#!/usr/bin/perl -w

=head1 NAME

truncate-eol-whitespace - truncate white spaces at end of line.

=head1 SYNOPSIS

  truncate-eol-whitespace [-total] [-truncated] [-nontruncated] [-dry-run] \
	[file ...] [-f files-from]

=head1 DESCRIPTION

This program truncates extra white spaces just before end of line
in specified files. File names can be specified as parameters
and/or readed from specified file, '-' for STDIN.

=head1 EXAMPLE

Truncate all text files under DIR:

 find-text-files DIR -total | truncate-eol-whitespace -total -f -

=cut

require 5.004;
use strict;
use integer;
use Getopt::Long;


sub usage {
	warn "\n".join(" ", @_)."\n" if @_;
	warn <<EOF;

Usage:
  truncate-eol-whitespace [-total] [-truncated] [-nontruncated] [-dry-run] \
	[file ...] [-f files-from]

Warning: this script truncates files! Use -dry-run for test first.

EOF
	exit(1);
}


=head1 OPTIONS

=over 4

=item -total

print statistic counters to STDERR.

=item -truncated

print to STDERR what files was truncated;

=item -nontruncated

print to STDERR what files was not truncated;

=item -dry-run

Do not write files, report only

=item file ...

Files to truncate (optional)

=item -f files-from

File name with file names to truncate, one name per line.
Use '-' for STDIN.

=back

=cut


my $        help_option = 0;
my $     dry_run_option = 0;
my $  files_from_option;
my $       total_option = 0;
my $   truncated_option = 0;
my $nontruncated_option = 0;

GetOptions(
	'help'		=> \$        help_option,
	'total'		=> \$       total_option,
	'truncated'	=> \$   truncated_option,
	'nontruncated'	=> \$nontruncated_option,
	'dry-run'	=> \$     dry_run_option,
	'f=s'		=> \$  files_from_option,
) or usage;

usage if $help_option;


usage("no files specified")
	if (! defined $files_from_option) and scalar(@ARGV) < 1;


my (
    $total_files_checked,
    $total_files_empty,
    $total_files_truncated,
    $total_files_no_chars_truncated
   ) = (0,0,0,0,0,0,0,0,0,0);

my ( $total_chars_readed, $total_chars_truncated ) = (0,0);


sub truncate_file($)
{
    my $fname = shift;

    $total_files_checked++;

    if ( ! -f $fname )
    {
	print STDERR "is not a plain file: ".$fname."\n";
	return;
    }

    if ( ! -s $fname )
    {
	print STDERR "zero size file: ".$fname."\n";
	$total_files_empty++;
	return;
    }


    local $/ = undef;	# no records, slurp mode

    local *IN;
    open   IN, "< $fname"
	or die "Can't open $fname: $!";

    my $file = <IN>;
    defined $file or die "Can't read $fname: $!";

    close IN;


    my $length_before = length $file;
    $total_chars_readed += $length_before;

    $file =~ s/[\000-\011\013-\040]+\n/\n/mg;

    my $length_after  = length $file;


    my $chars_truncated = $length_before - $length_after;

    die "size become greater after truncating: ".$fname
	if $chars_truncated < 0;


    if ( $chars_truncated > 0 )
    {
	$total_files_truncated++;
	$total_chars_truncated += $chars_truncated;
    }
    else
    {
	$total_files_no_chars_truncated++;
    }

    if    ( $chars_truncated >0 and $truncated_option )
    {
	printf(STDOUT  "%6u of %6u chars truncated from $fname\n",
		       $chars_truncated, $length_before);
    }
    elsif ( $chars_truncated==0 and $nontruncated_option )
    {
	printf(STDOUT       "%-16s chars truncated from $fname\n", 'no');
    }

    if ( ! $dry_run_option and $chars_truncated > 0 )
    {
	local *OUT;
	open   OUT, "> $fname" or die "Can't open $fname: $!";
	print  OUT $file or die "Can't write $fname: $!";
	close  OUT  or die "Error on closing $fname: $!";
    }
}



#+ main work

# do process file names from the @ARGV first
truncate_file($_) while defined ($_ = shift);

if (defined $files_from_option) # do process file names from file|STDIN
{
    local *IN;
    open  (IN, $files_from_option) or die "Can't open $files_from_option: $!";

    while ( my $fname = <IN> )
    {
	chomp $fname;
	next if length($fname) < 1; # skip empty lines

	truncate_file($fname);
    }
}

#- main work


format STDERR =

Total files: checked   empty   truncated  non-truncated
             -------  -------  -------   -------
             @>>>>>>  @>>>>>>  @>>>>>>   @>>>>>>
$total_files_checked, $total_files_empty, $total_files_truncated, $total_files_no_chars_truncated

Total chars truncated: @>>>>>> of @<<<<<<<<<<<<<<<<<
$total_chars_truncated, $total_chars_readed

.

write STDERR if $total_option;

exit 0;


=head1 AUTHOR

Dmitry Fedorov <dm.fedorov@gmail.com>

=head1 COPYRIGHT

Copyright (C) 2003 Dmitry Fedorov <dm.fedorov@gmail.com>

=head1 LICENSE

This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License,
or (at your option) any later version.

=head1 DISCLAIMER

The author disclaims any responsibility for any mangling of your system
etc, that this script may cause.

=cut