suggestion for rm

[terry jones]

Hi. It occurred to me last night (20 years after first typing rm) that
it would be useful in some (limited, I admit) situations to be able to
give rm a numeric count of the number of things you were passing it to
be removed. If rm did not find exactly (or found more) that number of
file arguments, it would print a warning message and exit without
removing anything.

For example, when using rm interactively from the shell, you might
want to do something like this:

  rm -1 *.old

believing that there is a single file matching *.old in your
directory.  From a shell script or Makefile (etc) where it is harder
to confirm the number of items matched by a regex, a numeric count
might be useful too. The count could either be exact or an upper
limit.

The advantage is pretty clear, even if slight. The main question is
whether it's worth the effort to put it in and whether rm should have
an option like this in the first place, etc. Of course, people should
be more careful and we shouldn't encourage them to be lazy. But,
people will be lazy no matter what (I even very occasionally type rm
*~ or put it in a Make rule, dreading the day when I accidentally hit
RETURN without the ~ or put a space between * and ~). For times when
you want to be lazy but you're not 100% sure that your regex is going
to match exactly one (or two, etc) things, a simple extra arg to rm
would help.

Regards,
Terry.