Subject: Re: Case-insensitive filesystems
To: Bill Studenmund <>
From: Frederick Bruckman <>
List: tech-pkg
Date: 02/04/2002 20:27:38
On Mon, 4 Feb 2002, Bill Studenmund wrote:
> On Mon, 4 Feb 2002, Frederick Bruckman wrote:
> > On Mon, 4 Feb 2002, Bill Studenmund wrote:
> >
> > > I realize there are going to be other problems with packages that
> > > internally won't work because of case problems, but those packages won't
> > > work on MacOS X anyway. The authors will have to change things to make
> > > them work. So while I think we could help this, we don't have to do it all
> > > ourselves. :-)
> >
> > There was a long thread on this on comp.unix.programmer. The most
> > compelling reason I've heard not to bother to try to support
> > case-insensitivity, is that the make rules get borked (.C vs. .c, .S
> > vs. .s). <URI:news:r6%58.60$7J2.221549@burlma1-snr2>.
> That's not really a problem with hfs. While hfs ignores case for lookup,
> it preserves case for creation. So if you create foo.C, it will always be
> foo.C. It will be found if you look for [fF][oO][oO].[cC], but it will
> always be foo.C.

The poster's in the URI above theorize that it would be a problem for
make's implicit rules. Has anyone tried it? E.g:

fredb@tautology-> touch foo.C
fredb@tautology-> make -n foo
c++ -O2   -o foo foo.C

That's on NetBSD (ffs). What happens when you do the same with "bmake"
on HFS?