Hi, Thanks for the tip about
> The book Mastering Perl/Tk has a pretty good Chapter 3 on Geometry Management. > You can probably read it for free with an sign up at http://safari.oreilly.com/ about the piping question, it's not an either/or situation. It could be both. A good example is the unix wc command. First create a small file echo "hello" > hello.txt # 6 chars "hello" + "\n" then do: # piped input but no command line options (so uses "default" options) cat hello.txt | wc 1 1 6 # both piped input and command line options cat hello.txt | wc -c 6 # only command line options wc -c hello.txt 6 I have tried similar to your suggestion (many different combinations): > This will test for a pipe > #!/usr/bin/perl > # -t tests if a tty is input, else it's a pipe > > if (@ARGV == 0 and -t) { > die "Usage: $0 INPUT\n"; > } > > while (<>) { > ## process input file(s) here > } > __END__ The problem(s) with using -t and <> (as above) are: if you give only command line line options, the program will hang waiting for input. If I change it to if-else type decision making, -t causes the piped input to be ignored. As I said, I hacked together a subroutine from examples in the Perl Cookbook, but it's not very elegant: sub piper { my @_ARGV = @ARGV; @ARGV = (); # save command line options (if any) my @input = (); $SIG{ALRM} = sub { die "timeout" }; my $pid; if ($pid = open(CHILD, "-|")) { # parent code while(<CHILD>) { # gets piped input from CHILD (if any) push @input, $_; } close CHILD; alarm(0); } else { # child code die "cannot fork: $!" unless defined $pid; eval { alarm(1); while(<>) { print; } # piped input goes to parent (if any) alarm(0); }; exit; } @ARGV = @_ARGV; # restore so we can parse any options return @input; # this is any data which was piped in } any comments/suggestions about this solution are greatly appreciated. As I said, surely there is a module out there that does this type of thing better? regards, John -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] <http://learn.perl.org/> <http://learn.perl.org/first-response>
