#!/usr/local/bin/perl -w

eval 'exec /usr/local/bin/perl -w -S $0 ${1+"$@"}'
    if 0; # not running under some shell
# $Id: asa,v 1.1.1.1 2001/06/06 08:54:00 sdague Exp $

use strict;

exit 1 if grep {!-r} @ARGV;		# traditional

if (grep /-/, @ARGV) {
	$0 =~ s(.*/)();
	warn "usage: $0 [filename ...]\n";
	exit 2;						# traditional
}

while (<>) {
	chomp;
	s/^$/ /;
	s/^[^10+-]/\n/;
	s/^1/\f/;
	s/^\+/\r/;
	s/^0/\n\n/;
	s/^-/\n\n\n/;
	print
		or exit 1;			# traditional
}

=head1 NAME

asa - interpret ASA/FORTRAN carriage-controls

=head1 SYNOPSIS

asa [I<filename> ...]

=head1 DESCRIPTION

=over 2

Traditional FORTRAN programs put carriage-control characters
in the first columns of their output,
which were interpreted by older lineprinters
according to the ASA vertical format control standard.
(ASA was the American Standards Association -- now ANSI.)

Under this standard, the first character of each printable record (line)
determines vertical spacing, as follows:

=over 2

I<blank>    carriage return
0           two carriage returns
1           Formfeed
+           overprint
-           three carriage returns (IBM extension)

=back

All other characters are discarded, and empty lines behave as though
they have a leading blank.

B<asa> interprets these characters.

=back

=head1 EXIT VALUES

=over 2

0 normal exit

1 inability to write on stdout or to read an input file

2 bad argument

Exit status values chosen from MKS toolkit.

=back

=head1 AUTHOR

Jeffrey S. Haemer

=head1 BUGS

Currently, B<asa> just looks at the readability of its input files
at startup time.  It should really do it a file at a time,
but that makes the code look gross.

The carriage-control '-' is an IBM extension.
Perhaps the default should ignore it
and there should be a '-i' option to interpret it.

=head1 SEE ALSO

Communications of the ACM, Vol 7, No. 10, p. 606, October
1964.

NWG/RFC 189, Appendix C

=cut