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

Comparing Coro/Coro.pm (file contents):
Revision 1.29 by root, Sat Aug 11 00:37:31 2001 UTC vs.
Revision 1.268 by root, Thu Oct 1 23:16:27 2009 UTC

1	=head1 NAME	1	=head1 NAME
2		2
3	Coro - coroutine process abstraction	3	Coro - the only real threads in perl
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	use Coro;	7	use Coro;
8		8
9	async {	9	async {
10	# some asynchronous thread of execution	10	# some asynchronous thread of execution
		11	print "2\n";
		12	cede; # yield back to main
		13	print "4\n";
11	};	14	};
12		15	print "1\n";
13	# alternatively create an async process like this:	16	cede; # yield to coro
14		17	print "3\n";
15	sub some_func : Coro {	18	cede; # and again
16	# some more async code	19
17	}	20	# use locking
18		21	use Coro::Semaphore;
19	cede;	22	my $lock = new Coro::Semaphore;
		23	my $locked;
		24
		25	$lock->down;
		26	$locked = 1;
		27	$lock->up;
20		28
21	=head1 DESCRIPTION	29	=head1 DESCRIPTION
22		30
23	This module collection manages coroutines. Coroutines are similar to	31	For a tutorial-style introduction, please read the L<Coro::Intro>
24	Threads but don't run in parallel.	32	manpage. This manpage mainly contains reference information.
25		33
26	This module is still experimental, see the BUGS section below.	34	This module collection manages continuations in general, most often in
		35	the form of cooperative threads (also called coros, or simply "coro"
		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.
27		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 more
		46	details) ported to unix, and as such act as processes), Coro provides
		47	a full shared address space, which makes communication between threads
		48	very easy. And Coro's 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 runs over 300 times faster on a single core than
		52	perl's pseudo-threads on a quad core using all four cores.
		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
28	In this module, coroutines are defined as "callchain + lexical variables	60	In this module, a thread is defined as "callchain + lexical variables +
29	+ @_ + $_ + $@ + $^W + C stack), that is, a coroutine has it's own	61	some package variables + C stack), that is, a thread has its own callchain,
30	callchain, it's own set of lexicals and it's own set of perl's most	62	its own set of lexicals and its own set of perls most important global
31	important global variables.	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.
32		67
33	=cut	68	=cut
34		69
35	package Coro;	70	package Coro;
36		71
		72	use common::sense;
		73
		74	use Carp ();
		75
		76	use Guard ();
		77
37	use Coro::State;	78	use Coro::State;
38		79
39	use base Exporter;	80	use base qw(Coro::State Exporter);
40		81
41	$VERSION = 0.45;	82	our $idle; # idle handler
		83	our $main; # main coro
		84	our $current; # current coro
42		85
		86	our $VERSION = 5.17;
		87
43	@EXPORT = qw(async cede schedule terminate current);	88	our @EXPORT = qw(async async_pool cede schedule terminate current unblock_sub);
44	@EXPORT_OK = qw($current);	89	our %EXPORT_TAGS = (
		90	prio => [qw(PRIO_MAX PRIO_HIGH PRIO_NORMAL PRIO_LOW PRIO_IDLE PRIO_MIN)],
		91	);
		92	our @EXPORT_OK = (@{$EXPORT_TAGS{prio}}, qw(nready));
45		93
46	{	94	=head1 GLOBAL VARIABLES
47	my @async;
48	my $init;
49		95
50	# this way of handling attributes simply is NOT scalable ;()	96	=over 4
51	sub import {
52	Coro->export_to_level(1, @_);
53	my $old = *{(caller)[0]."::MODIFY_CODE_ATTRIBUTES"}{CODE};
54	*{(caller)[0]."::MODIFY_CODE_ATTRIBUTES"} = sub {
55	my ($package, $ref) = (shift, shift);
56	my @attrs;
57	for (@_) {
58	if ($_ eq "Coro") {
59	push @async, $ref;
60	unless ($init++) {
61	eval q{
62	sub INIT {
63	&async(pop @async) while @async;
64	}
65	};
66	}
67	} else {
68	push @attrs, $_;
69	}
70	}
71	return $old ? $old->($package, $ref, @attrs) : @attrs;
72	};
73	}
74		97
75	}
76
77	=item $main	98	=item $Coro::main
78		99
79	This coroutine represents the main program.	100	This variable stores the Coro object that represents the main
		101	program. While you cna C<ready> it and do most other things you can do to
		102	coro, it is mainly useful to compare again C<$Coro::current>, to see
		103	whether you are running in the main program or not.
80		104
81	=cut	105	=cut
82		106
83	our $main = new Coro;	107	# $main is now being initialised by Coro::State
84		108
85	=item $current (or as function: current)	109	=item $Coro::current
86		110
87	The current coroutine (the last coroutine switched to). The initial value is C<$main> (of course).	111	The Coro object representing the current coro (the last
		112	coro that the Coro scheduler switched to). The initial value is
		113	C<$Coro::main> (of course).
88		114
89	=cut	115	This variable is B<strictly> I<read-only>. You can take copies of the
		116	value stored in it and use it as any other Coro object, but you must
		117	not otherwise modify the variable itself.
90		118
91	# maybe some other module used Coro::Specific before...	119	=cut
92	if ($current) {
93	$main->{specific} = $current->{specific};
94	}
95		120
96	our $current = $main;
97
98	sub current() { $current }	121	sub current() { $current } # [DEPRECATED]
99		122
100	=item $idle	123	=item $Coro::idle
101		124
102	The coroutine to switch to when no other coroutine is running. The default	125	This variable is mainly useful to integrate Coro into event loops. It is
103	implementation prints "FATAL: deadlock detected" and exits.	126	usually better to rely on L<Coro::AnyEvent> or L<Coro::EV>, as this is
		127	pretty low-level functionality.
104		128
105	=cut	129	This variable stores either a Coro object or a callback.
106		130
107	# should be done using priorities :(	131	If it is a callback, the it is called whenever the scheduler finds no
108	our $idle = new Coro sub {	132	ready coros to run. The default implementation prints "FATAL:
109	print STDERR "FATAL: deadlock detected\n";	133	deadlock detected" and exits, because the program has no other way to
110	exit(51);	134	continue.
		135
		136	If it is a coro object, then this object will be readied (without
		137	invoking any ready hooks, however) when the scheduler finds no other ready
		138	coros to run.
		139
		140	This hook is overwritten by modules such as C<Coro::EV> and
		141	C<Coro::AnyEvent> to wait on an external event that hopefully wake up a
		142	coro so the scheduler can run it.
		143
		144	Note that the callback I<must not>, under any circumstances, block
		145	the current coro. Normally, this is achieved by having an "idle
		146	coro" that calls the event loop and then blocks again, and then
		147	readying that coro in the idle handler, or by simply placing the idle
		148	coro in this variable.
		149
		150	See L<Coro::Event> or L<Coro::AnyEvent> for examples of using this
		151	technique.
		152
		153	Please note that if your callback recursively invokes perl (e.g. for event
		154	handlers), then it must be prepared to be called recursively itself.
		155
		156	=cut
		157
		158	$idle = sub {
		159	warn "oi\n";#d#
		160	Carp::confess ("FATAL: deadlock detected");
111	};	161	};
112		162
113	# this coroutine is necessary because a coroutine	163	# this coro is necessary because a coro
114	# cannot destroy itself.	164	# cannot destroy itself.
115	my @destroy;	165	our @destroy;
		166	our $manager;
		167
116	my $manager = new Coro sub {	168	$manager = new Coro sub {
117	while() {	169	while () {
118	delete ((pop @destroy)->{_coro_state}) while @destroy;	170	Coro::State::cancel shift @destroy
		171	while @destroy;
		172
119	&schedule;	173	&schedule;
120	}	174	}
121	};	175	};
		176	$manager->{desc} = "[coro manager]";
		177	$manager->prio (PRIO_MAX);
122		178
123	# we really need priorities...	179	=back
124	my @ready; # the ready queue. hehe, rather broken ;)
125		180
126	# static methods. not really.	181	=head1 SIMPLE CORO CREATION
127
128	=head2 STATIC METHODS
129
130	Static methods are actually functions that operate on the current process only.
131		182
132	=over 4	183	=over 4
133		184
134	=item async { ... } [@args...]	185	=item async { ... } [@args...]
135		186
136	Create a new asynchronous process and return it's process object	187	Create a new coro and return its Coro object (usually
137	(usually unused). When the sub returns the new process is automatically	188	unused). The coro will be put into the ready queue, so
		189	it will start running automatically on the next scheduler run.
		190
		191	The first argument is a codeblock/closure that should be executed in the
		192	coro. When it returns argument returns the coro is automatically
138	terminated.	193	terminated.
139		194
		195	The remaining arguments are passed as arguments to the closure.
		196
		197	See the C<Coro::State::new> constructor for info about the coro
		198	environment in which coro are executed.
		199
		200	Calling C<exit> in a coro will do the same as calling exit outside
		201	the coro. Likewise, when the coro dies, the program will exit,
		202	just as it would in the main program.
		203
		204	If you do not want that, you can provide a default C<die> handler, or
		205	simply avoid dieing (by use of C<eval>).
		206
140	# create a new coroutine that just prints its arguments	207	Example: Create a new coro that just prints its arguments.
		208
141	async {	209	async {
142	print "@_\n";	210	print "@_\n";
143	} 1,2,3,4;	211	} 1,2,3,4;
144		212
145	The coderef you submit MUST NOT be a closure that refers to variables	213	=item async_pool { ... } [@args...]
146	in an outer scope. This does NOT work. Pass arguments into it instead.
147		214
148	=cut	215	Similar to C<async>, but uses a coro pool, so you should not call
		216	terminate or join on it (although you are allowed to), and you get a
		217	coro that might have executed other code already (which can be good
		218	or bad :).
149		219
150	sub async(&@) {	220	On the plus side, this function is about twice as fast as creating (and
151	my $pid = new Coro @_;	221	destroying) a completely new coro, so if you need a lot of generic
152	$manager->ready; # this ensures that the stack is cloned from the manager	222	coros in quick successsion, use C<async_pool>, not C<async>.
153	$pid->ready;	223
154	$pid;	224	The code block is executed in an C<eval> context and a warning will be
		225	issued in case of an exception instead of terminating the program, as
		226	C<async> does. As the coro is being reused, stuff like C<on_destroy>
		227	will not work in the expected way, unless you call terminate or cancel,
		228	which somehow defeats the purpose of pooling (but is fine in the
		229	exceptional case).
		230
		231	The priority will be reset to C<0> after each run, tracing will be
		232	disabled, the description will be reset and the default output filehandle
		233	gets restored, so you can change all these. Otherwise the coro will
		234	be re-used "as-is": most notably if you change other per-coro global
		235	stuff such as C<$/> you I<must needs> revert that change, which is most
		236	simply done by using local as in: C<< local $/ >>.
		237
		238	The idle pool size is limited to C<8> idle coros (this can be
		239	adjusted by changing $Coro::POOL_SIZE), but there can be as many non-idle
		240	coros as required.
		241
		242	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
		244	{ 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 32kb
		246	(adjustable via $Coro::POOL_RSS) it will also be destroyed.
		247
		248	=cut
		249
		250	our $POOL_SIZE = 8;
		251	our $POOL_RSS = 32 * 1024;
		252	our @async_pool;
		253
		254	sub pool_handler {
		255	while () {
		256	eval {
		257	&{&_pool_handler} while 1;
		258	};
		259
		260	warn $@ if $@;
		261	}
155	}	262	}
156		263
		264	=back
		265
		266	=head1 STATIC METHODS
		267
		268	Static methods are actually functions that implicitly operate on the
		269	current coro.
		270
		271	=over 4
		272
157	=item schedule	273	=item schedule
158		274
159	Calls the scheduler. Please note that the current process will not be put	275	Calls the scheduler. The scheduler will find the next coro that is
		276	to be run from the ready queue and switches to it. The next coro
		277	to be run is simply the one with the highest priority that is longest
		278	in its ready queue. If there is no coro ready, it will clal the
		279	C<$Coro::idle> hook.
		280
		281	Please note that the current coro will I<not> be put into the ready
160	into the ready queue, so calling this function usually means you will	282	queue, so calling this function usually means you will never be called
161	never be called again.	283	again unless something else (e.g. an event handler) calls C<< ->ready >>,
		284	thus waking you up.
162		285
163	=cut	286	This makes C<schedule> I<the> generic method to use to block the current
		287	coro and wait for events: first you remember the current coro in
		288	a variable, then arrange for some callback of yours to call C<< ->ready
		289	>> on that once some event happens, and last you call C<schedule> to put
		290	yourself to sleep. Note that a lot of things can wake your coro up,
		291	so you need to check whether the event indeed happened, e.g. by storing the
		292	status in a variable.
164		293
165	my $prev;	294	See B<HOW TO WAIT FOR A CALLBACK>, below, for some ways to wait for callbacks.
166		295
167	sub schedule {	296	=item cede
168	# should be done using priorities :(	297
169	($prev, $current) = ($current, shift @ready \|\| $idle);	298	"Cede" to other coros. This function puts the current coro into
170	Coro::State::transfer($prev, $current);	299	the ready queue and calls C<schedule>, which has the effect of giving
		300	up the current "timeslice" to other coros of the same or higher
		301	priority. Once your coro gets its turn again it will automatically be
		302	resumed.
		303
		304	This function is often called C<yield> in other languages.
		305
		306	=item Coro::cede_notself
		307
		308	Works like cede, but is not exported by default and will cede to I<any>
		309	coro, regardless of priority. This is useful sometimes to ensure
		310	progress is made.
		311
		312	=item terminate [arg...]
		313
		314	Terminates the current coro with the given status values (see L<cancel>).
		315
		316	=item Coro::on_enter BLOCK, Coro::on_leave BLOCK
		317
		318	These function install enter and leave winders in the current scope. The
		319	enter block will be executed when on_enter is called and whenever the
		320	current coro is re-entered by the scheduler, while the leave block is
		321	executed whenever the current coro is blocked by the scheduler, and
		322	also when the containing scope is exited (by whatever means, be it exit,
		323	die, last etc.).
		324
		325	I<Neither invoking the scheduler, nor exceptions, are allowed within those
		326	BLOCKs>. That means: do not even think about calling C<die> without an
		327	eval, and do not even think of entering the scheduler in any way.
		328
		329	Since both BLOCKs are tied to the current scope, they will automatically
		330	be removed when the current scope exits.
		331
		332	These functions implement the same concept as C<dynamic-wind> in scheme
		333	does, and are useful when you want to localise some resource to a specific
		334	coro.
		335
		336	They slow down thread switching considerably for coros that use them
		337	(about 40% for a BLOCK with a single assignment, so thread switching is
		338	still reasonably fast if the handlers are fast).
		339
		340	These functions are best understood by an example: The following function
		341	will change the current timezone to "Antarctica/South_Pole", which
		342	requires a call to C<tzset>, but by using C<on_enter> and C<on_leave>,
		343	which remember/change the current timezone and restore the previous
		344	value, respectively, the timezone is only changed for the coro that
		345	installed those handlers.
		346
		347	use POSIX qw(tzset);
		348
		349	async {
		350	my $old_tz; # store outside TZ value here
		351
		352	Coro::on_enter {
		353	$old_tz = $ENV{TZ}; # remember the old value
		354
		355	$ENV{TZ} = "Antarctica/South_Pole";
		356	tzset; # enable new value
		357	};
		358
		359	Coro::on_leave {
		360	$ENV{TZ} = $old_tz;
		361	tzset; # restore old value
		362	};
		363
		364	# at this place, the timezone is Antarctica/South_Pole,
		365	# without disturbing the TZ of any other coro.
		366	};
		367
		368	This can be used to localise about any resource (locale, uid, current
		369	working directory etc.) to a block, despite the existance of other
		370	coros.
		371
		372	Another interesting example implements time-sliced multitasking using
		373	interval timers (this could obviously be optimised, but does the job):
		374
		375	# "timeslice" the given block
		376	sub timeslice(&) {
		377	use Time::HiRes ();
		378
		379	Coro::on_enter {
		380	# on entering the thread, we set an VTALRM handler to cede
		381	$SIG{VTALRM} = sub { cede };
		382	# and then start the interval timer
		383	Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0.01, 0.01;
		384	};
		385	Coro::on_leave {
		386	# on leaving the thread, we stop the interval timer again
		387	Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0, 0;
		388	};
		389
		390	&{+shift};
		391	}
		392
		393	# use like this:
		394	timeslice {
		395	# The following is an endless loop that would normally
		396	# monopolise the process. Since it runs in a timesliced
		397	# environment, it will regularly cede to other threads.
		398	while () { }
		399	};
		400
		401
		402	=item killall
		403
		404	Kills/terminates/cancels all coros except the currently running one.
		405
		406	Note that while this will try to free some of the main interpreter
		407	resources if the calling coro isn't the main coro, but one
		408	cannot free all of them, so if a coro that is not the main coro
		409	calls this function, there will be some one-time resource leak.
		410
		411	=cut
		412
		413	sub killall {
		414	for (Coro::State::list) {
		415	$_->cancel
		416	if $_ != $current && UNIVERSAL::isa $_, "Coro";
		417	}
171	}	418	}
172		419
173	=item cede
174
175	"Cede" to other processes. This function puts the current process into the
176	ready queue and calls C<schedule>, which has the effect of giving up the
177	current "timeslice" to other coroutines of the same or higher priority.
178
179	=cut
180
181	sub cede {
182	$current->ready;
183	&schedule;
184	}
185
186	=item terminate
187
188	Terminates the current process.
189
190	Future versions of this function will allow result arguments.
191
192	=cut
193
194	sub terminate {
195	$current->cancel;
196	&schedule;
197	die; # NORETURN
198	}
199
200	=back	420	=back
201		421
202	# dynamic methods	422	=head1 CORO OBJECT METHODS
203		423
204	=head2 PROCESS METHODS
205
206	These are the methods you can call on process objects.	424	These are the methods you can call on coro objects (or to create
		425	them).
207		426
208	=over 4	427	=over 4
209		428
210	=item new Coro \&sub [, @args...]	429	=item new Coro \&sub [, @args...]
211		430
212	Create a new process and return it. When the sub returns the process	431	Create a new coro and return it. When the sub returns, the coro
213	automatically terminates. To start the process you must first put it into	432	automatically terminates as if C<terminate> with the returned values were
		433	called. To make the coro run you must first put it into the ready
214	the ready queue by calling the ready method.	434	queue by calling the ready method.
215		435
216	The coderef you submit MUST NOT be a closure that refers to variables	436	See C<async> and C<Coro::State::new> for additional info about the
217	in an outer scope. This does NOT work. Pass arguments into it instead.	437	coro environment.
218		438
219	=cut	439	=cut
220		440
221	sub _newcoro {	441	sub _coro_run {
222	terminate &{+shift};	442	terminate &{+shift};
223	}	443	}
224		444
		445	=item $success = $coro->ready
		446
		447	Put the given coro into the end of its ready queue (there is one
		448	queue for each priority) and return true. If the coro is already in
		449	the ready queue, do nothing and return false.
		450
		451	This ensures that the scheduler will resume this coro automatically
		452	once all the coro of higher priority and all coro of the same
		453	priority that were put into the ready queue earlier have been resumed.
		454
		455	=item $coro->suspend
		456
		457	Suspends the specified coro. A suspended coro works just like any other
		458	coro, except that the scheduler will not select a suspended coro for
		459	execution.
		460
		461	Suspending a coro can be useful when you want to keep the coro from
		462	running, but you don't want to destroy it, or when you want to temporarily
		463	freeze a coro (e.g. for debugging) to resume it later.
		464
		465	A scenario for the former would be to suspend all (other) coros after a
		466	fork and keep them alive, so their destructors aren't called, but new
		467	coros can be created.
		468
		469	=item $coro->resume
		470
		471	If the specified coro was suspended, it will be resumed. Note that when
		472	the coro was in the ready queue when it was suspended, it might have been
		473	unreadied by the scheduler, so an activation might have been lost.
		474
		475	To avoid this, it is best to put a suspended coro into the ready queue
		476	unconditionally, as every synchronisation mechanism must protect itself
		477	against spurious wakeups, and the one in the Coro family certainly do
		478	that.
		479
		480	=item $is_ready = $coro->is_ready
		481
		482	Returns true iff the Coro object is in the ready queue. Unless the Coro
		483	object gets destroyed, it will eventually be scheduled by the scheduler.
		484
		485	=item $is_running = $coro->is_running
		486
		487	Returns true iff the Coro object is currently running. Only one Coro object
		488	can ever be in the running state (but it currently is possible to have
		489	multiple running Coro::States).
		490
		491	=item $is_suspended = $coro->is_suspended
		492
		493	Returns true iff this Coro object has been suspended. Suspended Coros will
		494	not ever be scheduled.
		495
		496	=item $coro->cancel (arg...)
		497
		498	Terminates the given Coro and makes it return the given arguments as
		499	status (default: the empty list). Never returns if the Coro is the
		500	current Coro.
		501
		502	=cut
		503
225	sub new {	504	sub cancel {
226	my $class = shift;	505	my $self = shift;
227	bless {	506
228	_coro_state => (new Coro::State $_[0] && \&_newcoro, @_),	507	if ($current == $self) {
229	}, $class;	508	terminate @_;
		509	} else {
		510	$self->{_status} = [@_];
		511	Coro::State::cancel $self;
		512	}
230	}	513	}
231		514
232	=item $process->ready	515	=item $coro->schedule_to
233		516
234	Put the current process into the ready queue.	517	Puts the current coro to sleep (like C<Coro::schedule>), but instead
		518	of continuing with the next coro from the ready queue, always switch to
		519	the given coro object (regardless of priority etc.). The readyness
		520	state of that coro isn't changed.
235		521
236	=cut	522	This is an advanced method for special cases - I'd love to hear about any
		523	uses for this one.
237		524
238	sub ready {	525	=item $coro->cede_to
239	push @ready, $_[0];	526
		527	Like C<schedule_to>, but puts the current coro into the ready
		528	queue. This has the effect of temporarily switching to the given
		529	coro, and continuing some time later.
		530
		531	This is an advanced method for special cases - I'd love to hear about any
		532	uses for this one.
		533
		534	=item $coro->throw ([$scalar])
		535
		536	If C<$throw> is specified and defined, it will be thrown as an exception
		537	inside the coro at the next convenient point in time. Otherwise
		538	clears the exception object.
		539
		540	Coro will check for the exception each time a schedule-like-function
		541	returns, i.e. after each C<schedule>, C<cede>, C<< Coro::Semaphore->down
		542	>>, C<< Coro::Handle->readable >> and so on. Most of these functions
		543	detect this case and return early in case an exception is pending.
		544
		545	The exception object will be thrown "as is" with the specified scalar in
		546	C<$@>, i.e. if it is a string, no line number or newline will be appended
		547	(unlike with C<die>).
		548
		549	This can be used as a softer means than C<cancel> to ask a coro to
		550	end itself, although there is no guarantee that the exception will lead to
		551	termination, and if the exception isn't caught it might well end the whole
		552	program.
		553
		554	You might also think of C<throw> as being the moral equivalent of
		555	C<kill>ing a coro with a signal (in this case, a scalar).
		556
		557	=item $coro->join
		558
		559	Wait until the coro terminates and return any values given to the
		560	C<terminate> or C<cancel> functions. C<join> can be called concurrently
		561	from multiple coro, and all will be resumed and given the status
		562	return once the C<$coro> terminates.
		563
		564	=cut
		565
		566	sub join {
		567	my $self = shift;
		568
		569	unless ($self->{_status}) {
		570	my $current = $current;
		571
		572	push @{$self->{_on_destroy}}, sub {
		573	$current->ready;
		574	undef $current;
		575	};
		576
		577	&schedule while $current;
		578	}
		579
		580	wantarray ? @{$self->{_status}} : $self->{_status}[0];
240	}	581	}
241		582
242	=item $process->cancel	583	=item $coro->on_destroy (\&cb)
243		584
244	Like C<terminate>, but terminates the specified process instead.	585	Registers a callback that is called when this coro gets destroyed,
		586	but before it is joined. The callback gets passed the terminate arguments,
		587	if any, and I<must not> die, under any circumstances.
245		588
246	=cut	589	=cut
247		590
248	sub cancel {	591	sub on_destroy {
249	push @destroy, $_[0];	592	my ($self, $cb) = @_;
250	$manager->ready;	593
		594	push @{ $self->{_on_destroy} }, $cb;
251	}	595	}
252		596
		597	=item $oldprio = $coro->prio ($newprio)
		598
		599	Sets (or gets, if the argument is missing) the priority of the
		600	coro. Higher priority coro get run before lower priority
		601	coro. Priorities are small signed integers (currently -4 .. +3),
		602	that you can refer to using PRIO_xxx constants (use the import tag :prio
		603	to get then):
		604
		605	PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN
		606	3 > 1 > 0 > -1 > -3 > -4
		607
		608	# set priority to HIGH
		609	current->prio (PRIO_HIGH);
		610
		611	The idle coro ($Coro::idle) always has a lower priority than any
		612	existing coro.
		613
		614	Changing the priority of the current coro will take effect immediately,
		615	but changing the priority of coro in the ready queue (but not
		616	running) will only take effect after the next schedule (of that
		617	coro). This is a bug that will be fixed in some future version.
		618
		619	=item $newprio = $coro->nice ($change)
		620
		621	Similar to C<prio>, but subtract the given value from the priority (i.e.
		622	higher values mean lower priority, just as in unix).
		623
		624	=item $olddesc = $coro->desc ($newdesc)
		625
		626	Sets (or gets in case the argument is missing) the description for this
		627	coro. This is just a free-form string you can associate with a
		628	coro.
		629
		630	This method simply sets the C<< $coro->{desc} >> member to the given
		631	string. You can modify this member directly if you wish.
		632
		633	=cut
		634
		635	sub desc {
		636	my $old = $_[0]{desc};
		637	$_[0]{desc} = $_[1] if @_ > 1;
		638	$old;
		639	}
		640
		641	sub transfer {
		642	require Carp;
		643	Carp::croak ("You must not call ->transfer on Coro objects. Use Coro::State objects or the ->schedule_to method. Caught");
		644	}
		645
253	=back	646	=back
254		647
		648	=head1 GLOBAL FUNCTIONS
		649
		650	=over 4
		651
		652	=item Coro::nready
		653
		654	Returns the number of coro that are currently in the ready state,
		655	i.e. that can be switched to by calling C<schedule> directory or
		656	indirectly. The value C<0> means that the only runnable coro is the
		657	currently running one, so C<cede> would have no effect, and C<schedule>
		658	would cause a deadlock unless there is an idle handler that wakes up some
		659	coro.
		660
		661	=item my $guard = Coro::guard { ... }
		662
		663	This function still exists, but is deprecated. Please use the
		664	C<Guard::guard> function instead.
		665
		666	=cut
		667
		668	BEGIN { *guard = \&Guard::guard }
		669
		670	=item unblock_sub { ... }
		671
		672	This utility function takes a BLOCK or code reference and "unblocks" it,
		673	returning a new coderef. Unblocking means that calling the new coderef
		674	will return immediately without blocking, returning nothing, while the
		675	original code ref will be called (with parameters) from within another
		676	coro.
		677
		678	The reason this function exists is that many event libraries (such as the
		679	venerable L<Event\|Event> module) are not thread-safe (a weaker form
		680	of reentrancy). This means you must not block within event callbacks,
		681	otherwise you might suffer from crashes or worse. The only event library
		682	currently known that is safe to use without C<unblock_sub> is L<EV>.
		683
		684	This function allows your callbacks to block by executing them in another
		685	coro where it is safe to block. One example where blocking is handy
		686	is when you use the L<Coro::AIO\|Coro::AIO> functions to save results to
		687	disk, for example.
		688
		689	In short: simply use C<unblock_sub { ... }> instead of C<sub { ... }> when
		690	creating event callbacks that want to block.
		691
		692	If your handler does not plan to block (e.g. simply sends a message to
		693	another coro, or puts some other coro into the ready queue), there is
		694	no reason to use C<unblock_sub>.
		695
		696	Note that you also need to use C<unblock_sub> for any other callbacks that
		697	are indirectly executed by any C-based event loop. For example, when you
		698	use a module that uses L<AnyEvent> (and you use L<Coro::AnyEvent>) and it
		699	provides callbacks that are the result of some event callback, then you
		700	must not block either, or use C<unblock_sub>.
		701
		702	=cut
		703
		704	our @unblock_queue;
		705
		706	# we create a special coro because we want to cede,
		707	# to reduce pressure on the coro pool (because most callbacks
		708	# return immediately and can be reused) and because we cannot cede
		709	# inside an event callback.
		710	our $unblock_scheduler = new Coro sub {
		711	while () {
		712	while (my $cb = pop @unblock_queue) {
		713	&async_pool (@$cb);
		714
		715	# for short-lived callbacks, this reduces pressure on the coro pool
		716	# as the chance is very high that the async_poll coro will be back
		717	# in the idle state when cede returns
		718	cede;
		719	}
		720	schedule; # sleep well
		721	}
		722	};
		723	$unblock_scheduler->{desc} = "[unblock_sub scheduler]";
		724
		725	sub unblock_sub(&) {
		726	my $cb = shift;
		727
		728	sub {
		729	unshift @unblock_queue, [$cb, @_];
		730	$unblock_scheduler->ready;
		731	}
		732	}
		733
		734	=item $cb = Coro::rouse_cb
		735
		736	Create and return a "rouse callback". That's a code reference that,
		737	when called, will remember a copy of its arguments and notify the owner
		738	coro of the callback.
		739
		740	See the next function.
		741
		742	=item @args = Coro::rouse_wait [$cb]
		743
		744	Wait for the specified rouse callback (or the last one that was created in
		745	this coro).
		746
		747	As soon as the callback is invoked (or when the callback was invoked
		748	before C<rouse_wait>), it will return the arguments originally passed to
		749	the rouse callback. In scalar context, that means you get the I<last>
		750	argument, just as if C<rouse_wait> had a C<return ($a1, $a2, $a3...)>
		751	statement at the end.
		752
		753	See the section B<HOW TO WAIT FOR A CALLBACK> for an actual usage example.
		754
		755	=back
		756
255	=cut	757	=cut
256		758
257	1;	759	1;
258		760
		761	=head1 HOW TO WAIT FOR A CALLBACK
		762
		763	It is very common for a coro to wait for some callback to be
		764	called. This occurs naturally when you use coro in an otherwise
		765	event-based program, or when you use event-based libraries.
		766
		767	These typically register a callback for some event, and call that callback
		768	when the event occured. In a coro, however, you typically want to
		769	just wait for the event, simplyifying things.
		770
		771	For example C<< AnyEvent->child >> registers a callback to be called when
		772	a specific child has exited:
		773
		774	my $child_watcher = AnyEvent->child (pid => $pid, cb => sub { ... });
		775
		776	But from within a coro, you often just want to write this:
		777
		778	my $status = wait_for_child $pid;
		779
		780	Coro offers two functions specifically designed to make this easy,
		781	C<Coro::rouse_cb> and C<Coro::rouse_wait>.
		782
		783	The first function, C<rouse_cb>, generates and returns a callback that,
		784	when invoked, will save its arguments and notify the coro that
		785	created the callback.
		786
		787	The second function, C<rouse_wait>, waits for the callback to be called
		788	(by calling C<schedule> to go to sleep) and returns the arguments
		789	originally passed to the callback.
		790
		791	Using these functions, it becomes easy to write the C<wait_for_child>
		792	function mentioned above:
		793
		794	sub wait_for_child($) {
		795	my ($pid) = @_;
		796
		797	my $watcher = AnyEvent->child (pid => $pid, cb => Coro::rouse_cb);
		798
		799	my ($rpid, $rstatus) = Coro::rouse_wait;
		800	$rstatus
		801	}
		802
		803	In the case where C<rouse_cb> and C<rouse_wait> are not flexible enough,
		804	you can roll your own, using C<schedule>:
		805
		806	sub wait_for_child($) {
		807	my ($pid) = @_;
		808
		809	# store the current coro in $current,
		810	# and provide result variables for the closure passed to ->child
		811	my $current = $Coro::current;
		812	my ($done, $rstatus);
		813
		814	# pass a closure to ->child
		815	my $watcher = AnyEvent->child (pid => $pid, cb => sub {
		816	$rstatus = $_[1]; # remember rstatus
		817	$done = 1; # mark $rstatus as valud
		818	});
		819
		820	# wait until the closure has been called
		821	schedule while !$done;
		822
		823	$rstatus
		824	}
		825
		826
259	=head1 BUGS/LIMITATIONS	827	=head1 BUGS/LIMITATIONS
260		828
261	- could be faster, especially when the core would introduce special	829	=over 4
262	support for coroutines (like it does for threads).	830
263	- there is still a memleak on coroutine termination that I could not	831	=item fork with pthread backend
264	identify. Could be as small as a single SV.	832
265	- this module is not well-tested.	833	When Coro is compiled using the pthread backend (which isn't recommended
266	- if variables or arguments "disappear" (become undef) or become	834	but required on many BSDs as their libcs are completely broken), then
267	corrupted please contact the author so he cen iron out the	835	coro will not survive a fork. There is no known workaround except to
268	remaining bugs.	836	fix your libc and use a saner backend.
269	- this module is not thread-safe. You must only ever use this module from	837
270	the same thread (this requirement might be loosened in the future to	838	=item perl process emulation ("threads")
		839
		840	This module is not perl-pseudo-thread-safe. You should only ever use this
		841	module from the first thread (this requirement might be removed in the
271	allow per-thread schedulers, but Coro::State does not yet allow this).	842	future to allow per-thread schedulers, but Coro::State does not yet allow
		843	this). I recommend disabling thread support and using processes, as having
		844	the windows process emulation enabled under unix roughly halves perl
		845	performance, even when not used.
		846
		847	=item coro switching is not signal safe
		848
		849	You must not switch to another coro from within a signal handler
		850	(only relevant with %SIG - most event libraries provide safe signals).
		851
		852	That means you I<MUST NOT> call any function that might "block" the
		853	current coro - C<cede>, C<schedule> C<< Coro::Semaphore->down >> or
		854	anything that calls those. Everything else, including calling C<ready>,
		855	works.
		856
		857	=back
		858
		859
		860	=head1 WINDOWS PROCESS EMULATION
		861
		862	A great many people seem to be confused about ithreads (for example, Chip
		863	Salzenberg called me unintelligent, incapable, stupid and gullible,
		864	while in the same mail making rather confused statements about perl
		865	ithreads (for example, that memory or files would be shared), showing his
		866	lack of understanding of this area - if it is hard to understand for Chip,
		867	it is probably not obvious to everybody).
		868
		869	What follows is an ultra-condensed version of my talk about threads in
		870	scripting languages given onthe perl workshop 2009:
		871
		872	The so-called "ithreads" were originally implemented for two reasons:
		873	first, to (badly) emulate unix processes on native win32 perls, and
		874	secondly, to replace the older, real thread model ("5.005-threads").
		875
		876	It does that by using threads instead of OS processes. The difference
		877	between processes and threads is that threads share memory (and other
		878	state, such as files) between threads within a single process, while
		879	processes do not share anything (at least not semantically). That
		880	means that modifications done by one thread are seen by others, while
		881	modifications by one process are not seen by other processes.
		882
		883	The "ithreads" work exactly like that: when creating a new ithreads
		884	process, all state is copied (memory is copied physically, files and code
		885	is copied logically). Afterwards, it isolates all modifications. On UNIX,
		886	the same behaviour can be achieved by using operating system processes,
		887	except that UNIX typically uses hardware built into the system to do this
		888	efficiently, while the windows process emulation emulates this hardware in
		889	software (rather efficiently, but of course it is still much slower than
		890	dedicated hardware).
		891
		892	As mentioned before, loading code, modifying code, modifying data
		893	structures and so on is only visible in the ithreads process doing the
		894	modification, not in other ithread processes within the same OS process.
		895
		896	This is why "ithreads" do not implement threads for perl at all, only
		897	processes. What makes it so bad is that on non-windows platforms, you can
		898	actually take advantage of custom hardware for this purpose (as evidenced
		899	by the forks module, which gives you the (i-) threads API, just much
		900	faster).
		901
		902	Sharing data is in the i-threads model is done by transfering data
		903	structures between threads using copying semantics, which is very slow -
		904	shared data simply does not exist. Benchmarks using i-threads which are
		905	communication-intensive show extremely bad behaviour with i-threads (in
		906	fact, so bad that Coro, which cannot take direct advantage of multiple
		907	CPUs, is often orders of magnitude faster because it shares data using
		908	real threads, refer to my talk for details).
		909
		910	As summary, i-threads use threads to implement processes, while
		911	the compatible forks module uses processes to emulate, uhm,
		912	processes. I-threads slow down every perl program when enabled, and
		913	outside of windows, serve no (or little) practical purpose, but
		914	disadvantages every single-threaded Perl program.
		915
		916	This is the reason that I try to avoid the name "ithreads", as it is
		917	misleading as it implies that it implements some kind of thread model for
		918	perl, and prefer the name "windows process emulation", which describes the
		919	actual use and behaviour of it much better.
272		920
273	=head1 SEE ALSO	921	=head1 SEE ALSO
274		922
275	L<Coro::Channel>, L<Coro::Cont>, L<Coro::Specific>, L<Coro::Semaphore>,	923	Event-Loop integration: L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>.
276	L<Coro::Signal>, L<Coro::State>, L<Coro::Event>, L<Coro::RWLock>,	924
277	L<Coro::Handle>, L<Coro::Socket>.	925	Debugging: L<Coro::Debug>.
		926
		927	Support/Utility: L<Coro::Specific>, L<Coro::Util>.
		928
		929	Locking and IPC: L<Coro::Signal>, L<Coro::Channel>, L<Coro::Semaphore>,
		930	L<Coro::SemaphoreSet>, L<Coro::RWLock>.
		931
		932	I/O and Timers: L<Coro::Timer>, L<Coro::Handle>, L<Coro::Socket>, L<Coro::AIO>.
		933
		934	Compatibility with other modules: L<Coro::LWP> (but see also L<AnyEvent::HTTP> for
		935	a better-working alternative), L<Coro::BDB>, L<Coro::Storable>,
		936	L<Coro::Select>.
		937
		938	XS API: L<Coro::MakeMaker>.
		939
		940	Low level Configuration, Thread Environment, Continuations: L<Coro::State>.
278		941
279	=head1 AUTHOR	942	=head1 AUTHOR
280		943
281	Marc Lehmann <pcg@goof.com>	944	Marc Lehmann <schmorp@schmorp.de>
282	http://www.goof.com/pcg/marc/	945	http://home.schmorp.de/
283		946
284	=cut	947	=cut
285		948

Diff Legend

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

Comparing Coro/Coro.pm (file contents): Revision 1.29 by root, Sat Aug 11 00:37:31 2001 UTC vs. Revision 1.268 by root, Thu Oct 1 23:16:27 2009 UTC

Diff Legend

Comparing Coro/Coro.pm (file contents):
Revision 1.29 by root, Sat Aug 11 00:37:31 2001 UTC vs.
Revision 1.268 by root, Thu Oct 1 23:16:27 2009 UTC