Path: utzoo!utgpu!utstat!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!decwrl!sun!pitstop!sundc!seismo!uunet!mcvax!ukc!etive!hwcs!zen!frank
From: frank@zen.co.uk (Frank Wales)
Newsgroups: comp.unix.questions
Subject: Re: Computational complexity of rm & ls
Message-ID: <1541@zen.UUCP>
Date: 12 Mar 89 18:59:43 GMT
References: <9000012@m.cs.uiuc.edu> <4461@pt.cs.cmu.edu>
Reply-To: frank@zen.co.uk (Frank Wales)
Organization: Zengrange Limited, Leeds, England
Lines: 18

In article <4461@pt.cs.cmu.edu> marcoz@MARCOZ.BOLTZ.CS.CMU.EDU (Marco Zagha) writes:
>In article <9000012@m.cs.uiuc.edu>, wsmith@m.cs.uiuc.edu writes:
>> [...]   "rm *" seems to be a common
>> enough idiom that rm could be optimized to handle that case better.
>
>Don't forget that the shell expands the "*", not rm.  All rm sees
>is a huge argv with one name for every file (except .* files).

This was my first thought upon reading what Bill posted, and then I thought:
"maybe he means the act of deleting all the files in a directory is a
common one, and should be handled specially."  Certainly, rm never sees
the '*', only the results of its (sorted, hidden-file-less) expansion.

Maybe adding a '-A' (for All, harder to type than '-a') option to rm would
be justified.  No doubt doing so would break jillions of scripts.  :-)
--
Frank Wales, Systems Manager,        [frank@zen.co.uk<->mcvax!zen.co.uk!frank]
Zengrange Ltd., Greenfield Rd., Leeds, ENGLAND, LS9 8DB. (+44) 532 489048 x217