Message-ID: <kd_vuqC00jeiM_iVhA@cs.cmu.edu>
Date: Thu, 21 Nov 1991 09:12:06 -0500 (EST)
From: William.Lott@cs.cmu.edu
To: cmucl-bugs@cs.cmu.edu, Steve Olson <olson@juliet.LL.MIT.EDU>
Subject: Re: questions and comments
In-Reply-To: <9111201656.AA04067@goneril>
References: <9111201656.AA04067@goneril>

olson@juliet.ll.mit.edu ( Steve Olson) writes:
> - Can an image be built from the sources?  Is it intended that we be able to
>   do so?  There is no build documentation, but in the sources there is a
>   directory with several promising looking shell and lisp scripts.

Depends on exactly what you mean by ``from the sources.''  If you
mean, can someone take the sources and nothing else and build a
system, the answer is no.  You need a working copy of CMUCL to compile
the sources and build a kernel core.

> - There seems to be a circular dependancy between ldb/lisp.h and ldb/ldb.map
>   that causes bootstrapping problems.  How do I get around this problem?

Actually, the easiest thing to do is you make a fake ldb.map file that
has nothing in it by a version number:
	echo "Map file for ldb version 0" > ldb.map
and then run genesis with NIL for the list of files.  It will generate
a whole bunch of warnings about things being undefined, but ignore
that, because it will also generate a correct lisp.h.  You can then
compile ldb producing a correct ldb.map, and then run genesis for
real.

This is better than the way Rob suggested, because you don't need to
find a lisp.h for a different system, and running genesis with no
files is faster than compiling all of ldb.  But either way would work.

> - This is irritating
> 
> * (setq x 1)
> 
> Warning: This variable is undefined:
>   X
> 
> 1

Count me in the Rob camp.  But I think warn-once is a reasonable
default.

-William Lott
CMU Common Lisp Group