recursive chmod

[Paul Melson]

On Mon, Nov 27, 2000 at 05:21:27PM -0500, Edward Glowacki wrote:
> I'm not doubting you on this, but could you explain why it's safer?
> Just curious...

	xargs gives you bounds checking for creating
	the command(s) you wish to execute.  In Ben's
	example, find would put a null character at
	the end of each file & path that are output
	and then require xargs to use that null as
	a delimiter of sorts.  It might not be such
	a big deal when using chmod (although I can
	 conceive of some situation where it could),
	but take this example:

find / -type f -name core ! -empty -print0 | xargs -0 rm -f

	Now, by using '-type f' we eliminate using
	symlinks to delete system files or whatever,
	but you could still create a file with a
	large number of spaces between 'core' and,
	oh, say '/vmlinuz' in its name.  Without
	using '-print0' and '-0', to endicate the
	end of the file's name to xargs, it would
	be trivial to get a clean-up script like
	this to delete system files when it runs.

	You could also use find's '-printf' switch
	to submit another type of "delimiter" into
	its output, and this can be paired with xargs'
	'-e' switch, but since xargs can use nulls
	with the '-0' switch, it's an easy way to 
	get the job done.


PaulM

-- 
							_____________________
							melson@holt.k12.mi.us