NAME

Command::Run - Execute external command or code reference

SYNOPSIS

use Command::Run;

# Simple usage
my $result = Command::Run->new(
    command => ['ls', '-l'],
    stderr  => 'redirect',  # merge stderr to stdout
)->run;
print $result->{data};

# Method chaining style
my $runner = Command::Run->new;
$runner->command('cat', '-n')->with(stdin => $data)->run;
print $runner->data;

# Separate stdout/stderr capture
my $result = Command::Run->new(
    command => ['some_command'],
    stderr  => 'capture',
)->run;
print "data: ", $result->{data};
print "error: ", $result->{error};

# Access output via file descriptor path
my $cmd = Command::Run->new(command => ['date']);
$cmd->update;
system("cat", $cmd->path);  # /dev/fd/N

# Code reference execution
my $result = Command::Run->new(
    command => [\&some_function, @args],
    stdin   => $input_data,
)->run;

# Using with() method
my ($out, $err);
Command::Run->new->command("command", @args)
    ->with(stdin => $input, stdout => \$out, stderr => \$err)
    ->run;

VERSION

Version 0.9902

DESCRIPTION

This module provides a simple interface to execute external commands or code references and capture their output.

When a code reference is passed as the first element of the command array, it is called in a forked child process instead of executing an external command. This avoids the overhead of loading Perl and modules for each invocation.

This module inherits from Command::Run::Tmpfile, which provides temporary file functionality. The captured output is stored in this temporary file, accessible via the path method as /dev/fd/N, which can be used as a file argument to external commands.

CONSTRUCTOR

• new(%parameters)

  Create a new Command::Run object. Parameters are passed as key-value pairs (see "PARAMETERS"):

    my $runner = Command::Run->new(
        command => \@command,
        stdin   => $input_data,
        stderr  => 'redirect',
    );
  
    # Or use method chaining
    my $runner = Command::Run->new->command('ls', '-l');
  

PARAMETERS

The following parameters can be used with new, with, and run. With new and with, parameters are stored in the object. With run, parameters are temporary and do not modify the object.

• command => \@command

  The command to execute. Can be an array reference of command and arguments, or a code reference with arguments.

• stdin => data

  Input data to be fed to the command's STDIN.

• stdout => \$scalar

  Scalar reference to capture STDOUT.

• stderr => \$scalar | 'redirect' | 'capture'

  Controls STDERR handling:

  • \$scalar - Capture STDERR into the referenced variable
  • 'redirect' - Merge STDERR into STDOUT
  • 'capture' - Capture STDERR separately (accessible via error method)
  • undef (default) - STDERR passes through to terminal

METHODS

• command(@command)

  Set the command to execute. The argument can be:

  • External command and arguments: 'ls', '-l'
  • Code reference and arguments: \&func, @args

  Returns the object for method chaining.

• with(%parameters)

  Set parameters (see "PARAMETERS"). Settings are stored in the object and persist across multiple run calls. Returns the object for method chaining.

    my ($out, $err);
    Command::Run->new("command")
        ->with(stdin => $data, stdout => \$out, stderr => \$err)
        ->run;
  

• run(%parameters)

  Execute the command and return the result hash reference. Accepts the same parameters as with, but parameters are temporary and do not modify the object state.

    # All-in-one style
    my $result = Command::Run->new->run(
        command => ['cat', '-n'],
        stdin   => $data,
        stderr  => 'redirect',
    );
  
    # Reuse runner with different input
    my $runner = Command::Run->new('cat');
    $runner->run(stdin => $input1);
    $runner->run(stdin => $input2);  # object state unchanged
  

• update()

  Execute the command and store the output. Returns the object for method chaining.

• result()

  Return the result hash reference from the last execution.

• data()

  Return the captured output (stdout) from the last execution.

• error()

  Return the captured error output (stderr) from the last execution.

• path()

  Return the file descriptor path (e.g., /dev/fd/3) for the captured output. This can be passed to external commands.

• rewind()

  Seek to the beginning of the output temp file.

• date()

  Return the timestamp of the last execution.

RETURN VALUE

The run method returns a hash reference containing:

• result

  The exit status of the command ($?).

• data

  The captured output (stdout).

• error

  The captured error output (stderr, empty string unless stderr is 'capture').

• pid

  The process ID of the executed command.

COMPARISON WITH SIMILAR MODULES

There are many modules on CPAN for executing external commands. This module is designed to be simple and lightweight, with minimal dependencies.

This module was originally developed as App::cdif::Command and has been used in production as part of the App::cdif distribution since 2014. It has also been adopted by several unrelated modules, which motivated its release as an independent distribution.

• IPC::Run

  Full-featured module for running processes with support for pipelines, pseudo-ttys, and timeouts. Very powerful but large (135KB+) with non-core dependencies (IO::Pty). Overkill for simple command execution.

• Capture::Tiny

  Excellent for capturing STDOUT/STDERR from Perl code or external commands. Does not provide stdin input functionality.

• IPC::Run3

  Simpler than IPC::Run, handles stdin/stdout/stderr. Good alternative but does not support code reference execution.

• Command::Runner

  Modern interface with timeout support and code reference execution. Has non-core dependencies.

• Proc::Simple

  Simple process management with background execution support. Focused on process control rather than I/O capture.

Command::Run differs from these modules in several ways:

• Core modules only - No non-core dependencies
• Code reference support - Execute Perl code with $0 and @ARGV setup
• File descriptor path - Output accessible via /dev/fd/N
• Minimal footprint - About 200 lines of code
• Method chaining - Fluent interface for readability

SEE ALSO

Command::Run::Tmpfile, IPC::Run, Capture::Tiny, IPC::Run3, Command::Runner

AUTHOR

Kazumasa Utashiro

LICENSE

Copyright 2026 Kazumasa Utashiro.

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.