ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/Coro/Coro.pm

Comparing Coro/Coro.pm (file contents):
Revision 1.202 by root, Tue Sep 30 17:12:34 2008 UTC vs.
Revision 1.359 by root, Wed Jul 21 06:40:13 2021 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 its 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 its 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 >>, the
		194	thread-equivalent of C<exit>, which works at any subroutine call nesting
		195	level:
		196
		197	async {
		198	Coro::terminate "return value 1", "return value 2";
		199	};
		200
		201	Yet another way is to C<< ->cancel >> (or C<< ->safe_cancel >>) the coro
		202	thread from another thread:
		203
		204	my $coro = async {
		205	exit 1;
		206	};
		207
		208	$coro->cancel; # also accepts values for ->join to retrieve
		209
		210	Cancellation I<can> be dangerous - it's a bit like calling C<exit> without
		211	actually exiting, and might leave C libraries and XS modules in a weird
		212	state. Unlike other thread implementations, however, Coro is exceptionally
		213	safe with regards to cancellation, as perl will always be in a consistent
		214	state, and for those cases where you want to do truly marvellous things
		215	with your coro while it is being cancelled - that is, make sure all
		216	cleanup code is executed from the thread being cancelled - there is even a
		217	C<< ->safe_cancel >> method.
		218
		219	So, cancelling a thread that runs in an XS event loop might not be the
		220	best idea, but any other combination that deals with perl only (cancelling
		221	when a thread is in a C<tie> method or an C<AUTOLOAD> for example) is
		222	safe.
		223
		224	Last not least, a coro thread object that isn't referenced is C<<
		225	->cancel >>'ed automatically - just like other objects in Perl. This
		226	is not such a common case, however - a running thread is referencedy by
		227	C<$Coro::current>, a thread ready to run is referenced by the ready queue,
		228	a thread waiting on a lock or semaphore is referenced by being in some
		229	wait list and so on. But a thread that isn't in any of those queues gets
		230	cancelled:
		231
		232	async {
		233	schedule; # cede to other coros, don't go into the ready queue
		234	};
		235
		236	cede;
		237	# now the async above is destroyed, as it is not referenced by anything.
		238
		239	A slightly embellished example might make it clearer:
		240
		241	async {
		242	my $guard = Guard::guard { print "destroyed\n" };
		243	schedule while 1;
		244	};
		245
		246	cede;
		247
		248	Superficially one might not expect any output - since the C<async>
		249	implements an endless loop, the C<$guard> will not be cleaned up. However,
		250	since the thread object returned by C<async> is not stored anywhere, the
		251	thread is initially referenced because it is in the ready queue, when it
		252	runs it is referenced by C<$Coro::current>, but when it calls C<schedule>,
		253	it gets C<cancel>ed causing the guard object to be destroyed (see the next
		254	section), and printing its message.
		255
		256	If this seems a bit drastic, remember that this only happens when nothing
		257	references the thread anymore, which means there is no way to further
		258	execute it, ever. The only options at this point are leaking the thread,
		259	or cleaning it up, which brings us to...
		260
		261	=item 5. Cleanup
		262
		263	Threads will allocate various resources. Most but not all will be returned
		264	when a thread terminates, during clean-up.
		265
		266	Cleanup is quite similar to throwing an uncaught exception: perl will
		267	work its way up through all subroutine calls and blocks. On its way, it
		268	will release all C<my> variables, undo all C<local>'s and free any other
		269	resources truly local to the thread.
		270
		271	So, a common way to free resources is to keep them referenced only by my
		272	variables:
		273
		274	async {
		275	my $big_cache = new Cache ...;
		276	};
		277
		278	If there are no other references, then the C<$big_cache> object will be
		279	freed when the thread terminates, regardless of how it does so.
		280
		281	What it does C<NOT> do is unlock any Coro::Semaphores or similar
		282	resources, but that's where the C<guard> methods come in handy:
		283
		284	my $sem = new Coro::Semaphore;
		285
		286	async {
		287	my $lock_guard = $sem->guard;
		288	# if we return, or die or get cancelled, here,
		289	# then the semaphore will be "up"ed.
		290	};
		291
		292	The C<Guard::guard> function comes in handy for any custom cleanup you
		293	might want to do (but you cannot switch to other coroutines from those
		294	code blocks):
		295
		296	async {
		297	my $window = new Gtk2::Window "toplevel";
		298	# The window will not be cleaned up automatically, even when $window
		299	# gets freed, so use a guard to ensure its destruction
		300	# in case of an error:
		301	my $window_guard = Guard::guard { $window->destroy };
		302
		303	# we are safe here
		304	};
		305
		306	Last not least, C<local> can often be handy, too, e.g. when temporarily
		307	replacing the coro thread description:
		308
		309	sub myfunction {
		310	local $Coro::current->{desc} = "inside myfunction(@_)";
		311
		312	# if we return or die here, the description will be restored
		313	}
		314
		315	=item 6. Viva La Zombie Muerte
		316
		317	Even after a thread has terminated and cleaned up its resources, the Coro
		318	object still is there and stores the return values of the thread.
		319
		320	When there are no other references, it will simply be cleaned up and
		321	freed.
		322
		323	If there areany references, the Coro object will stay around, and you
		324	can call C<< ->join >> as many times as you wish to retrieve the result
		325	values:
		326
		327	async {
		328	print "hi\n";
		329	1
		330	};
		331
		332	# run the async above, and free everything before returning
		333	# from Coro::cede:
		334	Coro::cede;
		335
		336	{
		337	my $coro = async {
		338	print "hi\n";
		339	1
		340	};
		341
		342	# run the async above, and clean up, but do not free the coro
		343	# object:
		344	Coro::cede;
		345
		346	# optionally retrieve the result values
		347	my @results = $coro->join;
		348
		349	# now $coro goes out of scope, and presumably gets freed
		350	};
		351
		352	=back
56		353
57	=cut	354	=cut
58		355
59	package Coro;	356	package Coro;
60		357
61	use strict;	358	use common::sense;
62	no warnings "uninitialized";	359
		360	use Carp ();
		361
		362	use Guard ();
63		363
64	use Coro::State;	364	use Coro::State;
65		365
66	use base qw(Coro::State Exporter);	366	use base qw(Coro::State Exporter);
67		367
68	our $idle; # idle handler	368	our $idle; # idle handler
69	our $main; # main coroutine	369	our $main; # main coro
70	our $current; # current coroutine	370	our $current; # current coro
71		371
72	our $VERSION = 4.8;	372	our $VERSION = 6.57;
73		373
74	our @EXPORT = qw(async async_pool cede schedule terminate current unblock_sub);	374	our @EXPORT = qw(async async_pool cede schedule terminate current unblock_sub rouse_cb rouse_wait);
75	our %EXPORT_TAGS = (	375	our %EXPORT_TAGS = (
76	prio => [qw(PRIO_MAX PRIO_HIGH PRIO_NORMAL PRIO_LOW PRIO_IDLE PRIO_MIN)],	376	prio => [qw(PRIO_MAX PRIO_HIGH PRIO_NORMAL PRIO_LOW PRIO_IDLE PRIO_MIN)],
77	);	377	);
78	our @EXPORT_OK = (@{$EXPORT_TAGS{prio}}, qw(nready));	378	our @EXPORT_OK = (@{$EXPORT_TAGS{prio}}, qw(nready));
79		379
		380	=head1 GLOBAL VARIABLES
		381
80	=over 4	382	=over 4
81		383
82	=item $Coro::main	384	=item $Coro::main
83		385
84	This variable stores the coroutine object that represents the main	386	This variable stores the Coro object that represents the main
85	program. While you cna C<ready> it and do most other things you can do to	387	program. While you can C<ready> it and do most other things you can do to
86	coroutines, it is mainly useful to compare again C<$Coro::current>, to see	388	coro, it is mainly useful to compare again C<$Coro::current>, to see
87	whether you are running in the main program or not.	389	whether you are running in the main program or not.
88		390
89	=cut	391	=cut
90		392
91	$main = new Coro;	393	# $main is now being initialised by Coro::State
92		394
93	=item $Coro::current	395	=item $Coro::current
94		396
95	The coroutine object representing the current coroutine (the last	397	The Coro object representing the current coro (the last
96	coroutine that the Coro scheduler switched to). The initial value is	398	coro that the Coro scheduler switched to). The initial value is
97	C<$main> (of course).	399	C<$Coro::main> (of course).
98		400
99	This variable is B<strictly> I<read-only>. You can take copies of the	401	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	402	value stored in it and use it as any other Coro object, but you must
101	not otherwise modify the variable itself.	403	not otherwise modify the variable itself.
102		404
103	=cut	405	=cut
104		406
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]	407	sub current() { $current } # [DEPRECATED]
114		408
115	=item $Coro::idle	409	=item $Coro::idle
116		410
117	This variable is mainly useful to integrate Coro into event loops. It is	411	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	412	usually better to rely on L<Coro::AnyEvent> or L<Coro::EV>, as this is
119	pretty low-level functionality.	413	pretty low-level functionality.
120		414
121	This variable stores a callback that is called whenever the scheduler	415	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	416	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		417
		418	The default implementation dies with "FATAL: deadlock detected.", followed
		419	by a thread listing, because the program has no other way to continue.
		420
126	This hook is overwritten by modules such as C<Coro::Timer> and	421	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	422	C<Coro::AnyEvent> to wait on an external event that hopefully wakes up a
128	coroutine so the scheduler can run it.	423	coro so the scheduler can run it.
129		424
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	425	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		426
141	=cut	427	=cut
142		428
143	$idle = sub {	429	# ||= because other modules could have provided their own by now
144	require Carp;	430	$idle ||= new Coro sub {
145	Carp::croak ("FATAL: deadlock detected");	431	require Coro::Debug;
		432	die "FATAL: deadlock detected.\n"
		433	. Coro::Debug::ps_listing ();
146	};	434	};
147		435
148	sub _cancel {
149	my ($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	436	# this coro is necessary because a coro
161	# cannot destroy itself.	437	# cannot destroy itself.
162	my @destroy;	438	our @destroy;
163	my $manager;	439	our $manager;
164		440
165	$manager = new Coro sub {	441	$manager = new Coro sub {
166	while () {	442	while () {
167	(shift @destroy)->_cancel	443	_destroy shift @destroy
168	while @destroy;	444	while @destroy;
169		445
170	&schedule;	446	&schedule;
171	}	447	}
172	};	448	};
173	$manager->desc ("[coro manager]");	449	$manager->{desc} = "[coro manager]";
174	$manager->prio (PRIO_MAX);	450	$manager->prio (PRIO_MAX);
175		451
176	=back	452	=back
177		453
178	=head2 SIMPLE COROUTINE CREATION	454	=head1 SIMPLE CORO CREATION
179		455
180	=over 4	456	=over 4
181		457
182	=item async { ... } [@args...]	458	=item async { ... } [@args...]
183		459
184	Create a new coroutine and return it's coroutine object (usually	460	Create a new coro and return its Coro object (usually
185	unused). The coroutine will be put into the ready queue, so	461	unused). The coro will be put into the ready queue, so
186	it will start running automatically on the next scheduler run.	462	it will start running automatically on the next scheduler run.
187		463
188	The first argument is a codeblock/closure that should be executed in the	464	The first argument is a codeblock/closure that should be executed in the
189	coroutine. When it returns argument returns the coroutine is automatically	465	coro. When it returns argument returns the coro is automatically
190	terminated.	466	terminated.
191		467
192	The remaining arguments are passed as arguments to the closure.	468	The remaining arguments are passed as arguments to the closure.
193		469
194	See the C<Coro::State::new> constructor for info about the coroutine	470	See the C<Coro::State::new> constructor for info about the coro
195	environment in which coroutines are executed.	471	environment in which coro are executed.
196		472
197	Calling C<exit> in a coroutine will do the same as calling exit outside	473	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,	474	the coro. Likewise, when the coro dies, the program will exit,
199	just as it would in the main program.	475	just as it would in the main program.
200		476
201	If you do not want that, you can provide a default C<die> handler, or	477	If you do not want that, you can provide a default C<die> handler, or
202	simply avoid dieing (by use of C<eval>).	478	simply avoid dieing (by use of C<eval>).
203		479
204	Example: Create a new coroutine that just prints its arguments.	480	Example: Create a new coro that just prints its arguments.
205		481
206	async {	482	async {
207	print "@_\n";	483	print "@_\n";
208	} 1,2,3,4;	484	} 1,2,3,4;
209		485
210	=cut
211
212	sub async(&@) {
213	my $coro = new Coro @_;
214	$coro->ready;
215	$coro
216	}
217
218	=item async_pool { ... } [@args...]	486	=item async_pool { ... } [@args...]
219		487
220	Similar to C<async>, but uses a coroutine pool, so you should not call	488	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	489	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	490	coro that might have executed other code already (which can be good
223	or bad :).	491	or bad :).
224		492
225	On the plus side, this function is faster than creating (and destroying)	493	On the plus side, this function is about twice as fast as creating (and
226	a completely new coroutine, so if you need a lot of generic coroutines in	494	destroying) a completely new coro, so if you need a lot of generic
227	quick successsion, use C<async_pool>, not C<async>.	495	coros in quick successsion, use C<async_pool>, not C<async>.
228		496
229	The code block is executed in an C<eval> context and a warning will be	497	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	498	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>	499	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,	500	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	501	which somehow defeats the purpose of pooling (but is fine in the
234	exceptional case).	502	exceptional case).
235		503
236	The priority will be reset to C<0> after each run, tracing will be	504	The priority will be reset to C<0> after each run, all C<swap_sv> calls
237	disabled, the description will be reset and the default output filehandle	505	will be undone, tracing will be disabled, the description will be reset
238	gets restored, so you can change all these. Otherwise the coroutine will	506	and the default output filehandle gets restored, so you can change all
239	be re-used "as-is": most notably if you change other per-coroutine global	507	these. Otherwise the coro will be re-used "as-is": most notably if you
240	stuff such as C<$/> you I<must needs> to revert that change, which is most	508	change other per-coro global stuff such as C<$/> you I<must needs> revert
241	simply done by using local as in: C< local $/ >.	509	that change, which is most simply done by using local as in: C<< local $/
		510	>>.
242		511
243	The pool size is limited to C<8> idle coroutines (this can be adjusted by	512	The idle pool size is limited to C<8> idle coros (this can be
244	changing $Coro::POOL_SIZE), and there can be as many non-idle coros as	513	adjusted by changing $Coro::POOL_SIZE), but there can be as many non-idle
245	required.	514	coros as required.
246		515
247	If you are concerned about pooled coroutines growing a lot because a	516	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	517	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	518	{ 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	519	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.	520	(adjustable via $Coro::POOL_RSS) it will also be destroyed.
252		521
253	=cut	522	=cut
254		523
255	our $POOL_SIZE = 8;	524	our $POOL_SIZE = 8;
256	our $POOL_RSS = 16 * 1024;	525	our $POOL_RSS = 32 * 1024;
257	our @async_pool;	526	our @async_pool;
258		527
259	sub pool_handler {	528	sub pool_handler {
260	my $cb;
261
262	while () {	529	while () {
263	eval {	530	eval {
264	while () {	531	&{&_pool_handler} while 1;
265	_pool_1 $cb;
266	&$cb;
267	_pool_2 $cb;
268	&schedule;
269	}
270	};	532	};
271		533
272	if ($@) {
273	last if $@ eq "\3async_pool terminate\2\n";
274	warn $@;	534	warn $@ if $@;
275	}
276	}	535	}
277	}	536	}
278		537
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	$coro
287	}
288
289	=back	538	=back
290		539
291	=head2 STATIC METHODS	540	=head1 STATIC METHODS
292		541
293	Static methods are actually functions that operate on the current coroutine.	542	Static methods are actually functions that implicitly operate on the
		543	current coro.
294		544
295	=over 4	545	=over 4
296		546
297	=item schedule	547	=item schedule
298		548
299	Calls the scheduler. The scheduler will find the next coroutine that is	549	Calls the scheduler. The scheduler will find the next coro that is
300	to be run from the ready queue and switches to it. The next coroutine	550	to be run from the ready queue and switches to it. The next coro
301	to be run is simply the one with the highest priority that is longest	551	to be run is simply the one with the highest priority that is longest
302	in its ready queue. If there is no coroutine ready, it will clal the	552	in its ready queue. If there is no coro ready, it will call the
303	C<$Coro::idle> hook.	553	C<$Coro::idle> hook.
304		554
305	Please note that the current coroutine will I<not> be put into the ready	555	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	556	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 >>,	557	again unless something else (e.g. an event handler) calls C<< ->ready >>,
308	thus waking you up.	558	thus waking you up.
309		559
310	This makes C<schedule> I<the> generic method to use to block the current	560	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	561	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	562	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	563	>> 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,	564	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	565	so you need to check whether the event indeed happened, e.g. by storing the
316	status in a variable.	566	status in a variable.
317		567
318	The canonical way to wait on external events is this:	568	See B<HOW TO WAIT FOR A CALLBACK>, below, for some ways to wait for callbacks.
319		569
320	{	570	=item cede
321	# remember current coroutine
322	my $current = $Coro::current;
323		571
324	# register a hypothetical event handler	572	"Cede" to other coros. This function puts the current coro into
325	on_event_invoke sub {	573	the ready queue and calls C<schedule>, which has the effect of giving
326	# wake up sleeping coroutine	574	up the current "timeslice" to other coros of the same or higher
327	$current->ready;	575	priority. Once your coro gets its turn again it will automatically be
328	undef $current;	576	resumed.
		577
		578	This function is often called C<yield> in other languages.
		579
		580	=item Coro::cede_notself
		581
		582	Works like cede, but is not exported by default and will cede to I<any>
		583	coro, regardless of priority. This is useful sometimes to ensure
		584	progress is made.
		585
		586	=item terminate [arg...]
		587
		588	Terminates the current coro with the given status values (see
		589	L<cancel>). The values will not be copied, but referenced directly.
		590
		591	=item Coro::on_enter BLOCK, Coro::on_leave BLOCK
		592
		593	These function install enter and leave winders in the current scope. The
		594	enter block will be executed when on_enter is called and whenever the
		595	current coro is re-entered by the scheduler, while the leave block is
		596	executed whenever the current coro is blocked by the scheduler, and
		597	also when the containing scope is exited (by whatever means, be it exit,
		598	die, last etc.).
		599
		600	I<Neither invoking the scheduler, nor exceptions, are allowed within those
		601	BLOCKs>. That means: do not even think about calling C<die> without an
		602	eval, and do not even think of entering the scheduler in any way.
		603
		604	Since both BLOCKs are tied to the current scope, they will automatically
		605	be removed when the current scope exits.
		606
		607	These functions implement the same concept as C<dynamic-wind> in scheme
		608	does, and are useful when you want to localise some resource to a specific
		609	coro.
		610
		611	They slow down thread switching considerably for coros that use them
		612	(about 40% for a BLOCK with a single assignment, so thread switching is
		613	still reasonably fast if the handlers are fast).
		614
		615	These functions are best understood by an example: The following function
		616	will change the current timezone to "Antarctica/South_Pole", which
		617	requires a call to C<tzset>, but by using C<on_enter> and C<on_leave>,
		618	which remember/change the current timezone and restore the previous
		619	value, respectively, the timezone is only changed for the coro that
		620	installed those handlers.
		621
		622	use POSIX qw(tzset);
		623
		624	async {
		625	my $old_tz; # store outside TZ value here
		626
		627	Coro::on_enter {
		628	$old_tz = $ENV{TZ}; # remember the old value
		629
		630	$ENV{TZ} = "Antarctica/South_Pole";
		631	tzset; # enable new value
329	};	632	};
330		633
331	# call schedule until event occurred.	634	Coro::on_leave {
332	# in case we are woken up for other reasons	635	$ENV{TZ} = $old_tz;
333	# (current still defined), loop.	636	tzset; # restore old value
334	Coro::schedule while $current;	637	};
		638
		639	# at this place, the timezone is Antarctica/South_Pole,
		640	# without disturbing the TZ of any other coro.
		641	};
		642
		643	This can be used to localise about any resource (locale, uid, current
		644	working directory etc.) to a block, despite the existence of other
		645	coros.
		646
		647	Another interesting example implements time-sliced multitasking using
		648	interval timers (this could obviously be optimised, but does the job):
		649
		650	# "timeslice" the given block
		651	sub timeslice(&) {
		652	use Time::HiRes ();
		653
		654	Coro::on_enter {
		655	# on entering the thread, we set an VTALRM handler to cede
		656	$SIG{VTALRM} = sub { cede };
		657	# and then start the interval timer
		658	Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0.01, 0.01;
		659	};
		660	Coro::on_leave {
		661	# on leaving the thread, we stop the interval timer again
		662	Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0, 0;
		663	};
		664
		665	&{+shift};
335	}	666	}
336		667
337	=item cede	668	# use like this:
		669	timeslice {
		670	# The following is an endless loop that would normally
		671	# monopolise the process. Since it runs in a timesliced
		672	# environment, it will regularly cede to other threads.
		673	while () { }
		674	};
338		675
339	"Cede" to other coroutines. This function puts the current coroutine into
340	the ready queue and calls C<schedule>, which has the effect of giving
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
345	This function is often called C<yield> in other languages.
346
347	=item Coro::cede_notself
348
349	Works like cede, but is not exported by default and will cede to I<any>
350	coroutine, regardless of priority. This is useful sometimes to ensure
351	progress is made.
352
353	=item terminate [arg...]
354
355	Terminates the current coroutine with the given status values (see L<cancel>).
356		676
357	=item killall	677	=item killall
358		678
359	Kills/terminates/cancels all coroutines except the currently running	679	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		680
363	Note that while this will try to free some of the main programs resources,	681	Note that while this will try to free some of the main interpreter
		682	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	683	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.	684	calls this function, there will be some one-time resource leak.
366		685
367	=cut	686	=cut
368
369	sub terminate {
370	$current->cancel (@_);
371	}
372		687
373	sub killall {	688	sub killall {
374	for (Coro::State::list) {	689	for (Coro::State::list) {
375	$_->cancel	690	$_->cancel
376	if $_ != $current && UNIVERSAL::isa $_, "Coro";	691	if $_ != $current && UNIVERSAL::isa $_, "Coro";
377	}	692	}
378	}	693	}
379		694
380	=back	695	=back
381		696
382	=head2 COROUTINE METHODS	697	=head1 CORO OBJECT METHODS
383		698
384	These are the methods you can call on coroutine objects (or to create	699	These are the methods you can call on coro objects (or to create
385	them).	700	them).
386		701
387	=over 4	702	=over 4
388		703
389	=item new Coro \&sub [, @args...]	704	=item new Coro \&sub [, @args...]
390		705
391	Create a new coroutine and return it. When the sub returns, the coroutine	706	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	707	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	708	called. To make the coro run you must first put it into the ready
394	queue by calling the ready method.	709	queue by calling the ready method.
395		710
396	See C<async> and C<Coro::State::new> for additional info about the	711	See C<async> and C<Coro::State::new> for additional info about the
397	coroutine environment.	712	coro environment.
398		713
399	=cut	714	=cut
400		715
401	sub _run_coro {	716	sub _coro_run {
402	terminate &{+shift};	717	terminate &{+shift};
403	}	718	}
404		719
405	sub new {
406	my $class = shift;
407
408	$class->SUPER::new (\&_run_coro, @_)
409	}
410
411	=item $success = $coroutine->ready	720	=item $success = $coro->ready
412		721
413	Put the given coroutine into the end of its ready queue (there is one	722	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	723	queue for each priority) and return true. If the coro is already in
415	the ready queue, do nothing and return false.	724	the ready queue, do nothing and return false.
416		725
417	This ensures that the scheduler will resume this coroutine automatically	726	This ensures that the scheduler will resume this coro automatically
418	once all the coroutines of higher priority and all coroutines of the same	727	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.	728	priority that were put into the ready queue earlier have been resumed.
420		729
		730	=item $coro->suspend
		731
		732	Suspends the specified coro. A suspended coro works just like any other
		733	coro, except that the scheduler will not select a suspended coro for
		734	execution.
		735
		736	Suspending a coro can be useful when you want to keep the coro from
		737	running, but you don't want to destroy it, or when you want to temporarily
		738	freeze a coro (e.g. for debugging) to resume it later.
		739
		740	A scenario for the former would be to suspend all (other) coros after a
		741	fork and keep them alive, so their destructors aren't called, but new
		742	coros can be created.
		743
		744	=item $coro->resume
		745
		746	If the specified coro was suspended, it will be resumed. Note that when
		747	the coro was in the ready queue when it was suspended, it might have been
		748	unreadied by the scheduler, so an activation might have been lost.
		749
		750	To avoid this, it is best to put a suspended coro into the ready queue
		751	unconditionally, as every synchronisation mechanism must protect itself
		752	against spurious wakeups, and the one in the Coro family certainly do
		753	that.
		754
		755	=item $state->is_new
		756
		757	Returns true iff this Coro object is "new", i.e. has never been run
		758	yet. Those states basically consist of only the code reference to call and
		759	the arguments, but consumes very little other resources. New states will
		760	automatically get assigned a perl interpreter when they are transferred to.
		761
		762	=item $state->is_zombie
		763
		764	Returns true iff the Coro object has been cancelled, i.e.
		765	its resources freed because they were C<cancel>'ed, C<terminate>'d,
		766	C<safe_cancel>'ed or simply went out of scope.
		767
		768	The name "zombie" stems from UNIX culture, where a process that has
		769	exited and only stores and exit status and no other resources is called a
		770	"zombie".
		771
421	=item $is_ready = $coroutine->is_ready	772	=item $is_ready = $coro->is_ready
422		773
423	Return whether the coroutine is currently the ready queue or not,	774	Returns true iff the Coro object is in the ready queue. Unless the Coro
		775	object gets destroyed, it will eventually be scheduled by the scheduler.
424		776
		777	=item $is_running = $coro->is_running
		778
		779	Returns true iff the Coro object is currently running. Only one Coro object
		780	can ever be in the running state (but it currently is possible to have
		781	multiple running Coro::States).
		782
		783	=item $is_suspended = $coro->is_suspended
		784
		785	Returns true iff this Coro object has been suspended. Suspended Coros will
		786	not ever be scheduled.
		787
425	=item $coroutine->cancel (arg...)	788	=item $coro->cancel ($arg...)
426		789
427	Terminates the given coroutine and makes it return the given arguments as	790	Terminate the given Coro thread and make it return the given arguments as
428	status (default: the empty list). Never returns if the coroutine is the	791	status (default: an empty list). Never returns if the Coro is the
429	current coroutine.	792	current Coro.
430		793
431	=cut	794	This is a rather brutal way to free a coro, with some limitations - if
		795	the thread is inside a C callback that doesn't expect to be canceled,
		796	bad things can happen, or if the cancelled thread insists on running
		797	complicated cleanup handlers that rely on its thread context, things will
		798	not work.
432		799
433	sub cancel {	800	Any cleanup code being run (e.g. from C<guard> blocks, destructors and so
434	my $self = shift;	801	on) will be run without a thread context, and is not allowed to switch
435	$self->{_status} = [@_];	802	to other threads. A common mistake is to call C<< ->cancel >> from a
		803	destructor called by die'ing inside the thread to be cancelled for
		804	example.
436		805
437	if ($current == $self) {	806	On the plus side, C<< ->cancel >> will always clean up the thread, no
438	push @destroy, $self;	807	matter what. If your cleanup code is complex or you want to avoid
439	$manager->ready;	808	cancelling a C-thread that doesn't know how to clean up itself, it can be
440	&schedule while 1;	809	better to C<< ->throw >> an exception, or use C<< ->safe_cancel >>.
441	} else {	810
442	$self->_cancel;	811	The arguments to C<< ->cancel >> are not copied, but instead will
		812	be referenced directly (e.g. if you pass C<$var> and after the call
		813	change that variable, then you might change the return values passed to
		814	e.g. C<join>, so don't do that).
		815
		816	The resources of the Coro are usually freed (or destructed) before this
		817	call returns, but this can be delayed for an indefinite amount of time, as
		818	in some cases the manager thread has to run first to actually destruct the
		819	Coro object.
		820
		821	=item $coro->safe_cancel ($arg...)
		822
		823	Works mostly like C<< ->cancel >>, but is inherently "safer", and
		824	consequently, can fail with an exception in cases the thread is not in a
		825	cancellable state. Essentially, C<< ->safe_cancel >> is a C<< ->cancel >>
		826	with extra checks before canceling.
		827
		828	It works a bit like throwing an exception that cannot be caught -
		829	specifically, it will clean up the thread from within itself, so all
		830	cleanup handlers (e.g. C<guard> blocks) are run with full thread
		831	context and can block if they wish. The downside is that there is no
		832	guarantee that the thread can be cancelled when you call this method, and
		833	therefore, it might fail. It is also considerably slower than C<cancel> or
		834	C<terminate>.
		835
		836	A thread is in a safe-cancellable state if it either has never been run
		837	yet, has already been canceled/terminated or otherwise destroyed, or has
		838	no C context attached and is inside an SLF function.
		839
		840	The first two states are trivial - a thread that has not started or has
		841	already finished is safe to cancel.
		842
		843	The last state basically means that the thread isn't currently inside a
		844	perl callback called from some C function (usually via some XS modules)
		845	and isn't currently executing inside some C function itself (via Coro's XS
		846	API).
		847
		848	This call returns true when it could cancel the thread, or croaks with an
		849	error otherwise (i.e. it either returns true or doesn't return at all).
		850
		851	Why the weird interface? Well, there are two common models on how and
		852	when to cancel things. In the first, you have the expectation that your
		853	coro thread can be cancelled when you want to cancel it - if the thread
		854	isn't cancellable, this would be a bug somewhere, so C<< ->safe_cancel >>
		855	croaks to notify of the bug.
		856
		857	In the second model you sometimes want to ask nicely to cancel a thread,
		858	but if it's not a good time, well, then don't cancel. This can be done
		859	relatively easy like this:
		860
		861	if (! eval { $coro->safe_cancel }) {
		862	warn "unable to cancel thread: $@";
443	}	863	}
444	}
445		864
446	=item $coroutine->join	865	However, what you never should do is first try to cancel "safely" and
		866	if that fails, cancel the "hard" way with C<< ->cancel >>. That makes
		867	no sense: either you rely on being able to execute cleanup code in your
		868	thread context, or you don't. If you do, then C<< ->safe_cancel >> is the
		869	only way, and if you don't, then C<< ->cancel >> is always faster and more
		870	direct.
447		871
448	Wait until the coroutine terminates and return any values given to the	872	=item $coro->schedule_to
449	C<terminate> or C<cancel> functions. C<join> can be called concurrently
450	from multiple coroutines, and all will be resumed and given the status
451	return once the C<$coroutine> terminates.
452		873
453	=cut	874	Puts the current coro to sleep (like C<Coro::schedule>), but instead
		875	of continuing with the next coro from the ready queue, always switch to
		876	the given coro object (regardless of priority etc.). The readyness
		877	state of that coro isn't changed.
454		878
455	sub join {	879	This is an advanced method for special cases - I'd love to hear about any
456	my $self = shift;	880	uses for this one.
457		881
458	unless ($self->{_status}) {	882	=item $coro->cede_to
459	my $current = $current;
460		883
461	push @{$self->{_on_destroy}}, sub {	884	Like C<schedule_to>, but puts the current coro into the ready
462	$current->ready;	885	queue. This has the effect of temporarily switching to the given
463	undef $current;	886	coro, and continuing some time later.
464	};
465		887
466	&schedule while $current;	888	This is an advanced method for special cases - I'd love to hear about any
467	}	889	uses for this one.
468		890
469	wantarray ? @{$self->{_status}} : $self->{_status}[0];
470	}
471
472	=item $coroutine->on_destroy (\&cb)
473
474	Registers a callback that is called when this coroutine gets destroyed,
475	but before it is joined. The callback gets passed the terminate arguments,
476	if any, and I<must not> die, under any circumstances.
477
478	=cut
479
480	sub on_destroy {
481	my ($self, $cb) = @_;
482
483	push @{ $self->{_on_destroy} }, $cb;
484	}
485
486	=item $oldprio = $coroutine->prio ($newprio)
487
488	Sets (or gets, if the argument is missing) the priority of the
489	coroutine. Higher priority coroutines get run before lower priority
490	coroutines. Priorities are small signed integers (currently -4 .. +3),
491	that you can refer to using PRIO_xxx constants (use the import tag :prio
492	to get then):
493
494	PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN
495	3 > 1 > 0 > -1 > -3 > -4
496
497	# set priority to HIGH
498	current->prio(PRIO_HIGH);
499
500	The idle coroutine ($Coro::idle) always has a lower priority than any
501	existing coroutine.
502
503	Changing the priority of the current coroutine will take effect immediately,
504	but changing the priority of coroutines in the ready queue (but not
505	running) will only take effect after the next schedule (of that
506	coroutine). This is a bug that will be fixed in some future version.
507
508	=item $newprio = $coroutine->nice ($change)
509
510	Similar to C<prio>, but subtract the given value from the priority (i.e.
511	higher values mean lower priority, just as in unix).
512
513	=item $olddesc = $coroutine->desc ($newdesc)
514
515	Sets (or gets in case the argument is missing) the description for this
516	coroutine. This is just a free-form string you can associate with a coroutine.
517
518	This method simply sets the C<< $coroutine->{desc} >> member to the given string. You
519	can modify this member directly if you wish.
520
521	=item $coroutine->throw ([$scalar])	891	=item $coro->throw ([$scalar])
522		892
523	If C<$throw> is specified and defined, it will be thrown as an exception	893	If C<$throw> is specified and defined, it will be thrown as an exception
524	inside the coroutine at the next convinient point in time (usually after	894	inside the coro at the next convenient point in time. Otherwise
525	it gains control at the next schedule/transfer/cede). Otherwise clears the
526	exception object.	895	clears the exception object.
		896
		897	Coro will check for the exception each time a schedule-like-function
		898	returns, i.e. after each C<schedule>, C<cede>, C<< Coro::Semaphore->down
		899	>>, C<< Coro::Handle->readable >> and so on. Most of those functions (all
		900	that are part of Coro itself) detect this case and return early in case an
		901	exception is pending.
527		902
528	The exception object will be thrown "as is" with the specified scalar in	903	The exception object will be thrown "as is" with the specified scalar in
529	C<$@>, i.e. if it is a string, no line number or newline will be appended	904	C<$@>, i.e. if it is a string, no line number or newline will be appended
530	(unlike with C<die>).	905	(unlike with C<die>).
531		906
532	This can be used as a softer means than C<cancel> to ask a coroutine to	907	This can be used as a softer means than either C<cancel> or C<safe_cancel
533	end itself, although there is no guarentee that the exception will lead to	908	>to ask a coro to end itself, although there is no guarantee that the
534	termination, and if the exception isn't caught it might well end the whole	909	exception will lead to termination, and if the exception isn't caught it
535	program.	910	might well end the whole program.
		911
		912	You might also think of C<throw> as being the moral equivalent of
		913	C<kill>ing a coro with a signal (in this case, a scalar).
		914
		915	=item $coro->join
		916
		917	Wait until the coro terminates and return any values given to the
		918	C<terminate> or C<cancel> functions. C<join> can be called concurrently
		919	from multiple threads, and all will be resumed and given the status
		920	return once the C<$coro> terminates.
		921
		922	=item $coro->on_destroy (\&cb)
		923
		924	Registers a callback that is called when this coro thread gets destroyed,
		925	that is, after its resources have been freed but before it is joined. The
		926	callback gets passed the terminate/cancel arguments, if any, and I<must
		927	not> die, under any circumstances.
		928
		929	There can be any number of C<on_destroy> callbacks per coro, and there is
		930	currently no way to remove a callback once added.
		931
		932	=item $oldprio = $coro->prio ($newprio)
		933
		934	Sets (or gets, if the argument is missing) the priority of the
		935	coro thread. Higher priority coro get run before lower priority
		936	coros. Priorities are small signed integers (currently -4 .. +3),
		937	that you can refer to using PRIO_xxx constants (use the import tag :prio
		938	to get then):
		939
		940	PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN
		941	3 > 1 > 0 > -1 > -3 > -4
		942
		943	# set priority to HIGH
		944	current->prio (PRIO_HIGH);
		945
		946	The idle coro thread ($Coro::idle) always has a lower priority than any
		947	existing coro.
		948
		949	Changing the priority of the current coro will take effect immediately,
		950	but changing the priority of a coro in the ready queue (but not running)
		951	will only take effect after the next schedule (of that coro). This is a
		952	bug that will be fixed in some future version.
		953
		954	=item $newprio = $coro->nice ($change)
		955
		956	Similar to C<prio>, but subtract the given value from the priority (i.e.
		957	higher values mean lower priority, just as in UNIX's nice command).
		958
		959	=item $olddesc = $coro->desc ($newdesc)
		960
		961	Sets (or gets in case the argument is missing) the description for this
		962	coro thread. This is just a free-form string you can associate with a
		963	coro.
		964
		965	This method simply sets the C<< $coro->{desc} >> member to the given
		966	string. You can modify this member directly if you wish, and in fact, this
		967	is often preferred to indicate major processing states that can then be
		968	seen for example in a L<Coro::Debug> session:
		969
		970	sub my_long_function {
		971	local $Coro::current->{desc} = "now in my_long_function";
		972	...
		973	$Coro::current->{desc} = "my_long_function: phase 1";
		974	...
		975	$Coro::current->{desc} = "my_long_function: phase 2";
		976	...
		977	}
536		978
537	=cut	979	=cut
538		980
539	sub desc {	981	sub desc {
540	my $old = $_[0]{desc};	982	my $old = $_[0]{desc};
541	$_[0]{desc} = $_[1] if @_ > 1;	983	$_[0]{desc} = $_[1] if @_ > 1;
542	$old;	984	$old;
543	}	985	}
544		986
		987	sub transfer {
		988	require Carp;
		989	Carp::croak ("You must not call ->transfer on Coro objects. Use Coro::State objects or the ->schedule_to method. Caught");
		990	}
		991
545	=back	992	=back
546		993
547	=head2 GLOBAL FUNCTIONS	994	=head1 GLOBAL FUNCTIONS
548		995
549	=over 4	996	=over 4
550		997
551	=item Coro::nready	998	=item Coro::nready
552		999
553	Returns the number of coroutines that are currently in the ready state,	1000	Returns the number of coro that are currently in the ready state,
554	i.e. that can be switched to by calling C<schedule> directory or	1001	i.e. that can be switched to by calling C<schedule> directory or
555	indirectly. The value C<0> means that the only runnable coroutine is the	1002	indirectly. The value C<0> means that the only runnable coro is the
556	currently running one, so C<cede> would have no effect, and C<schedule>	1003	currently running one, so C<cede> would have no effect, and C<schedule>
557	would cause a deadlock unless there is an idle handler that wakes up some	1004	would cause a deadlock unless there is an idle handler that wakes up some
558	coroutines.	1005	coro.
559		1006
560	=item my $guard = Coro::guard { ... }	1007	=item my $guard = Coro::guard { ... }
561		1008
562	This creates and returns a guard object. Nothing happens until the object	1009	This function still exists, but is deprecated. Please use the
563	gets destroyed, in which case the codeblock given as argument will be	1010	C<Guard::guard> function instead.
564	executed. This is useful to free locks or other resources in case of a
565	runtime error or when the coroutine gets canceled, as in both cases the
566	guard block will be executed. The guard object supports only one method,
567	C<< ->cancel >>, which will keep the codeblock from being executed.
568
569	Example: set some flag and clear it again when the coroutine gets canceled
570	or the function returns:
571
572	sub do_something {
573	my $guard = Coro::guard { $busy = 0 };
574	$busy = 1;
575
576	# do something that requires $busy to be true
577	}
578		1011
579	=cut	1012	=cut
580		1013
581	sub guard(&) {	1014	BEGIN { *guard = \&Guard::guard }
582	bless \(my $cb = $_[0]), "Coro::guard"
583	}
584
585	sub Coro::guard::cancel {
586	${$_[0]} = sub { };
587	}
588
589	sub Coro::guard::DESTROY {
590	${$_[0]}->();
591	}
592
593		1015
594	=item unblock_sub { ... }	1016	=item unblock_sub { ... }
595		1017
596	This utility function takes a BLOCK or code reference and "unblocks" it,	1018	This utility function takes a BLOCK or code reference and "unblocks" it,
597	returning a new coderef. Unblocking means that calling the new coderef	1019	returning a new coderef. Unblocking means that calling the new coderef
598	will return immediately without blocking, returning nothing, while the	1020	will return immediately without blocking, returning nothing, while the
599	original code ref will be called (with parameters) from within another	1021	original code ref will be called (with parameters) from within another
600	coroutine.	1022	coro.
601		1023
602	The reason this function exists is that many event libraries (such as the	1024	The reason this function exists is that many event libraries (such as
603	venerable L<Event|Event> module) are not coroutine-safe (a weaker form	1025	the venerable L<Event|Event> module) are not thread-safe (a weaker form
604	of thread-safety). This means you must not block within event callbacks,	1026	of reentrancy). This means you must not block within event callbacks,
605	otherwise you might suffer from crashes or worse. The only event library	1027	otherwise you might suffer from crashes or worse. The only event library
606	currently known that is safe to use without C<unblock_sub> is L<EV>.	1028	currently known that is safe to use without C<unblock_sub> is L<EV> (but
		1029	you might still run into deadlocks if all event loops are blocked).
		1030
		1031	Coro will try to catch you when you block in the event loop
		1032	("FATAL: $Coro::idle blocked itself"), but this is just best effort and
		1033	only works when you do not run your own event loop.
607		1034
608	This function allows your callbacks to block by executing them in another	1035	This function allows your callbacks to block by executing them in another
609	coroutine where it is safe to block. One example where blocking is handy	1036	coro where it is safe to block. One example where blocking is handy
610	is when you use the L<Coro::AIO|Coro::AIO> functions to save results to	1037	is when you use the L<Coro::AIO|Coro::AIO> functions to save results to
611	disk, for example.	1038	disk, for example.
612		1039
613	In short: simply use C<unblock_sub { ... }> instead of C<sub { ... }> when	1040	In short: simply use C<unblock_sub { ... }> instead of C<sub { ... }> when
614	creating event callbacks that want to block.	1041	creating event callbacks that want to block.
615		1042
616	If your handler does not plan to block (e.g. simply sends a message to	1043	If your handler does not plan to block (e.g. simply sends a message to
617	another coroutine, or puts some other coroutine into the ready queue),	1044	another coro, or puts some other coro into the ready queue), there is
618	there is no reason to use C<unblock_sub>.	1045	no reason to use C<unblock_sub>.
619		1046
620	Note that you also need to use C<unblock_sub> for any other callbacks that	1047	Note that you also need to use C<unblock_sub> for any other callbacks that
621	are indirectly executed by any C-based event loop. For example, when you	1048	are indirectly executed by any C-based event loop. For example, when you
622	use a module that uses L<AnyEvent> (and you use L<Coro::AnyEvent>) and it	1049	use a module that uses L<AnyEvent> (and you use L<Coro::AnyEvent>) and it
623	provides callbacks that are the result of some event callback, then you	1050	provides callbacks that are the result of some event callback, then you
…		…
632	# return immediately and can be reused) and because we cannot cede	1059	# return immediately and can be reused) and because we cannot cede
633	# inside an event callback.	1060	# inside an event callback.
634	our $unblock_scheduler = new Coro sub {	1061	our $unblock_scheduler = new Coro sub {
635	while () {	1062	while () {
636	while (my $cb = pop @unblock_queue) {	1063	while (my $cb = pop @unblock_queue) {
637	# this is an inlined copy of async_pool	1064	&async_pool (@$cb);
638	my $coro = (pop @async_pool) || new Coro \&pool_handler;
639		1065
640	$coro->{_invoke} = $cb;
641	$coro->ready;
642	cede; # for short-lived callbacks, this reduces pressure on the coro pool	1066	# for short-lived callbacks, this reduces pressure on the coro pool
		1067	# as the chance is very high that the async_poll coro will be back
		1068	# in the idle state when cede returns
		1069	cede;
643	}	1070	}
644	schedule; # sleep well	1071	schedule; # sleep well
645	}	1072	}
646	};	1073	};
647	$unblock_scheduler->desc ("[unblock_sub scheduler]");	1074	$unblock_scheduler->{desc} = "[unblock_sub scheduler]";
648		1075
649	sub unblock_sub(&) {	1076	sub unblock_sub(&) {
650	my $cb = shift;	1077	my $cb = shift;
651		1078
652	sub {	1079	sub {
653	unshift @unblock_queue, [$cb, @_];	1080	unshift @unblock_queue, [$cb, @_];
654	$unblock_scheduler->ready;	1081	$unblock_scheduler->ready;
655	}	1082	}
656	}	1083	}
657		1084
		1085	=item $cb = rouse_cb
		1086
		1087	Create and return a "rouse callback". That's a code reference that,
		1088	when called, will remember a copy of its arguments and notify the owner
		1089	coro of the callback.
		1090
		1091	Only the first invocation will store agruments and signal any waiter -
		1092	further calls will effectively be ignored, but it is ok to try.
		1093
		1094	Also see the next function.
		1095
		1096	=item @args = rouse_wait [$cb]
		1097
		1098	Wait for the specified rouse callback to be invoked (or if the argument is
		1099	missing, use the most recently created callback in the current coro).
		1100
		1101	As soon as the callback is invoked (or when the callback was invoked
		1102	before C<rouse_wait>), it will return the arguments originally passed to
		1103	the rouse callback. In scalar context, that means you get the I<last>
		1104	argument, just as if C<rouse_wait> had a C<return ($a1, $a2, $a3...)>
		1105	statement at the end.
		1106
		1107	You are only allowed to wait once for a given rouse callback.
		1108
		1109	See the section B<HOW TO WAIT FOR A CALLBACK> for an actual usage example.
		1110
		1111	As of Coro 6.57, you can reliably wait for a rouse callback in a different
		1112	thread than from where it was created.
		1113
658	=back	1114	=back
659		1115
660	=cut	1116	=cut
661		1117
		1118	for my $module (qw(Channel RWLock Semaphore SemaphoreSet Signal Specific)) {
		1119	my $old = defined &{"Coro::$module\::new"} && \&{"Coro::$module\::new"};
		1120
		1121	*{"Coro::$module\::new"} = sub {
		1122	require "Coro/$module.pm";
		1123
		1124	# some modules have their new predefined in State.xs, some don't
		1125	*{"Coro::$module\::new"} = $old
		1126	if $old;
		1127
		1128	goto &{"Coro::$module\::new"}
		1129	};
		1130	}
		1131
662	1;	1132	1;
663		1133
		1134	=head1 HOW TO WAIT FOR A CALLBACK
		1135
		1136	It is very common for a coro to wait for some callback to be
		1137	called. This occurs naturally when you use coro in an otherwise
		1138	event-based program, or when you use event-based libraries.
		1139
		1140	These typically register a callback for some event, and call that callback
		1141	when the event occurred. In a coro, however, you typically want to
		1142	just wait for the event, simplyifying things.
		1143
		1144	For example C<< AnyEvent->child >> registers a callback to be called when
		1145	a specific child has exited:
		1146
		1147	my $child_watcher = AnyEvent->child (pid => $pid, cb => sub { ... });
		1148
		1149	But from within a coro, you often just want to write this:
		1150
		1151	my $status = wait_for_child $pid;
		1152
		1153	Coro offers two functions specifically designed to make this easy,
		1154	C<rouse_cb> and C<rouse_wait>.
		1155
		1156	The first function, C<rouse_cb>, generates and returns a callback that,
		1157	when invoked, will save its arguments and notify the coro that
		1158	created the callback.
		1159
		1160	The second function, C<rouse_wait>, waits for the callback to be called
		1161	(by calling C<schedule> to go to sleep) and returns the arguments
		1162	originally passed to the callback.
		1163
		1164	Using these functions, it becomes easy to write the C<wait_for_child>
		1165	function mentioned above:
		1166
		1167	sub wait_for_child($) {
		1168	my ($pid) = @_;
		1169
		1170	my $watcher = AnyEvent->child (pid => $pid, cb => rouse_cb);
		1171
		1172	my ($rpid, $rstatus) = rouse_wait;
		1173	$rstatus
		1174	}
		1175
		1176	In the case where C<rouse_cb> and C<rouse_wait> are not flexible enough,
		1177	you can roll your own, using C<schedule> and C<ready>:
		1178
		1179	sub wait_for_child($) {
		1180	my ($pid) = @_;
		1181
		1182	# store the current coro in $current,
		1183	# and provide result variables for the closure passed to ->child
		1184	my $current = $Coro::current;
		1185	my ($done, $rstatus);
		1186
		1187	# pass a closure to ->child
		1188	my $watcher = AnyEvent->child (pid => $pid, cb => sub {
		1189	$rstatus = $_[1]; # remember rstatus
		1190	$done = 1; # mark $rstatus as valid
		1191	$current->ready; # wake up the waiting thread
		1192	});
		1193
		1194	# wait until the closure has been called
		1195	schedule while !$done;
		1196
		1197	$rstatus
		1198	}
		1199
		1200
664	=head1 BUGS/LIMITATIONS	1201	=head1 BUGS/LIMITATIONS
665		1202
		1203	=over 4
		1204
		1205	=item fork with pthread backend
		1206
		1207	When Coro is compiled using the pthread backend (which isn't recommended
		1208	but required on many BSDs as their libcs are completely broken), then
		1209	coro will not survive a fork. There is no known workaround except to
		1210	fix your libc and use a saner backend.
		1211
		1212	=item perl process emulation ("threads")
		1213
666	This module is not perl-pseudo-thread-safe. You should only ever use this	1214	This module is not perl-pseudo-thread-safe. You should only ever use this
667	module from the same thread (this requirement might be removed in the	1215	module from the first thread (this requirement might be removed in the
668	future to allow per-thread schedulers, but Coro::State does not yet allow	1216	future to allow per-thread schedulers, but Coro::State does not yet allow
669	this). I recommend disabling thread support and using processes, as this	1217	this). I recommend disabling thread support and using processes, as having
670	is much faster and uses less memory.	1218	the windows process emulation enabled under unix roughly halves perl
		1219	performance, even when not used.
		1220
		1221	Attempts to use threads created in another emulated process will crash
		1222	("cleanly", with a null pointer exception).
		1223
		1224	=item coro switching is not signal safe
		1225
		1226	You must not switch to another coro from within a signal handler (only
		1227	relevant with %SIG - most event libraries provide safe signals), I<unless>
		1228	you are sure you are not interrupting a Coro function.
		1229
		1230	That means you I<MUST NOT> call any function that might "block" the
		1231	current coro - C<cede>, C<schedule> C<< Coro::Semaphore->down >> or
		1232	anything that calls those. Everything else, including calling C<ready>,
		1233	works.
		1234
		1235	=back
		1236
		1237
		1238	=head1 WINDOWS PROCESS EMULATION
		1239
		1240	A great many people seem to be confused about ithreads (for example, Chip
		1241	Salzenberg called me unintelligent, incapable, stupid and gullible,
		1242	while in the same mail making rather confused statements about perl
		1243	ithreads (for example, that memory or files would be shared), showing his
		1244	lack of understanding of this area - if it is hard to understand for Chip,
		1245	it is probably not obvious to everybody).
		1246
		1247	What follows is an ultra-condensed version of my talk about threads in
		1248	scripting languages given on the perl workshop 2009:
		1249
		1250	The so-called "ithreads" were originally implemented for two reasons:
		1251	first, to (badly) emulate unix processes on native win32 perls, and
		1252	secondly, to replace the older, real thread model ("5.005-threads").
		1253
		1254	It does that by using threads instead of OS processes. The difference
		1255	between processes and threads is that threads share memory (and other
		1256	state, such as files) between threads within a single process, while
		1257	processes do not share anything (at least not semantically). That
		1258	means that modifications done by one thread are seen by others, while
		1259	modifications by one process are not seen by other processes.
		1260
		1261	The "ithreads" work exactly like that: when creating a new ithreads
		1262	process, all state is copied (memory is copied physically, files and code
		1263	is copied logically). Afterwards, it isolates all modifications. On UNIX,
		1264	the same behaviour can be achieved by using operating system processes,
		1265	except that UNIX typically uses hardware built into the system to do this
		1266	efficiently, while the windows process emulation emulates this hardware in
		1267	software (rather efficiently, but of course it is still much slower than
		1268	dedicated hardware).
		1269
		1270	As mentioned before, loading code, modifying code, modifying data
		1271	structures and so on is only visible in the ithreads process doing the
		1272	modification, not in other ithread processes within the same OS process.
		1273
		1274	This is why "ithreads" do not implement threads for perl at all, only
		1275	processes. What makes it so bad is that on non-windows platforms, you can
		1276	actually take advantage of custom hardware for this purpose (as evidenced
		1277	by the forks module, which gives you the (i-) threads API, just much
		1278	faster).
		1279
		1280	Sharing data is in the i-threads model is done by transferring data
		1281	structures between threads using copying semantics, which is very slow -
		1282	shared data simply does not exist. Benchmarks using i-threads which are
		1283	communication-intensive show extremely bad behaviour with i-threads (in
		1284	fact, so bad that Coro, which cannot take direct advantage of multiple
		1285	CPUs, is often orders of magnitude faster because it shares data using
		1286	real threads, refer to my talk for details).
		1287
		1288	As summary, i-threads *use* threads to implement processes, while
		1289	the compatible forks module *uses* processes to emulate, uhm,
		1290	processes. I-threads slow down every perl program when enabled, and
		1291	outside of windows, serve no (or little) practical purpose, but
		1292	disadvantages every single-threaded Perl program.
		1293
		1294	This is the reason that I try to avoid the name "ithreads", as it is
		1295	misleading as it implies that it implements some kind of thread model for
		1296	perl, and prefer the name "windows process emulation", which describes the
		1297	actual use and behaviour of it much better.
671		1298
672	=head1 SEE ALSO	1299	=head1 SEE ALSO
673		1300
674	Event-Loop integration: L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>.	1301	Event-Loop integration: L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>.
675		1302
676	Debugging: L<Coro::Debug>.	1303	Debugging: L<Coro::Debug>.
677		1304
678	Support/Utility: L<Coro::Specific>, L<Coro::Util>.	1305	Support/Utility: L<Coro::Specific>, L<Coro::Util>.
679		1306
680	Locking/IPC: L<Coro::Signal>, L<Coro::Channel>, L<Coro::Semaphore>, L<Coro::SemaphoreSet>, L<Coro::RWLock>.	1307	Locking and IPC: L<Coro::Signal>, L<Coro::Channel>, L<Coro::Semaphore>,
		1308	L<Coro::SemaphoreSet>, L<Coro::RWLock>.
681		1309
682	IO/Timers: L<Coro::Timer>, L<Coro::Handle>, L<Coro::Socket>, L<Coro::AIO>.	1310	I/O and Timers: L<Coro::Timer>, L<Coro::Handle>, L<Coro::Socket>, L<Coro::AIO>.
683		1311
684	Compatibility: L<Coro::LWP>, L<Coro::BDB>, L<Coro::Storable>, L<Coro::Select>.	1312	Compatibility with other modules: L<Coro::LWP> (but see also L<AnyEvent::HTTP> for
		1313	a better-working alternative), L<Coro::BDB>, L<Coro::Storable>,
		1314	L<Coro::Select>.
685		1315
686	XS API: L<Coro::MakeMaker>.	1316	XS API: L<Coro::MakeMaker>.
687		1317
688	Low level Configuration, Coroutine Environment: L<Coro::State>.	1318	Low level Configuration, Thread Environment, Continuations: L<Coro::State>.
689		1319
690	=head1 AUTHOR	1320	=head1 AUTHOR/SUPPORT/CONTACT
691		1321
692	Marc Lehmann <schmorp@schmorp.de>	1322	Marc A. Lehmann <schmorp@schmorp.de>
693	http://home.schmorp.de/	1323	http://software.schmorp.de/pkg/Coro.html
694		1324
695	=cut	1325	=cut
696		1326

Diff Legend

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