Date: Mon, 15 Oct 2001 21:42:04 +0200 From: "Eli Zaretskii" Sender: halo1 AT zahav DOT net DOT il To: Tim Van Holder Message-Id: <8296-Mon15Oct2001214204+0200-eliz@is.elta.co.il> X-Mailer: Emacs 20.6 (via feedmail 8.3.emacs20_6 I) and Blat ver 1.8.9 CC: djgpp-workers AT delorie DOT com, sandmann AT clio DOT rice DOT edu In-reply-to: <1003159030.28450.105.camel@bender.falconsoft.be> (message from Tim Van Holder on 15 Oct 2001 17:16:58 +0200) Subject: Re: W2K/XP fncase References: <1003159030 DOT 28450 DOT 105 DOT camel AT bender DOT falconsoft DOT be> Reply-To: djgpp-workers AT delorie DOT com Errors-To: nobody AT delorie DOT com X-Mailing-List: djgpp-workers AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk > From: Tim Van Holder > Date: 15 Oct 2001 17:16:58 +0200 > > > > Unfortunately, raw COFF files produced by the linker don't have > > extensions, as well as shell scripts. > > Right. Though if you have a file like 'autoconf' under WinME, you'll > probably have an LFN entry for it (all lowercase), while 'README' would > not have an LFN entry. So maybe we could skip downcasing for > extensionless files on LFN=y only? I _was_ talking about LFN=y. The problem is that a file could come from DOS or be created when LFN=n was in effect. Then you turn LFN=y back again, and all h*ll breaks lose... > > If we don't downcase COFF files, Make will always want to remake all > > targets like this: > > > > foo: foo.o > > $(CC) $(LDFLAGS) $< -o $@ $(LOADLIBES) > > True - but that's the same as the situation where gcc creates foo.exe > instead of foo. GCC only creates foo.exe without foo if you say "-o foo.exe". Makefiles which do that should say "foo.exe: foo.o", not "foo: foo.o". If they do this, there's no problem. > > Even worse, targets like below will always fail: > > > > all: foo > > Why? Because if we don't downcase `foo', Make might see `FOO' in the directory, not `foo'.