From:	Martin Str|mberg <ams AT ludd DOT luth DOT se>
Message-Id:	<199904221710.TAA11748@father.ludd.luth.se>
Subject:	Re: [binkley AT sst DOT ncsl DOT nist DOT gov: djgpp and ld]
To:	eliz AT is DOT elta DOT co DOT il (Eli Zaretskii)
Date:	Thu, 22 Apr 1999 19:10:52 +0200 (MET DST)
Cc:	djgpp-workers AT delorie DOT com (DJGPP-WORKERS)
In-Reply-To:	<Pine.SUN.3.91.990422112410.15748F-100000@is> from Eli Zaretskii at "Apr 22, 99 11:24:26 am"
X-Mailer:	ELM [version 2.4ME+ PL15 (25)]
MIME-Version:	1.0
Reply-To:	djgpp-workers AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

According to Eli Zaretskii:
> 
> On Tue, 20 Apr 1999, Martin Str|mberg wrote:
> 
> > > My only purpose in this suggestion is to ensure that the hardcoded
> > > paths fail to accidentally pick up some non-djgpp package.
> > 
> > If that is the purpose, I suggest we hardcode them to sometihng that
> > will always _fail_.
> 
> Sorry, I disagree.  Although every path name we can come up with
> could, in theory, be used by some other package, in practice neither
> ${DJDIR} nor "c:/djgpp" has ever been reported as a problem.  The pain
> of going through all the Makefile's and fixing an invalid prefix is
> too much to ask.  In contrast, today people simply need to type "make"
> and sit back.
> 
> Why should we punish 99.99% of users to satisfy 0.01%?

I'm obviuosly missing something here. I thought the hard coded default
path was never used. My idea was to make sure that the path would
fail, would mean everybody had the same environment (i. e. just
because somebody happened to have C:/USR/LOCAL, this or that
happened), hence easier to diagnose problems.

So suppose I had C:/DJGPP which is a previous version of DJGPP v1,
D:/DJGPP which is DJGPP v2.0, D:/DJGPP.GOD which is the known latest
stable version and E:/DJGPP which is experimental DJGPP version. What
would happen? (Actually, I'm not sure what should happen myself.)


Right,

							MartinS

- Raw text -