> -----Original Message-----
> From: Andrew G.McArthur [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, August 21, 2001 3:44 PM
> To: [EMAIL PROTECTED]
> Subject: Detecting if a pipe exists
>
>
> Dear all,
>
> I'm developing a perl program in UNIX that I want to
> sometimes work on
> data from the pipe, sometimes from the command line, and sometimes
> both. For example, consider the hypothetical program
> "squareroot". I
> would like to use it in the following three ways from the
> command prompt:
>
> > squareroot 100 1000 25
>
> > cat numbers.list | squareroot
>
> > cat numbers.list | squareroot 25 144
>
> I've had no problem with the second two examples, but the first waits
> for input from the STDIN. How can I get my perl program to know that
> I'm not piping in data from the STDIN?
I don't think this can really be done.
You might use -p to see if STDIN is a pipe, but that wouldn't work for
squareroot <numbers.list
You might use -t to see if STDIN is a tty, but that wouldn't work for
$ cat >sq.sh
squareroot 100
$ sh sq.sh # STDIN is a tty, but should I read input?
My recommendation would be to use a -f option to supply input from
a file (use Getopt::Std to parse the argument):
squareroot -f numbers.list 25 144
You can still use this in a pipeline by using '-' as the filename:
cat numbers.list | squareroot -f - 25 144
squareroot -f - 25 144 <numbers.list
perl treats a filename of '-' as STDIN. (see perldoc -f open)
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
