Date: Thu, 30 Jan 2003 22:19:30 +0200
From: "Eli Zaretskii" <eliz AT is DOT elta DOT co DOT il>
Sender: halo1 AT zahav DOT net DOT il
To: djgpp AT delorie DOT com
Message-Id: <2427-Thu30Jan2003221929+0200-eliz@is.elta.co.il>
X-Mailer: emacs 21.3.50 (via feedmail 8 I) and Blat ver 1.8.9
In-reply-to: <3e394a9a.sandmann@clio.rice.edu> (message from Charles Sandmann
	on Thu, 30 Jan 2003 09:54:02 CST)
Subject: Re: Command line wildcard expansion under Win2K
References: <Pine DOT SUN DOT 3 DOT 91 DOT 1030130095126 DOT 2797A-100000 AT is> <3e394a9a DOT sandmann AT clio DOT rice DOT edu>
Reply-To: djgpp AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> From: Charles Sandmann <sandmann AT clio DOT rice DOT edu>
> Newsgroups: comp.os.msdos.djgpp
> Date: Thu, 30 Jan 2003 09:54:02 CST
> 
> "A.vec" exists (A upper case):
> 
> C:\>djecho a*.vec A*.vec A*.VEC a*.VEC
> a.vec A.vec A*.VEC a*.VEC                    (current behavior)
> a.vec A.vec A.vec a.vec                      (intuitive behavior)

On a Windows 98 system, I get:

 A.vec A.vec A*.VEC a*.VEC

That is, "A.vec", the real file's name, matches itself and the
all-lowercase "a*.vec", but not the others.  This is
``case-sensitivity on LFN system except for all-lowercase patterns''
paradigm in action.

> In other words, act as if you had passed these strings to the 
> "dir" command.
> 
> "B.VEC" exists (all upper case):
> 
> C:\>djecho b*.vec B*.vec b*.VEC B*.VEC
> b.VEC B*.vec b*.VEC B.VEC                    (current behavior)
> b.vec b.vec b.vec B.VEC                      (intuitive behavior)

On my Windows 98, I get:

  b.vec B*.vec B.VEC b*.VEC

What version of djecho did you use?  Was it stock DJGPP v2.03 or
something newer?

> In other words, I think forcing an exact match on any upper case
> characters is not intuitive - that all matches should be case
> insensitive - at least as an option.

I support such a behavior as an option, but I think that having that
as the default is not a good idea.  Think about interactive programs
such as Bash and Emacs, which support file-name completion: there, if
you type "Make TAB", you don't want to see a file "makefile" in the
list of possible completions.

> I think this behavior should be decoupled from FNCASE.

Why?  What function will be left with FNCASE, then?