Subject: Re: kern/15397: regress/sys/fs/mkdir fails
To: None <thorpej@wasabisystems.com>
From: Christos Zoulas <christos@zoulas.com>
List: netbsd-bugs
Date: 01/29/2002 12:07:24
On Jan 27,  3:33pm, thorpej@wasabisystems.com (Jason R Thorpe) wrote:
-- Subject: Re: kern/15397: regress/sys/fs/mkdir fails

| On Sun, Jan 27, 2002 at 11:17:08PM +0000, Christos Zoulas wrote:
| 
|  > >>How-To-Repeat:
|  > >	Run test in regress/sys/fs/mkdir.
|  > 
|  > So that I understand how far this is going to go...
|  > will open("foo/", O_CREAT, 0666); create a plain file? Or will
|  > stat("foo/", &st); work after that?
| 
| Read the regression test.  In particular:
| 
|         /*
|          * IEEE 1003.1 second ed. 2.2.2.78:
|          *
|          *      If the pathname refers to a directory, it may also have
|          *      one or more trailing slashes.  Multiple successive slashes
|          *      are considered to be the same as one slash.
|          */
| 
| Since the pathname argument to mkdir(2) and rmdir(2) refer to a directory,
| trailing /s must be allowed.

Yes, I know and the reason I am asking is that I recall that another
section -probably from an earlier revision- used to say that any
number of trailing slashes can be ignored in any pathname, that is
why we see today's lossage on solaris and other svr4 variants where
the above two examples work. Since this involves changes to namei(9)
I just want to make sure we are not going to go the same way if we
don't have to.

christos