Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!uunet!brunix!no!jsb
From: jsb@no.cs.brown.edu (John Bazik)
Newsgroups: comp.lang.perl
Subject: Re: pwd in perl
Message-ID: <46905@brunix.UUCP>
Date: 9 Aug 90 14:59:43 GMT
References: <1141@etnibsd.UUCP>
Sender: news@brunix.UUCP
Reply-To: jsb@no.cs.brown.edu (John Bazik)
Organization: Brown Computer Science Dept.
Lines: 18


In article <1141@etnibsd.UUCP>, vsh@etnibsd.UUCP (Steve Harris) writes:
|> Here is a simple-minded implementation of pwd in perl.
[stuff deleted]

Yeah, I did that, too.  It was significantly faster than `pwd`, but,
unfortunately, breaks under SunOS 4.1.  The latter has "virtual
filesystems" (or whatever they call it), which reside in memory rather
than on disk.  By default /tmp is one.  For some braindead reason,
the inode number of /tmp matches that of /.

Actually, I suppose
	if ($dinode == $pinode && $dname eq $pname) {

would do it...hmmm.  Anyway, I guess the moral is that any algorithm
has to anticipate weirdness but a getwd built-in would just always work.

John