[ViewVC] Annotation of: cvs/Coro/README

NAME
    Coro - coroutine process abstraction

SYNOPSIS
     use Coro;

     async {
        # some asynchronous thread of execution
     };

     # alternatively create an async coroutine like this:

     sub some_func : Coro {
        # some more async code
     }

     cede;

DESCRIPTION
    This module collection manages coroutines. Coroutines are similar to
    threads but don't run in parallel at the same time even on SMP machines.
    The specific flavor of coroutine used in this module also guarantees you
    that it will not switch between coroutines unless necessary, at
    easily-identified points in your program, so locking and parallel access
    are rarely an issue, making coroutine programming much safer than
    threads programming.

    (Perl, however, does not natively support real threads but instead does
    a very slow and memory-intensive emulation of processes using threads.
    This is a performance win on Windows machines, and a loss everywhere
    else).

    In this module, coroutines are defined as "callchain + lexical variables
    + @_ + $_ + $@ + $/ + C stack), that is, a coroutine has its own
    callchain, its own set of lexicals and its own set of perls most
    important global variables.

    $main
        This coroutine represents the main program.

    $current (or as function: current)
        The current coroutine (the last coroutine switched to). The initial
        value is $main (of course).

        This variable is strictly *read-only*. It is provided for
        performance reasons. If performance is not essential you are
        encouraged to use the "Coro::current" function instead.

    $idle
        A callback that is called whenever the scheduler finds no ready
        coroutines to run. The default implementation prints "FATAL:
        deadlock detected" and exits, because the program has no other way
        to continue.

        This hook is overwritten by modules such as "Coro::Timer" and
        "Coro::Event" to wait on an external event that hopefully wake up a
        coroutine so the scheduler can run it.

        Please note that if your callback recursively invokes perl (e.g. for
        event handlers), then it must be prepared to be called recursively.

  STATIC METHODS
    Static methods are actually functions that operate on the current
    coroutine only.

    async { ... } [@args...]
        Create a new asynchronous coroutine and return it's coroutine object
        (usually unused). When the sub returns the new coroutine is
        automatically terminated.

        See the "Coro::State::new" constructor for info about the coroutine
        environment.

        Calling "exit" in a coroutine will do the same as calling exit
        outside the coroutine. Likewise, when the coroutine dies, the
        program will exit, just as it would in the main program.

           # create a new coroutine that just prints its arguments
           async {
              print "@_\n";
           } 1,2,3,4;

    async_pool { ... } [@args...]
        Similar to "async", but uses a coroutine pool, so you should not
        call terminate or join (although you are allowed to), and you get a
        coroutine that might have executed other code already (which can be
        good or bad :).

        Also, the block is executed in an "eval" context and a warning will
        be issued in case of an exception instead of terminating the
        program, as "async" does. As the coroutine is being reused, stuff
        like "on_destroy" will not work in the expected way, unless you call
        terminate or cancel, which somehow defeats the purpose of pooling.

        The priority will be reset to 0 after each job, tracing will be
        disabled, the description will be reset and the default output
        filehandle gets restored, so you can change alkl these. Otherwise
        the coroutine will be re-used "as-is": most notably if you change
        other per-coroutine global stuff such as $/ you need to revert that
        change, which is most simply done by using local as in " local $/ ".

        The pool size is limited to 8 idle coroutines (this can be adjusted
        by changing $Coro::POOL_SIZE), and there can be as many non-idle
        coros as required.

        If you are concerned about pooled coroutines growing a lot because a
        single "async_pool" used a lot of stackspace you can e.g.
        "async_pool { terminate }" once per second or so to slowly replenish
        the pool. In addition to that, when the stacks used by a handler
        grows larger than 16kb (adjustable with $Coro::POOL_RSS) it will
        also exit.

    schedule
        Calls the scheduler. Please note that the current coroutine will not
        be put into the ready queue, so calling this function usually means
        you will never be called again unless something else (e.g. an event
        handler) calls ready.

        The canonical way to wait on external events is this:

           {
              # remember current coroutine
              my $current = $Coro::current;

              # register a hypothetical event handler
              on_event_invoke sub {
                 # wake up sleeping coroutine
                 $current->ready;
                 undef $current;
              };

              # call schedule until event occurred.
              # in case we are woken up for other reasons
              # (current still defined), loop.
              Coro::schedule while $current;
           }

    cede
        "Cede" to other coroutines. This function puts the current coroutine
        into the ready queue and calls "schedule", which has the effect of
        giving up the current "timeslice" to other coroutines of the same or
        higher priority.

        Returns true if at least one coroutine switch has happened.

    Coro::cede_notself
        Works like cede, but is not exported by default and will cede to any
        coroutine, regardless of priority, once.

        Returns true if at least one coroutine switch has happened.

    terminate [arg...]
        Terminates the current coroutine with the given status values (see
        cancel).

    killall
        Kills/terminates/cancels all coroutines except the currently running
        one. This is useful after a fork, either in the child or the parent,
        as usually only one of them should inherit the running coroutines.

    # dynamic methods

  COROUTINE METHODS
    These are the methods you can call on coroutine objects.

    new Coro \&sub [, @args...]
        Create a new coroutine and return it. When the sub returns the
        coroutine automatically terminates as if "terminate" with the
        returned values were called. To make the coroutine run you must
        first put it into the ready queue by calling the ready method.

        See "async" and "Coro::State::new" for additional info about the
        coroutine environment.

    $success = $coroutine->ready
        Put the given coroutine into the ready queue (according to it's
        priority) and return true. If the coroutine is already in the ready
        queue, do nothing and return false.

    $is_ready = $coroutine->is_ready
        Return wether the coroutine is currently the ready queue or not,

    $coroutine->cancel (arg...)
        Terminates the given coroutine and makes it return the given
        arguments as status (default: the empty list). Never returns if the
        coroutine is the current coroutine.

    $coroutine->join
        Wait until the coroutine terminates and return any values given to
        the "terminate" or "cancel" functions. "join" can be called
        concurrently from multiple coroutines.

    $coroutine->on_destroy (\&cb)
        Registers a callback that is called when this coroutine gets
        destroyed, but before it is joined. The callback gets passed the
        terminate arguments, if any.

    $oldprio = $coroutine->prio ($newprio)
        Sets (or gets, if the argument is missing) the priority of the
        coroutine. Higher priority coroutines get run before lower priority
        coroutines. Priorities are small signed integers (currently -4 ..
        +3), that you can refer to using PRIO_xxx constants (use the import
        tag :prio to get then):

           PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN
               3    >     1     >      0      >    -1    >    -3     >    -4

           # set priority to HIGH
           current->prio(PRIO_HIGH);

        The idle coroutine ($Coro::idle) always has a lower priority than
        any existing coroutine.

        Changing the priority of the current coroutine will take effect
        immediately, but changing the priority of coroutines in the ready
        queue (but not running) will only take effect after the next
        schedule (of that coroutine). This is a bug that will be fixed in
        some future version.

    $newprio = $coroutine->nice ($change)
        Similar to "prio", but subtract the given value from the priority
        (i.e. higher values mean lower priority, just as in unix).

    $olddesc = $coroutine->desc ($newdesc)
        Sets (or gets in case the argument is missing) the description for
        this coroutine. This is just a free-form string you can associate
        with a coroutine.

        This method simply sets the "$coroutine->{desc}" member to the given
        string. You can modify this member directly if you wish.

    $coroutine->throw ([$scalar])
        If $throw is specified and defined, it will be thrown as an
        exception inside the coroutine at the next convinient point in time
        (usually after it gains control at the next schedule/transfer/cede).
        Otherwise clears the exception object.

        The exception object will be thrown "as is" with the specified
        scalar in $@, i.e. if it is a string, no line number or newline will
        be appended (unlike with "die").

        This can be used as a softer means than "cancel" to ask a coroutine
        to end itself, although there is no guarentee that the exception
        will lead to termination, and if the exception isn't caught it might
        well end the whole program.

  GLOBAL FUNCTIONS
    Coro::nready
        Returns the number of coroutines that are currently in the ready
        state, i.e. that can be switched to. The value 0 means that the only
        runnable coroutine is the currently running one, so "cede" would
        have no effect, and "schedule" would cause a deadlock unless there
        is an idle handler that wakes up some coroutines.

    my $guard = Coro::guard { ... }
        This creates and returns a guard object. Nothing happens until the
        object gets destroyed, in which case the codeblock given as argument
        will be executed. This is useful to free locks or other resources in
        case of a runtime error or when the coroutine gets canceled, as in
        both cases the guard block will be executed. The guard object
        supports only one method, "->cancel", which will keep the codeblock
        from being executed.

        Example: set some flag and clear it again when the coroutine gets
        canceled or the function returns:

           sub do_something {
              my $guard = Coro::guard { $busy = 0 };
              $busy = 1;

              # do something that requires $busy to be true
           }

    unblock_sub { ... }
        This utility function takes a BLOCK or code reference and "unblocks"
        it, returning the new coderef. This means that the new coderef will
        return immediately without blocking, returning nothing, while the
        original code ref will be called (with parameters) from within its
        own coroutine.

        The reason this function exists is that many event libraries (such
        as the venerable Event module) are not coroutine-safe (a weaker form
        of thread-safety). This means you must not block within event
        callbacks, otherwise you might suffer from crashes or worse.

        This function allows your callbacks to block by executing them in
        another coroutine where it is safe to block. One example where
        blocking is handy is when you use the Coro::AIO functions to save
        results to disk.

        In short: simply use "unblock_sub { ... }" instead of "sub { ... }"
        when creating event callbacks that want to block.

BUGS/LIMITATIONS
     - you must make very sure that no coro is still active on global
       destruction. very bad things might happen otherwise (usually segfaults).

     - this module is not thread-safe. You should only ever use this module
       from the same thread (this requirement might be loosened in the future
       to allow per-thread schedulers, but Coro::State does not yet allow
       this).

SEE ALSO
    Support/Utility: Coro::Specific, Coro::State, Coro::Util.

    Locking/IPC: Coro::Signal, Coro::Channel, Coro::Semaphore,
    Coro::SemaphoreSet, Coro::RWLock.

    Event/IO: Coro::Timer, Coro::Event, Coro::Handle, Coro::Socket,
    Coro::Select.

    Embedding: <Coro:MakeMaker>

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

Revision:	1.11
Committed:	Sat Oct 6 19:25:00 2007 UTC (18 years, 11 months ago) by root
Branch:	MAIN
CVS Tags:	rel-4_03
Changes since 1.10:	+15 -0 lines
Log Message:	* empty log message *
#	User	Rev	Content
1	root	1.1	NAME
2			Coro - coroutine process abstraction
3
4			SYNOPSIS
5			use Coro;
6
7			async {
8			# some asynchronous thread of execution
9			};
10
11	root	1.4	# alternatively create an async coroutine like this:
12	root	1.1
13			sub some_func : Coro {
14			# some more async code
15			}
16
17			cede;
18
19			DESCRIPTION
20			This module collection manages coroutines. Coroutines are similar to
21	root	1.5	threads but don't run in parallel at the same time even on SMP machines.
22	root	1.8	The specific flavor of coroutine used in this module also guarantees you
23	root	1.5	that it will not switch between coroutines unless necessary, at
24			easily-identified points in your program, so locking and parallel access
25			are rarely an issue, making coroutine programming much safer than
26			threads programming.
27
28			(Perl, however, does not natively support real threads but instead does
29			a very slow and memory-intensive emulation of processes using threads.
30			This is a performance win on Windows machines, and a loss everywhere
31			else).
32	root	1.1
33			In this module, coroutines are defined as "callchain + lexical variables
34	root	1.5	+ @_ + $_ + $@ + $/ + C stack), that is, a coroutine has its own
35			callchain, its own set of lexicals and its own set of perls most
36	root	1.1	important global variables.
37
38			$main
39			This coroutine represents the main program.
40
41			$current (or as function: current)
42			The current coroutine (the last coroutine switched to). The initial
43			value is $main (of course).
44
45	root	1.4	This variable is strictly read-only. It is provided for
46	root	1.8	performance reasons. If performance is not essential you are
47	root	1.4	encouraged to use the "Coro::current" function instead.
48
49	root	1.1	$idle
50	root	1.4	A callback that is called whenever the scheduler finds no ready
51			coroutines to run. The default implementation prints "FATAL:
52			deadlock detected" and exits, because the program has no other way
53			to continue.
54
55			This hook is overwritten by modules such as "Coro::Timer" and
56			"Coro::Event" to wait on an external event that hopefully wake up a
57			coroutine so the scheduler can run it.
58
59			Please note that if your callback recursively invokes perl (e.g. for
60			event handlers), then it must be prepared to be called recursively.
61	root	1.1
62			STATIC METHODS
63			Static methods are actually functions that operate on the current
64	root	1.4	coroutine only.
65	root	1.1
66			async { ... } [@args...]
67	root	1.4	Create a new asynchronous coroutine and return it's coroutine object
68			(usually unused). When the sub returns the new coroutine is
69	root	1.1	automatically terminated.
70
71	root	1.10	See the "Coro::State::new" constructor for info about the coroutine
72			environment.
73
74	root	1.7	Calling "exit" in a coroutine will do the same as calling exit
75			outside the coroutine. Likewise, when the coroutine dies, the
76			program will exit, just as it would in the main program.
77	root	1.3
78	root	1.1	# create a new coroutine that just prints its arguments
79			async {
80			print "@_\n";
81			} 1,2,3,4;
82
83	root	1.6	async_pool { ... } [@args...]
84			Similar to "async", but uses a coroutine pool, so you should not
85			call terminate or join (although you are allowed to), and you get a
86			coroutine that might have executed other code already (which can be
87			good or bad :).
88
89			Also, the block is executed in an "eval" context and a warning will
90			be issued in case of an exception instead of terminating the
91			program, as "async" does. As the coroutine is being reused, stuff
92			like "on_destroy" will not work in the expected way, unless you call
93			terminate or cancel, which somehow defeats the purpose of pooling.
94
95	root	1.10	The priority will be reset to 0 after each job, tracing will be
96			disabled, the description will be reset and the default output
97			filehandle gets restored, so you can change alkl these. Otherwise
98			the coroutine will be re-used "as-is": most notably if you change
99			other per-coroutine global stuff such as $/ you need to revert that
100			change, which is most simply done by using local as in " local $/ ".
101	root	1.6
102			The pool size is limited to 8 idle coroutines (this can be adjusted
103			by changing $Coro::POOL_SIZE), and there can be as many non-idle
104			coros as required.
105
106			If you are concerned about pooled coroutines growing a lot because a
107			single "async_pool" used a lot of stackspace you can e.g.
108			"async_pool { terminate }" once per second or so to slowly replenish
109	root	1.9	the pool. In addition to that, when the stacks used by a handler
110			grows larger than 16kb (adjustable with $Coro::POOL_RSS) it will
111			also exit.
112	root	1.6
113	root	1.1	schedule
114	root	1.4	Calls the scheduler. Please note that the current coroutine will not
115	root	1.1	be put into the ready queue, so calling this function usually means
116	root	1.4	you will never be called again unless something else (e.g. an event
117			handler) calls ready.
118
119			The canonical way to wait on external events is this:
120
121			{
122			# remember current coroutine
123			my $current = $Coro::current;
124
125			# register a hypothetical event handler
126			on_event_invoke sub {
127			# wake up sleeping coroutine
128			$current->ready;
129			undef $current;
130			};
131
132	root	1.8	# call schedule until event occurred.
133	root	1.4	# in case we are woken up for other reasons
134			# (current still defined), loop.
135			Coro::schedule while $current;
136			}
137	root	1.1
138			cede
139	root	1.4	"Cede" to other coroutines. This function puts the current coroutine
140	root	1.1	into the ready queue and calls "schedule", which has the effect of
141			giving up the current "timeslice" to other coroutines of the same or
142			higher priority.
143
144	root	1.6	Returns true if at least one coroutine switch has happened.
145
146			Coro::cede_notself
147			Works like cede, but is not exported by default and will cede to any
148			coroutine, regardless of priority, once.
149
150			Returns true if at least one coroutine switch has happened.
151
152	root	1.1	terminate [arg...]
153	root	1.4	Terminates the current coroutine with the given status values (see
154	root	1.1	cancel).
155
156	root	1.10	killall
157			Kills/terminates/cancels all coroutines except the currently running
158			one. This is useful after a fork, either in the child or the parent,
159			as usually only one of them should inherit the running coroutines.
160
161	root	1.1	# dynamic methods
162
163	root	1.4	COROUTINE METHODS
164			These are the methods you can call on coroutine objects.
165	root	1.1
166			new Coro \&sub [, @args...]
167	root	1.4	Create a new coroutine and return it. When the sub returns the
168			coroutine automatically terminates as if "terminate" with the
169			returned values were called. To make the coroutine run you must
170			first put it into the ready queue by calling the ready method.
171
172	root	1.10	See "async" and "Coro::State::new" for additional info about the
173			coroutine environment.
174	root	1.4
175			$success = $coroutine->ready
176			Put the given coroutine into the ready queue (according to it's
177			priority) and return true. If the coroutine is already in the ready
178			queue, do nothing and return false.
179
180			$is_ready = $coroutine->is_ready
181			Return wether the coroutine is currently the ready queue or not,
182
183			$coroutine->cancel (arg...)
184			Terminates the given coroutine and makes it return the given
185	root	1.6	arguments as status (default: the empty list). Never returns if the
186			coroutine is the current coroutine.
187	root	1.1
188	root	1.4	$coroutine->join
189	root	1.1	Wait until the coroutine terminates and return any values given to
190	root	1.10	the "terminate" or "cancel" functions. "join" can be called
191			concurrently from multiple coroutines.
192	root	1.1
193	root	1.6	$coroutine->on_destroy (\&cb)
194			Registers a callback that is called when this coroutine gets
195			destroyed, but before it is joined. The callback gets passed the
196			terminate arguments, if any.
197
198	root	1.4	$oldprio = $coroutine->prio ($newprio)
199	root	1.1	Sets (or gets, if the argument is missing) the priority of the
200	root	1.4	coroutine. Higher priority coroutines get run before lower priority
201			coroutines. Priorities are small signed integers (currently -4 ..
202	root	1.1	+3), that you can refer to using PRIO_xxx constants (use the import
203			tag :prio to get then):
204
205			PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN
206			3 > 1 > 0 > -1 > -3 > -4
207
208			# set priority to HIGH
209			current->prio(PRIO_HIGH);
210
211			The idle coroutine ($Coro::idle) always has a lower priority than
212			any existing coroutine.
213
214	root	1.4	Changing the priority of the current coroutine will take effect
215			immediately, but changing the priority of coroutines in the ready
216	root	1.1	queue (but not running) will only take effect after the next
217	root	1.4	schedule (of that coroutine). This is a bug that will be fixed in
218			some future version.
219	root	1.1
220	root	1.4	$newprio = $coroutine->nice ($change)
221	root	1.1	Similar to "prio", but subtract the given value from the priority
222			(i.e. higher values mean lower priority, just as in unix).
223
224	root	1.4	$olddesc = $coroutine->desc ($newdesc)
225	root	1.1	Sets (or gets in case the argument is missing) the description for
226	root	1.4	this coroutine. This is just a free-form string you can associate
227			with a coroutine.
228
229	root	1.10	This method simply sets the "$coroutine->{desc}" member to the given
230			string. You can modify this member directly if you wish.
231
232	root	1.11	$coroutine->throw ([$scalar])
233			If $throw is specified and defined, it will be thrown as an
234			exception inside the coroutine at the next convinient point in time
235			(usually after it gains control at the next schedule/transfer/cede).
236			Otherwise clears the exception object.
237
238			The exception object will be thrown "as is" with the specified
239			scalar in $@, i.e. if it is a string, no line number or newline will
240			be appended (unlike with "die").
241
242			This can be used as a softer means than "cancel" to ask a coroutine
243			to end itself, although there is no guarentee that the exception
244			will lead to termination, and if the exception isn't caught it might
245			well end the whole program.
246
247	root	1.5	GLOBAL FUNCTIONS
248			Coro::nready
249			Returns the number of coroutines that are currently in the ready
250	root	1.8	state, i.e. that can be switched to. The value 0 means that the only
251	root	1.5	runnable coroutine is the currently running one, so "cede" would
252			have no effect, and "schedule" would cause a deadlock unless there
253			is an idle handler that wakes up some coroutines.
254
255	root	1.6	my $guard = Coro::guard { ... }
256			This creates and returns a guard object. Nothing happens until the
257	root	1.7	object gets destroyed, in which case the codeblock given as argument
258	root	1.6	will be executed. This is useful to free locks or other resources in
259			case of a runtime error or when the coroutine gets canceled, as in
260			both cases the guard block will be executed. The guard object
261			supports only one method, "->cancel", which will keep the codeblock
262			from being executed.
263
264			Example: set some flag and clear it again when the coroutine gets
265			canceled or the function returns:
266
267			sub do_something {
268			my $guard = Coro::guard { $busy = 0 };
269			$busy = 1;
270
271			# do something that requires $busy to be true
272			}
273
274	root	1.4	unblock_sub { ... }
275			This utility function takes a BLOCK or code reference and "unblocks"
276			it, returning the new coderef. This means that the new coderef will
277			return immediately without blocking, returning nothing, while the
278			original code ref will be called (with parameters) from within its
279			own coroutine.
280
281	root	1.8	The reason this function exists is that many event libraries (such
282	root	1.4	as the venerable Event module) are not coroutine-safe (a weaker form
283			of thread-safety). This means you must not block within event
284			callbacks, otherwise you might suffer from crashes or worse.
285
286			This function allows your callbacks to block by executing them in
287			another coroutine where it is safe to block. One example where
288			blocking is handy is when you use the Coro::AIO functions to save
289			results to disk.
290
291			In short: simply use "unblock_sub { ... }" instead of "sub { ... }"
292			when creating event callbacks that want to block.
293	root	1.1
294			BUGS/LIMITATIONS
295			- you must make very sure that no coro is still active on global
296			destruction. very bad things might happen otherwise (usually segfaults).
297
298			- this module is not thread-safe. You should only ever use this module
299	root	1.8	from the same thread (this requirement might be loosened in the future
300	root	1.1	to allow per-thread schedulers, but Coro::State does not yet allow
301			this).
302
303			SEE ALSO
304	root	1.10	Support/Utility: Coro::Specific, Coro::State, Coro::Util.
305	root	1.2
306			Locking/IPC: Coro::Signal, Coro::Channel, Coro::Semaphore,
307			Coro::SemaphoreSet, Coro::RWLock.
308
309			Event/IO: Coro::Timer, Coro::Event, Coro::Handle, Coro::Socket,
310			Coro::Select.
311
312			Embedding: <Coro:MakeMaker>
313	root	1.1
314			AUTHOR
315			Marc Lehmann <schmorp@schmorp.de>
316			http://home.schmorp.de/
317