From: Martin Str|mberg <ams AT ludd DOT luth DOT se>
Message-Id: <199807081623.SAA23335@father.ludd.luth.se>
Subject: Re: compiling syslinux-1.40
To: eliz AT is DOT elta DOT co DOT il (Eli Zaretskii)
Date: Wed, 8 Jul 1998 18:23:09 +0200 (MET DST)
Cc: djgpp-workers AT delorie DOT com (DJGPP-WORKERS)
In-Reply-To: <Pine.SUN.3.91.980708115412.5286R-100000@is> from Eli Zaretskii at "Jul 8, 98 11:54:43 am"
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

According to Eli Zaretskii:
> > Well I disagree; there's no standard DJGPP directory, hardcoding them
> > to c:/djgpp is BAD.
> 
> There's no other way to do it, given that DOS/Windows lack
> standardized directory trees.  Almost all DJGPP port of GNU packages
> have been doing this for years, albeit without paths.h: they all have
> some path like c:/djgpp/something hard-wired into them, and then rely
> on environment variables set from DJGPP.ENV to take care of the actual
> place where to look for files.
> 
> > Consider that when there are there people will use
> > them and then when this program is moved to another machine, it wont
> > work because the path which is compiled into the program doesn't
> > exist!
> 
> We do it already, as I said above.

Ok. If there's already an incorrect hardcoded path used in the
programs then it doesn't matter. 

> And if a program is compiled on Linux, then moved to another Linux
> machine, it runs the risk of not working as well.

No, wrong. If you use a descent distribution, there are no
problems. Otherwise you are used to compile the programs yourself and
then your path.h is presumably correct.

> Please keep in mind that paths.h is a last fallback in case no
> environment variable overrides it.

Ok. Then I have no objections.


Gorecki, Symphony No. 3,

							MartinS