From: Bennett E. Todd (bet@cyclone.sbi.com)
Date: 06/29/92 

From: bet@cyclone.sbi.com (Bennett E. Todd)
Subject: Re: intalling > 1.2M tar files
Date: 29 Jun 1992 21:42:25 GMT

In article <2401@nlsun1.oracle.nl> jhelberg@nl.oracle.com (Joost Helberg) writes:
>Use split -b 1474560 for 1.44 MB floppies and
> split -b 1228800 for 1.2 MB floppies
>
>
>So: cat <file> | split -b <amount>;
> for i in x??
> do
> dd if=$i of=/dev/PS0
> read dummy
> done
>
>To read again, dd all floppies to flop1 -- flopn and type:
>cat flop? | tar xvfz -

I have been using much the above procedure to bust up big archives on my Sun
and write 'em out, but I have got an alternative to the proposed mechanism
for reading 'em back in, that saves the extra copy. I use the rdflops script
from tsx-11.mit.edu:pub/linux/sources/usr.bin/floptools.shar.Z. It looks
like this:

        #!/bin/sh

        progname=`basename $0`
        syntax="syntax: $progname device"
        die(){
                echo "$progname: $*" 1>&2
                exit 1
        }

        case $# in
                1) device="$1" ;;
                *) die "$syntax" ;;
        esac

        while cat "$device"; do
                echo -n 'Insert another floppy: ' >/dev/tty
                read ans </dev/tty >/dev/null
        done

So instead of reading all the floppies onto the hard disk then using cat(1)
to concatenate them, I'd just go something like:

        rdflops /dev/fd0 | compress -d | tar xf -

-Bennett
bet@sbi.com