=pod =encoding UTF-8 =head1 NAME Parallel::WorkUnit - Provide multi-paradigm forking with ability to pass back data =head1 VERSION version 2.202270 =head1 SYNOPSIS # # Standard Interface # my $wu = Parallel::WorkUnit->new(); $wu->async( sub { ... }, \&callback ); $wu->waitall(); # # Limiting Maximum Parallelization # $wu->max_children(5); $wu->queue( sub { ... }, \&callback ); $wu->waitall(); # # Ordered Responses # my $wu = Parallel::WorkUnit->new(); $wu->async( sub { ... } ); @results = $wu->waitall(); # # Spawning off X number of workers # (Ordered Response paradigm shown with 10 children) # my $wu = Parallel::WorkUnit->new(); $wu->asyncs( 10, sub { ... } ); @results = $wu->waitall(); # # AnyEvent Interface (only usable if AnyEvent is installed) # use AnyEvent; my $wu = Parallel::WorkUnit->new(); $wu->use_anyevent(1); $wu->async( sub { ... }, \&callback ); $wu->waitall(); # Not strictly necessary # # Just spawn something into another process, don't capture return # values, don't allow waiting on process, etc. # my $wu = Parallel::Workunit->new(); $wu->start( { ... } ); =head1 DESCRIPTION This is a very simple forking implementation of parallelism, with the ability to pass data back from the asyncronous child process in a relatively efficient way (with the limitation of using a pipe to pass the information, serialized, back). It was designed to be very simple for a developer to use, with the ability to pass reasonably large amounts of data back to the parent process. This module is also designed to work with AnyEvent when desired, but it does not require AnyEvent to be installed for other functionality to work. There are many other Parallel::* applications in CPAN - it would be worth any developer's time to look through those and choose the best one. =head1 ATTRIBUTES =head2 use_anyevent $wu->use_anyevent(1); If set to a value that is true, creates AnyEvent watchers for each asyncronous or queued job. The equivilent of an C condition variable C, used when all processes finish executing, is the C method. However, the processes are integrated into a standard C loop, so it isn't strictly necessary to call C. In addition, a call to C will execute other processes in the C event loop. The default value is false. =head2 max_children $wu->max_children(5); $wu->max_children(undef); say "Max number of children: " . $wu->max_children(); If set to a value other than zero or undef, limits the number of outstanding queue children (created by the C method) that can be executing at any given time. This defaults to 5. This attribute does not impact the C method's ability to create children, but these children will count against the limit used by C. Calling without any parameters will return the number of children. =head1 METHODS =head2 new Create a new workunit class. Optionally, takes a list that corresponds to a hashref, in the form of key and value. This accepts the key C, which, if present (and not undef) will limit the number of spawned subprocesses that can be active when using the C method. Defaults to 5. See the C method for additional information. =head2 async( sub { ... }, \&callback ) $wu->async( sub { return 1 }, \&callback ); # To get back results in "ordered" return mode $wu->async( sub { return 1 } ); @results = $wu->waitall(); Spawns work on a new forked process. The forked process inherits all Perl state from the parent process, as would be expected with a standard C call. The child shares nothing with the parent, other than the return value of the work done. The work is specified either as a subroutine reference or an anonymous sub (C) and should return a scalar. Any scalar that L's C method can deal with is acceptable (for instance, a hash reference or C). The result is serialized and streamed back to the parent process via a pipe. The parent, in a C call, will call the callback function (if provided) with the unserialized return value. If a callback is not provided, the parent, in the C call, will gather these results and return them as an ordered list. In all modes, should the child process C, the parent process will also die (inside the C method). The PID of the child is returned to the parent process when this method is executed. The C attribute is not examined in this method - you can spawn a new child regardless of the number of children already spawned. However, you children started with this method still count against the limit used by C. =head2 asyncs( $children, sub { ... }, \&callback ) $wu->asyncs( 10, sub { return 1 }, \&callback ); # To get back results in "ordered" return mode $wu->asyncs( 10, sub { return 1 } ); @results = $wu->waitall(); Added in 1.117. This functions similarly to the C method, with a couple key differences. First, it takes an additional parameter, C<$children>, that specifies the number of child processes to spawn. Like the C method, the children are spawned immediately, regardless of the value of the C attribute. In addition, when the sub/coderef is executed, it is called with a single parameter representing the child number in that particular instance (between zero to C<$children-1>). Returns the number of children spawned. See C for more details on how this function works. =head2 waitall() Called from the parent method while waiting for the children to exit. This method handles children that C or return a serializable data structure. When all children return, this method will return. If a child dies unexpectedly, this method will C and propagate a modified exception. In the standard (not ordered) mode, I.E. where a callback was passed to C, this will return nothing. In the ordered mode, I.E. where no callbacks were provided to C, this will return the results of the async calls in an ordered list. The list will be ordered by the order in which the async calls were executed. =head2 waitone() This method similarly to C, but only waits for a single PID. It will return after any PID exits. If this method is called when there is no processes executing, it will simply return undef. Otherwise, it will wait and then return 1. =head2 wait($pid) This functions simiarly to C, but waits only for a specific PID. See the C documentation above for details. If C is called on a process that is already done executing, it simply returns. Otherwise, it waits until the child process's work unit is complete and executes the callback routine, then returns. =head2 count() This method returns the number of currently outstanding processes (in either a running state or a waiting to send their output). =head2 queue( sub { ... }, \&callback ) Spawns work on a new forked process, doing so immediately if less than C are running. If there are already C are running, this will run the process once a slot becomes available. This method should be treated as nearly identical to C, with the only difference being the above behavior (limiting to C) and not returning a PID. Instead, a value of 1 is returned if the process is immediately started, C otherwise. The result is serialized and streamed back to the parent process via a pipe. The parent, in a C call, will call the callback function (if provided) with the unserialized return value. If a callback is not provided, the parent, in the C call, will gather these results and return them as an ordered list. =head2 start( sub { ... } ); Added in 1.191810. Spawns work on a new forked process, doing so immediately regardless of how many other children are running. This method is similar to C, but unlike C, no provision to receive return value or wait on the child is made. This is somewhat similar to C in Raku (but differs as this starts a subprocess, not a new thread, and there is thus no shared data (changes to data in the child process will not be seen in the parent process). Note that the child inherits all open file descriptors. Not also that the child process will be part of the same process group as the parent process. Additional work is required to daemonize the child. =head1 AUTHOR Joelle Maslak =head1 COPYRIGHT AND LICENSE This software is copyright (c) 2020 by Joelle Maslak. This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself. =cut