ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/Coro/Coro.pm

Comparing Coro/Coro.pm (file contents):
Revision 1.163 by root, Mon Dec 17 06:36:24 2007 UTC vs.
Revision 1.296 by root, Thu May 12 23:24:28 2011 UTC

1	=head1 NAME	1	=head1 NAME
2		2
3	Coro - coroutine process abstraction	3	Coro - the only real threads in perl
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	use Coro;	7	use Coro;
8		8
9	async {	9	async {
10	# some asynchronous thread of execution	10	# some asynchronous thread of execution
11	print "2\n";	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;
21	my $lock = new Coro::Semaphore;	22	my $lock = new Coro::Semaphore;
22	my $locked;	23	my $locked;
23		24
24	$lock->down;	25	$lock->down;
25	$locked = 1;	26	$locked = 1;
26	$lock->up;	27	$lock->up;
27		28
28	=head1 DESCRIPTION	29	=head1 DESCRIPTION
29		30
30	This module collection manages coroutines. Coroutines are similar	31	For a tutorial-style introduction, please read the L<Coro::Intro>
31	to threads but don't run in parallel at the same time even on SMP	32	manpage. This manpage mainly contains reference information.
32	machines. The specific flavor of coroutine used in this module also
33	guarantees you that it will not switch between coroutines unless
34	necessary, at easily-identified points in your program, so locking and
35	parallel access are rarely an issue, making coroutine programming much
36	safer than threads programming.
37		33
38	(Perl, however, does not natively support real threads but instead does a	34	This module collection manages continuations in general, most often in
39	very slow and memory-intensive emulation of processes using threads. This	35	the form of cooperative threads (also called coros, or simply "coro"
40	is a performance win on Windows machines, and a loss everywhere else).	36	in the documentation). They are similar to kernel threads but don't (in
		37	general) run in parallel at the same time even on SMP machines. The
		38	specific flavor of thread offered by this module also guarantees you that
		39	it will not switch between threads unless necessary, at easily-identified
		40	points in your program, so locking and parallel access are rarely an
		41	issue, making thread programming much safer and easier than using other
		42	thread models.
41		43
		44	Unlike the so-called "Perl threads" (which are not actually real threads
		45	but only the windows process emulation (see section of same name for
		46	more details) ported to UNIX, and as such act as processes), Coro
		47	provides a full shared address space, which makes communication between
		48	threads very easy. And coro threads are fast, too: disabling the Windows
		49	process emulation code in your perl and using Coro can easily result in
		50	a two to four times speed increase for your programs. A parallel matrix
		51	multiplication benchmark (very communication-intensive) runs over 300
		52	times faster on a single core than perls pseudo-threads on a quad core
		53	using all four cores.
		54
		55	Coro achieves that by supporting multiple running interpreters that share
		56	data, which is especially useful to code pseudo-parallel processes and
		57	for event-based programming, such as multiple HTTP-GET requests running
		58	concurrently. See L<Coro::AnyEvent> to learn more on how to integrate Coro
		59	into an event-based environment.
		60
42	In this module, coroutines are defined as "callchain + lexical variables +	61	In this module, a thread is defined as "callchain + lexical variables +
43	@_ + $_ + $@ + $/ + C stack), that is, a coroutine has its own callchain,	62	some package variables + C stack), that is, a thread has its own callchain,
44	its own set of lexicals and its own set of perls most important global	63	its own set of lexicals and its own set of perls most important global
45	variables (see L<Coro::State> for more configuration).	64	variables (see L<Coro::State> for more configuration and background info).
		65
		66	See also the C<SEE ALSO> section at the end of this document - the Coro
		67	module family is quite large.
		68
		69	=head1 CORO THREAD LIFE CYCLE
		70
		71	During the long and exciting (or not) life of a coro thread, it goes
		72	through a number of states:
		73
		74	=over 4
		75
		76	=item 1. Creation
		77
		78	The first thing in the life of a coro thread is it's creation -
		79	obviously. The typical way to create a thread is to call the C<async
		80	BLOCK> function:
		81
		82	async {
		83	# thread code goes here
		84	};
		85
		86	You can also pass arguments, which are put in C<@_>:
		87
		88	async {
		89	print $_[1]; # prints 2
		90	} 1, 2, 3;
		91
		92	This creates a new coro thread and puts it into the ready queue, meaning
		93	it will run as soon as the CPU is free for it.
		94
		95	C<async> will return a coro object - you can store this for future
		96	reference or ignore it, the thread itself will keep a reference to it's
		97	thread object - threads are alive on their own.
		98
		99	Another way to create a thread is to call the C<new> constructor with a
		100	code-reference:
		101
		102	new Coro sub {
		103	# thread code goes here
		104	}, @optional_arguments;
		105
		106	This is quite similar to calling C<async>, but the important difference is
		107	that the new thread is not put into the ready queue, so the thread will
		108	not run until somebody puts it there. C<async> is, therefore, identical to
		109	this sequence:
		110
		111	my $coro = new Coro sub {
		112	# thread code goes here
		113	};
		114	$coro->ready;
		115	return $coro;
		116
		117	=item 2. Startup
		118
		119	When a new coro thread is created, only a copy of the code reference
		120	and the arguments are stored, no extra memory for stacks and so on is
		121	allocated, keeping the coro thread in a low-memory state.
		122
		123	Only when it actually starts executing will all the resources be finally
		124	allocated.
		125
		126	The optional arguments specified at coro creation are available in C<@_>,
		127	similar to function calls.
		128
		129	=item 3. Running / Blocking
		130
		131	A lot can happen after the coro thread has started running. Quite usually,
		132	it will not run to the end in one go (because you could use a function
		133	instead), but it will give up the CPU regularly because it waits for
		134	external events.
		135
		136	As long as a coro thread runs, it's coro object is available in the global
		137	variable C<$Coro::current>.
		138
		139	The low-level way to give up the CPU is to call the scheduler, which
		140	selects a new coro thread to run:
		141
		142	Coro::schedule;
		143
		144	Since running threads are not in the ready queue, calling the scheduler
		145	without doing anything else will block the coro thread forever - you need
		146	to arrange either for the coro to put woken up (readied) by some other
		147	event or some other thread, or you can put it into the ready queue before
		148	scheduling:
		149
		150	# this is exactly what Coro::cede does
		151	$Coro::current->ready;
		152	Coro::schedule;
		153
		154	All the higher-level synchronisation methods (Coro::Semaphore,
		155	Coro::rouse_*...) are actually implemented via C<< ->ready >> and C<<
		156	Coro::schedule >>.
		157
		158	While the coro thread is running it also might get assigned a C-level
		159	thread, or the C-level thread might be unassigned from it, as the Coro
		160	runtime wishes. A C-level thread needs to be assigned when your perl
		161	thread calls into some C-level function and that function in turn calls
		162	perl and perl then wants to switch coroutines. This happens most often
		163	when you run an event loop and block in the callback, or when perl
		164	itself calls some function such as C<AUTOLOAD> or methods via the C<tie>
		165	mechanism.
		166
		167	=item 4. Termination
		168
		169	Many threads actually terminate after some time. There are a number of
		170	ways to terminate a coro thread, the simplest is returning from the
		171	top-level code reference:
		172
		173	async {
		174	# after returning from here, the coro thread is terminated
		175	};
		176
		177	async {
		178	return if 0.5 < rand; # terminate a little earlier, maybe
		179	print "got a chance to print this\n";
		180	# or here
		181	};
		182
		183	Any values returned from the coroutine can be recovered using C<< ->join
		184	>>:
		185
		186	my $coro = async {
		187	"hello, world\n" # return a string
		188	};
		189
		190	my $hello_world = $coro->join;
		191
		192	print $hello_world;
		193
		194	Another way to terminate is to call C<< Coro::terminate >>, which at any
		195	subroutine call nesting level:
		196
		197	async {
		198	Coro::terminate "return value 1", "return value 2";
		199	};
		200
		201	And yet another way is to C<< ->cancel >> (or C<< ->safe_cancel >>) the
		202	coro 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	=item 5. Cleanup
		225
		226	Threads will allocate various resources. Most but not all will be returned
		227	when a thread terminates, during clean-up.
		228
		229	Cleanup is quite similar to throwing an uncaught exception: perl will
		230	work it's way up through all subroutine calls and blocks. On it's way, it
		231	will release all C<my> variables, undo all C<local>'s and free any other
		232	resources truly local to the thread.
		233
		234	So, a common way to free resources is to keep them referenced only by my
		235	variables:
		236
		237	async {
		238	my $big_cache = new Cache ...;
		239	};
		240
		241	If there are no other references, then the C<$big_cache> object will be
		242	freed when the thread terminates, regardless of how it does so.
		243
		244	What it does C<NOT> do is unlock any Coro::Semaphores or similar
		245	resources, but that's where the C<guard> methods come in handy:
		246
		247	my $sem = new Coro::Semaphore;
		248
		249	async {
		250	my $lock_guard = $sem->guard;
		251	# if we reutrn, or die or get cancelled, here,
		252	# then the semaphore will be "up"ed.
		253	};
		254
		255	The C<Guard::guard> function comes in handy for any custom cleanup you
		256	might want to do:
		257
		258	async {
		259	my $window = new Gtk2::Window "toplevel";
		260	# The window will not be cleaned up automatically, even when $window
		261	# gets freed, so use a guard to ensure it's destruction
		262	# in case of an error:
		263	my $window_guard = Guard::guard { $window->destroy };
		264
		265	# we are safe here
		266	};
		267
		268	Last not least, C<local> can often be handy, too, e.g. when temporarily
		269	replacing the coro thread description:
		270
		271	sub myfunction {
		272	local $Coro::current->{desc} = "inside myfunction(@_)";
		273
		274	# if we return or die here, the description will be restored
		275	}
		276
		277	=item 6. Viva La Zombie Muerte
		278
		279	Even after a thread has terminated and cleaned up it's resources, the coro
		280	object still is there and stores the return values of the thread. Only in
		281	this state will the coro object be "reference counted" in the normal perl
		282	sense: the thread code keeps a reference to it when it is active, but not
		283	after it has terminated.
		284
		285	The means the coro object gets freed automatically when the thread has
		286	terminated and cleaned up and there arenot other references.
		287
		288	If there are, the coro object will stay around, and you can call C<<
		289	->join >> as many times as you wish to retrieve the result values:
		290
		291	async {
		292	print "hi\n";
		293	1
		294	};
		295
		296	# run the async above, and free everything before returning
		297	# from Coro::cede:
		298	Coro::cede;
		299
		300	{
		301	my $coro = async {
		302	print "hi\n";
		303	1
		304	};
		305
		306	# run the async above, and clean up, but do not free the coro
		307	# object:
		308	Coro::cede;
		309
		310	# optionally retrieve the result values
		311	my @results = $coro->join;
		312
		313	# now $coro goes out of scope, and presumably gets freed
		314	};
		315
		316	=back
46		317
47	=cut	318	=cut
48		319
49	package Coro;	320	package Coro;
50		321
51	use strict;	322	use common::sense;
52	no warnings "uninitialized";	323
		324	use Carp ();
		325
		326	use Guard ();
53		327
54	use Coro::State;	328	use Coro::State;
55		329
56	use base qw(Coro::State Exporter);	330	use base qw(Coro::State Exporter);
57		331
58	our $idle; # idle handler	332	our $idle; # idle handler
59	our $main; # main coroutine	333	our $main; # main coro
60	our $current; # current coroutine	334	our $current; # current coro
61		335
62	our $VERSION = '4.32';	336	our $VERSION = 5.372;
63		337
64	our @EXPORT = qw(async async_pool cede schedule terminate current unblock_sub);	338	our @EXPORT = qw(async async_pool cede schedule terminate current unblock_sub rouse_cb rouse_wait);
65	our %EXPORT_TAGS = (	339	our %EXPORT_TAGS = (
66	prio => [qw(PRIO_MAX PRIO_HIGH PRIO_NORMAL PRIO_LOW PRIO_IDLE PRIO_MIN)],	340	prio => [qw(PRIO_MAX PRIO_HIGH PRIO_NORMAL PRIO_LOW PRIO_IDLE PRIO_MIN)],
67	);	341	);
68	our @EXPORT_OK = (@{$EXPORT_TAGS{prio}}, qw(nready));	342	our @EXPORT_OK = (@{$EXPORT_TAGS{prio}}, qw(nready));
69		343
70	{	344	=head1 GLOBAL VARIABLES
71	my @async;
72	my $init;
73
74	# this way of handling attributes simply is NOT scalable ;()
75	sub import {
76	no strict 'refs';
77
78	Coro->export_to_level (1, @_);
79
80	my $old = *{(caller)[0]."::MODIFY_CODE_ATTRIBUTES"}{CODE};
81	*{(caller)[0]."::MODIFY_CODE_ATTRIBUTES"} = sub {
82	my ($package, $ref) = (shift, shift);
83	my @attrs;
84	for (@_) {
85	if ($_ eq "Coro") {
86	push @async, $ref;
87	unless ($init++) {
88	eval q{
89	sub INIT {
90	&async(pop @async) while @async;
91	}
92	};
93	}
94	} else {
95	push @attrs, $_;
96	}
97	}
98	return $old ? $old->($package, $ref, @attrs) : @attrs;
99	};
100	}
101
102	}
103		345
104	=over 4	346	=over 4
105		347
106	=item $main	348	=item $Coro::main
107		349
108	This coroutine represents the main program.	350	This variable stores the Coro object that represents the main
		351	program. While you cna C<ready> it and do most other things you can do to
		352	coro, it is mainly useful to compare again C<$Coro::current>, to see
		353	whether you are running in the main program or not.
109		354
110	=cut	355	=cut
111		356
112	$main = new Coro;	357	# $main is now being initialised by Coro::State
113		358
114	=item $current (or as function: current)	359	=item $Coro::current
115		360
116	The current coroutine (the last coroutine switched to). The initial value	361	The Coro object representing the current coro (the last
		362	coro that the Coro scheduler switched to). The initial value is
117	is C<$main> (of course).	363	C<$Coro::main> (of course).
118		364
119	This variable is B<strictly> I<read-only>. It is provided for performance	365	This variable is B<strictly> I<read-only>. You can take copies of the
120	reasons. If performance is not essential you are encouraged to use the	366	value stored in it and use it as any other Coro object, but you must
121	C<Coro::current> function instead.	367	not otherwise modify the variable itself.
122		368
123	=cut	369	=cut
124		370
125	$main->{desc} = "[main::]";
126
127	# maybe some other module used Coro::Specific before...
128	$main->{_specific} = $current->{_specific}
129	if $current;
130
131	_set_current $main;
132
133	sub current() { $current }	371	sub current() { $current } # [DEPRECATED]
134		372
135	=item $idle	373	=item $Coro::idle
136		374
137	A callback that is called whenever the scheduler finds no ready coroutines	375	This variable is mainly useful to integrate Coro into event loops. It is
		376	usually better to rely on L<Coro::AnyEvent> or L<Coro::EV>, as this is
		377	pretty low-level functionality.
		378
		379	This variable stores a Coro object that is put into the ready queue when
		380	there are no other ready threads (without invoking any ready hooks).
		381
138	to run. The default implementation prints "FATAL: deadlock detected" and	382	The default implementation dies with "FATAL: deadlock detected.", followed
139	exits, because the program has no other way to continue.	383	by a thread listing, because the program has no other way to continue.
140		384
141	This hook is overwritten by modules such as C<Coro::Timer> and	385	This hook is overwritten by modules such as C<Coro::EV> and
142	C<Coro::Event> to wait on an external event that hopefully wake up a	386	C<Coro::AnyEvent> to wait on an external event that hopefully wakes up a
143	coroutine so the scheduler can run it.	387	coro so the scheduler can run it.
144		388
145	Please note that if your callback recursively invokes perl (e.g. for event	389	See L<Coro::EV> or L<Coro::AnyEvent> for examples of using this technique.
146	handlers), then it must be prepared to be called recursively itself.
147		390
148	=cut	391	=cut
149		392
150	$idle = sub {	393	# ||= because other modules could have provided their own by now
151	require Carp;	394	$idle ||= new Coro sub {
152	Carp::croak ("FATAL: deadlock detected");	395	require Coro::Debug;
		396	die "FATAL: deadlock detected.\n"
		397	. Coro::Debug::ps_listing ();
153	};	398	};
154		399
155	sub _cancel {
156	my ($self) = @_;
157
158	# free coroutine data and mark as destructed
159	$self->_destroy
160	or return;
161
162	# call all destruction callbacks
163	$_->(@{$self->{_status}})
164	for @{(delete $self->{_on_destroy}) || []};
165	}
166
167	# this coroutine is necessary because a coroutine	400	# this coro is necessary because a coro
168	# cannot destroy itself.	401	# cannot destroy itself.
169	my @destroy;	402	our @destroy;
170	my $manager;	403	our $manager;
171		404
172	$manager = new Coro sub {	405	$manager = new Coro sub {
173	while () {	406	while () {
174	(shift @destroy)->_cancel	407	_destroy shift @destroy
175	while @destroy;	408	while @destroy;
176		409
177	&schedule;	410	&schedule;
178	}	411	}
179	};	412	};
180	$manager->desc ("[coro manager]");	413	$manager->{desc} = "[coro manager]";
181	$manager->prio (PRIO_MAX);	414	$manager->prio (PRIO_MAX);
182		415
183	# static methods. not really.
184
185	=back	416	=back
186		417
187	=head2 STATIC METHODS	418	=head1 SIMPLE CORO CREATION
188
189	Static methods are actually functions that operate on the current coroutine only.
190		419
191	=over 4	420	=over 4
192		421
193	=item async { ... } [@args...]	422	=item async { ... } [@args...]
194		423
195	Create a new asynchronous coroutine and return it's coroutine object	424	Create a new coro and return its Coro object (usually
196	(usually unused). When the sub returns the new coroutine is automatically	425	unused). The coro will be put into the ready queue, so
		426	it will start running automatically on the next scheduler run.
		427
		428	The first argument is a codeblock/closure that should be executed in the
		429	coro. When it returns argument returns the coro is automatically
197	terminated.	430	terminated.
198		431
		432	The remaining arguments are passed as arguments to the closure.
		433
199	See the C<Coro::State::new> constructor for info about the coroutine	434	See the C<Coro::State::new> constructor for info about the coro
200	environment in which coroutines run.	435	environment in which coro are executed.
201		436
202	Calling C<exit> in a coroutine will do the same as calling exit outside	437	Calling C<exit> in a coro will do the same as calling exit outside
203	the coroutine. Likewise, when the coroutine dies, the program will exit,	438	the coro. Likewise, when the coro dies, the program will exit,
204	just as it would in the main program.	439	just as it would in the main program.
205		440
		441	If you do not want that, you can provide a default C<die> handler, or
		442	simply avoid dieing (by use of C<eval>).
		443
206	# create a new coroutine that just prints its arguments	444	Example: Create a new coro that just prints its arguments.
		445
207	async {	446	async {
208	print "@_\n";	447	print "@_\n";
209	} 1,2,3,4;	448	} 1,2,3,4;
210		449
211	=cut
212
213	sub async(&@) {
214	my $coro = new Coro @_;
215	$coro->ready;
216	$coro
217	}
218
219	=item async_pool { ... } [@args...]	450	=item async_pool { ... } [@args...]
220		451
221	Similar to C<async>, but uses a coroutine pool, so you should not call	452	Similar to C<async>, but uses a coro pool, so you should not call
222	terminate or join (although you are allowed to), and you get a coroutine	453	terminate or join on it (although you are allowed to), and you get a
223	that might have executed other code already (which can be good or bad :).	454	coro that might have executed other code already (which can be good
		455	or bad :).
224		456
		457	On the plus side, this function is about twice as fast as creating (and
		458	destroying) a completely new coro, so if you need a lot of generic
		459	coros in quick successsion, use C<async_pool>, not C<async>.
		460
225	Also, the block is executed in an C<eval> context and a warning will be	461	The code block is executed in an C<eval> context and a warning will be
226	issued in case of an exception instead of terminating the program, as	462	issued in case of an exception instead of terminating the program, as
227	C<async> does. As the coroutine is being reused, stuff like C<on_destroy>	463	C<async> does. As the coro is being reused, stuff like C<on_destroy>
228	will not work in the expected way, unless you call terminate or cancel,	464	will not work in the expected way, unless you call terminate or cancel,
229	which somehow defeats the purpose of pooling.	465	which somehow defeats the purpose of pooling (but is fine in the
		466	exceptional case).
230		467
231	The priority will be reset to C<0> after each job, tracing will be	468	The priority will be reset to C<0> after each run, tracing will be
232	disabled, the description will be reset and the default output filehandle	469	disabled, the description will be reset and the default output filehandle
233	gets restored, so you can change alkl these. Otherwise the coroutine will	470	gets restored, so you can change all these. Otherwise the coro will
234	be re-used "as-is": most notably if you change other per-coroutine global	471	be re-used "as-is": most notably if you change other per-coro global
235	stuff such as C<$/> you need to revert that change, which is most simply	472	stuff such as C<$/> you I<must needs> revert that change, which is most
236	done by using local as in C< local $/ >.	473	simply done by using local as in: C<< local $/ >>.
237		474
238	The pool size is limited to 8 idle coroutines (this can be adjusted by	475	The idle pool size is limited to C<8> idle coros (this can be
239	changing $Coro::POOL_SIZE), and there can be as many non-idle coros as	476	adjusted by changing $Coro::POOL_SIZE), but there can be as many non-idle
240	required.	477	coros as required.
241		478
242	If you are concerned about pooled coroutines growing a lot because a	479	If you are concerned about pooled coros growing a lot because a
243	single C<async_pool> used a lot of stackspace you can e.g. C<async_pool	480	single C<async_pool> used a lot of stackspace you can e.g. C<async_pool
244	{ terminate }> once per second or so to slowly replenish the pool. In	481	{ terminate }> once per second or so to slowly replenish the pool. In
245	addition to that, when the stacks used by a handler grows larger than 16kb	482	addition to that, when the stacks used by a handler grows larger than 32kb
246	(adjustable with $Coro::POOL_RSS) it will also exit.	483	(adjustable via $Coro::POOL_RSS) it will also be destroyed.
247		484
248	=cut	485	=cut
249		486
250	our $POOL_SIZE = 8;	487	our $POOL_SIZE = 8;
251	our $POOL_RSS = 16 * 1024;	488	our $POOL_RSS = 32 * 1024;
252	our @async_pool;	489	our @async_pool;
253		490
254	sub pool_handler {	491	sub pool_handler {
255	my $cb;
256
257	while () {	492	while () {
258	eval {	493	eval {
259	while () {	494	&{&_pool_handler} while 1;
260	_pool_1 $cb;
261	&$cb;
262	_pool_2 $cb;
263	&schedule;
264	}
265	};	495	};
266		496
267	last if $@ eq "\3async_pool terminate\2\n";
268	warn $@ if $@;	497	warn $@ if $@;
269	}	498	}
270	}	499	}
271		500
272	sub async_pool(&@) {	501	=back
273	# this is also inlined into the unlock_scheduler
274	my $coro = (pop @async_pool) || new Coro \&pool_handler;
275		502
276	$coro->{_invoke} = [@_];	503	=head1 STATIC METHODS
277	$coro->ready;
278		504
279	$coro	505	Static methods are actually functions that implicitly operate on the
280	}	506	current coro.
		507
		508	=over 4
281		509
282	=item schedule	510	=item schedule
283		511
284	Calls the scheduler. Please note that the current coroutine will not be put	512	Calls the scheduler. The scheduler will find the next coro that is
		513	to be run from the ready queue and switches to it. The next coro
		514	to be run is simply the one with the highest priority that is longest
		515	in its ready queue. If there is no coro ready, it will call the
		516	C<$Coro::idle> hook.
		517
		518	Please note that the current coro will I<not> be put into the ready
285	into the ready queue, so calling this function usually means you will	519	queue, so calling this function usually means you will never be called
286	never be called again unless something else (e.g. an event handler) calls	520	again unless something else (e.g. an event handler) calls C<< ->ready >>,
287	ready.	521	thus waking you up.
288		522
289	The canonical way to wait on external events is this:	523	This makes C<schedule> I<the> generic method to use to block the current
		524	coro and wait for events: first you remember the current coro in
		525	a variable, then arrange for some callback of yours to call C<< ->ready
		526	>> on that once some event happens, and last you call C<schedule> to put
		527	yourself to sleep. Note that a lot of things can wake your coro up,
		528	so you need to check whether the event indeed happened, e.g. by storing the
		529	status in a variable.
290		530
291	{	531	See B<HOW TO WAIT FOR A CALLBACK>, below, for some ways to wait for callbacks.
292	# remember current coroutine
293	my $current = $Coro::current;
294		532
295	# register a hypothetical event handler	533	=item cede
296	on_event_invoke sub {	534
297	# wake up sleeping coroutine	535	"Cede" to other coros. This function puts the current coro into
298	$current->ready;	536	the ready queue and calls C<schedule>, which has the effect of giving
299	undef $current;	537	up the current "timeslice" to other coros of the same or higher
		538	priority. Once your coro gets its turn again it will automatically be
		539	resumed.
		540
		541	This function is often called C<yield> in other languages.
		542
		543	=item Coro::cede_notself
		544
		545	Works like cede, but is not exported by default and will cede to I<any>
		546	coro, regardless of priority. This is useful sometimes to ensure
		547	progress is made.
		548
		549	=item terminate [arg...]
		550
		551	Terminates the current coro with the given status values (see
		552	L<cancel>). The values will not be copied, but referenced directly.
		553
		554	=item Coro::on_enter BLOCK, Coro::on_leave BLOCK
		555
		556	These function install enter and leave winders in the current scope. The
		557	enter block will be executed when on_enter is called and whenever the
		558	current coro is re-entered by the scheduler, while the leave block is
		559	executed whenever the current coro is blocked by the scheduler, and
		560	also when the containing scope is exited (by whatever means, be it exit,
		561	die, last etc.).
		562
		563	I<Neither invoking the scheduler, nor exceptions, are allowed within those
		564	BLOCKs>. That means: do not even think about calling C<die> without an
		565	eval, and do not even think of entering the scheduler in any way.
		566
		567	Since both BLOCKs are tied to the current scope, they will automatically
		568	be removed when the current scope exits.
		569
		570	These functions implement the same concept as C<dynamic-wind> in scheme
		571	does, and are useful when you want to localise some resource to a specific
		572	coro.
		573
		574	They slow down thread switching considerably for coros that use them
		575	(about 40% for a BLOCK with a single assignment, so thread switching is
		576	still reasonably fast if the handlers are fast).
		577
		578	These functions are best understood by an example: The following function
		579	will change the current timezone to "Antarctica/South_Pole", which
		580	requires a call to C<tzset>, but by using C<on_enter> and C<on_leave>,
		581	which remember/change the current timezone and restore the previous
		582	value, respectively, the timezone is only changed for the coro that
		583	installed those handlers.
		584
		585	use POSIX qw(tzset);
		586
		587	async {
		588	my $old_tz; # store outside TZ value here
		589
		590	Coro::on_enter {
		591	$old_tz = $ENV{TZ}; # remember the old value
		592
		593	$ENV{TZ} = "Antarctica/South_Pole";
		594	tzset; # enable new value
300	};	595	};
301		596
302	# call schedule until event occurred.	597	Coro::on_leave {
303	# in case we are woken up for other reasons	598	$ENV{TZ} = $old_tz;
304	# (current still defined), loop.	599	tzset; # restore old value
305	Coro::schedule while $current;	600	};
		601
		602	# at this place, the timezone is Antarctica/South_Pole,
		603	# without disturbing the TZ of any other coro.
306	}	604	};
307		605
308	=item cede	606	This can be used to localise about any resource (locale, uid, current
		607	working directory etc.) to a block, despite the existance of other
		608	coros.
309		609
310	"Cede" to other coroutines. This function puts the current coroutine into the	610	Another interesting example implements time-sliced multitasking using
311	ready queue and calls C<schedule>, which has the effect of giving up the	611	interval timers (this could obviously be optimised, but does the job):
312	current "timeslice" to other coroutines of the same or higher priority.
313		612
314	=item Coro::cede_notself	613	# "timeslice" the given block
		614	sub timeslice(&) {
		615	use Time::HiRes ();
315		616
316	Works like cede, but is not exported by default and will cede to any	617	Coro::on_enter {
317	coroutine, regardless of priority, once.	618	# on entering the thread, we set an VTALRM handler to cede
		619	$SIG{VTALRM} = sub { cede };
		620	# and then start the interval timer
		621	Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0.01, 0.01;
		622	};
		623	Coro::on_leave {
		624	# on leaving the thread, we stop the interval timer again
		625	Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0, 0;
		626	};
318		627
319	=item terminate [arg...]	628	&{+shift};
		629	}
320		630
321	Terminates the current coroutine with the given status values (see L<cancel>).	631	# use like this:
		632	timeslice {
		633	# The following is an endless loop that would normally
		634	# monopolise the process. Since it runs in a timesliced
		635	# environment, it will regularly cede to other threads.
		636	while () { }
		637	};
		638
322		639
323	=item killall	640	=item killall
324		641
325	Kills/terminates/cancels all coroutines except the currently running	642	Kills/terminates/cancels all coros except the currently running one.
326	one. This is useful after a fork, either in the child or the parent, as
327	usually only one of them should inherit the running coroutines.
328		643
329	=cut	644	Note that while this will try to free some of the main interpreter
		645	resources if the calling coro isn't the main coro, but one
		646	cannot free all of them, so if a coro that is not the main coro
		647	calls this function, there will be some one-time resource leak.
330		648
331	sub terminate {	649	=cut
332	$current->cancel (@_);
333	}
334		650
335	sub killall {	651	sub killall {
336	for (Coro::State::list) {	652	for (Coro::State::list) {
337	$_->cancel	653	$_->cancel
338	if $_ != $current && UNIVERSAL::isa $_, "Coro";	654	if $_ != $current && UNIVERSAL::isa $_, "Coro";
339	}	655	}
340	}	656	}
341		657
342	=back	658	=back
343		659
344	# dynamic methods
345
346	=head2 COROUTINE METHODS	660	=head1 CORO OBJECT METHODS
347		661
348	These are the methods you can call on coroutine objects.	662	These are the methods you can call on coro objects (or to create
		663	them).
349		664
350	=over 4	665	=over 4
351		666
352	=item new Coro \&sub [, @args...]	667	=item new Coro \&sub [, @args...]
353		668
354	Create a new coroutine and return it. When the sub returns the coroutine	669	Create a new coro and return it. When the sub returns, the coro
355	automatically terminates as if C<terminate> with the returned values were	670	automatically terminates as if C<terminate> with the returned values were
356	called. To make the coroutine run you must first put it into the ready queue	671	called. To make the coro run you must first put it into the ready
357	by calling the ready method.	672	queue by calling the ready method.
358		673
359	See C<async> and C<Coro::State::new> for additional info about the	674	See C<async> and C<Coro::State::new> for additional info about the
360	coroutine environment.	675	coro environment.
361		676
362	=cut	677	=cut
363		678
364	sub _run_coro {	679	sub _coro_run {
365	terminate &{+shift};	680	terminate &{+shift};
366	}	681	}
367		682
368	sub new {
369	my $class = shift;
370
371	$class->SUPER::new (\&_run_coro, @_)
372	}
373
374	=item $success = $coroutine->ready	683	=item $success = $coro->ready
375		684
376	Put the given coroutine into the ready queue (according to it's priority)	685	Put the given coro into the end of its ready queue (there is one
377	and return true. If the coroutine is already in the ready queue, do nothing	686	queue for each priority) and return true. If the coro is already in
378	and return false.	687	the ready queue, do nothing and return false.
379		688
		689	This ensures that the scheduler will resume this coro automatically
		690	once all the coro of higher priority and all coro of the same
		691	priority that were put into the ready queue earlier have been resumed.
		692
		693	=item $coro->suspend
		694
		695	Suspends the specified coro. A suspended coro works just like any other
		696	coro, except that the scheduler will not select a suspended coro for
		697	execution.
		698
		699	Suspending a coro can be useful when you want to keep the coro from
		700	running, but you don't want to destroy it, or when you want to temporarily
		701	freeze a coro (e.g. for debugging) to resume it later.
		702
		703	A scenario for the former would be to suspend all (other) coros after a
		704	fork and keep them alive, so their destructors aren't called, but new
		705	coros can be created.
		706
		707	=item $coro->resume
		708
		709	If the specified coro was suspended, it will be resumed. Note that when
		710	the coro was in the ready queue when it was suspended, it might have been
		711	unreadied by the scheduler, so an activation might have been lost.
		712
		713	To avoid this, it is best to put a suspended coro into the ready queue
		714	unconditionally, as every synchronisation mechanism must protect itself
		715	against spurious wakeups, and the one in the Coro family certainly do
		716	that.
		717
380	=item $is_ready = $coroutine->is_ready	718	=item $is_ready = $coro->is_ready
381		719
382	Return wether the coroutine is currently the ready queue or not,	720	Returns true iff the Coro object is in the ready queue. Unless the Coro
		721	object gets destroyed, it will eventually be scheduled by the scheduler.
383		722
		723	=item $is_running = $coro->is_running
		724
		725	Returns true iff the Coro object is currently running. Only one Coro object
		726	can ever be in the running state (but it currently is possible to have
		727	multiple running Coro::States).
		728
		729	=item $is_suspended = $coro->is_suspended
		730
		731	Returns true iff this Coro object has been suspended. Suspended Coros will
		732	not ever be scheduled.
		733
384	=item $coroutine->cancel (arg...)	734	=item $coro->cancel (arg...)
385		735
386	Terminates the given coroutine and makes it return the given arguments as	736	Terminates the given Coro thread and makes it return the given arguments as
387	status (default: the empty list). Never returns if the coroutine is the	737	status (default: an empty list). Never returns if the Coro is the
388	current coroutine.	738	current Coro.
389		739
390	=cut	740	This is a rather brutal way to free a coro, with some limitations - if
		741	the thread is inside a C callback that doesn't expect to be canceled,
		742	bad things can happen, or if the cancelled thread insists on running
		743	complicated cleanup handlers that rely on it'S thread context, things will
		744	not work.
391		745
392	sub cancel {	746	Any cleanup code being run (e.g. from C<guard> blocks) will be run without
393	my $self = shift;	747	a thread context, and is not allowed to switch to other threads. On the
394	$self->{_status} = [@_];	748	plus side, C<< ->cancel >> will always clean up the thread, no matter
		749	what. If your cleanup code is complex or you want to avoid cancelling a
		750	C-thread that doesn't know how to clean up itself, it can be better to C<<
		751	->throw >> an exception, or use C<< ->safe_cancel >>.
395		752
396	if ($current == $self) {	753	The arguments to C<< ->cancel >> are not copied, but instead will
397	push @destroy, $self;	754	be referenced directly (e.g. if you pass C<$var> and after the call
398	$manager->ready;	755	change that variable, then you might change the return values passed to
399	&schedule while 1;	756	e.g. C<join>, so don't do that).
400	} else {	757
401	$self->_cancel;	758	The resources of the Coro are usually freed (or destructed) before this
		759	call returns, but this can be delayed for an indefinite amount of time, as
		760	in some cases the manager thread has to run first to actually destruct the
		761	Coro object.
		762
		763	=item $coro->safe_cancel ($arg...)
		764
		765	Works mostly like C<< ->cancel >>, but is inherently "safer", and
		766	consequently, can fail with an exception in cases the thread is not in a
		767	cancellable state.
		768
		769	This method works a bit like throwing an exception that cannot be caught
		770	- specifically, it will clean up the thread from within itself, so
		771	all cleanup handlers (e.g. C<guard> blocks) are run with full thread
		772	context and can block if they wish. The downside is that there is no
		773	guarantee that the thread can be cancelled when you call this method, and
		774	therefore, it might fail. It is also considerably slower than C<cancel> or
		775	C<terminate>.
		776
		777	A thread is in a safe-cancellable state if it either hasn't been run yet,
		778	or it has no C context attached and is inside an SLF function.
		779
		780	The latter two basically mean that the thread isn't currently inside a
		781	perl callback called from some C function (usually via some XS modules)
		782	and isn't currently executing inside some C function itself (via Coro's XS
		783	API).
		784
		785	This call returns true when it could cancel the thread, or croaks with an
		786	error otherwise (i.e. it either returns true or doesn't return at all).
		787
		788	Why the weird interface? Well, there are two common models on how and
		789	when to cancel things. In the first, you have the expectation that your
		790	coro thread can be cancelled when you want to cancel it - if the thread
		791	isn't cancellable, this would be a bug somewhere, so C<< ->safe_cancel >>
		792	croaks to notify of the bug.
		793
		794	In the second model you sometimes want to ask nicely to cancel a thread,
		795	but if it's not a good time, well, then don't cancel. This can be done
		796	relatively easy like this:
		797
		798	if (! eval { $coro->safe_cancel }) {
		799	warn "unable to cancel thread: $@";
402	}	800	}
403	}
404		801
		802	However, what you never should do is first try to cancel "safely" and
		803	if that fails, cancel the "hard" way with C<< ->cancel >>. That makes
		804	no sense: either you rely on being able to execute cleanup code in your
		805	thread context, or you don't. If you do, then C<< ->safe_cancel >> is the
		806	only way, and if you don't, then C<< ->cancel >> is always faster and more
		807	direct.
		808
		809	=item $coro->schedule_to
		810
		811	Puts the current coro to sleep (like C<Coro::schedule>), but instead
		812	of continuing with the next coro from the ready queue, always switch to
		813	the given coro object (regardless of priority etc.). The readyness
		814	state of that coro isn't changed.
		815
		816	This is an advanced method for special cases - I'd love to hear about any
		817	uses for this one.
		818
		819	=item $coro->cede_to
		820
		821	Like C<schedule_to>, but puts the current coro into the ready
		822	queue. This has the effect of temporarily switching to the given
		823	coro, and continuing some time later.
		824
		825	This is an advanced method for special cases - I'd love to hear about any
		826	uses for this one.
		827
		828	=item $coro->throw ([$scalar])
		829
		830	If C<$throw> is specified and defined, it will be thrown as an exception
		831	inside the coro at the next convenient point in time. Otherwise
		832	clears the exception object.
		833
		834	Coro will check for the exception each time a schedule-like-function
		835	returns, i.e. after each C<schedule>, C<cede>, C<< Coro::Semaphore->down
		836	>>, C<< Coro::Handle->readable >> and so on. Most of those functions (all
		837	that are part of Coro itself) detect this case and return early in case an
		838	exception is pending.
		839
		840	The exception object will be thrown "as is" with the specified scalar in
		841	C<$@>, i.e. if it is a string, no line number or newline will be appended
		842	(unlike with C<die>).
		843
		844	This can be used as a softer means than either C<cancel> or C<safe_cancel
		845	>to ask a coro to end itself, although there is no guarantee that the
		846	exception will lead to termination, and if the exception isn't caught it
		847	might well end the whole program.
		848
		849	You might also think of C<throw> as being the moral equivalent of
		850	C<kill>ing a coro with a signal (in this case, a scalar).
		851
405	=item $coroutine->join	852	=item $coro->join
406		853
407	Wait until the coroutine terminates and return any values given to the	854	Wait until the coro terminates and return any values given to the
408	C<terminate> or C<cancel> functions. C<join> can be called concurrently	855	C<terminate> or C<cancel> functions. C<join> can be called concurrently
409	from multiple coroutines.	856	from multiple threads, and all will be resumed and given the status
		857	return once the C<$coro> terminates.
410		858
411	=cut	859	=cut
412		860
413	sub join {	861	sub xjoin {
414	my $self = shift;	862	my $self = shift;
415		863
416	unless ($self->{_status}) {	864	unless ($self->{_status}) {
417	my $current = $current;	865	my $current = $current;
418		866
…		…
422	};	870	};
423		871
424	&schedule while $current;	872	&schedule while $current;
425	}	873	}
426		874
427	wantarray ? @{$self->{_status}} : $self->{_status}[0];	875	wantarray ? @{$self->{_status}} : $self->{_status}[0]
428	}	876	}
429		877
430	=item $coroutine->on_destroy (\&cb)	878	=item $coro->on_destroy (\&cb)
431		879
432	Registers a callback that is called when this coroutine gets destroyed,	880	Registers a callback that is called when this coro thread gets destroyed,
433	but before it is joined. The callback gets passed the terminate arguments,	881	that is, after it's resources have been freed but before it is joined. The
434	if any.	882	callback gets passed the terminate/cancel arguments, if any, and I<must
		883	not> die, under any circumstances.
435		884
436	=cut	885	There can be any number of C<on_destroy> callbacks per coro, and there is
		886	no way currently to remove a callback once added.
437		887
		888	=cut
		889
438	sub on_destroy {	890	sub xon_destroy {
439	my ($self, $cb) = @_;	891	my ($self, $cb) = @_;
440		892
441	push @{ $self->{_on_destroy} }, $cb;	893	push @{ $self->{_on_destroy} }, $cb;
442	}	894	}
443		895
444	=item $oldprio = $coroutine->prio ($newprio)	896	=item $oldprio = $coro->prio ($newprio)
445		897
446	Sets (or gets, if the argument is missing) the priority of the	898	Sets (or gets, if the argument is missing) the priority of the
447	coroutine. Higher priority coroutines get run before lower priority	899	coro thread. Higher priority coro get run before lower priority
448	coroutines. Priorities are small signed integers (currently -4 .. +3),	900	coros. Priorities are small signed integers (currently -4 .. +3),
449	that you can refer to using PRIO_xxx constants (use the import tag :prio	901	that you can refer to using PRIO_xxx constants (use the import tag :prio
450	to get then):	902	to get then):
451		903
452	PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN	904	PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN
453	3 > 1 > 0 > -1 > -3 > -4	905	3 > 1 > 0 > -1 > -3 > -4
454		906
455	# set priority to HIGH	907	# set priority to HIGH
456	current->prio(PRIO_HIGH);	908	current->prio (PRIO_HIGH);
457		909
458	The idle coroutine ($Coro::idle) always has a lower priority than any	910	The idle coro thread ($Coro::idle) always has a lower priority than any
459	existing coroutine.	911	existing coro.
460		912
461	Changing the priority of the current coroutine will take effect immediately,	913	Changing the priority of the current coro will take effect immediately,
462	but changing the priority of coroutines in the ready queue (but not	914	but changing the priority of a coro in the ready queue (but not running)
463	running) will only take effect after the next schedule (of that	915	will only take effect after the next schedule (of that coro). This is a
464	coroutine). This is a bug that will be fixed in some future version.	916	bug that will be fixed in some future version.
465		917
466	=item $newprio = $coroutine->nice ($change)	918	=item $newprio = $coro->nice ($change)
467		919
468	Similar to C<prio>, but subtract the given value from the priority (i.e.	920	Similar to C<prio>, but subtract the given value from the priority (i.e.
469	higher values mean lower priority, just as in unix).	921	higher values mean lower priority, just as in UNIX's nice command).
470		922
471	=item $olddesc = $coroutine->desc ($newdesc)	923	=item $olddesc = $coro->desc ($newdesc)
472		924
473	Sets (or gets in case the argument is missing) the description for this	925	Sets (or gets in case the argument is missing) the description for this
474	coroutine. This is just a free-form string you can associate with a coroutine.	926	coro thread. This is just a free-form string you can associate with a
		927	coro.
475		928
476	This method simply sets the C<< $coroutine->{desc} >> member to the given string. You	929	This method simply sets the C<< $coro->{desc} >> member to the given
477	can modify this member directly if you wish.	930	string. You can modify this member directly if you wish, and in fact, this
		931	is often preferred to indicate major processing states that cna then be
		932	seen for example in a L<Coro::Debug> session:
478		933
479	=item $coroutine->throw ([$scalar])	934	sub my_long_function {
480		935	local $Coro::current->{desc} = "now in my_long_function";
481	If C<$throw> is specified and defined, it will be thrown as an exception	936	...
482	inside the coroutine at the next convinient point in time (usually after	937	$Coro::current->{desc} = "my_long_function: phase 1";
483	it gains control at the next schedule/transfer/cede). Otherwise clears the	938	...
484	exception object.	939	$Coro::current->{desc} = "my_long_function: phase 2";
485		940	...
486	The exception object will be thrown "as is" with the specified scalar in	941	}
487	C<$@>, i.e. if it is a string, no line number or newline will be appended
488	(unlike with C<die>).
489
490	This can be used as a softer means than C<cancel> to ask a coroutine to
491	end itself, although there is no guarentee that the exception will lead to
492	termination, and if the exception isn't caught it might well end the whole
493	program.
494		942
495	=cut	943	=cut
496		944
497	sub desc {	945	sub desc {
498	my $old = $_[0]{desc};	946	my $old = $_[0]{desc};
499	$_[0]{desc} = $_[1] if @_ > 1;	947	$_[0]{desc} = $_[1] if @_ > 1;
500	$old;	948	$old;
501	}	949	}
502		950
		951	sub transfer {
		952	require Carp;
		953	Carp::croak ("You must not call ->transfer on Coro objects. Use Coro::State objects or the ->schedule_to method. Caught");
		954	}
		955
503	=back	956	=back
504		957
505	=head2 GLOBAL FUNCTIONS	958	=head1 GLOBAL FUNCTIONS
506		959
507	=over 4	960	=over 4
508		961
509	=item Coro::nready	962	=item Coro::nready
510		963
511	Returns the number of coroutines that are currently in the ready state,	964	Returns the number of coro that are currently in the ready state,
512	i.e. that can be switched to. The value C<0> means that the only runnable	965	i.e. that can be switched to by calling C<schedule> directory or
		966	indirectly. The value C<0> means that the only runnable coro is the
513	coroutine is the currently running one, so C<cede> would have no effect,	967	currently running one, so C<cede> would have no effect, and C<schedule>
514	and C<schedule> would cause a deadlock unless there is an idle handler	968	would cause a deadlock unless there is an idle handler that wakes up some
515	that wakes up some coroutines.	969	coro.
516		970
517	=item my $guard = Coro::guard { ... }	971	=item my $guard = Coro::guard { ... }
518		972
519	This creates and returns a guard object. Nothing happens until the object	973	This function still exists, but is deprecated. Please use the
520	gets destroyed, in which case the codeblock given as argument will be	974	C<Guard::guard> function instead.
521	executed. This is useful to free locks or other resources in case of a
522	runtime error or when the coroutine gets canceled, as in both cases the
523	guard block will be executed. The guard object supports only one method,
524	C<< ->cancel >>, which will keep the codeblock from being executed.
525		975
526	Example: set some flag and clear it again when the coroutine gets canceled
527	or the function returns:
528
529	sub do_something {
530	my $guard = Coro::guard { $busy = 0 };
531	$busy = 1;
532
533	# do something that requires $busy to be true
534	}
535
536	=cut	976	=cut
537		977
538	sub guard(&) {	978	BEGIN { *guard = \&Guard::guard }
539	bless \(my $cb = $_[0]), "Coro::guard"
540	}
541
542	sub Coro::guard::cancel {
543	${$_[0]} = sub { };
544	}
545
546	sub Coro::guard::DESTROY {
547	${$_[0]}->();
548	}
549
550		979
551	=item unblock_sub { ... }	980	=item unblock_sub { ... }
552		981
553	This utility function takes a BLOCK or code reference and "unblocks" it,	982	This utility function takes a BLOCK or code reference and "unblocks" it,
554	returning the new coderef. This means that the new coderef will return	983	returning a new coderef. Unblocking means that calling the new coderef
555	immediately without blocking, returning nothing, while the original code	984	will return immediately without blocking, returning nothing, while the
556	ref will be called (with parameters) from within its own coroutine.	985	original code ref will be called (with parameters) from within another
		986	coro.
557		987
558	The reason this function exists is that many event libraries (such as the	988	The reason this function exists is that many event libraries (such as
559	venerable L<Event|Event> module) are not coroutine-safe (a weaker form	989	the venerable L<Event|Event> module) are not thread-safe (a weaker form
560	of thread-safety). This means you must not block within event callbacks,	990	of reentrancy). This means you must not block within event callbacks,
561	otherwise you might suffer from crashes or worse.	991	otherwise you might suffer from crashes or worse. The only event library
		992	currently known that is safe to use without C<unblock_sub> is L<EV> (but
		993	you might still run into deadlocks if all event loops are blocked).
		994
		995	Coro will try to catch you when you block in the event loop
		996	("FATAL:$Coro::IDLE blocked itself"), but this is just best effort and
		997	only works when you do not run your own event loop.
562		998
563	This function allows your callbacks to block by executing them in another	999	This function allows your callbacks to block by executing them in another
564	coroutine where it is safe to block. One example where blocking is handy	1000	coro where it is safe to block. One example where blocking is handy
565	is when you use the L<Coro::AIO|Coro::AIO> functions to save results to	1001	is when you use the L<Coro::AIO|Coro::AIO> functions to save results to
566	disk.	1002	disk, for example.
567		1003
568	In short: simply use C<unblock_sub { ... }> instead of C<sub { ... }> when	1004	In short: simply use C<unblock_sub { ... }> instead of C<sub { ... }> when
569	creating event callbacks that want to block.	1005	creating event callbacks that want to block.
		1006
		1007	If your handler does not plan to block (e.g. simply sends a message to
		1008	another coro, or puts some other coro into the ready queue), there is
		1009	no reason to use C<unblock_sub>.
		1010
		1011	Note that you also need to use C<unblock_sub> for any other callbacks that
		1012	are indirectly executed by any C-based event loop. For example, when you
		1013	use a module that uses L<AnyEvent> (and you use L<Coro::AnyEvent>) and it
		1014	provides callbacks that are the result of some event callback, then you
		1015	must not block either, or use C<unblock_sub>.
570		1016
571	=cut	1017	=cut
572		1018
573	our @unblock_queue;	1019	our @unblock_queue;
574		1020
…		…
577	# return immediately and can be reused) and because we cannot cede	1023	# return immediately and can be reused) and because we cannot cede
578	# inside an event callback.	1024	# inside an event callback.
579	our $unblock_scheduler = new Coro sub {	1025	our $unblock_scheduler = new Coro sub {
580	while () {	1026	while () {
581	while (my $cb = pop @unblock_queue) {	1027	while (my $cb = pop @unblock_queue) {
582	# this is an inlined copy of async_pool	1028	&async_pool (@$cb);
583	my $coro = (pop @async_pool) || new Coro \&pool_handler;
584		1029
585	$coro->{_invoke} = $cb;
586	$coro->ready;
587	cede; # for short-lived callbacks, this reduces pressure on the coro pool	1030	# for short-lived callbacks, this reduces pressure on the coro pool
		1031	# as the chance is very high that the async_poll coro will be back
		1032	# in the idle state when cede returns
		1033	cede;
588	}	1034	}
589	schedule; # sleep well	1035	schedule; # sleep well
590	}	1036	}
591	};	1037	};
592	$unblock_scheduler->desc ("[unblock_sub scheduler]");	1038	$unblock_scheduler->{desc} = "[unblock_sub scheduler]";
593		1039
594	sub unblock_sub(&) {	1040	sub unblock_sub(&) {
595	my $cb = shift;	1041	my $cb = shift;
596		1042
597	sub {	1043	sub {
598	unshift @unblock_queue, [$cb, @_];	1044	unshift @unblock_queue, [$cb, @_];
599	$unblock_scheduler->ready;	1045	$unblock_scheduler->ready;
600	}	1046	}
601	}	1047	}
602		1048
		1049	=item $cb = rouse_cb
		1050
		1051	Create and return a "rouse callback". That's a code reference that,
		1052	when called, will remember a copy of its arguments and notify the owner
		1053	coro of the callback.
		1054
		1055	See the next function.
		1056
		1057	=item @args = rouse_wait [$cb]
		1058
		1059	Wait for the specified rouse callback (or the last one that was created in
		1060	this coro).
		1061
		1062	As soon as the callback is invoked (or when the callback was invoked
		1063	before C<rouse_wait>), it will return the arguments originally passed to
		1064	the rouse callback. In scalar context, that means you get the I<last>
		1065	argument, just as if C<rouse_wait> had a C<return ($a1, $a2, $a3...)>
		1066	statement at the end.
		1067
		1068	See the section B<HOW TO WAIT FOR A CALLBACK> for an actual usage example.
		1069
603	=back	1070	=back
604		1071
605	=cut	1072	=cut
606		1073
		1074	for my $module (qw(Channel RWLock Semaphore SemaphoreSet Signal Specific)) {
		1075	my $old = defined &{"Coro::$module\::new"} && \&{"Coro::$module\::new"};
		1076
		1077	*{"Coro::$module\::new"} = sub {
		1078	require "Coro/$module.pm";
		1079
		1080	# some modules have their new predefined in State.xs, some don't
		1081	*{"Coro::$module\::new"} = $old
		1082	if $old;
		1083
		1084	goto &{"Coro::$module\::new"};
		1085	};
		1086	}
		1087
607	1;	1088	1;
608		1089
		1090	=head1 HOW TO WAIT FOR A CALLBACK
		1091
		1092	It is very common for a coro to wait for some callback to be
		1093	called. This occurs naturally when you use coro in an otherwise
		1094	event-based program, or when you use event-based libraries.
		1095
		1096	These typically register a callback for some event, and call that callback
		1097	when the event occured. In a coro, however, you typically want to
		1098	just wait for the event, simplyifying things.
		1099
		1100	For example C<< AnyEvent->child >> registers a callback to be called when
		1101	a specific child has exited:
		1102
		1103	my $child_watcher = AnyEvent->child (pid => $pid, cb => sub { ... });
		1104
		1105	But from within a coro, you often just want to write this:
		1106
		1107	my $status = wait_for_child $pid;
		1108
		1109	Coro offers two functions specifically designed to make this easy,
		1110	C<Coro::rouse_cb> and C<Coro::rouse_wait>.
		1111
		1112	The first function, C<rouse_cb>, generates and returns a callback that,
		1113	when invoked, will save its arguments and notify the coro that
		1114	created the callback.
		1115
		1116	The second function, C<rouse_wait>, waits for the callback to be called
		1117	(by calling C<schedule> to go to sleep) and returns the arguments
		1118	originally passed to the callback.
		1119
		1120	Using these functions, it becomes easy to write the C<wait_for_child>
		1121	function mentioned above:
		1122
		1123	sub wait_for_child($) {
		1124	my ($pid) = @_;
		1125
		1126	my $watcher = AnyEvent->child (pid => $pid, cb => Coro::rouse_cb);
		1127
		1128	my ($rpid, $rstatus) = Coro::rouse_wait;
		1129	$rstatus
		1130	}
		1131
		1132	In the case where C<rouse_cb> and C<rouse_wait> are not flexible enough,
		1133	you can roll your own, using C<schedule>:
		1134
		1135	sub wait_for_child($) {
		1136	my ($pid) = @_;
		1137
		1138	# store the current coro in $current,
		1139	# and provide result variables for the closure passed to ->child
		1140	my $current = $Coro::current;
		1141	my ($done, $rstatus);
		1142
		1143	# pass a closure to ->child
		1144	my $watcher = AnyEvent->child (pid => $pid, cb => sub {
		1145	$rstatus = $_[1]; # remember rstatus
		1146	$done = 1; # mark $rstatus as valud
		1147	});
		1148
		1149	# wait until the closure has been called
		1150	schedule while !$done;
		1151
		1152	$rstatus
		1153	}
		1154
		1155
609	=head1 BUGS/LIMITATIONS	1156	=head1 BUGS/LIMITATIONS
610		1157
611	- you must make very sure that no coro is still active on global	1158	=over 4
612	destruction. very bad things might happen otherwise (usually segfaults).
613		1159
		1160	=item fork with pthread backend
		1161
		1162	When Coro is compiled using the pthread backend (which isn't recommended
		1163	but required on many BSDs as their libcs are completely broken), then
		1164	coro will not survive a fork. There is no known workaround except to
		1165	fix your libc and use a saner backend.
		1166
		1167	=item perl process emulation ("threads")
		1168
614	- this module is not thread-safe. You should only ever use this module	1169	This module is not perl-pseudo-thread-safe. You should only ever use this
615	from the same thread (this requirement might be loosened in the future	1170	module from the first thread (this requirement might be removed in the
616	to allow per-thread schedulers, but Coro::State does not yet allow	1171	future to allow per-thread schedulers, but Coro::State does not yet allow
617	this).	1172	this). I recommend disabling thread support and using processes, as having
		1173	the windows process emulation enabled under unix roughly halves perl
		1174	performance, even when not used.
		1175
		1176	Attempts to use threads created in another emulated process will crash
		1177	("cleanly", with a null pointer exception).
		1178
		1179	=item coro switching is not signal safe
		1180
		1181	You must not switch to another coro from within a signal handler (only
		1182	relevant with %SIG - most event libraries provide safe signals), I<unless>
		1183	you are sure you are not interrupting a Coro function.
		1184
		1185	That means you I<MUST NOT> call any function that might "block" the
		1186	current coro - C<cede>, C<schedule> C<< Coro::Semaphore->down >> or
		1187	anything that calls those. Everything else, including calling C<ready>,
		1188	works.
		1189
		1190	=back
		1191
		1192
		1193	=head1 WINDOWS PROCESS EMULATION
		1194
		1195	A great many people seem to be confused about ithreads (for example, Chip
		1196	Salzenberg called me unintelligent, incapable, stupid and gullible,
		1197	while in the same mail making rather confused statements about perl
		1198	ithreads (for example, that memory or files would be shared), showing his
		1199	lack of understanding of this area - if it is hard to understand for Chip,
		1200	it is probably not obvious to everybody).
		1201
		1202	What follows is an ultra-condensed version of my talk about threads in
		1203	scripting languages given on the perl workshop 2009:
		1204
		1205	The so-called "ithreads" were originally implemented for two reasons:
		1206	first, to (badly) emulate unix processes on native win32 perls, and
		1207	secondly, to replace the older, real thread model ("5.005-threads").
		1208
		1209	It does that by using threads instead of OS processes. The difference
		1210	between processes and threads is that threads share memory (and other
		1211	state, such as files) between threads within a single process, while
		1212	processes do not share anything (at least not semantically). That
		1213	means that modifications done by one thread are seen by others, while
		1214	modifications by one process are not seen by other processes.
		1215
		1216	The "ithreads" work exactly like that: when creating a new ithreads
		1217	process, all state is copied (memory is copied physically, files and code
		1218	is copied logically). Afterwards, it isolates all modifications. On UNIX,
		1219	the same behaviour can be achieved by using operating system processes,
		1220	except that UNIX typically uses hardware built into the system to do this
		1221	efficiently, while the windows process emulation emulates this hardware in
		1222	software (rather efficiently, but of course it is still much slower than
		1223	dedicated hardware).
		1224
		1225	As mentioned before, loading code, modifying code, modifying data
		1226	structures and so on is only visible in the ithreads process doing the
		1227	modification, not in other ithread processes within the same OS process.
		1228
		1229	This is why "ithreads" do not implement threads for perl at all, only
		1230	processes. What makes it so bad is that on non-windows platforms, you can
		1231	actually take advantage of custom hardware for this purpose (as evidenced
		1232	by the forks module, which gives you the (i-) threads API, just much
		1233	faster).
		1234
		1235	Sharing data is in the i-threads model is done by transfering data
		1236	structures between threads using copying semantics, which is very slow -
		1237	shared data simply does not exist. Benchmarks using i-threads which are
		1238	communication-intensive show extremely bad behaviour with i-threads (in
		1239	fact, so bad that Coro, which cannot take direct advantage of multiple
		1240	CPUs, is often orders of magnitude faster because it shares data using
		1241	real threads, refer to my talk for details).
		1242
		1243	As summary, i-threads *use* threads to implement processes, while
		1244	the compatible forks module *uses* processes to emulate, uhm,
		1245	processes. I-threads slow down every perl program when enabled, and
		1246	outside of windows, serve no (or little) practical purpose, but
		1247	disadvantages every single-threaded Perl program.
		1248
		1249	This is the reason that I try to avoid the name "ithreads", as it is
		1250	misleading as it implies that it implements some kind of thread model for
		1251	perl, and prefer the name "windows process emulation", which describes the
		1252	actual use and behaviour of it much better.
618		1253
619	=head1 SEE ALSO	1254	=head1 SEE ALSO
620		1255
621	Lower level Configuration, Coroutine Environment: L<Coro::State>.	1256	Event-Loop integration: L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>.
622		1257
623	Debugging: L<Coro::Debug>.	1258	Debugging: L<Coro::Debug>.
624		1259
625	Support/Utility: L<Coro::Specific>, L<Coro::Util>.	1260	Support/Utility: L<Coro::Specific>, L<Coro::Util>.
626		1261
627	Locking/IPC: L<Coro::Signal>, L<Coro::Channel>, L<Coro::Semaphore>, L<Coro::SemaphoreSet>, L<Coro::RWLock>.	1262	Locking and IPC: L<Coro::Signal>, L<Coro::Channel>, L<Coro::Semaphore>,
		1263	L<Coro::SemaphoreSet>, L<Coro::RWLock>.
628		1264
629	Event/IO: L<Coro::Timer>, L<Coro::Event>, L<Coro::Handle>, L<Coro::Socket>.	1265	I/O and Timers: L<Coro::Timer>, L<Coro::Handle>, L<Coro::Socket>, L<Coro::AIO>.
630		1266
631	Compatibility: L<Coro::LWP>, L<Coro::Storable>, L<Coro::Select>.	1267	Compatibility with other modules: L<Coro::LWP> (but see also L<AnyEvent::HTTP> for
		1268	a better-working alternative), L<Coro::BDB>, L<Coro::Storable>,
		1269	L<Coro::Select>.
632		1270
633	Embedding: L<Coro::MakeMaker>.	1271	XS API: L<Coro::MakeMaker>.
		1272
		1273	Low level Configuration, Thread Environment, Continuations: L<Coro::State>.
634		1274
635	=head1 AUTHOR	1275	=head1 AUTHOR
636		1276
637	Marc Lehmann <schmorp@schmorp.de>	1277	Marc Lehmann <schmorp@schmorp.de>
638	http://home.schmorp.de/	1278	http://home.schmorp.de/

Diff Legend

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