From new-httpd-owner-new-httpd-archive=hyperreal.org@apache.org Fri Oct 02 17:06:20 1998
Return-Path: <new-httpd-owner-new-httpd-archive=hyperreal.org@apache.org>
Delivered-To: new-httpd-archive@hyperreal.org
Received: (qmail 13584 invoked by uid 6000); 2 Oct 1998 17:06:19 -0000
Received: (qmail 13571 invoked from network); 2 Oct 1998 17:06:16 -0000
Received: from redfish.go2net.com (207.178.55.5)
  by taz.hyperreal.org with SMTP; 2 Oct 1998 17:06:16 -0000
Received: from marcs by redfish.go2net.com with smtp (Exim 1.82 #2)
	id 0zP8dN-0002TL-00; Fri, 2 Oct 1998 10:04:49 -0700
Date: Fri, 2 Oct 1998 10:04:49 -0700 (PDT)
From: Marc Slemko <marcs@znep.com>
X-Sender: marcs@redfish
To: new-httpd@apache.org
Subject: Re: Win32: Why strlwr os_canonical_filename?
In-Reply-To: <3615062C.19DCBE71@algroup.co.uk>
Message-ID: <Pine.GSO.4.02A.9810021002420.9400-100000@redfish>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: new-httpd-owner@apache.org
Precedence: bulk
Reply-To: new-httpd@apache.org
Status: O

On Fri, 2 Oct 1998, Ben Laurie wrote:

> Marc Slemko wrote:
> > 
> > On Wed, 30 Sep 1998, Ken Parzygnat wrote:
> > 
> > > I have a question regarding the following code
> > > segment from sub_canonical_filename:
> > >
> > > if (h == INVALID_HANDLE_VALUE) {
> > >     ap_assert(strlen(szCanon) + strlen(szFilePart) + nSlashes < nCanon);
> > >     for (n = 0; n < nSlashes; ++n) {
> > >         strcat(szCanon, "/");
> > >     }
> > >     strcat(szCanon, szFilePart);
> > >     return FALSE;
> > > }
> > > else {
> > >     ap_assert(strlen(szCanon)+strlen(d.cFileName) < nCanon);
> > >     strlwr(d.cFileName);
> > >     strcat(szCanon, d.cFileName);
> > >     return TRUE;
> > > }
> > >
> > > Why is strlwr being called in the else leg?  Can we remove
> > > this call?
> > 
> > Because we need some way to be able to try to compare filenames, for
> > numerous purposes such as access control.
> > 
> > Because, due to the greatness of Win32, there are around 9543 different
> > names for any file, and differing case is one of those ways, we have to
> > lowercase things.
> > 
> > This should, however, only be done on the actual filename part, nothing
> > else.  I seem to remember this being fixed three or four times.
> 
> It is only being done on the actual filename part.

Except that PATH_TRANSLATED is a filename but it isn't a filename.

See PR#mumble, which likely prompted this question.