely on L<Coro::AnyEvent> or LC<Coro::EV>, as this is	411	usually better to rely on L<Coro::AnyEvent> or L<Coro::EV>, as this is
119	pretty low-level functionality.	412	pretty low-level functionality.
120		413
121	This variable stores a callback that is called whenever the scheduler	414	This variable stores a Coro object that is put into the ready queue when
122	finds no ready coroutines to run. The default implementation prints	415	there are no other ready threads (without invoking any ready hooks).
123	"FATAL: deadlock detected" and exits, because the program has no other way
124	to continue.
125		416
		417	The default implementation dies with "FATAL: deadlock detected.", followed
		418	by a thread listing, because the program has no other way to continue.
		419
126	This hook is overwritten by modules such as C<Coro::Timer> and	420	This hook is overwritten by modules such as C<Coro::EV> and
127	C<Coro::AnyEvent> to wait on an external event that hopefully wake up a	421	C<Coro::AnyEvent> to wait on an external event that hopefully wakes up a
128	coroutine so the scheduler can run it.	422	coro so the scheduler can run it.
129		423
130	Note that the callback I<must not>, under any circumstances, block
131	the current coroutine. Normally, this is achieved by having an "idle
132	coroutine" that calls the event loop and then blocks again, and then
133	readying that coroutine in the idle handler.
134
135	See L<Coro::Event> or L<Coro::AnyEvent> for examples of using this	424	See L<Coro::EV> or L<Coro::AnyEvent> for examples of using this technique.
136	technique.
137
138	Please note that if your callback recursively invokes perl (e.g. for event
139	handlers), then it must be prepared to be called recursively itself.
140		425
141	=cut	426	=cut
142		427
143	$idle = sub {	428	# ||= because other modules could have provided their own by now
144	require Carp;	429	$idle ||= new Coro sub {
145	Carp::croak ("FATAL: deadlock detected");	430	require Coro::Debug;
		431	die "FATAL: deadlock detected.\n"
		432	. Coro::Debug::ps_listing ();
146	};	433	};
147		434
148	sub _cancel {
149	my ($self) = @_;
150
151	# free coroutine data and mark as destructed
152	$self->_destroy
153	or return;
154
155	# call all destruction callbacks
156	$_->(@{$self->{_status}})
157	for @{ delete $self->{_on_destroy} || [] };
158	}
159
160	# this coroutine is necessary because a coroutine	435	# this coro is necessary because a coro
161	# cannot destroy itself.	436	# cannot destroy itself.
162	my @destroy;	437	our @destroy;
163	my $manager;	438	our $manager;
164		439
165	$manager = new Coro sub {	440	$manager = new Coro sub {
166	while () {	441	while () {
167	(shift @destroy)->_cancel	442	_destroy shift @destroy
168	while @destroy;	443	while @destroy;
169		444
170	&schedule;	445	&schedule;
171	}	446	}
172	};	447	};
173	$manager->{desc} = "[coro manager]";	448	$manager->{desc} = "[coro manager]";
174	$manager->prio (PRIO_MAX);	449	$manager->prio (PRIO_MAX);
175		450
176	=back	451	=back
177		452
178	=head2 SIMPLE COROUTINE CREATION	453	=head1 SIMPLE CORO CREATION
179		454
180	=over 4	455	=over 4
181		456
182	=item async { ... } [@args...]	457	=item async { ... } [@args...]
183		458
184	Create a new coroutine and return it's coroutine object (usually	459	Create a new coro and return its Coro object (usually
185	unused). The coroutine will be put into the ready queue, so	460	unused). The coro will be put into the ready queue, so
186	it will start running automatically on the next scheduler run.	461	it will start running automatically on the next scheduler run.
187		462
188	The first argument is a codeblock/closure that should be executed in the	463	The first argument is a codeblock/closure that should be executed in the
189	coroutine. When it returns argument returns the coroutine is automatically	464	coro. When it returns argument returns the coro is automatically
190	terminated.	465	terminated.
191		466
192	The remaining arguments are passed as arguments to the closure.	467	The remaining arguments are passed as arguments to the closure.
193		468
194	See the C<Coro::State::new> constructor for info about the coroutine	469	See the C<Coro::State::new> constructor for info about the coro
195	environment in which coroutines are executed.	470	environment in which coro are executed.
196		471
197	Calling C<exit> in a coroutine will do the same as calling exit outside	472	Calling C<exit> in a coro will do the same as calling exit outside
198	the coroutine. Likewise, when the coroutine dies, the program will exit,	473	the coro. Likewise, when the coro dies, the program will exit,
199	just as it would in the main program.	474	just as it would in the main program.
200		475
201	If you do not want that, you can provide a default C<die> handler, or	476	If you do not want that, you can provide a default C<die> handler, or
202	simply avoid dieing (by use of C<eval>).	477	simply avoid dieing (by use of C<eval>).
203		478
204	Example: Create a new coroutine that just prints its arguments.	479	Example: Create a new coro that just prints its arguments.
205		480
206	async {	481	async {
207	print "@_\n";	482	print "@_\n";
208	} 1,2,3,4;	483	} 1,2,3,4;
209		484
210	=cut
211
212	sub async(&@) {
213	my $coro = new Coro @_;
214	$coro->ready;
215	$coro
216	}
217
218	=item async_pool { ... } [@args...]	485	=item async_pool { ... } [@args...]
219		486
220	Similar to C<async>, but uses a coroutine pool, so you should not call	487	Similar to C<async>, but uses a coro pool, so you should not call
221	terminate or join on it (although you are allowed to), and you get a	488	terminate or join on it (although you are allowed to), and you get a
222	coroutine that might have executed other code already (which can be good	489	coro that might have executed other code already (which can be good
223	or bad :).	490	or bad :).
224		491
225	On the plus side, this function is faster than creating (and destroying)	492	On the plus side, this function is about twice as fast as creating (and
226	a completly new coroutine, so if you need a lot of generic coroutines in	493	destroying) a completely new coro, so if you need a lot of generic
227	quick successsion, use C<async_pool>, not C<async>.	494	coros in quick successsion, use C<async_pool>, not C<async>.
228		495
229	The code block is executed in an C<eval> context and a warning will be	496	The code block is executed in an C<eval> context and a warning will be
230	issued in case of an exception instead of terminating the program, as	497	issued in case of an exception instead of terminating the program, as
231	C<async> does. As the coroutine is being reused, stuff like C<on_destroy>	498	C<async> does. As the coro is being reused, stuff like C<on_destroy>
232	will not work in the expected way, unless you call terminate or cancel,	499	will not work in the expected way, unless you call terminate or cancel,
233	which somehow defeats the purpose of pooling (but is fine in the	500	which somehow defeats the purpose of pooling (but is fine in the
234	exceptional case).	501	exceptional case).
235		502
236	The priority will be reset to C<0> after each run, tracing will be	503	The priority will be reset to C<0> after each run, tracing will be
237	disabled, the description will be reset and the default output filehandle	504	disabled, the description will be reset and the default output filehandle
238	gets restored, so you can change all these. Otherwise the coroutine will	505	gets restored, so you can change all these. Otherwise the coro will
239	be re-used "as-is": most notably if you change other per-coroutine global	506	be re-used "as-is": most notably if you change other per-coro global
240	stuff such as C<$/> you I<must needs> revert that change, which is most	507	stuff such as C<$/> you I<must needs> revert that change, which is most
241	simply done by using local as in: C<< local $/ >>.	508	simply done by using local as in: C<< local $/ >>.
242		509
243	The idle pool size is limited to C<8> idle coroutines (this can be	510	The idle pool size is limited to C<8> idle coros (this can be
244	adjusted by changing $Coro::POOL_SIZE), but there can be as many non-idle	511	adjusted by changing $Coro::POOL_SIZE), but there can be as many non-idle
245	coros as required.	512	coros as required.
246		513
247	If you are concerned about pooled coroutines growing a lot because a	514	If you are concerned about pooled coros growing a lot because a
248	single C<async_pool> used a lot of stackspace you can e.g. C<async_pool	515	single C<async_pool> used a lot of stackspace you can e.g. C<async_pool
249	{ terminate }> once per second or so to slowly replenish the pool. In	516	{ terminate }> once per second or so to slowly replenish the pool. In
250	addition to that, when the stacks used by a handler grows larger than 16kb	517	addition to that, when the stacks used by a handler grows larger than 32kb
251	(adjustable via $Coro::POOL_RSS) it will also be destroyed.	518	(adjustable via $Coro::POOL_RSS) it will also be destroyed.
252		519
253	=cut	520	=cut
254		521
255	our $POOL_SIZE = 8;	522	our $POOL_SIZE = 8;
256	our $POOL_RSS = 16 * 1024;	523	our $POOL_RSS = 32 * 1024;
257	our @async_pool;	524	our @async_pool;
258		525
259	sub pool_handler {	526	sub pool_handler {
260	my $cb;
261
262	while () {	527	while () {
263	eval {	528	eval {
264	while () {	529	&{&_pool_handler} while 1;
265	_pool_1 $cb;
266	&$cb;
267	_pool_2 $cb;
268	&schedule;
269	}
270	};	530	};
271		531
272	if ($@) {
273	last if $@ eq "\3async_pool terminate\2\n";
274	warn $@;	532	warn $@ if $@;
275	}
276	}	533	}
277	}	534	}
278		535
279	sub async_pool(&@) {
280	# this is also inlined into the unlock_scheduler
281	my $coro = (pop @async_pool) || new Coro \&pool_handler;
282
283	$coro->{_invoke} = [@_];
284	$coro->ready;
285
286	$coro
287	}
288
289	=back	536	=back
290		537
291	=head2 STATIC METHODS	538	=head1 STATIC METHODS
292		539
293	Static methods are actually functions that operate on the current coroutine.	540	Static methods are actually functions that implicitly operate on the
		541	current coro.
294		542
295	=over 4	543	=over 4
296		544
297	=item schedule	545	=item schedule
298		546
299	Calls the scheduler. The scheduler will find the next coroutine that is	547	Calls the scheduler. The scheduler will find the next coro that is
300	to be run from the ready queue and switches to it. The next coroutine	548	to be run from the ready queue and switches to it. The next coro
301	to be run is simply the one with the highest priority that is longest	549	to be run is simply the one with the highest priority that is longest
302	in its ready queue. If there is no coroutine ready, it will clal the	550	in its ready queue. If there is no coro ready, it will call the
303	C<$Coro::idle> hook.	551	C<$Coro::idle> hook.
304		552
305	Please note that the current coroutine will I<not> be put into the ready	553	Please note that the current coro will I<not> be put into the ready
306	queue, so calling this function usually means you will never be called	554	queue, so calling this function usually means you will never be called
307	again unless something else (e.g. an event handler) calls C<< ->ready >>,	555	again unless something else (e.g. an event handler) calls C<< ->ready >>,
308	thus waking you up.	556	thus waking you up.
309		557
310	This makes C<schedule> I<the> generic method to use to block the current	558	This makes C<schedule> I<the> generic method to use to block the current
311	coroutine and wait for events: first you remember the current coroutine in	559	coro and wait for events: first you remember the current coro in
312	a variable, then arrange for some callback of yours to call C<< ->ready	560	a variable, then arrange for some callback of yours to call C<< ->ready
313	>> on that once some event happens, and last you call C<schedule> to put	561	>> on that once some event happens, and last you call C<schedule> to put
314	yourself to sleep. Note that a lot of things can wake your coroutine up,	562	yourself to sleep. Note that a lot of things can wake your coro up,
315	so you need to check whether the event indeed happened, e.g. by storing the	563	so you need to check whether the event indeed happened, e.g. by storing the
316	status in a variable.	564	status in a variable.
317		565
318	The canonical way to wait on external events is this:	566	See B<HOW TO WAIT FOR A CALLBACK>, below, for some ways to wait for callbacks.
319		567
320	{	568	=item cede
321	# remember current coroutine
322	my $current = $Coro::current;
323		569
324	# register a hypothetical event handler	570	"Cede" to other coros. This function puts the current coro into
325	on_event_invoke sub {	571	the ready queue and calls C<schedule>, which has the effect of giving
326	# wake up sleeping coroutine	572	up the current "timeslice" to other coros of the same or higher
327	$current->ready;	573	priority. Once your coro gets its turn again it will automatically be
328	undef $current;	574	resumed.
		575
		576	This function is often called C<yield> in other languages.
		577
		578	=item Coro::cede_notself
		579
		580	Works like cede, but is not exported by default and will cede to I<any>
		581	coro, regardless of priority. This is useful sometimes to ensure
		582	progress is made.
		583
		584	=item terminate [arg...]
		585
		586	Terminates the current coro with the given status values (see
		587	L<cancel>). The values will not be copied, but referenced directly.
		588
		589	=item Coro::on_enter BLOCK, Coro::on_leave BLOCK
		590
		591	These function install enter and leave winders in the current scope. The
		592	enter block will be executed when on_enter is called and whenever the
		593	current coro is re-entered by the scheduler, while the leave block is
		594	executed whenever the current coro is blocked by the scheduler, and
		595	also when the containing scope is exited (by whatever means, be it exit,
		596	die, last etc.).
		597
		598	I<Neither invoking the scheduler, nor exceptions, are allowed within those
		599	BLOCKs>. That means: do not even think about calling C<die> without an
		600	eval, and do not even think of entering the scheduler in any way.
		601
		602	Since both BLOCKs are tied to the current scope, they will automatically
		603	be removed when the current scope exits.
		604
		605	These functions implement the same concept as C<dynamic-wind> in scheme
		606	does, and are useful when you want to localise some resource to a specific
		607	coro.
		608
		609	They slow down thread switching considerably for coros that use them
		610	(about 40% for a BLOCK with a single assignment, so thread switching is
		611	still reasonably fast if the handlers are fast).
		612
		613	These functions are best understood by an example: The following function
		614	will change the current timezone to "Antarctica/South_Pole", which
		615	requires a call to C<tzset>, but by using C<on_enter> and C<on_leave>,
		616	which remember/change the current timezone and restore the previous
		617	value, respectively, the timezone is only changed for the coro that
		618	installed those handlers.
		619
		620	use POSIX qw(tzset);
		621
		622	async {
		623	my $old_tz; # store outside TZ value here
		624
		625	Coro::on_enter {
		626	$old_tz = $ENV{TZ}; # remember the old value
		627
		628	$ENV{TZ} = "Antarctica/South_Pole";
		629	tzset; # enable new value
329	};	630	};
330		631
331	# call schedule until event occurred.	632	Coro::on_leave {
332	# in case we are woken up for other reasons	633	$ENV{TZ} = $old_tz;
333	# (current still defined), loop.	634	tzset; # restore old value
334	Coro::schedule while $current;	635	};
		636
		637	# at this place, the timezone is Antarctica/South_Pole,
		638	# without disturbing the TZ of any other coro.
335	}	639	};
336		640
337	=item cede	641	This can be used to localise about any resource (locale, uid, current
		642	working directory etc.) to a block, despite the existance of other
		643	coros.
338		644
339	"Cede" to other coroutines. This function puts the current coroutine into	645	Another interesting example implements time-sliced multitasking using
340	the ready queue and calls C<schedule>, which has the effect of giving	646	interval timers (this could obviously be optimised, but does the job):
341	up the current "timeslice" to other coroutines of the same or higher
342	priority. Once your coroutine gets its turn again it will automatically be
343	resumed.
344		647
345	This function is often called C<yield> in other languages.	648	# "timeslice" the given block
		649	sub timeslice(&) {
		650	use Time::HiRes ();
346		651
347	=item Coro::cede_notself	652	Coro::on_enter {
		653	# on entering the thread, we set an VTALRM handler to cede
		654	$SIG{VTALRM} = sub { cede };
		655	# and then start the interval timer
		656	Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0.01, 0.01;
		657	};
		658	Coro::on_leave {
		659	# on leaving the thread, we stop the interval timer again
		660	Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0, 0;
		661	};
348		662
349	Works like cede, but is not exported by default and will cede to I<any>	663	&{+shift};
350	coroutine, regardless of priority. This is useful sometimes to ensure	664	}
351	progress is made.
352		665
353	=item terminate [arg...]	666	# use like this:
		667	timeslice {
		668	# The following is an endless loop that would normally
		669	# monopolise the process. Since it runs in a timesliced
		670	# environment, it will regularly cede to other threads.
		671	while () { }
		672	};
354		673
355	Terminates the current coroutine with the given status values (see L<cancel>).
356		674
357	=item killall	675	=item killall
358		676
359	Kills/terminates/cancels all coroutines except the currently running	677	Kills/terminates/cancels all coros except the currently running one.
360	one. This is useful after a fork, either in the child or the parent, as
361	usually only one of them should inherit the running coroutines.
362		678
363	Note that while this will try to free some of the main programs resources,	679	Note that while this will try to free some of the main interpreter
		680	resources if the calling coro isn't the main coro, but one
364	you cannot free all of them, so if a coroutine that is not the main	681	cannot free all of them, so if a coro that is not the main coro
365	program calls this function, there will be some one-time resource leak.	682	calls this function, there will be some one-time resource leak.
366		683
367	=cut	684	=cut
368
369	sub terminate {
370	$current->cancel (@_);
371	}
372		685
373	sub killall {	686	sub killall {
374	for (Coro::State::list) {	687	for (Coro::State::list) {
375	$_->cancel	688	$_->cancel
376	if $_ != $current && UNIVERSAL::isa $_, "Coro";	689	if $_ != $current && UNIVERSAL::isa $_, "Coro";
377	}	690	}
378	}	691	}
379		692
380	=back	693	=back
381		694
382	=head2 COROUTINE METHODS	695	=head1 CORO OBJECT METHODS
383		696
384	These are the methods you can call on coroutine objects (or to create	697	These are the methods you can call on coro objects (or to create
385	them).	698	them).
386		699
387	=over 4	700	=over 4
388		701
389	=item new Coro \&sub [, @args...]	702	=item new Coro \&sub [, @args...]
390		703
391	Create a new coroutine and return it. When the sub returns, the coroutine	704	Create a new coro and return it. When the sub returns, the coro
392	automatically terminates as if C<terminate> with the returned values were	705	automatically terminates as if C<terminate> with the returned values were
393	called. To make the coroutine run you must first put it into the ready	706	called. To make the coro run you must first put it into the ready
394	queue by calling the ready method.	707	queue by calling the ready method.
395		708
396	See C<async> and C<Coro::State::new> for additional info about the	709	See C<async> and C<Coro::State::new> for additional info about the
397	coroutine environment.	710	coro environment.
398		711
399	=cut	712	=cut
400		713
401	sub _run_coro {	714	sub _coro_run {
402	terminate &{+shift};	715	terminate &{+shift};
403	}	716	}
404		717
405	sub new {
406	my $class = shift;
407
408	$class->SUPER::new (\&_run_coro, @_)
409	}
410
411	=item $success = $coroutine->ready	718	=item $success = $coro->ready
412		719
413	Put the given coroutine into the end of its ready queue (there is one	720	Put the given coro into the end of its ready queue (there is one
414	queue for each priority) and return true. If the coroutine is already in	721	queue for each priority) and return true. If the coro is already in
415	the ready queue, do nothing and return false.	722	the ready queue, do nothing and return false.
416		723
417	This ensures that the scheduler will resume this coroutine automatically	724	This ensures that the scheduler will resume this coro automatically
418	once all the coroutines of higher priority and all coroutines of the same	725	once all the coro of higher priority and all coro of the same
419	priority that were put into the ready queue earlier have been resumed.	726	priority that were put into the ready queue earlier have been resumed.
420		727
		728	=item $coro->suspend
		729
		730	Suspends the specified coro. A suspended coro works just like any other
		731	coro, except that the scheduler will not select a suspended coro for
		732	execution.
		733
		734	Suspending a coro can be useful when you want to keep the coro from
		735	running, but you don't want to destroy it, or when you want to temporarily
		736	freeze a coro (e.g. for debugging) to resume it later.
		737
		738	A scenario for the former would be to suspend all (other) coros after a
		739	fork and keep them alive, so their destructors aren't called, but new
		740	coros can be created.
		741
		742	=item $coro->resume
		743
		744	If the specified coro was suspended, it will be resumed. Note that when
		745	the coro was in the ready queue when it was suspended, it might have been
		746	unreadied by the scheduler, so an activation might have been lost.
		747
		748	To avoid this, it is best to put a suspended coro into the ready queue
		749	unconditionally, as every synchronisation mechanism must protect itself
		750	against spurious wakeups, and the one in the Coro family certainly do
		751	that.
		752
		753	=item $state->is_new
		754
		755	Returns true iff this Coro object is "new", i.e. has never been run
		756	yet. Those states basically consist of only the code reference to call and
		757	the arguments, but consumes very little other resources. New states will
		758	automatically get assigned a perl interpreter when they are transfered to.
		759
		760	=item $state->is_zombie
		761
		762	Returns true iff the Coro object has been cancelled, i.e.
		763	it's resources freed because they were C<cancel>'ed, C<terminate>'d,
		764	C<safe_cancel>'ed or simply went out of scope.
		765
		766	The name "zombie" stems from UNIX culture, where a process that has
		767	exited and only stores and exit status and no other resources is called a
		768	"zombie".
		769
421	=item $is_ready = $coroutine->is_ready	770	=item $is_ready = $coro->is_ready
422		771
423	Return whether the coroutine is currently the ready queue or not,	772	Returns true iff the Coro object is in the ready queue. Unless the Coro
		773	object gets destroyed, it will eventually be scheduled by the scheduler.
424		774
		775	=item $is_running = $coro->is_running
		776
		777	Returns true iff the Coro object is currently running. Only one Coro object
		778	can ever be in the running state (but it currently is possible to have
		779	multiple running Coro::States).
		780
		781	=item $is_suspended = $coro->is_suspended
		782
		783	Returns true iff this Coro object has been suspended. Suspended Coros will
		784	not ever be scheduled.
		785
425	=item $coroutine->cancel (arg...)	786	=item $coro->cancel (arg...)
426		787
427	Terminates the given coroutine and makes it return the given arguments as	788	Terminates the given Coro thread and makes it return the given arguments as
428	status (default: the empty list). Never returns if the coroutine is the	789	status (default: an empty list). Never returns if the Coro is the
429	current coroutine.	790	current Coro.
430		791
431	=cut	792	This is a rather brutal way to free a coro, with some limitations - if
		793	the thread is inside a C callback that doesn't expect to be canceled,
		794	bad things can happen, or if the cancelled thread insists on running
		795	complicated cleanup handlers that rely on its thread context, things will
		796	not work.
432		797
433	sub cancel {	798	Any cleanup code being run (e.g. from C<guard> blocks) will be run without
434	my $self = shift;	799	a thread context, and is not allowed to switch to other threads. On the
435	$self->{_status} = [@_];	800	plus side, C<< ->cancel >> will always clean up the thread, no matter
		801	what. If your cleanup code is complex or you want to avoid cancelling a
		802	C-thread that doesn't know how to clean up itself, it can be better to C<<
		803	->throw >> an exception, or use C<< ->safe_cancel >>.
436		804
437	if ($current == $self) {	805	The arguments to C<< ->cancel >> are not copied, but instead will
438	push @destroy, $self;	806	be referenced directly (e.g. if you pass C<$var> and after the call
439	$manager->ready;	807	change that variable, then you might change the return values passed to
440	&schedule while 1;	808	e.g. C<join>, so don't do that).
441	} else {	809
442	$self->_cancel;	810	The resources of the Coro are usually freed (or destructed) before this
		811	call returns, but this can be delayed for an indefinite amount of time, as
		812	in some cases the manager thread has to run first to actually destruct the
		813	Coro object.
		814
		815	=item $coro->safe_cancel ($arg...)
		816
		817	Works mostly like C<< ->cancel >>, but is inherently "safer", and
		818	consequently, can fail with an exception in cases the thread is not in a
		819	cancellable state.
		820
		821	This method works a bit like throwing an exception that cannot be caught
		822	- specifically, it will clean up the thread from within itself, so
		823	all cleanup handlers (e.g. C<guard> blocks) are run with full thread
		824	context and can block if they wish. The downside is that there is no
		825	guarantee that the thread can be cancelled when you call this method, and
		826	therefore, it might fail. It is also considerably slower than C<cancel> or
		827	C<terminate>.
		828
		829	A thread is in a safe-cancellable state if it either hasn't been run yet,
		830	or it has no C context attached and is inside an SLF function.
		831
		832	The latter two basically mean that the thread isn't currently inside a
		833	perl callback called from some C function (usually via some XS modules)
		834	and isn't currently executing inside some C function itself (via Coro's XS
		835	API).
		836
		837	This call returns true when it could cancel the thread, or croaks with an
		838	error otherwise (i.e. it either returns true or doesn't return at all).
		839
		840	Why the weird interface? Well, there are two common models on how and
		841	when to cancel things. In the first, you have the expectation that your
		842	coro thread can be cancelled when you want to cancel it - if the thread
		843	isn't cancellable, this would be a bug somewhere, so C<< ->safe_cancel >>
		844	croaks to notify of the bug.
		845
		846	In the second model you sometimes want to ask nicely to cancel a thread,
		847	but if it's not a good time, well, then don't cancel. This can be done
		848	relatively easy like this:
		849
		850	if (! eval { $coro->safe_cancel }) {
		851	warn "unable to cancel thread: $@";
443	}	852	}
444	}
445		853
		854	However, what you never should do is first try to cancel "safely" and
		855	if that fails, cancel the "hard" way with C<< ->cancel >>. That makes
		856	no sense: either you rely on being able to execute cleanup code in your
		857	thread context, or you don't. If you do, then C<< ->safe_cancel >> is the
		858	only way, and if you don't, then C<< ->cancel >> is always faster and more
		859	direct.
		860
		861	=item $coro->schedule_to
		862
		863	Puts the current coro to sleep (like C<Coro::schedule>), but instead
		864	of continuing with the next coro from the ready queue, always switch to
		865	the given coro object (regardless of priority etc.). The readyness
		866	state of that coro isn't changed.
		867
		868	This is an advanced method for special cases - I'd love to hear about any
		869	uses for this one.
		870
		871	=item $coro->cede_to
		872
		873	Like C<schedule_to>, but puts the current coro into the ready
		874	queue. This has the effect of temporarily switching to the given
		875	coro, and continuing some time later.
		876
		877	This is an advanced method for special cases - I'd love to hear about any
		878	uses for this one.
		879
446	=item $coroutine->throw ([$scalar])	880	=item $coro->throw ([$scalar])
447		881
448	If C<$throw> is specified and defined, it will be thrown as an exception	882	If C<$throw> is specified and defined, it will be thrown as an exception
449	inside the coroutine at the next convenient point in time (usually after	883	inside the coro at the next convenient point in time. Otherwise
450	it gains control at the next schedule/transfer/cede). Otherwise clears the
451	exception object.	884	clears the exception object.
		885
		886	Coro will check for the exception each time a schedule-like-function
		887	returns, i.e. after each C<schedule>, C<cede>, C<< Coro::Semaphore->down
		888	>>, C<< Coro::Handle->readable >> and so on. Most of those functions (all
		889	that are part of Coro itself) detect this case and return early in case an
		890	exception is pending.
452		891
453	The exception object will be thrown "as is" with the specified scalar in	892	The exception object will be thrown "as is" with the specified scalar in
454	C<$@>, i.e. if it is a string, no line number or newline will be appended	893	C<$@>, i.e. if it is a string, no line number or newline will be appended
455	(unlike with C<die>).	894	(unlike with C<die>).
456		895
457	This can be used as a softer means than C<cancel> to ask a coroutine to	896	This can be used as a softer means than either C<cancel> or C<safe_cancel
458	end itself, although there is no guarantee that the exception will lead to	897	>to ask a coro to end itself, although there is no guarantee that the
459	termination, and if the exception isn't caught it might well end the whole	898	exception will lead to termination, and if the exception isn't caught it
460	program.	899	might well end the whole program.
461		900
462	You might also think of C<throw> as being the moral equivalent of	901	You might also think of C<throw> as being the moral equivalent of
463	C<kill>ing a coroutine with a signal (in this case, a scalar).	902	C<kill>ing a coro with a signal (in this case, a scalar).
464		903
465	=item $coroutine->join	904	=item $coro->join
466		905
467	Wait until the coroutine terminates and return any values given to the	906	Wait until the coro terminates and return any values given to the
468	C<terminate> or C<cancel> functions. C<join> can be called concurrently	907	C<terminate> or C<cancel> functions. C<join> can be called concurrently
469	from multiple coroutines, and all will be resumed and given the status	908	from multiple threads, and all will be resumed and given the status
470	return once the C<$coroutine> terminates.	909	return once the C<$coro> terminates.
471		910
472	=cut
473
474	sub join {
475	my $self = shift;
476
477	unless ($self->{_status}) {
478	my $current = $current;
479
480	push @{$self->{_on_destroy}}, sub {
481	$current->ready;
482	undef $current;
483	};
484
485	&schedule while $current;
486	}
487
488	wantarray ? @{$self->{_status}} : $self->{_status}[0];
489	}
490
491	=item $coroutine->on_destroy (\&cb)	911	=item $coro->on_destroy (\&cb)
492		912
493	Registers a callback that is called when this coroutine gets destroyed,	913	Registers a callback that is called when this coro thread gets destroyed,
494	but before it is joined. The callback gets passed the terminate arguments,	914	that is, after it's resources have been freed but before it is joined. The
		915	callback gets passed the terminate/cancel arguments, if any, and I<must
495	if any, and I<must not> die, under any circumstances.	916	not> die, under any circumstances.
496		917
497	=cut	918	There can be any number of C<on_destroy> callbacks per coro, and there is
		919	currently no way to remove a callback once added.
498		920
499	sub on_destroy {
500	my ($self, $cb) = @_;
501
502	push @{ $self->{_on_destroy} }, $cb;
503	}
504
505	=item $oldprio = $coroutine->prio ($newprio)	921	=item $oldprio = $coro->prio ($newprio)
506		922
507	Sets (or gets, if the argument is missing) the priority of the	923	Sets (or gets, if the argument is missing) the priority of the
508	coroutine. Higher priority coroutines get run before lower priority	924	coro thread. Higher priority coro get run before lower priority
509	coroutines. Priorities are small signed integers (currently -4 .. +3),	925	coros. Priorities are small signed integers (currently -4 .. +3),
510	that you can refer to using PRIO_xxx constants (use the import tag :prio	926	that you can refer to using PRIO_xxx constants (use the import tag :prio
511	to get then):	927	to get then):
512		928
513	PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN	929	PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN
514	3 > 1 > 0 > -1 > -3 > -4	930	3 > 1 > 0 > -1 > -3 > -4
515		931
516	# set priority to HIGH	932	# set priority to HIGH
517	current->prio(PRIO_HIGH);	933	current->prio (PRIO_HIGH);
518		934
519	The idle coroutine ($Coro::idle) always has a lower priority than any	935	The idle coro thread ($Coro::idle) always has a lower priority than any
520	existing coroutine.	936	existing coro.
521		937
522	Changing the priority of the current coroutine will take effect immediately,	938	Changing the priority of the current coro will take effect immediately,
523	but changing the priority of coroutines in the ready queue (but not	939	but changing the priority of a coro in the ready queue (but not running)
524	running) will only take effect after the next schedule (of that	940	will only take effect after the next schedule (of that coro). This is a
525	coroutine). This is a bug that will be fixed in some future version.	941	bug that will be fixed in some future version.
526		942
527	=item $newprio = $coroutine->nice ($change)	943	=item $newprio = $coro->nice ($change)
528		944
529	Similar to C<prio>, but subtract the given value from the priority (i.e.	945	Similar to C<prio>, but subtract the given value from the priority (i.e.
530	higher values mean lower priority, just as in unix).	946	higher values mean lower priority, just as in UNIX's nice command).
531		947
532	=item $olddesc = $coroutine->desc ($newdesc)	948	=item $olddesc = $coro->desc ($newdesc)
533		949
534	Sets (or gets in case the argument is missing) the description for this	950	Sets (or gets in case the argument is missing) the description for this
535	coroutine. This is just a free-form string you can associate with a	951	coro thread. This is just a free-form string you can associate with a
536	coroutine.	952	coro.
537		953
538	This method simply sets the C<< $coroutine->{desc} >> member to the given	954	This method simply sets the C<< $coro->{desc} >> member to the given
539	string. You can modify this member directly if you wish.	955	string. You can modify this member directly if you wish, and in fact, this
		956	is often preferred to indicate major processing states that can then be
		957	seen for example in a L<Coro::Debug> session:
		958
		959	sub my_long_function {
		960	local $Coro::current->{desc} = "now in my_long_function";
		961	...
		962	$Coro::current->{desc} = "my_long_function: phase 1";
		963	...
		964	$Coro::current->{desc} = "my_long_function: phase 2";
		965	...
		966	}
540		967
541	=cut	968	=cut
542		969
543	sub desc {	970	sub desc {
544	my $old = $_[0]{desc};	971	my $old = $_[0]{desc};
545	$_[0]{desc} = $_[1] if @_ > 1;	972	$_[0]{desc} = $_[1] if @_ > 1;
546	$old;	973	$old;
547	}	974	}
548		975
		976	sub transfer {
		977	require Carp;
		978	Carp::croak ("You must not call ->transfer on Coro objects. Use Coro::State objects or the ->schedule_to method. Caught");
		979	}
		980
549	=back	981	=back
550		982
551	=head2 GLOBAL FUNCTIONS	983	=head1 GLOBAL FUNCTIONS
552		984
553	=over 4	985	=over 4
554		986
555	=item Coro::nready	987	=item Coro::nready
556		988
557	Returns the number of coroutines that are currently in the ready state,	989	Returns the number of coro that are currently in the ready state,
558	i.e. that can be switched to by calling C<schedule> directory or	990	i.e. that can be switched to by calling C<schedule> directory or
559	indirectly. The value C<0> means that the only runnable coroutine is the	991	indirectly. The value C<0> means that the only runnable coro is the
560	currently running one, so C<cede> would have no effect, and C<schedule>	992	currently running one, so C<cede> would have no effect, and C<schedule>
561	would cause a deadlock unless there is an idle handler that wakes up some	993	would cause a deadlock unless there is an idle handler that wakes up some
562	coroutines.	994	coro.
563		995
564	=item my $guard = Coro::guard { ... }	996	=item my $guard = Coro::guard { ... }
565		997
566	This creates and returns a guard object. Nothing happens until the object	998	This function still exists, but is deprecated. Please use the
567	gets destroyed, in which case the codeblock given as argument will be	999	C<Guard::guard> function instead.
568	executed. This is useful to free locks or other resources in case of a
569	runtime error or when the coroutine gets canceled, as in both cases the
570	guard block will be executed. The guard object supports only one method,
571	C<< ->cancel >>, which will keep the codeblock from being executed.
572
573	Example: set some flag and clear it again when the coroutine gets canceled
574	or the function returns:
575
576	sub do_something {
577	my $guard = Coro::guard { $busy = 0 };
578	$busy = 1;
579
580	# do something that requires $busy to be true
581	}
582		1000
583	=cut	1001	=cut
584		1002
585	sub guard(&) {	1003	BEGIN { *guard = \&Guard::guard }
586	bless \(my $cb = $_[0]), "Coro::guard"
587	}
588
589	sub Coro::guard::cancel {
590	${$_[0]} = sub { };
591	}
592
593	sub Coro::guard::DESTROY {
594	${$_[0]}->();
595	}
596
597		1004
598	=item unblock_sub { ... }	1005	=item unblock_sub { ... }
599		1006
600	This utility function takes a BLOCK or code reference and "unblocks" it,	1007	This utility function takes a BLOCK or code reference and "unblocks" it,
601	returning a new coderef. Unblocking means that calling the new coderef	1008	returning a new coderef. Unblocking means that calling the new coderef
602	will return immediately without blocking, returning nothing, while the	1009	will return immediately without blocking, returning nothing, while the
603	original code ref will be called (with parameters) from within another	1010	original code ref will be called (with parameters) from within another
604	coroutine.	1011	coro.
605		1012
606	The reason this function exists is that many event libraries (such as the	1013	The reason this function exists is that many event libraries (such as
607	venerable L<Event|Event> module) are not coroutine-safe (a weaker form	1014	the venerable L<Event|Event> module) are not thread-safe (a weaker form
608	of thread-safety). This means you must not block within event callbacks,	1015	of reentrancy). This means you must not block within event callbacks,
609	otherwise you might suffer from crashes or worse. The only event library	1016	otherwise you might suffer from crashes or worse. The only event library
610	currently known that is safe to use without C<unblock_sub> is L<EV>.	1017	currently known that is safe to use without C<unblock_sub> is L<EV> (but
		1018	you might still run into deadlocks if all event loops are blocked).
		1019
		1020	Coro will try to catch you when you block in the event loop
		1021	("FATAL:$Coro::IDLE blocked itself"), but this is just best effort and
		1022	only works when you do not run your own event loop.
611		1023
612	This function allows your callbacks to block by executing them in another	1024	This function allows your callbacks to block by executing them in another
613	coroutine where it is safe to block. One example where blocking is handy	1025	coro where it is safe to block. One example where blocking is handy
614	is when you use the L<Coro::AIO|Coro::AIO> functions to save results to	1026	is when you use the L<Coro::AIO|Coro::AIO> functions to save results to
615	disk, for example.	1027	disk, for example.
616		1028
617	In short: simply use C<unblock_sub { ... }> instead of C<sub { ... }> when	1029	In short: simply use C<unblock_sub { ... }> instead of C<sub { ... }> when
618	creating event callbacks that want to block.	1030	creating event callbacks that want to block.
619		1031
620	If your handler does not plan to block (e.g. simply sends a message to	1032	If your handler does not plan to block (e.g. simply sends a message to
621	another coroutine, or puts some other coroutine into the ready queue),	1033	another coro, or puts some other coro into the ready queue), there is
622	there is no reason to use C<unblock_sub>.	1034	no reason to use C<unblock_sub>.
623		1035
624	Note that you also need to use C<unblock_sub> for any other callbacks that	1036	Note that you also need to use C<unblock_sub> for any other callbacks that
625	are indirectly executed by any C-based event loop. For example, when you	1037	are indirectly executed by any C-based event loop. For example, when you
626	use a module that uses L<AnyEvent> (and you use L<Coro::AnyEvent>) and it	1038	use a module that uses L<AnyEvent> (and you use L<Coro::AnyEvent>) and it
627	provides callbacks that are the result of some event callback, then you	1039	provides callbacks that are the result of some event callback, then you
…		…
636	# return immediately and can be reused) and because we cannot cede	1048	# return immediately and can be reused) and because we cannot cede
637	# inside an event callback.	1049	# inside an event callback.
638	our $unblock_scheduler = new Coro sub {	1050	our $unblock_scheduler = new Coro sub {
639	while () {	1051	while () {
640	while (my $cb = pop @unblock_queue) {	1052	while (my $cb = pop @unblock_queue) {
641	# this is an inlined copy of async_pool	1053	&async_pool (@$cb);
642	my $coro = (pop @async_pool) || new Coro \&pool_handler;
643		1054
644	$coro->{_invoke} = $cb;
645	$coro->ready;
646	cede; # for short-lived callbacks, this reduces pressure on the coro pool	1055	# for short-lived callbacks, this reduces pressure on the coro pool
		1056	# as the chance is very high that the async_poll coro will be back
		1057	# in the idle state when cede returns
		1058	cede;
647	}	1059	}
648	schedule; # sleep well	1060	schedule; # sleep well
649	}	1061	}
650	};	1062	};
651	$unblock_scheduler->{desc} = "[unblock_sub scheduler]";	1063	$unblock_scheduler->{desc} = "[unblock_sub scheduler]";
…		…
657	unshift @unblock_queue, [$cb, @_];	1069	unshift @unblock_queue, [$cb, @_];
658	$unblock_scheduler->ready;	1070	$unblock_scheduler->ready;
659	}	1071	}
660	}	1072	}
661		1073
		1074	=item $cb = rouse_cb
		1075
		1076	Create and return a "rouse callback". That's a code reference that,
		1077	when called, will remember a copy of its arguments and notify the owner
		1078	coro of the callback.
		1079
		1080	See the next function.
		1081
		1082	=item @args = rouse_wait [$cb]
		1083
		1084	Wait for the specified rouse callback (or the last one that was created in
		1085	this coro).
		1086
		1087	As soon as the callback is invoked (or when the callback was invoked
		1088	before C<rouse_wait>), it will return the arguments originally passed to
		1089	the rouse callback. In scalar context, that means you get the I<last>
		1090	argument, just as if C<rouse_wait> had a C<return ($a1, $a2, $a3...)>
		1091	statement at the end.
		1092
		1093	See the section B<HOW TO WAIT FOR A CALLBACK> for an actual usage example.
		1094
662	=back	1095	=back
663		1096
664	=cut	1097	=cut
665		1098
		1099	for my $module (qw(Channel RWLock Semaphore SemaphoreSet Signal Specific)) {
		1100	my $old = defined &{"Coro::$module\::new"} && \&{"Coro::$module\::new"};
		1101
		1102	*{"Coro::$module\::new"} = sub {
		1103	require "Coro/$module.pm";
		1104
		1105	# some modules have their new predefined in State.xs, some don't
		1106	*{"Coro::$module\::new"} = $old
		1107	if $old;
		1108
		1109	goto &{"Coro::$module\::new"};
		1110	};
		1111	}
		1112
666	1;	1113	1;
667		1114
		1115	=head1 HOW TO WAIT FOR A CALLBACK
		1116
		1117	It is very common for a coro to wait for some callback to be
		1118	called. This occurs naturally when you use coro in an otherwise
		1119	event-based program, or when you use event-based libraries.
		1120
		1121	These typically register a callback for some event, and call that callback
		1122	when the event occured. In a coro, however, you typically want to
		1123	just wait for the event, simplyifying things.
		1124
		1125	For example C<< AnyEvent->child >> registers a callback to be called when
		1126	a specific child has exited:
		1127
		1128	my $child_watcher = AnyEvent->child (pid => $pid, cb => sub { ... });
		1129
		1130	But from within a coro, you often just want to write this:
		1131
		1132	my $status = wait_for_child $pid;
		1133
		1134	Coro offers two functions specifically designed to make this easy,
		1135	C<rouse_cb> and C<rouse_wait>.
		1136
		1137	The first function, C<rouse_cb>, generates and returns a callback that,
		1138	when invoked, will save its arguments and notify the coro that
		1139	created the callback.
		1140
		1141	The second function, C<rouse_wait>, waits for the callback to be called
		1142	(by calling C<schedule> to go to sleep) and returns the arguments
		1143	originally passed to the callback.
		1144
		1145	Using these functions, it becomes easy to write the C<wait_for_child>
		1146	function mentioned above:
		1147
		1148	sub wait_for_child($) {
		1149	my ($pid) = @_;
		1150
		1151	my $watcher = AnyEvent->child (pid => $pid, cb => rouse_cb);
		1152
		1153	my ($rpid, $rstatus) = rouse_wait;
		1154	$rstatus
		1155	}
		1156
		1157	In the case where C<rouse_cb> and C<rouse_wait> are not flexible enough,
		1158	you can roll your own, using C<schedule> and C<ready>:
		1159
		1160	sub wait_for_child($) {
		1161	my ($pid) = @_;
		1162
		1163	# store the current coro in $current,
		1164	# and provide result variables for the closure passed to ->child
		1165	my $current = $Coro::current;
		1166	my ($done, $rstatus);
		1167
		1168	# pass a closure to ->child
		1169	my $watcher = AnyEvent->child (pid => $pid, cb => sub {
		1170	$rstatus = $_[1]; # remember rstatus
		1171	$done = 1; # mark $rstatus as valid
		1172	$current->ready; # wake up the waiting thread
		1173	});
		1174
		1175	# wait until the closure has been called
		1176	schedule while !$done;
		1177
		1178	$rstatus
		1179	}
		1180
		1181
668	=head1 BUGS/LIMITATIONS	1182	=head1 BUGS/LIMITATIONS
669		1183
		1184	=over 4
		1185
		1186	=item fork with pthread backend
		1187
		1188	When Coro is compiled using the pthread backend (which isn't recommended
		1189	but required on many BSDs as their libcs are completely broken), then
		1190	coro will not survive a fork. There is no known workaround except to
		1191	fix your libc and use a saner backend.
		1192
		1193	=item perl process emulation ("threads")
		1194
670	This module is not perl-pseudo-thread-safe. You should only ever use this	1195	This module is not perl-pseudo-thread-safe. You should only ever use this
671	module from the same thread (this requirement might be removed in the	1196	module from the first thread (this requirement might be removed in the
672	future to allow per-thread schedulers, but Coro::State does not yet allow	1197	future to allow per-thread schedulers, but Coro::State does not yet allow
673	this). I recommend disabling thread support and using processes, as this	1198	this). I recommend disabling thread support and using processes, as having
674	is much faster and uses less memory.	1199	the windows process emulation enabled under unix roughly halves perl
		1200	performance, even when not used.
		1201
		1202	Attempts to use threads created in another emulated process will crash
		1203	("cleanly", with a null pointer exception).
		1204
		1205	=item coro switching is not signal safe
		1206
		1207	You must not switch to another coro from within a signal handler (only
		1208	relevant with %SIG - most event libraries provide safe signals), I<unless>
		1209	you are sure you are not interrupting a Coro function.
		1210
		1211	That means you I<MUST NOT> call any function that might "block" the
		1212	current coro - C<cede>, C<schedule> C<< Coro::Semaphore->down >> or
		1213	anything that calls those. Everything else, including calling C<ready>,
		1214	works.
		1215
		1216	=back
		1217
		1218
		1219	=head1 WINDOWS PROCESS EMULATION
		1220
		1221	A great many people seem to be confused about ithreads (for example, Chip
		1222	Salzenberg called me unintelligent, incapable, stupid and gullible,
		1223	while in the same mail making rather confused statements about perl
		1224	ithreads (for example, that memory or files would be shared), showing his
		1225	lack of understanding of this area - if it is hard to understand for Chip,
		1226	it is probably not obvious to everybody).
		1227
		1228	What follows is an ultra-condensed version of my talk about threads in
		1229	scripting languages given on the perl workshop 2009:
		1230
		1231	The so-called "ithreads" were originally implemented for two reasons:
		1232	first, to (badly) emulate unix processes on native win32 perls, and
		1233	secondly, to replace the older, real thread model ("5.005-threads").
		1234
		1235	It does that by using threads instead of OS processes. The difference
		1236	between processes and threads is that threads share memory (and other
		1237	state, such as files) between threads within a single process, while
		1238	processes do not share anything (at least not semantically). That
		1239	means that modifications done by one thread are seen by others, while
		1240	modifications by one process are not seen by other processes.
		1241
		1242	The "ithreads" work exactly like that: when creating a new ithreads
		1243	process, all state is copied (memory is copied physically, files and code
		1244	is copied logically). Afterwards, it isolates all modifications. On UNIX,
		1245	the same behaviour can be achieved by using operating system processes,
		1246	except that UNIX typically uses hardware built into the system to do this
		1247	efficiently, while the windows process emulation emulates this hardware in
		1248	software (rather efficiently, but of course it is still much slower than
		1249	dedicated hardware).
		1250
		1251	As mentioned before, loading code, modifying code, modifying data
		1252	structures and so on is only visible in the ithreads process doing the
		1253	modification, not in other ithread processes within the same OS process.
		1254
		1255	This is why "ithreads" do not implement threads for perl at all, only
		1256	processes. What makes it so bad is that on non-windows platforms, you can
		1257	actually take advantage of custom hardware for this purpose (as evidenced
		1258	by the forks module, which gives you the (i-) threads API, just much
		1259	faster).
		1260
		1261	Sharing data is in the i-threads model is done by transfering data
		1262	structures between threads using copying semantics, which is very slow -
		1263	shared data simply does not exist. Benchmarks using i-threads which are
		1264	communication-intensive show extremely bad behaviour with i-threads (in
		1265	fact, so bad that Coro, which cannot take direct advantage of multiple
		1266	CPUs, is often orders of magnitude faster because it shares data using
		1267	real threads, refer to my talk for details).
		1268
		1269	As summary, i-threads *use* threads to implement processes, while
		1270	the compatible forks module *uses* processes to emulate, uhm,
		1271	processes. I-threads slow down every perl program when enabled, and
		1272	outside of windows, serve no (or little) practical purpose, but
		1273	disadvantages every single-threaded Perl program.
		1274
		1275	This is the reason that I try to avoid the name "ithreads", as it is
		1276	misleading as it implies that it implements some kind of thread model for
		1277	perl, and prefer the name "windows process emulation", which describes the
		1278	actual use and behaviour of it much better.
675		1279
676	=head1 SEE ALSO	1280	=head1 SEE ALSO
677		1281
678	Event-Loop integration: L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>.	1282	Event-Loop integration: L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>.
679		1283
680	Debugging: L<Coro::Debug>.	1284	Debugging: L<Coro::Debug>.
681		1285
682	Support/Utility: L<Coro::Specific>, L<Coro::Util>.	1286	Support/Utility: L<Coro::Specific>, L<Coro::Util>.
683		1287
684	Locking/IPC: L<Coro::Signal>, L<Coro::Channel>, L<Coro::Semaphore>, L<Coro::SemaphoreSet>, L<Coro::RWLock>.	1288	Locking and IPC: L<Coro::Signal>, L<Coro::Channel>, L<Coro::Semaphore>,
		1289	L<Coro::SemaphoreSet>, L<Coro::RWLock>.
685		1290
686	IO/Timers: L<Coro::Timer>, L<Coro::Handle>, L<Coro::Socket>, L<Coro::AIO>.	1291	I/O and Timers: L<Coro::Timer>, L<Coro::Handle>, L<Coro::Socket>, L<Coro::AIO>.
687		1292
688	Compatibility: L<Coro::LWP>, L<Coro::BDB>, L<Coro::Storable>, L<Coro::Select>.	1293	Compatibility with other modules: L<Coro::LWP> (but see also L<AnyEvent::HTTP> for
		1294	a better-working alternative), L<Coro::BDB>, L<Coro::Storable>,
		1295	L<Coro::Select>.
689		1296
690	XS API: L<Coro::MakeMaker>.	1297	XS API: L<Coro::MakeMaker>.
691		1298
692	Low level Configuration, Coroutine Environment: L<Coro::State>.	1299	Low level Configuration, Thread Environment, Continuations: L<Coro::State>.
693		1300
694	=head1 AUTHOR	1301	=head1 AUTHOR
695		1302
696	Marc Lehmann <schmorp@schmorp.de>	1303	Marc Lehmann <schmorp@schmorp.de>
697	http://home.schmorp.de/	1304	http://home.schmorp.de/

Diff Legend

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