🚀 go-pugleaf

In article <1993Jan12.131934.20309@shannon.ee.wits.ac.za> philip@concave.cs.wits.ac.za (Philip Machanick) writes:
>I tried compiling a public domain MIPS architecture (SPIM), and ran
>into several problems. Here is one that appears to be a bug in an
>SGI header file, syms.h (4.0.5F; in 4.0.1, the "__mips" is replaced by
>"mips"):
>
>There is a line
>
>#if __mips
>
>followed a few lines down by
>
>#else  /* !defined(__mips) */
>
>Surely "#if X" is not the same as "#if defined(X)" - so either the
>"#if" or the comment is wrong.

This appears in several files. Ugh.  I fixed this just now
for the next major release. Thanks for noticing the (minor) ugliness..

The compiler driver does -D__mips, which has the same meaning as
-D__mips=1.  Since the value of __mips is 1, the result is the same, so
you need not change anything (unless you wish to).  Changed to
#ifdef __mips for the next major release for readability.

>More important - between the "#if" and the "#else" is
>
>#include "cmplrs/stsupport.h"
>
>which header file doesn't exist (nor does the directory cmplrs).

cmplrs headers are not installed by default, but they are on the
compilers installation media for the machine.

Install
        dev.cmplrs_hdr.internal
Or if the compilers on that machine came from a 'maint' tape install:
        maint_dev.dev_cmplrs_hdr.internal

Either of those will get you cmplrs/stsupport.h

Use
        versions -I
to see what is currently installed.


>Any hints on how to work around this would be welcome. I would rather
>not patch the header file unless reassured that this won't break
>anything else.

Fixing the .h files won't break anything.

>For that matter if anyone has built SPIM or any other architecture
>simulator it would be useful for a course I am giving soon.

Someone else will have to deal with this one.

Regards,
[ David B. Anderson             (415)390-4263             davea@sgi.com ]