Oliver Schieche


Getopt::O2 - Command line argument processing and automated help generation, object oriented


  package MyPackage;
  use parent 'Getopt::O2';

  # return a short descriptive string about the program (appears in --help)
  sub get_program_description
          return 'A sample program';

  # return rules about parameters
  sub get_option_rules
          return shift->SUPER::get_option_rules(),
                  'length=i' => ['A numeric argument', 'default' => 33],
                  'file=s'   => ['A mandatory argument', 'required' => 1],
                  'quiet'    => ['A "flag" argument'];

  # read options
  new MyPackage->getopt(\my %options, \my @values);


The Getopt::O2 module implements an extended Getopt class which parses the command line from @ARGV, recognizing and removing specified options and their possible values.

This module adheres to the POSIX syntax for command line options, with GNU extensions. In general, this means that options have long names instead of single letters, and are introduced with a double dash "--". Support for bundling of command line options, as was the case with the more traditional single-letter approach, is provided.

Getopt::O2 stands out for its extensive usage generation feature; anything printed in its "usage" output is generated from the input options and saves the users the time to write usage output by themselves.



Processes command line options and stores their values in the hash reference passed as its argument. Anything not recognized as parameters or their values is pushed into the second (optional) ARRAYREF.


Returns a list of rules of command line options. The base package provides two options --help and --verbose by default. The former calls usage(); the latter is an incremental option. See "Writing Rules" for what your implementation should return.


Returns the program name for display in usage.


Returns a short descriptive string about the program's functionality. This string is used as a caption of the generated program usage text and should be implemented by sub-modules using this module.


Display program usage summary and exit with status CODE. Without any further arguments it will show the program's description text. If given, MESSAGE will be treated as an sprintf()-like formatter string followed by its arguments and prefixed with "Error: ".


This method is called internally when processing or validation of options failed and does nothing but passing its arguments to usage() (along with an exit code of 1). Override this method if you require other methods of error handling.

Writing Rules

Command line options are processed using rules returned the getOptionsRules() implementation. Rules are expressed much like with Getopt::Long. A rule expression is followed by the rule's help string and possible options.

The options must be represented as either a string (used as help string) or an ARRAYREF. The first element of the latter is used as the options' help string. Its second element can be a CODEREF which is called when the option was seen. The rest are key-value-pairs that are coerced to a hash.

A single undef can be used to separate option categories by producing an empty line in usage() output.

      # Short variant. Define flag and its help string
      'q|quiet' => 'Suppresses informational program output'
      # Actual implementation of "--help" parameter
      'h|help' => ['Display this help message', sub {
      # Enumeration with allowed values
      'o|output=?' => ['Use ARG as output format', 'values' => [qw(xml html json)]]
      # One or more occurences of a value (result is ARRAYREF)
      'i|input=s@' => 'Create result from input file ARG'
      # Use callback return value as option value
      'l|limit=i' => ['Limit amount of things', sub {
          my ($arg, $key) = @_;
          $arg = 100 if $arg > 100;
          return $arg; # make sure --limit is not larger than 100

Rule syntax


Defines a negatable option. The value of it will be a "boolean" in the resulting options hash reference depending on whether --warnings or --no-warnings was seen on the command line. There's no short negatable option.


Defines an incremental option. Depending on how often it's seen on the command line, the option's value will increase in the resulting hashref.


Defines a flag option. The flag will be set in the resulting hashref if this option was seen on the command line.


Defines an option with a mandatory value. The character after the = sign determines the expected value: s is a generic string, i is a numeric value (it uses Perl's "looks_like_number" in Scalar::Util) and ? is an enumeration. If the type specifier is suffixed with a @, the resulting value will be an ARRAYREF with all values.

Enumerations must provide a values option which must be an ARRAYREF of valid values for the option. They may use the keep_unique option which defaults to being set in order to control whether the resulting list contains unique values or all given values.

Contextual rules

Rules can be allowed in a given context and may change the context appropriately.

Consider the following ruleset:

      sub get_option_rules
              'q|quiet'     => ['Be quiet', 'context' => '-logging'],
              'v|verbose'   => ['Be verbose', 'context' => '+logging'],
              'l|logfile=s' => ['Log to file ARG', 'context' => 'logging']

The above example would introduce the logging context; an internal state which makes options appearing outside of that context invalid.

The --verbose flag would activate the context - allowing for the option --logfile, which would otherwise (without the context) be considered illegal.

Contexts can be comma separated. A context of -a,-b,+c,d would:

  • deactivate both contexts a and b

  • activate context c

  • restrict the option to the previously activated context d.



None special. Uses core perl libraries.


Oliver Schieche <schiecheo@cpan.org>



Copyright 2013-2019 Oliver Schieche.

This software is a free library. You can modify and/or distribute it under the same terms as Perl itself.