Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!sundc!pitstop!sun!amdcad!ames!ptsfa!ihnp4!homxb!mtuxo!mtune!codas!killer!jfh
From: jfh@killer.UUCP (The Beach Bum)
Newsgroups: comp.unix.wizards
Subject: Re: To . or not to .
Message-ID: <1973@killer.UUCP>
Date: Mon, 2-Nov-87 17:32:50 EST
Article-I.D.: killer.1973
Posted: Mon Nov  2 17:32:50 1987
Date-Received: Sat, 7-Nov-87 04:21:49 EST
References: <648@tut.cis.ohio-state.edu> <722@wacsvax.OZ> <9171@mimsy.UUCP>
Organization: Big "D" Home for Wayward Hackers
Lines: 64
Summary: Don't about about x/ using the zero length bug. Works just fine on 3B2 at killer.

In article <9171@mimsy.UUCP>, chris@mimsy.UUCP (Chris Torek) writes:
> In article <722@wacsvax.OZ> frank@wacsvax.OZ (Frank O'Connor) writes:
> >Please flame me if I am wrong but why wouldn't :
> >	for p in $x/*/
> >	do
> >		stuff
> >	done
> >be sufficient. ...
> 
> Well, for one thing, this uses the zero-length-string file name
> (right there after the second slash: it refers to the current
> directory as of that slash).
> 
> If System V's namei is consistent, this will fail in the same way
> as open("", 0) (and it will not accept names like a//b or //x).
> -- 
> In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7690)
> Domain:	chris@mimsy.umd.edu	Path:	uunet!mimsy!chris

Sorry to do this to you Chris, but that works just fine here.

I don't have sys/nami.c (or whatever the new name is) in front of me,
but my understanding of what is going on is this:

namei (func, mode)
char	(*func) ();
int	mode;
{
	int c;

	c = (func) ();

	/* test for c == 0 to set u.u_error = ENOENT, or make current
	   inode == root inode or current directory inode. */

	do
		if room left insert character in u-page dirent structure.
	while ((c = (func) ()) != '/' && c != 0)

	while ((c = (func) ()) == '/' && c != 0)
		toss character out because it was just a /.

	/* at this point, c is '\0' or first char in next path name. */

	/* go look in current directory for an entry that matches. */

	/* if c == 0, return the current i-node, else go back. */
}

Okay.  Did anyone not know this already?  Now, my understanding of the
original bug was if the first `c' was '\0', the current I-node, which
was the current directory, since c != '/'.  That's why "" == . in the
old world.  Consider "/", where the first character returned is '/',
and after that, the next is '\0'.  The test, if I understand, and unless
you have the sources, (which I don't, so I don't know ...), only happens
that one time.  A much better test is if u.u_dent.d_name[0] == 0, and
return ENOENT.  But '/' is still quite special ...

- John.
-- 
John F. Haugh II		HECI Exploration Co. Inc.
UUCP:	...!ihnp4!killer!jfh	11910 Greenville Ave, Suite 600
"Don't Have an Oil Well?"	Dallas, TX. 75243
" ... Then Buy One!"		(214) 231-0993