Perl 5 version 22.4 documentation

Pod::Usage

NAME

Pod::Usage - print a usage message from embedded pod documentation

SYNOPSIS

  1.   use Pod::Usage
  3.   my $message_text  = "This text precedes the usage message.";
  4.   my $exit_status   = 2;          ## The exit status to use
  5.   my $verbose_level = 0;          ## The verbose level to use
  6.   my $filehandle    = \*STDERR;   ## The filehandle to write to
  8.   pod2usage($message_text);
  10.   pod2usage($exit_status);
  12.   pod2usage( { -message => $message_text ,
  13.                -exitval => $exit_status  ,  
  14.                -verbose => $verbose_level,  
  15.                -output  => $filehandle } );
  17.   pod2usage(   -msg     => $message_text ,
  18.                -exitval => $exit_status  ,  
  19.                -verbose => $verbose_level,  
  20.                -output  => $filehandle   );
  22.   pod2usage(   -verbose => 2,
  23.                -noperldoc => 1  )

ARGUMENTS

pod2usage should be given either a single argument, or a list of arguments corresponding to an associative array (a "hash"). When a single argument is given, it should correspond to exactly one of the following:

  • A string containing the text of a message to print before printing the usage message

  • A numeric value corresponding to the desired exit status

  • A reference to a hash

If more than one argument is given then the entire argument list is assumed to be a hash. If a hash is supplied (either as a reference or as a list) it should contain one or more elements with the following keys:

  • -message
  • -msg

    The text of a message to print immediately prior to printing the program's usage message.

  • -exitval

    The desired exit status to pass to the exit() function. This should be an integer, or else the string "NOEXIT" to indicate that control should simply be returned without terminating the invoking process.

  • -verbose

    The desired level of "verboseness" to use when printing the usage message. If the corresponding value is 0, then only the "SYNOPSIS" section of the pod documentation is printed. If the corresponding value is 1, then the "SYNOPSIS" section, along with any section entitled "OPTIONS", "ARGUMENTS", or "OPTIONS AND ARGUMENTS" is printed. If the corresponding value is 2 or more then the entire manpage is printed.

    The special verbosity level 99 requires to also specify the -sections parameter; then these sections are extracted and printed.

  • -sections

    A string representing a selection list for sections to be printed when -verbose is set to 99, e.g. "NAME|SYNOPSIS|DESCRIPTION|VERSION" .

    Alternatively, an array reference of section specifications can be used:

    1.   pod2usage(-verbose => 99, 
    2.             -sections => [ qw(fred fred/subsection) ] );
  • -output

    A reference to a filehandle, or the pathname of a file to which the usage message should be written. The default is \*STDERR unless the exit value is less than 2 (in which case the default is \*STDOUT ).

  • -input

    A reference to a filehandle, or the pathname of a file from which the invoking script's pod documentation should be read. It defaults to the file indicated by $0 ($PROGRAM_NAME for users of English.pm).

    If you are calling pod2usage() from a module and want to display that module's POD, you can use this:

    1.   use Pod::Find qw(pod_where);
    2.   pod2usage( -input => pod_where({-inc => 1}, __PACKAGE__) );
  • -pathlist

    A list of directory paths. If the input file does not exist, then it will be searched for in the given directory list (in the order the directories appear in the list). It defaults to the list of directories implied by $ENV{PATH} . The list may be specified either by a reference to an array, or by a string of directory paths which use the same path separator as $ENV{PATH} on your system (e.g., : for Unix, ; for MSWin32 and DOS).

  • -noperldoc

    By default, Pod::Usage will call perldoc when -verbose >= 2 is specified. This does not work well e.g. if the script was packed with PAR. The -noperldoc option suppresses the external call to perldoc and uses the simple text formatter (Pod::Text) to output the POD.

Formatting base class

The default text formatter is Pod::Text. The base class for Pod::Usage can be defined by pre-setting $Pod::Usage::Formatter before loading Pod::Usage, e.g.:

  1.     BEGIN { $Pod::Usage::Formatter = 'Pod::Text::Termcap'; }
  2.     use Pod::Usage qw(pod2usage);

Pass-through options

The following options are passed through to the underlying text formatter. See the manual pages of these modules for more information.

  1.   alt code indent loose margin quotes sentence stderr utf8 width

DESCRIPTION

pod2usage will print a usage message for the invoking script (using its embedded pod documentation) and then exit the script with the desired exit status. The usage message printed may have any one of three levels of "verboseness": If the verbose level is 0, then only a synopsis is printed. If the verbose level is 1, then the synopsis is printed along with a description (if present) of the command line options and arguments. If the verbose level is 2, then the entire manual page is printed.

Unless they are explicitly specified, the default values for the exit status, verbose level, and output stream to use are determined as follows:

  • If neither the exit status nor the verbose level is specified, then the default is to use an exit status of 2 with a verbose level of 0.

  • If an exit status is specified but the verbose level is not, then the verbose level will default to 1 if the exit status is less than 2 and will default to 0 otherwise.

  • If an exit status is not specified but verbose level is given, then the exit status will default to 2 if the verbose level is 0 and will default to 1 otherwise.

  • If the exit status used is less than 2, then output is printed on STDOUT . Otherwise output is printed on STDERR .

Although the above may seem a bit confusing at first, it generally does "the right thing" in most situations. This determination of the default values to use is based upon the following typical Unix conventions:

  • An exit status of 0 implies "success". For example, diff(1) exits with a status of 0 if the two files have the same contents.

  • An exit status of 1 implies possibly abnormal, but non-defective, program termination. For example, grep(1) exits with a status of 1 if it did not find a matching line for the given regular expression.

  • An exit status of 2 or more implies a fatal error. For example, ls(1) exits with a status of 2 if you specify an illegal (unknown) option on the command line.

  • Usage messages issued as a result of bad command-line syntax should go to STDERR . However, usage messages issued due to an explicit request to print usage (like specifying -help on the command line) should go to STDOUT , just in case the user wants to pipe the output to a pager (such as more(1)).

  • If program usage has been explicitly requested by the user, it is often desirable to exit with a status of 1 (as opposed to 0) after issuing the user-requested usage message. It is also desirable to give a more verbose description of program usage in this case.

pod2usage doesn't force the above conventions upon you, but it will use them by default if you don't expressly tell it to do otherwise. The ability of pod2usage() to accept a single number or a string makes it convenient to use as an innocent looking error message handling function:

  1.     use strict;
  2.     use Pod::Usage;
  3.     use Getopt::Long;
  5.     ## Parse options
  6.     my %opt;
  7.     GetOptions(\%opt, "help|?", "man", "flag1")  ||  pod2usage(2);
  8.     pod2usage(1)  if ($opt{help});
  9.     pod2usage(-exitval => 0, -verbose => 2)  if ($opt{man});
  11.     ## Check for too many filenames
  12.     pod2usage("$0: Too many files given.\n")  if (@ARGV > 1);

Some user's however may feel that the above "economy of expression" is not particularly readable nor consistent and may instead choose to do something more like the following:

  1.     use strict;
  2.     use Pod::Usage qw(pod2usage);
  3.     use Getopt::Long qw(GetOptions);
  5.     ## Parse options
  6.     my %opt;
  7.     GetOptions(\%opt, "help|?", "man", "flag1")  ||
  8.       pod2usage(-verbose => 0);
  10.     pod2usage(-verbose => 1)  if ($opt{help});
  11.     pod2usage(-verbose => 2)  if ($opt{man});
  13.     ## Check for too many filenames
  14.     pod2usage(-verbose => 2, -message => "$0: Too many files given.\n")
  15.       if (@ARGV > 1);

As with all things in Perl, there's more than one way to do it, and pod2usage() adheres to this philosophy. If you are interested in seeing a number of different ways to invoke pod2usage (although by no means exhaustive), please refer to EXAMPLES.

Scripts

The Pod::Usage distribution comes with a script pod2usage which offers a command line interface to the functionality of Pod::Usage. See pod2usage.

EXAMPLES

Each of the following invocations of pod2usage() will print just the "SYNOPSIS" section to STDERR and will exit with a status of 2:

  1.     pod2usage();
  3.     pod2usage(2);
  5.     pod2usage(-verbose => 0);
  7.     pod2usage(-exitval => 2);
  9.     pod2usage({-exitval => 2, -output => \*STDERR});
  11.     pod2usage({-verbose => 0, -output  => \*STDERR});
  13.     pod2usage(-exitval => 2, -verbose => 0);
  15.     pod2usage(-exitval => 2, -verbose => 0, -output => \*STDERR);

Each of the following invocations of pod2usage() will print a message of "Syntax error." (followed by a newline) to STDERR , immediately followed by just the "SYNOPSIS" section (also printed to STDERR ) and will exit with a status of 2:

  1.     pod2usage("Syntax error.");
  3.     pod2usage(-message => "Syntax error.", -verbose => 0);
  5.     pod2usage(-msg  => "Syntax error.", -exitval => 2);
  7.     pod2usage({-msg => "Syntax error.", -exitval => 2, -output => \*STDERR});
  9.     pod2usage({-msg => "Syntax error.", -verbose => 0, -output => \*STDERR});
  11.     pod2usage(-msg  => "Syntax error.", -exitval => 2, -verbose => 0);
  13.     pod2usage(-message => "Syntax error.",
  14.               -exitval => 2,
  15.               -verbose => 0,
  16.               -output  => \*STDERR);

Each of the following invocations of pod2usage() will print the "SYNOPSIS" section and any "OPTIONS" and/or "ARGUMENTS" sections to STDOUT and will exit with a status of 1:

  1.     pod2usage(1);
  3.     pod2usage(-verbose => 1);
  5.     pod2usage(-exitval => 1);
  7.     pod2usage({-exitval => 1, -output => \*STDOUT});
  9.     pod2usage({-verbose => 1, -output => \*STDOUT});
  11.     pod2usage(-exitval => 1, -verbose => 1);
  13.     pod2usage(-exitval => 1, -verbose => 1, -output => \*STDOUT});

Each of the following invocations of pod2usage() will print the entire manual page to STDOUT and will exit with a status of 1:

  1.     pod2usage(-verbose  => 2);
  3.     pod2usage({-verbose => 2, -output => \*STDOUT});
  5.     pod2usage(-exitval  => 1, -verbose => 2);
  7.     pod2usage({-exitval => 1, -verbose => 2, -output => \*STDOUT});

Recommended Use

Most scripts should print some type of usage message to STDERR when a command line syntax error is detected. They should also provide an option (usually -H or -help ) to print a (possibly more verbose) usage message to STDOUT . Some scripts may even wish to go so far as to provide a means of printing their complete documentation to STDOUT (perhaps by allowing a -man option). The following complete example uses Pod::Usage in combination with Getopt::Long to do all of these things:

  1.     use strict;
  2.     use Getopt::Long qw(GetOptions);
  3.     use Pod::Usage qw(pod2usage);
  5.     my $man = 0;
  6.     my $help = 0;
  7.     ## Parse options and print usage if there is a syntax error,
  8.     ## or if usage was explicitly requested.
  9.     GetOptions('help|?' => \$help, man => \$man) or pod2usage(2);
  10.     pod2usage(1) if $help;
  11.     pod2usage(-verbose => 2) if $man;
  13.     ## If no arguments were given, then allow STDIN to be used only
  14.     ## if it's not connected to a terminal (otherwise print usage)
  15.     pod2usage("$0: No files given.")  if ((@ARGV == 0) && (-t STDIN));
  17.     __END__
  19.     =head1 NAME
  21.     sample - Using GetOpt::Long and Pod::Usage
  23.     =head1 SYNOPSIS
  25.     sample [options] [file ...]
  27.      Options:
  28.        -help            brief help message
  29.        -man             full documentation
  31.     =head1 OPTIONS
  33.     =over 4
  35.     =item B<-help>
  37.     Print a brief help message and exits.
  39.     =item B<-man>
  41.     Prints the manual page and exits.
  43.     =back
  45.     =head1 DESCRIPTION
  47.     B<This program> will read the given input file(s) and do something
  48.     useful with the contents thereof.
  50.     =cut

CAVEATS

By default, pod2usage() will use $0 as the path to the pod input file. Unfortunately, not all systems on which Perl runs will set $0 properly (although if $0 isn't found, pod2usage() will search $ENV{PATH} or else the list specified by the -pathlist option). If this is the case for your system, you may need to explicitly specify the path to the pod docs for the invoking script using something similar to the following:

  1.     pod2usage(-exitval => 2, -input => "/path/to/your/pod/docs");

In the pathological case that a script is called via a relative path and the script itself changes the current working directory (see chdir) before calling pod2usage, Pod::Usage will fail even on robust platforms. Don't do that. Or use FindBin to locate the script:

  1.     use FindBin;
  2.     pod2usage(-input => $FindBin::Bin . "/" . $FindBin::Script);

AUTHOR

Please report bugs using http://rt.cpan.org.

Marek Rouchal <marekr@cpan.org>

Brad Appleton <bradapp@enteract.com>

Based on code for Pod::Text::pod2text() written by Tom Christiansen <tchrist@mox.perl.com>

ACKNOWLEDGMENTS

rjbs for refactoring Pod::Usage to not use Pod::Parser any more.

Steven McDougall <swmcd@world.std.com> for his help and patience with re-writing this manpage.

SEE ALSO

Pod::Usage is now a standalone distribution, depending on Pod::Text which in turn depends on Pod::Simple.

Pod::Perldoc, Getopt::Long, Pod::Find, FindBin, Pod::Text, Pod::Text::Termcap, Pod::Simple