From mboxrd@z Thu Jan 1 00:00:00 1970 From: Steve Dickson Subject: Re: [Patch 3/8] NFS Mount Configuration File (Vers 2) Date: Mon, 03 Aug 2009 12:05:10 -0400 Message-ID: <4A770AB6.8010806@RedHat.com> References: <4A76F8E6.2010607@RedHat.com> <4A76F9E9.5020604@RedHat.com> <20090803152409.GB28838@fieldses.org> <4A7706E8.7060106@RedHat.com> <20090803155658.GF28838@fieldses.org> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Cc: Linux NFS Mailing list , Linux NFSv4 mailing list To: "J. Bruce Fields" Return-path: In-Reply-To: <20090803155658.GF28838@fieldses.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: nfsv4-bounces@linux-nfs.org Errors-To: nfsv4-bounces@linux-nfs.org List-ID: On 08/03/2009 11:56 AM, J. Bruce Fields wrote: > On Mon, Aug 03, 2009 at 11:48:56AM -0400, Steve Dickson wrote: >> >> On 08/03/2009 11:24 AM, J. Bruce Fields wrote: >>> On Mon, Aug 03, 2009 at 10:53:29AM -0400, Steve Dickson wrote: >>>> commit 528e270be6bb3e4a072cbbb7f3a8b378b64b7fba >>>> Author: Steve Dickson >>>> Date: Mon Mar 9 13:57:10 2009 -0400 >>>> >>>> Store values in lower case. This makes it easier to do string >>>> comparisons when reading lists from the configuration file. >>> Just curious--which parts of the configuration file (section headers, >>> keys, values--looks like maybe all 3?) are you making case-insensitive, >>> and why do they need to be? >> Yes all three of them... >>> E.g. the idmapd configuration file contains paths and user/group names, >>> and both those still need to be case sensitive. >> hmm... good point... I would suspect this might be the case for >> all users of these routines since pathnames and user/group names >> are always case-sensitive.... >> >> I would guess both the section and keys can be case-insensitive >> but the key probably say case-sensitive... > > If you're using mount paths, for example, in the section names, then > that part also needs to be case-sensitive. Point... > > Probably the right thing to do is to leave the core config-file code out > of the business of playing with case, and let the callers do it in > particular cases when it's appropriate. I guess... but that leaves the disconnect between what the caller expects the sections/keys to be called and what is actually in the config file. If the only difference is case, I was thinking routines should be smart enough to figure that out... steved.