Thread - manipulate threads in Perl (EXPERIMENTAL, subject to change)
The Thread extension requires Perl to be built in a particular way to enable the older 5.005 threading model. Just to confuse matters, there is an alternate threading model known as "ithreads" that does NOT support this extension. If you are using a binary distribution such as ActivePerl that is built with ithreads support, this extension CANNOT be used.
use Thread;
my $t = new Thread \&start_sub, @start_args;
$result = $t->join;
$result = $t->eval;
$t->detach;
if($t->equal($another_thread)) {
# ...
}
my $tid = Thread->self->tid;
my $tlist = Thread->list;
lock($scalar);
yield();
use Thread 'async';
WARNING: Threading is an experimental feature. Both the interface
and implementation are subject to change drastically. In fact, this
documentation describes the flavor of threads that was in version
5.005. Perl 5.6.0 and later have the beginnings of support for
interpreter threads, which (when finished) is expected to be
significantly different from what is described here. The information
contained here may therefore soon be obsolete. Use at your own risk!
The Thread
module provides multithreading support for perl.
new
starts a new thread of execution in the referenced subroutine. The optional list is passed as parameters to the subroutine. Execution continues in both the subroutine and the code after the new
call.
new Thread
returns a thread object representing the newly created thread.
lock
places a lock on a variable until the lock goes out of scope. If the variable is locked by another thread, the lock
call will block until it's available. lock
is recursive, so multiple calls to lock
are safe--the variable will remain locked until the outermost lock on the variable goes out of scope.
Locks on variables only affect lock
calls--they do not affect normal access to a variable. (Locks on subs are different, and covered in a bit) If you really, really want locks to block access, then go ahead and tie them to something and manage this yourself. This is done on purpose. While managing access to variables is a good thing, perl doesn't force you out of its living room...
If a container object, such as a hash or array, is locked, all the elements of that container are not locked. For example, if a thread does a lock @a
, any other thread doing a lock($a[12])
won't block.
You may also lock
a sub, using lock &sub
. Any calls to that sub from another thread will block until the lock is released. This behaviour is not equivalent to declaring the sub with the locked
attribute. The locked
attribute serializes access to a subroutine, but allows different threads non-simultaneous access. lock &sub
, on the other hand, will not allow any other thread access for the duration of the lock.
Finally, lock
will traverse up references exactly one level. lock(\$a)
is equivalent to lock($a)
, while lock(\\$a)
is not.
async
creates a thread to execute the block immediately following it. This block is treated as an anonymous sub, and so must have a semi-colon after the closing brace. Like new Thread
, async
returns a thread object.
The Thread->self
function returns a thread object that represents the thread making the Thread->self
call.
Thread->list
returns a list of thread objects for all running and finished but un-join
ed threads.
The cond_wait
function takes a locked variable as a parameter, unlocks the variable, and blocks until another thread does a cond_signal
or cond_broadcast
for that same locked variable. The variable that cond_wait
blocked on is relocked after the cond_wait
is satisfied. If there are multiple threads cond_wait
ing on the same variable, all but one will reblock waiting to reaquire the lock on the variable. (So if you're only using cond_wait
for synchronization, give up the lock as soon as possible)
The cond_signal
function takes a locked variable as a parameter and unblocks one thread that's cond_wait
ing on that variable. If more than one thread is blocked in a cond_wait
on that variable, only one (and which one is indeterminate) will be unblocked.
If there are no threads blocked in a cond_wait
on the variable, the signal is discarded.
The cond_broadcast
function works similarly to cond_signal
. cond_broadcast
, though, will unblock all the threads that are blocked in a cond_wait
on the locked variable, rather than only one.
The yield
function allows another thread to take control of the CPU. The exact results are implementation-dependent.
join
waits for a thread to end and returns any values the thread exited with. join
will block until the thread has ended, though it won't block if the thread has already terminated.
If the thread being join
ed die
d, the error it died with will be returned at this time. If you don't want the thread performing the join
to die as well, you should either wrap the join
in an eval
or use the eval
thread method instead of join
.
The eval
method wraps an eval
around a join
, and so waits for a thread to exit, passing along any values the thread might have returned. Errors, of course, get placed into $@
.
detach
tells a thread that it is never going to be joined i.e. that all traces of its existence can be removed once it stops running. Errors in detached threads will not be visible anywhere - if you want to catch them, you should use $SIG{__DIE__} or something like that.
equal
tests whether two thread objects represent the same thread and returns true if they do.
The tid
method returns the tid of a thread. The tid is a monotonically increasing integer assigned when a thread is created. The main thread of a program will have a tid of zero, while subsequent threads will have tids assigned starting with one.
The sequence number used to assign tids is a simple integer, and no checking is done to make sure the tid isn't currently in use. If a program creates more than 2^32 - 1 threads in a single run, threads may be assigned duplicate tids. This limitation may be lifted in a future version of Perl.
attributes, Thread::Queue, Thread::Semaphore, Thread::Specific.