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 |
# alternatively create an async coroutine like this: |
12 |
|
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 |
threads but don't run in parallel at the same time even on SMP machines. |
22 |
The specific flavor of coroutine use din this module also guarentees you |
23 |
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 |
|
33 |
In this module, coroutines are defined as "callchain + lexical variables |
34 |
+ @_ + $_ + $@ + $/ + C stack), that is, a coroutine has its own |
35 |
callchain, its own set of lexicals and its own set of perls most |
36 |
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 |
This variable is strictly *read-only*. It is provided for |
46 |
performance reasons. If performance is not essentiel you are |
47 |
encouraged to use the "Coro::current" function instead. |
48 |
|
49 |
$idle |
50 |
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 |
|
62 |
STATIC METHODS |
63 |
Static methods are actually functions that operate on the current |
64 |
coroutine only. |
65 |
|
66 |
async { ... } [@args...] |
67 |
Create a new asynchronous coroutine and return it's coroutine object |
68 |
(usually unused). When the sub returns the new coroutine is |
69 |
automatically terminated. |
70 |
|
71 |
Calling "exit" in a coroutine will not work correctly, so do not do |
72 |
that. |
73 |
|
74 |
When the coroutine dies, the program will exit, just as in the main |
75 |
program. |
76 |
|
77 |
# create a new coroutine that just prints its arguments |
78 |
async { |
79 |
print "@_\n"; |
80 |
} 1,2,3,4; |
81 |
|
82 |
async_pool { ... } [@args...] |
83 |
Similar to "async", but uses a coroutine pool, so you should not |
84 |
call terminate or join (although you are allowed to), and you get a |
85 |
coroutine that might have executed other code already (which can be |
86 |
good or bad :). |
87 |
|
88 |
Also, the block is executed in an "eval" context and a warning will |
89 |
be issued in case of an exception instead of terminating the |
90 |
program, as "async" does. As the coroutine is being reused, stuff |
91 |
like "on_destroy" will not work in the expected way, unless you call |
92 |
terminate or cancel, which somehow defeats the purpose of pooling. |
93 |
|
94 |
The priority will be reset to 0 after each job, otherwise the |
95 |
coroutine will be re-used "as-is". |
96 |
|
97 |
The pool size is limited to 8 idle coroutines (this can be adjusted |
98 |
by changing $Coro::POOL_SIZE), and there can be as many non-idle |
99 |
coros as required. |
100 |
|
101 |
If you are concerned about pooled coroutines growing a lot because a |
102 |
single "async_pool" used a lot of stackspace you can e.g. |
103 |
"async_pool { terminate }" once per second or so to slowly replenish |
104 |
the pool. |
105 |
|
106 |
schedule |
107 |
Calls the scheduler. Please note that the current coroutine will not |
108 |
be put into the ready queue, so calling this function usually means |
109 |
you will never be called again unless something else (e.g. an event |
110 |
handler) calls ready. |
111 |
|
112 |
The canonical way to wait on external events is this: |
113 |
|
114 |
{ |
115 |
# remember current coroutine |
116 |
my $current = $Coro::current; |
117 |
|
118 |
# register a hypothetical event handler |
119 |
on_event_invoke sub { |
120 |
# wake up sleeping coroutine |
121 |
$current->ready; |
122 |
undef $current; |
123 |
}; |
124 |
|
125 |
# call schedule until event occured. |
126 |
# in case we are woken up for other reasons |
127 |
# (current still defined), loop. |
128 |
Coro::schedule while $current; |
129 |
} |
130 |
|
131 |
cede |
132 |
"Cede" to other coroutines. This function puts the current coroutine |
133 |
into the ready queue and calls "schedule", which has the effect of |
134 |
giving up the current "timeslice" to other coroutines of the same or |
135 |
higher priority. |
136 |
|
137 |
Returns true if at least one coroutine switch has happened. |
138 |
|
139 |
Coro::cede_notself |
140 |
Works like cede, but is not exported by default and will cede to any |
141 |
coroutine, regardless of priority, once. |
142 |
|
143 |
Returns true if at least one coroutine switch has happened. |
144 |
|
145 |
terminate [arg...] |
146 |
Terminates the current coroutine with the given status values (see |
147 |
cancel). |
148 |
|
149 |
# dynamic methods |
150 |
|
151 |
COROUTINE METHODS |
152 |
These are the methods you can call on coroutine objects. |
153 |
|
154 |
new Coro \&sub [, @args...] |
155 |
Create a new coroutine and return it. When the sub returns the |
156 |
coroutine automatically terminates as if "terminate" with the |
157 |
returned values were called. To make the coroutine run you must |
158 |
first put it into the ready queue by calling the ready method. |
159 |
|
160 |
Calling "exit" in a coroutine will not work correctly, so do not do |
161 |
that. |
162 |
|
163 |
$success = $coroutine->ready |
164 |
Put the given coroutine into the ready queue (according to it's |
165 |
priority) and return true. If the coroutine is already in the ready |
166 |
queue, do nothing and return false. |
167 |
|
168 |
$is_ready = $coroutine->is_ready |
169 |
Return wether the coroutine is currently the ready queue or not, |
170 |
|
171 |
$coroutine->cancel (arg...) |
172 |
Terminates the given coroutine and makes it return the given |
173 |
arguments as status (default: the empty list). Never returns if the |
174 |
coroutine is the current coroutine. |
175 |
|
176 |
$coroutine->join |
177 |
Wait until the coroutine terminates and return any values given to |
178 |
the "terminate" or "cancel" functions. "join" can be called multiple |
179 |
times from multiple coroutine. |
180 |
|
181 |
$coroutine->on_destroy (\&cb) |
182 |
Registers a callback that is called when this coroutine gets |
183 |
destroyed, but before it is joined. The callback gets passed the |
184 |
terminate arguments, if any. |
185 |
|
186 |
$oldprio = $coroutine->prio ($newprio) |
187 |
Sets (or gets, if the argument is missing) the priority of the |
188 |
coroutine. Higher priority coroutines get run before lower priority |
189 |
coroutines. Priorities are small signed integers (currently -4 .. |
190 |
+3), that you can refer to using PRIO_xxx constants (use the import |
191 |
tag :prio to get then): |
192 |
|
193 |
PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN |
194 |
3 > 1 > 0 > -1 > -3 > -4 |
195 |
|
196 |
# set priority to HIGH |
197 |
current->prio(PRIO_HIGH); |
198 |
|
199 |
The idle coroutine ($Coro::idle) always has a lower priority than |
200 |
any existing coroutine. |
201 |
|
202 |
Changing the priority of the current coroutine will take effect |
203 |
immediately, but changing the priority of coroutines in the ready |
204 |
queue (but not running) will only take effect after the next |
205 |
schedule (of that coroutine). This is a bug that will be fixed in |
206 |
some future version. |
207 |
|
208 |
$newprio = $coroutine->nice ($change) |
209 |
Similar to "prio", but subtract the given value from the priority |
210 |
(i.e. higher values mean lower priority, just as in unix). |
211 |
|
212 |
$olddesc = $coroutine->desc ($newdesc) |
213 |
Sets (or gets in case the argument is missing) the description for |
214 |
this coroutine. This is just a free-form string you can associate |
215 |
with a coroutine. |
216 |
|
217 |
GLOBAL FUNCTIONS |
218 |
Coro::nready |
219 |
Returns the number of coroutines that are currently in the ready |
220 |
state, i.e. that can be swicthed to. The value 0 means that the only |
221 |
runnable coroutine is the currently running one, so "cede" would |
222 |
have no effect, and "schedule" would cause a deadlock unless there |
223 |
is an idle handler that wakes up some coroutines. |
224 |
|
225 |
my $guard = Coro::guard { ... } |
226 |
This creates and returns a guard object. Nothing happens until the |
227 |
objetc gets destroyed, in which case the codeblock given as argument |
228 |
will be executed. This is useful to free locks or other resources in |
229 |
case of a runtime error or when the coroutine gets canceled, as in |
230 |
both cases the guard block will be executed. The guard object |
231 |
supports only one method, "->cancel", which will keep the codeblock |
232 |
from being executed. |
233 |
|
234 |
Example: set some flag and clear it again when the coroutine gets |
235 |
canceled or the function returns: |
236 |
|
237 |
sub do_something { |
238 |
my $guard = Coro::guard { $busy = 0 }; |
239 |
$busy = 1; |
240 |
|
241 |
# do something that requires $busy to be true |
242 |
} |
243 |
|
244 |
unblock_sub { ... } |
245 |
This utility function takes a BLOCK or code reference and "unblocks" |
246 |
it, returning the new coderef. This means that the new coderef will |
247 |
return immediately without blocking, returning nothing, while the |
248 |
original code ref will be called (with parameters) from within its |
249 |
own coroutine. |
250 |
|
251 |
The reason this fucntion exists is that many event libraries (such |
252 |
as the venerable Event module) are not coroutine-safe (a weaker form |
253 |
of thread-safety). This means you must not block within event |
254 |
callbacks, otherwise you might suffer from crashes or worse. |
255 |
|
256 |
This function allows your callbacks to block by executing them in |
257 |
another coroutine where it is safe to block. One example where |
258 |
blocking is handy is when you use the Coro::AIO functions to save |
259 |
results to disk. |
260 |
|
261 |
In short: simply use "unblock_sub { ... }" instead of "sub { ... }" |
262 |
when creating event callbacks that want to block. |
263 |
|
264 |
BUGS/LIMITATIONS |
265 |
- you must make very sure that no coro is still active on global |
266 |
destruction. very bad things might happen otherwise (usually segfaults). |
267 |
|
268 |
- this module is not thread-safe. You should only ever use this module |
269 |
from the same thread (this requirement might be losened in the future |
270 |
to allow per-thread schedulers, but Coro::State does not yet allow |
271 |
this). |
272 |
|
273 |
SEE ALSO |
274 |
Support/Utility: Coro::Cont, Coro::Specific, Coro::State, Coro::Util. |
275 |
|
276 |
Locking/IPC: Coro::Signal, Coro::Channel, Coro::Semaphore, |
277 |
Coro::SemaphoreSet, Coro::RWLock. |
278 |
|
279 |
Event/IO: Coro::Timer, Coro::Event, Coro::Handle, Coro::Socket, |
280 |
Coro::Select. |
281 |
|
282 |
Embedding: <Coro:MakeMaker> |
283 |
|
284 |
AUTHOR |
285 |
Marc Lehmann <schmorp@schmorp.de> |
286 |
http://home.schmorp.de/ |
287 |
|