[Gllug] spaces and shell scripting

[Tethys]

Jon Dye writes:

>I'm trying to write a shell (bash) script

Try and avoid writing bash scripts. Unless you really need a bash
specific feature, write a generic bourne shell script (or korn shell
if you need more than /bin/sh provides).

>but I want to avoid multiple find statements because they are slow.  I 
>guess I want either to put the list of files in a variable with 
>delimiters between the file names or put them into an array somehow.
>
>If any can tell me how to achieve what I want (or suggest a better way 
>to do it) I'd be grateful.

Put the output of the find command into a temporary file, and work
on it there. This will handle files containing whitespace and other
metacharacters, provided you don't have filenames containing newlines[1].

	tmpfile=/tmp/files.$$
	trap "/bin/rm -f $tmpfile" 0 HUP INT QUIT TERM

	find /some/directory -type f -print > "$tmpfile"

	c=$(wc -l /tmp/files | cut -d' ' -f1)

	case $c in
		0)  do_somthing ;;

		1)  do_something_else "$(cat $tmpfile)" ;;

		*)
			while read file
			do
				some_other_command "$file"
			done < "$tmpfile"
	esac

Tet

[1] Not a common occurrence... unless they're somewhere that's mounted
    by a Windows or Mac box, at which point it becomes more likely.
-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug