#!/usr/bin/perl

=begin metadata

Name: colrm
Description: remove columns from a file
Author: Jeffrey S. Haemer
License: perl

=end metadata

=cut


use strict;

$0 =~ s(.*/)();
my $usage = "usage: $0 [startcol endcol]\n";
my ($startcol, $endcol);

# I could be more clever, but this will run faster

if (@ARGV > 2) {
	die $usage;
} elsif (@ARGV == 0) {
	print while(<>);
} elsif (@ARGV == 1) {
	$startcol = (shift() - 1);
	while (<>) {
		print substr $_, 0, $startcol;
		print "\n";
	}
} elsif (@ARGV == 2) {
	$startcol = (shift() - 1);
	$endcol = (shift() - 1);
	while (<>) {
		chomp;
		substr($_, $startcol, $endcol) = '';
		print "$_\n";
	}
}

=head1 NAME

colrm - remove columns from a file

=head1 SYNOPSIS

colrm [startcol [endcol]]

=head1 DESCRIPTION

B<colrm> removes the named columns from each line of its standard input
(one column = one character).
Column numbering starts at 1, not 0.

If a only I<startcol> is provided, removes all columns from I<startcol>
rightwards.

If both I<startcol> and I<endcol> are provided, removes all columns from
I<startcol> to I<endcol>, inclusive.

If neither is provided, acts just like B<cat>.

If I<firstcol> is greater than I<lastcol>, or the arguments aren't numeric,
it pretends it heard no arguments at all, just like the Linux version.

=head1 OPTIONS AND ARGUMENTS

=over 2

=item I<startcol>

The first column to remove.

=item I<endcol>

The last column to remove.

=back

=head1 AUTHOR

  Jeffrey S. Haemer

=head1 BUGS

Lacks the special-case handling of backspace and tab added in some
other versions.  Acts, instead, like the simpler Linux and SunOS versions.

=head1 SEE ALSO

  awk(1)