Coro/Coro/State.pm

=head1 NAME

Coro::State - first class continuations

=head1 SYNOPSIS

 use Coro::State;

 $new = new Coro::State sub {
    print "in coro (called with @_), switching back\n";
    $new->transfer ($main);
    print "in coro again, switching back\n";
    $new->transfer ($main);
 }, 5;

 $main = new Coro::State;

 print "in main, switching to coro\n";
 $main->transfer ($new);
 print "back in main, switch to coro again\n";
 $main->transfer ($new);
 print "back in main\n";

=head1 DESCRIPTION

This module implements coro. Coros, similar to threads and continuations,
allow you to run more than one "thread of execution" in parallel. Unlike
so-called "kernel" threads, there is no parallelism and only voluntary
switching is used so locking problems are greatly reduced. The latter is
called "cooperative" threading as opposed to "preemptive" threading.

This can be used to implement non-local jumps, exception handling,
continuation objects and more.

This module provides only low-level functionality. See L<Coro> and related
modules for a higher level threads abstraction including a scheduler.

=head2 MODEL

Coro::State implements two different thread models: Perl and C. The C
threads (called cctx's) are basically simplified perl interpreters
running/interpreting the Perl threads. A single interpreter can run any
number of Perl threads, so usually there are very few C threads.

When Perl code calls a C function (e.g. in an extension module) and that C
function then calls back into Perl or transfers control to another thread,
the C thread can no longer execute other Perl threads, so it stays tied to
the specific thread until it returns to the original Perl caller, after
which it is again available to run other Perl threads.

The main program always has its own "C thread" (which really is
*the* Perl interpreter running the whole program), so there will always
be at least one additional C thread. You can use the debugger (see
L<Coro::Debug>) to find out which threads are tied to their cctx and
which aren't.

=head2 MEMORY CONSUMPTION

A newly created Coro::State that has not been used only allocates a
relatively small (a hundred bytes) structure. Only on the first
C<transfer> will perl allocate stacks (a few kb, 64 bit architetcures
use twice as much, i.e. a few kb :) and optionally a C stack/thread
(cctx) for threads that recurse through C functions. All this is very
system-dependent. On my x86-pc-linux-gnu system this amounts to about 2k
per (non-trivial but simple) Coro::State.

You can view the actual memory consumption using Coro::Debug. Keep in mind
that a for loop or other block constructs can easily consume 100-200 bytes
per nesting level.

=cut

package Coro::State;

use common::sense;

use Carp;

use Time::HiRes (); # currently only used for PerlIO::cede

our $DIEHOOK;
our $WARNHOOK;

BEGIN {
   $DIEHOOK  = sub { };
   $WARNHOOK = sub { warn $_[0] };
}

sub diehook  { &$DIEHOOK  }
sub warnhook { &$WARNHOOK }

use XSLoader;

BEGIN {
   our $VERSION = 5.25;

   # must be done here because the xs part expects it to exist
   # it might exist already because Coro::Specific created it.
   $Coro::current ||= { };

   {
      # save/restore the handlers before/after overwriting %SIG magic
      local $SIG{__DIE__};
      local $SIG{__WARN__};

      XSLoader::load __PACKAGE__, $VERSION;
   }

   # need to do it after overwriting the %SIG magic
   $SIG{__DIE__}  ||= \&diehook;
   $SIG{__WARN__} ||= \&warnhook;
}

use Exporter;
use base Exporter::;

=head2 GLOBAL VARIABLES

=over 4

=item $Coro::State::DIEHOOK

This works similarly to C<$SIG{__DIE__}> and is used as the default die
hook for newly created Coro::States. This is useful if you want some generic
logging function that works for all threads that don't set their own
hook.

When Coro::State is first loaded it will install these handlers for the
main program, too, unless they have been overwritten already.

The default handlers provided will behave like the built-in ones (as if
they weren't there).

If you don't want to exit your program on uncaught exceptions, you must
not return from your die hook - call C<Coro::terminate> instead.

Note 1: You I<must> store a valid code reference in these variables,
C<undef> will I<not> do.

Note 2: The value of this variable will be shared among all threads, so
changing its value will change it in all threads that don't have their
own die handler.

=item $Coro::State::WARNHOOK

Similar to above die hook, but augments C<$SIG{__WARN__}>.

=back

=head2 FUNCTIONS

=over 4

=item $coro = new Coro::State [$coderef[, @args...]]

Create a new Coro::State thread object and return it. The first
C<transfer> call to this thread will start execution at the given
coderef, with the given arguments.

Note that the arguments will not be copied. Instead, as with normal
function calls, the thread receives passed arguments by reference, so
make sure you don't change them in unexpected ways.

Returning from such a thread is I<NOT> supported. Neither is calling
C<exit> or throwing an uncaught exception. The following paragraphs
describe what happens in current versions of Coro.

If the subroutine returns the program will be terminated as if execution
of the main program ended.

If it throws an exception the program will terminate unless the exception
is caught, exactly like in the main program.

Calling C<exit> in a thread does the same as calling it in the main
program, but due to libc bugs on many BSDs, this doesn't work reliable
everywhere.

If the coderef is omitted this function will create a new "empty"
thread, i.e. a thread that cannot be transfered to but can be used
to save the current thread state in (note that this is dangerous, as no
reference is taken to ensure that the "current thread state" survives,
the caller is responsible to ensure that the cloned state does not go
away).

The returned object is an empty hash which can be used for any purpose
whatsoever, for example when subclassing Coro::State.

Certain variables are "localised" to each thread, that is, certain
"global" variables are actually per thread. Not everything that would
sensibly be localised currently is, and not everything that is localised
makes sense for every application, and the future might bring changes.

The following global variables can have different values per thread,
and have the stated initial values:

   Variable       Initial Value
   @_             whatever arguments were passed to the Coro
   $_             undef
   $@             undef
   $/             "\n"
   $SIG{__DIE__}  aliased to $Coro::State::DIEHOOK(*)
   $SIG{__WARN__} aliased to $Coro::State::WARNHOOK(*)
   (default fh)   *STDOUT
   $^H, %^H       zero/empty.
   $1, $2...      all regex results are initially undefined

   (*) reading the value from %SIG is not supported, but local'ising is.

If you feel that something important is missing then tell me. Also
remember that every function call that might call C<transfer> (such
as C<Coro::Channel::put>) might clobber any global and/or special
variables. Yes, this is by design ;) You can always create your own
process abstraction model that saves these variables.

The easiest way to do this is to create your own scheduling primitive like
in the code below, and use it in your threads:

  sub my_cede {
     local ($;, ...);
     Coro::cede;
  }

Another way is to use dynamic winders, see C<Coro::on_enter> and
C<Coro::on_leave> for this.

=item $prev->transfer ($next)

Save the state of the current subroutine in C<$prev> and switch to the
thread saved in C<$next>.

The "state" of a subroutine includes the scope, i.e. lexical variables and
the current execution state (subroutine, stack).

=item $state->is_new

Returns true iff this Coro::State object is "new", i.e. has never been run
yet. Those states basically consist of only the code reference to call and
the arguments, but consumes very little other resources. New states will
automatically get assigned a perl interpreter when they are transfered to.

=item $state->is_destroyed

Returns true iff the Coro::State object has been destroyed (by
C<cancel>), i.e. it's resources freed because they were C<cancel>'d (or
C<terminate>'d).

=item $state->cancel

Forcefully destructs the given Coro::State. While you can keep the
reference, and some memory is still allocated, the Coro::State object is
effecticely dead, destructors have been freed, it cannot be transfered to
anymore.

=item $state->throw ([$scalar])

See L<< Coro->throw >>.

=item $state->call ($coderef)

Try to call the given C<$coderef> in the context of the given state. This
works even when the state is currently within an XS function, and can
be very dangerous. You can use it to acquire stack traces etc. (see the
Coro::Debug module for more details). The coderef MUST NOT EVER transfer
to another state.

=item $state->eval ($string)

Like C<call>, but eval's the string. Dangerous.

=item $state->swap_defsv

=item $state->swap_defav

Swap the current C<$_> (swap_defsv) or C<@_> (swap_defav) with the
equivalent in the saved state of C<$state>. This can be used to give the
coro a defined content for C<@_> and C<$_> before transfer'ing to it.

=item $state->swap_sv (\$sv, \$swap_sv)

This (very advanced) function can be used to make I<any> variable local to
a thread.

It works by swapping the contents of C<$sv> and C<$swap_sv> each time the
thread is entered and left again, i.e. it is similarly to:

   $tmp = $sv; $sv = $swap_sv; $swap_sv = $tmp;

Except that it doesn't make an copies and works on hashes and even more
exotic values (code references!).

Needless to say, this function can be very very dangerous: you can easily
swap a hash with a reference (i.e. C<%hash> I<becomes> a reference), and perl
will not like this at all.

It will also swap "magicalness" - so when swapping a builtin perl variable
(such as C<$.>), it will lose it's magicalness, which, again, perl will
not like, so don't do it.

Lastly, the C<$swap_sv> itself will be used, not a copy, so make sure you
give each thread it's own C<$swap_sv> instance.

It is, however, quite safe to swap some normal variable with
another. For example, L<PApp::SQL> stores the default database handle in
C<$PApp::SQL::DBH>. To make this a per-thread variable, use this:

   my $private_dbh = ...;
   $coro->swap_sv (\$PApp::SQL::DBH, \$private_dbh);

This results in C<$PApp::SQL::DBH> having the value of C<$private_dbh>
while it executes, and whatever other value it had when it doesn't
execute.

You can also swap hashes and other values:

   my %private_hash;
   $coro->swap_sv (\%some_hash, \%private_hash);

=item $state->trace ($flags)

Internal function to control tracing. I just mention this so you can stay
away from abusing it.

=item $bytes = $state->rss

Returns the memory allocated by the coro (which includes static
structures, various perl stacks but NOT local variables, arguments or any
C context data). This is a rough indication of how much memory it might
use.

=item $state->has_cctx

Returns whether the state currently uses a cctx/C context. An active
state always has a cctx, as well as the main program. Other states only
use a cctxts when needed.

=item Coro::State::force_cctx

Forces the allocation of a C context for the currently running coro
(if not already done). Apart from benchmarking there is little point
in doing so, however.

=item $ncctx = Coro::State::cctx_count

Returns the number of C-level coro allocated. If this number is
very high (more than a dozen) it might help to identify points of C-level
recursion in your code and moving this into a separate coro.

=item $nidle = Coro::State::cctx_idle

Returns the number of allocated but idle (free for reuse) C level
coro. Currently, Coro will limit the number of idle/unused cctxs to
8.

=item $old = Coro::State::cctx_stacksize [$new_stacksize]

Returns the current C stack size and optionally sets the new I<minimum>
stack size to C<$new_stacksize> I<long>s. Existing stacks will not
be changed, but Coro will try to replace smaller stacks as soon as
possible. Any Coro::State that starts to use a stack after this call is
guaranteed this minimum stack size.

Please note that coros will only need to use a C-level stack if the
interpreter recurses or calls a function in a module that calls back into
the interpreter, so use of this feature is usually never needed.

=item $old = Coro::State::cctx_max_idle [$new_count]

Coro caches C contexts that are not in use currently, as creating them
from scratch has some overhead.

This function returns the current maximum number of idle C contexts and
optionally sets the new amount. The count must be at least C<1>, with the
default being C<4>.

=item @states = Coro::State::list

Returns a list of all states currently allocated.

=item $was_enabled = Coro::State::enable_times [$enable]

Enables/disables/queries the current state of per-thread real and
cpu-time gathering.

When enabled, the real time and the cpu time (user + system time)
spent in each thread is accumulated. If disabled, then the accumulated
times will stay as they are (they start at 0).

Currently, cpu time is only measured on GNU/Linux systems, all other
systems only gather real time.

Enabling time profiling slows down thread switching by a factor of 2 to
10, depending on platform on hardware.

The times will be displayed when running C<Coro::Debug::command "ps">, and
cna be queried by calling C<< $state->times >>.

=item ($real, $cpu) = $state->times

Returns the real time and cpu times spent in the given C<$state>. See
C<Coro::State::enable_times> for more info.

=item $clone = $state->clone

This exciting method takes a Coro::State object and clones it, i.e., it
creates a copy. This makes it possible to restore a state more than once,
and even return to states that have returned or have been terminated.

Since its only known purpose is for intellectual self-gratification, and
because it is a difficult piece of code, it is not enabled by default, and
not supported.

Here are a few little-known facts: First, coros *are* full/true/real
continuations. Secondly Coro::State objects (without clone) *are* first
class continuations. Thirdly, nobody has ever found a use for the full
power of call/cc that isn't better (faster, easier, more efficiently)
implemented differently, and nobody has yet found a useful control
construct that can't be implemented without it already, just much faster
and with fewer resources. And lastly, Scheme's call/cc doesn't support
using call/cc to implement threads.

Among the games you can play with this is implementing a scheme-like
call-with-current-continuation, as the following code does (well, with
small differences).

   # perl disassociates from local lexicals on frame exit,
   # so use a global variable for return values.
   my @ret;

   sub callcc($@) {
      my ($func, @arg) = @_;

      my $continuation = new Coro::State;
      $continuation->transfer (new Coro::State sub {
         my $escape = sub {
            @ret = @_;
            Coro::State->new->transfer ($continuation->clone);
         };
         $escape->($func->($escape, @arg));
      });

      my @ret_ = @ret; @ret = ();
      wantarray ? @ret_ : pop @ret_
   }

Which could be used to implement a loop like this:

   async {
      my $n; 
      my $l = callcc sub { $_[0] };
     
      $n++; 
      print "iteration $n\n";

      $l->($l) unless $n == 10;
   };  

If you find this confusing, then you already understand the coolness of
call/cc: It can turn anything into spaghetti code real fast.

Besides, call/cc is much less useful in a Perl-like dynamic language (with
references, and its scoping rules) then in, say, scheme.

Now, the known limitations of C<clone>:

It probably only works on perl 5.10; it cannot clone a coro inside
the substition operator (but windows perl can't fork from there either)
and some other contexts, and C<abort ()> is the preferred mechanism to
signal errors. It cannot clone a state that has a c context attached
(implementing clone on the C level is too hard for me to even try),
which rules out calling call/cc from the main coro. It cannot
clone a context that hasn't even been started yet. It doesn't work with
C<-DDEBUGGING> (but what does). It probably also leaks, and sometimes
triggers a few assertions inside Coro. Most of these limitations *are*
fixable with some effort, but that's pointless just to make a point that
it could be done.

The current implementation could without doubt be optimised to be a
constant-time operation by doing lazy stack copying, if somebody were
insane enough to invest the time.

=cut

# used by Coro::Debug only atm.
sub debug_desc {
   $_[0]{desc}
}

# for very deep reasons, we must initialise $Coro::main here.

{
   package Coro;

   our $main;    # main coro
   our $current; # current coro

   $main = Coro::new Coro::;

   $main->{desc} = "[main::]";

   # maybe some other module used Coro::Specific before...
   $main->{_specific} = $current->{_specific}
      if $current;

   _set_current $main;
}

# we also make sure we have Coro::AnyEvent when AnyEvent is used,
# without loading or initialising AnyEvent
if (defined $AnyEvent::MODEL) {
   require Coro::AnyEvent;
} else {
   push @AnyEvent::post_detect, sub { require Coro::AnyEvent };
}

1;

=back

=head1 BUGS

This module is not thread-safe. You must only ever use this module from
the same thread (this requirement might be removed in the future).

=head1 SEE ALSO

L<Coro>.

=head1 AUTHOR

 Marc Lehmann <schmorp@schmorp.de>
 http://home.schmorp.de/

=cut

Revision:	1.159
Committed:	Thu Nov 11 00:09:09 2010 UTC (13 years, 6 months ago) by root
Branch:	MAIN
CVS Tags:	rel-5_25
Changes since 1.158:	+1 -1 lines
Log Message:	5.25
#	User	Rev	Content
1	root	1.1	=head1 NAME
2
3	root	1.132	Coro::State - first class continuations
4	root	1.1
5			=head1 SYNOPSIS
6
7			use Coro::State;
8
9			$new = new Coro::State sub {
10	root	1.140	print "in coro (called with @_), switching back\n";
11	root	1.43	$new->transfer ($main);
12	root	1.140	print "in coro again, switching back\n";
13	root	1.43	$new->transfer ($main);
14	root	1.3	}, 5;
15	root	1.1
16			$main = new Coro::State;
17
18	root	1.140	print "in main, switching to coro\n";
19	root	1.43	$main->transfer ($new);
20	root	1.140	print "back in main, switch to coro again\n";
21	root	1.43	$main->transfer ($new);
22	root	1.1	print "back in main\n";
23
24			=head1 DESCRIPTION
25
26	root	1.140	This module implements coro. Coros, similar to threads and continuations,
27	root	1.1	allow you to run more than one "thread of execution" in parallel. Unlike
28	root	1.140	so-called "kernel" threads, there is no parallelism and only voluntary
29			switching is used so locking problems are greatly reduced. The latter is
30			called "cooperative" threading as opposed to "preemptive" threading.
31	root	1.42
32			This can be used to implement non-local jumps, exception handling,
33	root	1.140	continuation objects and more.
34	root	1.1
35			This module provides only low-level functionality. See L<Coro> and related
36	root	1.140	modules for a higher level threads abstraction including a scheduler.
37	root	1.1
38	root	1.86	=head2 MODEL
39
40	root	1.140	Coro::State implements two different thread models: Perl and C. The C
41			threads (called cctx's) are basically simplified perl interpreters
42			running/interpreting the Perl threads. A single interpreter can run any
43			number of Perl threads, so usually there are very few C threads.
44
45			When Perl code calls a C function (e.g. in an extension module) and that C
46			function then calls back into Perl or transfers control to another thread,
47			the C thread can no longer execute other Perl threads, so it stays tied to
48			the specific thread until it returns to the original Perl caller, after
49			which it is again available to run other Perl threads.
50	root	1.86
51	root	1.140	The main program always has its own "C thread" (which really is
52	root	1.86	the Perl interpreter running the whole program), so there will always
53	root	1.140	be at least one additional C thread. You can use the debugger (see
54			L<Coro::Debug>) to find out which threads are tied to their cctx and
55	root	1.86	which aren't.
56
57	root	1.9	=head2 MEMORY CONSUMPTION
58
59	root	1.140	A newly created Coro::State that has not been used only allocates a
60	root	1.84	relatively small (a hundred bytes) structure. Only on the first
61	root	1.94	C<transfer> will perl allocate stacks (a few kb, 64 bit architetcures
62	root	1.140	use twice as much, i.e. a few kb :) and optionally a C stack/thread
63			(cctx) for threads that recurse through C functions. All this is very
64	root	1.94	system-dependent. On my x86-pc-linux-gnu system this amounts to about 2k
65	root	1.140	per (non-trivial but simple) Coro::State.
66	root	1.94
67			You can view the actual memory consumption using Coro::Debug. Keep in mind
68			that a for loop or other block constructs can easily consume 100-200 bytes
69			per nesting level.
70	root	1.1
71			=cut
72
73			package Coro::State;
74
75	root	1.153	use common::sense;
76	root	1.47
77	root	1.84	use Carp;
78	root	1.87
79	root	1.109	use Time::HiRes (); # currently only used for PerlIO::cede
80
81	root	1.87	our $DIEHOOK;
82			our $WARNHOOK;
83
84			BEGIN {
85			$DIEHOOK = sub { };
86			$WARNHOOK = sub { warn $_[0] };
87			}
88
89			sub diehook { &$DIEHOOK }
90			sub warnhook { &$WARNHOOK }
91	root	1.84
92	root	1.47	use XSLoader;
93	root	1.18
94	root	1.1	BEGIN {
95	root	1.159	our $VERSION = 5.25;
96	root	1.1
97	root	1.67	# must be done here because the xs part expects it to exist
98			# it might exist already because Coro::Specific created it.
99			$Coro::current \|\|= { };
100
101	root	1.101	{
102			# save/restore the handlers before/after overwriting %SIG magic
103			local $SIG{__DIE__};
104			local $SIG{__WARN__};
105
106			XSLoader::load __PACKAGE__, $VERSION;
107			}
108
109			# need to do it after overwriting the %SIG magic
110			$SIG{__DIE__} \|\|= \&diehook;
111			$SIG{__WARN__} \|\|= \&warnhook;
112	root	1.1	}
113
114	root	1.51	use Exporter;
115	root	1.47	use base Exporter::;
116	root	1.5
117	root	1.84	=head2 GLOBAL VARIABLES
118
119			=over 4
120
121			=item $Coro::State::DIEHOOK
122
123			This works similarly to C<$SIG{__DIE__}> and is used as the default die
124	root	1.140	hook for newly created Coro::States. This is useful if you want some generic
125			logging function that works for all threads that don't set their own
126	root	1.84	hook.
127
128			When Coro::State is first loaded it will install these handlers for the
129	root	1.101	main program, too, unless they have been overwritten already.
130	root	1.84
131	root	1.100	The default handlers provided will behave like the built-in ones (as if
132	root	1.84	they weren't there).
133
134	root	1.150	If you don't want to exit your program on uncaught exceptions, you must
135			not return from your die hook - call C<Coro::terminate> instead.
136	root	1.125
137	root	1.94	Note 1: You I<must> store a valid code reference in these variables,
138			C<undef> will I<not> do.
139	root	1.84
140	root	1.140	Note 2: The value of this variable will be shared among all threads, so
141			changing its value will change it in all threads that don't have their
142	root	1.100	own die handler.
143	root	1.84
144			=item $Coro::State::WARNHOOK
145
146			Similar to above die hook, but augments C<$SIG{__WARN__}>.
147
148			=back
149
150			=head2 FUNCTIONS
151
152			=over 4
153
154	root	1.65	=item $coro = new Coro::State [$coderef[, @args...]]
155	root	1.1
156	root	1.140	Create a new Coro::State thread object and return it. The first
157			C<transfer> call to this thread will start execution at the given
158			coderef, with the given arguments.
159	root	1.125
160			Note that the arguments will not be copied. Instead, as with normal
161	root	1.140	function calls, the thread receives passed arguments by reference, so
162	root	1.125	make sure you don't change them in unexpected ways.
163
164	root	1.140	Returning from such a thread is I<NOT> supported. Neither is calling
165	root	1.125	C<exit> or throwing an uncaught exception. The following paragraphs
166			describe what happens in current versions of Coro.
167	root	1.94
168			If the subroutine returns the program will be terminated as if execution
169			of the main program ended.
170
171			If it throws an exception the program will terminate unless the exception
172			is caught, exactly like in the main program.
173	root	1.74
174	root	1.140	Calling C<exit> in a thread does the same as calling it in the main
175	root	1.133	program, but due to libc bugs on many BSDs, this doesn't work reliable
176			everywhere.
177	root	1.1
178			If the coderef is omitted this function will create a new "empty"
179	root	1.140	thread, i.e. a thread that cannot be transfered to but can be used
180			to save the current thread state in (note that this is dangerous, as no
181			reference is taken to ensure that the "current thread state" survives,
182	root	1.104	the caller is responsible to ensure that the cloned state does not go
183			away).
184	root	1.1
185	root	1.55	The returned object is an empty hash which can be used for any purpose
186			whatsoever, for example when subclassing Coro::State.
187
188	root	1.140	Certain variables are "localised" to each thread, that is, certain
189			"global" variables are actually per thread. Not everything that would
190	root	1.79	sensibly be localised currently is, and not everything that is localised
191			makes sense for every application, and the future might bring changes.
192	root	1.1
193	root	1.140	The following global variables can have different values per thread,
194	root	1.84	and have the stated initial values:
195	root	1.5
196	root	1.83	Variable Initial Value
197			@_ whatever arguments were passed to the Coro
198			$_ undef
199			$@ undef
200			$/ "\n"
201	root	1.88	$SIG{__DIE__} aliased to $Coro::State::DIEHOOK(*)
202			$SIG{__WARN__} aliased to $Coro::State::WARNHOOK(*)
203	root	1.83	(default fh) *STDOUT
204	root	1.133	$^H, %^H zero/empty.
205	root	1.84	$1, $2... all regex results are initially undefined
206	root	1.2
207	root	1.88	(*) reading the value from %SIG is not supported, but local'ising is.
208
209	root	1.70	If you feel that something important is missing then tell me. Also
210	root	1.2	remember that every function call that might call C<transfer> (such
211			as C<Coro::Channel::put>) might clobber any global and/or special
212			variables. Yes, this is by design ;) You can always create your own
213			process abstraction model that saves these variables.
214	root	1.1
215	root	1.9	The easiest way to do this is to create your own scheduling primitive like
216	root	1.140	in the code below, and use it in your threads:
217	root	1.1
218	root	1.84	sub my_cede {
219	root	1.80	local ($;, ...);
220	root	1.84	Coro::cede;
221	root	1.1	}
222
223	root	1.140	Another way is to use dynamic winders, see C<Coro::on_enter> and
224			C<Coro::on_leave> for this.
225
226			=item $prev->transfer ($next)
227
228			Save the state of the current subroutine in C<$prev> and switch to the
229			thread saved in C<$next>.
230
231			The "state" of a subroutine includes the scope, i.e. lexical variables and
232			the current execution state (subroutine, stack).
233
234			=item $state->is_new
235
236			Returns true iff this Coro::State object is "new", i.e. has never been run
237			yet. Those states basically consist of only the code reference to call and
238			the arguments, but consumes very little other resources. New states will
239			automatically get assigned a perl interpreter when they are transfered to.
240
241			=item $state->is_destroyed
242
243			Returns true iff the Coro::State object has been destroyed (by
244			C<cancel>), i.e. it's resources freed because they were C<cancel>'d (or
245			C<terminate>'d).
246
247			=item $state->cancel
248
249			Forcefully destructs the given Coro::State. While you can keep the
250			reference, and some memory is still allocated, the Coro::State object is
251			effecticely dead, destructors have been freed, it cannot be transfered to
252			anymore.
253	root	1.79
254	root	1.119	=item $state->throw ([$scalar])
255
256			See L<< Coro->throw >>.
257
258	root	1.76	=item $state->call ($coderef)
259
260	root	1.94	Try to call the given C<$coderef> in the context of the given state. This
261	root	1.76	works even when the state is currently within an XS function, and can
262			be very dangerous. You can use it to acquire stack traces etc. (see the
263			Coro::Debug module for more details). The coderef MUST NOT EVER transfer
264			to another state.
265
266			=item $state->eval ($string)
267
268	root	1.94	Like C<call>, but eval's the string. Dangerous.
269	root	1.76
270	root	1.90	=item $state->swap_defsv
271
272			=item $state->swap_defav
273
274			Swap the current C<$_> (swap_defsv) or C<@_> (swap_defav) with the
275			equivalent in the saved state of C<$state>. This can be used to give the
276	root	1.140	coro a defined content for C<@_> and C<$_> before transfer'ing to it.
277	root	1.90
278	root	1.154	=item $state->swap_sv (\$sv, \$swap_sv)
279
280			This (very advanced) function can be used to make I<any> variable local to
281			a thread.
282
283			It works by swapping the contents of C<$sv> and C<$swap_sv> each time the
284			thread is entered and left again, i.e. it is similarly to:
285
286			$tmp = $sv; $sv = $swap_sv; $swap_sv = $tmp;
287
288			Except that it doesn't make an copies and works on hashes and even more
289			exotic values (code references!).
290
291			Needless to say, this function can be very very dangerous: you can easily
292			swap a hash with a reference (i.e. C<%hash> I<becomes> a reference), and perl
293			will not like this at all.
294
295			It will also swap "magicalness" - so when swapping a builtin perl variable
296			(such as C<$.>), it will lose it's magicalness, which, again, perl will
297			not like, so don't do it.
298
299			Lastly, the C<$swap_sv> itself will be used, not a copy, so make sure you
300			give each thread it's own C<$swap_sv> instance.
301
302			It is, however, quite safe to swap some normal variable with
303			another. For example, L<PApp::SQL> stores the default database handle in
304			C<$PApp::SQL::DBH>. To make this a per-thread variable, use this:
305
306			my $private_dbh = ...;
307			$coro->swap_sv (\$PApp::SQL::DBH, \$private_dbh);
308
309			This results in C<$PApp::SQL::DBH> having the value of C<$private_dbh>
310			while it executes, and whatever other value it had when it doesn't
311			execute.
312
313			You can also swap hashes and other values:
314
315			my %private_hash;
316			$coro->swap_sv (\%some_hash, \%private_hash);
317
318	root	1.77	=item $state->trace ($flags)
319
320			Internal function to control tracing. I just mention this so you can stay
321	root	1.90	away from abusing it.
322	root	1.77
323	root	1.117	=item $bytes = $state->rss
324
325	root	1.140	Returns the memory allocated by the coro (which includes static
326	root	1.117	structures, various perl stacks but NOT local variables, arguments or any
327			C context data). This is a rough indication of how much memory it might
328			use.
329
330	root	1.94	=item $state->has_cctx
331
332	root	1.117	Returns whether the state currently uses a cctx/C context. An active
333	root	1.94	state always has a cctx, as well as the main program. Other states only
334			use a cctxts when needed.
335
336	root	1.117	=item Coro::State::force_cctx
337	root	1.94
338	root	1.140	Forces the allocation of a C context for the currently running coro
339	root	1.117	(if not already done). Apart from benchmarking there is little point
340			in doing so, however.
341	root	1.94
342	root	1.117	=item $ncctx = Coro::State::cctx_count
343	root	1.64
344	root	1.140	Returns the number of C-level coro allocated. If this number is
345	root	1.64	very high (more than a dozen) it might help to identify points of C-level
346	root	1.140	recursion in your code and moving this into a separate coro.
347	root	1.64
348	root	1.117	=item $nidle = Coro::State::cctx_idle
349	root	1.64
350			Returns the number of allocated but idle (free for reuse) C level
351	root	1.140	coro. Currently, Coro will limit the number of idle/unused cctxs to
352	root	1.76	8.
353	root	1.64
354	root	1.117	=item $old = Coro::State::cctx_stacksize [$new_stacksize]
355	root	1.72
356			Returns the current C stack size and optionally sets the new I<minimum>
357			stack size to C<$new_stacksize> I<long>s. Existing stacks will not
358			be changed, but Coro will try to replace smaller stacks as soon as
359	root	1.91	possible. Any Coro::State that starts to use a stack after this call is
360	root	1.94	guaranteed this minimum stack size.
361
362	root	1.140	Please note that coros will only need to use a C-level stack if the
363	root	1.94	interpreter recurses or calls a function in a module that calls back into
364			the interpreter, so use of this feature is usually never needed.
365	root	1.72
366	root	1.117	=item $old = Coro::State::cctx_max_idle [$new_count]
367
368			Coro caches C contexts that are not in use currently, as creating them
369			from scratch has some overhead.
370	root	1.92
371	root	1.117	This function returns the current maximum number of idle C contexts and
372			optionally sets the new amount. The count must be at least C<1>, with the
373			default being C<4>.
374	root	1.92
375	root	1.76	=item @states = Coro::State::list
376
377			Returns a list of all states currently allocated.
378
379	root	1.144	=item $was_enabled = Coro::State::enable_times [$enable]
380
381			Enables/disables/queries the current state of per-thread real and
382			cpu-time gathering.
383
384			When enabled, the real time and the cpu time (user + system time)
385			spent in each thread is accumulated. If disabled, then the accumulated
386			times will stay as they are (they start at 0).
387
388			Currently, cpu time is only measured on GNU/Linux systems, all other
389			systems only gather real time.
390
391			Enabling time profiling slows down thread switching by a factor of 2 to
392			10, depending on platform on hardware.
393
394			The times will be displayed when running C<Coro::Debug::command "ps">, and
395			cna be queried by calling C<< $state->times >>.
396
397			=item ($real, $cpu) = $state->times
398
399			Returns the real time and cpu times spent in the given C<$state>. See
400			C<Coro::State::enable_times> for more info.
401
402	root	1.127	=item $clone = $state->clone
403
404	root	1.136	This exciting method takes a Coro::State object and clones it, i.e., it
405			creates a copy. This makes it possible to restore a state more than once,
406			and even return to states that have returned or have been terminated.
407	root	1.127
408	root	1.136	Since its only known purpose is for intellectual self-gratification, and
409	root	1.127	because it is a difficult piece of code, it is not enabled by default, and
410			not supported.
411
412	root	1.140	Here are a few little-known facts: First, coros are full/true/real
413	root	1.136	continuations. Secondly Coro::State objects (without clone) are first
414			class continuations. Thirdly, nobody has ever found a use for the full
415			power of call/cc that isn't better (faster, easier, more efficiently)
416			implemented differently, and nobody has yet found a useful control
417			construct that can't be implemented without it already, just much faster
418	root	1.138	and with fewer resources. And lastly, Scheme's call/cc doesn't support
419			using call/cc to implement threads.
420	root	1.136
421			Among the games you can play with this is implementing a scheme-like
422			call-with-current-continuation, as the following code does (well, with
423			small differences).
424
425			# perl disassociates from local lexicals on frame exit,
426			# so use a global variable for return values.
427			my @ret;
428	root	1.127
429	root	1.136	sub callcc($@) {
430	root	1.129	my ($func, @arg) = @_;
431	root	1.127
432	root	1.136	my $continuation = new Coro::State;
433			$continuation->transfer (new Coro::State sub {
434	root	1.129	my $escape = sub {
435	root	1.136	@ret = @_;
436			Coro::State->new->transfer ($continuation->clone);
437	root	1.129	};
438			$escape->($func->($escape, @arg));
439	root	1.136	});
440	root	1.127
441	root	1.136	my @ret_ = @ret; @ret = ();
442			wantarray ? @ret_ : pop @ret_
443	root	1.127	}
444
445	root	1.136	Which could be used to implement a loop like this:
446
447			async {
448			my $n;
449			my $l = callcc sub { $_[0] };
450
451			$n++;
452			print "iteration $n\n";
453
454			$l->($l) unless $n == 10;
455			};
456
457			If you find this confusing, then you already understand the coolness of
458			call/cc: It can turn anything into spaghetti code real fast.
459	root	1.127
460			Besides, call/cc is much less useful in a Perl-like dynamic language (with
461			references, and its scoping rules) then in, say, scheme.
462
463	root	1.130	Now, the known limitations of C<clone>:
464	root	1.127
465	root	1.140	It probably only works on perl 5.10; it cannot clone a coro inside
466	root	1.129	the substition operator (but windows perl can't fork from there either)
467			and some other contexts, and C<abort ()> is the preferred mechanism to
468			signal errors. It cannot clone a state that has a c context attached
469	root	1.131	(implementing clone on the C level is too hard for me to even try),
470	root	1.140	which rules out calling call/cc from the main coro. It cannot
471	root	1.131	clone a context that hasn't even been started yet. It doesn't work with
472	root	1.130	C<-DDEBUGGING> (but what does). It probably also leaks, and sometimes
473			triggers a few assertions inside Coro. Most of these limitations are
474			fixable with some effort, but that's pointless just to make a point that
475			it could be done.
476	root	1.127
477	root	1.136	The current implementation could without doubt be optimised to be a
478			constant-time operation by doing lazy stack copying, if somebody were
479			insane enough to invest the time.
480
481	root	1.1	=cut
482
483	root	1.115	# used by Coro::Debug only atm.
484	root	1.75	sub debug_desc {
485			$_[0]{desc}
486			}
487
488	root	1.122	# for very deep reasons, we must initialise $Coro::main here.
489
490			{
491			package Coro;
492
493	root	1.140	our $main; # main coro
494			our $current; # current coro
495	root	1.122
496	root	1.151	$main = Coro::new Coro::;
497	root	1.122
498			$main->{desc} = "[main::]";
499
500			# maybe some other module used Coro::Specific before...
501			$main->{_specific} = $current->{_specific}
502			if $current;
503
504			_set_current $main;
505			}
506
507	root	1.155	# we also make sure we have Coro::AnyEvent when AnyEvent is used,
508			# without loading or initialising AnyEvent
509			if (defined $AnyEvent::MODEL) {
510			require Coro::AnyEvent;
511			} else {
512			push @AnyEvent::post_detect, sub { require Coro::AnyEvent };
513			}
514
515	root	1.1	1;
516
517			=back
518
519			=head1 BUGS
520
521	root	1.5	This module is not thread-safe. You must only ever use this module from
522	root	1.94	the same thread (this requirement might be removed in the future).
523	root	1.1
524			=head1 SEE ALSO
525
526			L<Coro>.
527
528			=head1 AUTHOR
529
530	root	1.41	Marc Lehmann <schmorp@schmorp.de>
531	root	1.39	http://home.schmorp.de/
532	root	1.1
533			=cut
534