[gnu.ghostscript.bug] Building Ghostscript with gcc

ghost@aladdin.com (L. Peter Deutsch) (08/05/89)

Sigh.  Various people have built Ghostscript with gcc in the past,
but when I started to follow up on Stephen Gildea's problem report, I
discovered that, once again, both the Sun-3 AND the Sun-4 versions of
gcc are broken on the system that I have access to.  (The Sun-4
version blows up with the assembler not recognizing the switch
'-mc68020'; the Sun-3 version blows up during compilation with cpp
complaining about an unknown switch '-$'.)

If someone (preferably in the San Francisco Bay Area) can provide me
with an account on a Unix machine (presumably a Sun, or a VAX running
Ultrix) that's on the Internet and whose system software is better
maintained, I'd be very grateful, and it would be a big help in
improving the quality of Ghostscript releases.

I'm looking into getting Unix (and X) for my development machine, but
it may take a little while, and I'm resisting spending the $500-1000
that it looks like it will cost.  If anyone has any good leads on a
reasonably-priced, good-quality Unix (NOT Xenix) and X implementation
(server and client) for the 80386, I'd appreciate the information.
Oh, and of course it should run gcc.

Meanwhile, I have a fix, which I'm sure I got from bug-ghostscript,
to the specific problem Gildea mentioned (the inter-reference of the
declarations in gxdevice.h.)  It's too long to mail, and I've made
other changes to my copy of gxdevice.h in preparation for version
1.4.  The idea is to do both of the typedefs, and then both of the
actual struct definitions.  I haven't been able to test this on gcc
myself yet, obviously.

L. Peter Deutsch			ghost@aladdin.com
Aladdin Enterprises			...{uunet,parcplace}!aladdin!ghost
P. O. box 60264, Palo Alto, CA 94306