Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!uwm.edu!linac!att!ucbvax!bloom-beacon!dont-send-mail-to-path-lines
From: jinx@altdorf.ai.mit.EDU ("Guillermo J. Rozas")
Newsgroups: comp.lang.scheme.c
Subject: Scheme on SGi machines
Message-ID: <9105151950.AA09815@bloom-beacon.MIT.EDU>
Date: 15 May 91 19:34:12 GMT
References: <9105151544.AA13154@UVM.EDU>
Sender: daemon@athena.mit.edu (Mr Background)
Reply-To: jinx@zurich.ai.mit.edu
Organization: The Internet
Lines: 22


   Date: Wed, 15 May 91 11:44:22 -0400
   From: Tom.Emerson@UVM.EDU (Tom Emerson)

   Has anyone brought Scheme 7.1 on the SGi Iris under Irix 3.3.2?
   The SGi is a MIPS machine, so I wouldn't think it would be difficult.

You are right, it should not be difficult, but...

The main issues are:

- Making sure that the microcode compiles and runs correctly.  Each
version of Unix is different from every other, and the
conditionalization may not work correctly for yours.
- Compiling the system sources for the Iris machine.

Note that if the Iris machine uses little-endian byte ordering, it
should be able to use the MIPS scheme binaries.  If it uses big-endian
byte ordering, it cannot.  The compiler is conditionalized to handle
both byte orderings, but no one has tried it on a big-endian machine.
We probably will shortly because we are getting some MIPS-based
machines that (we believe) use big-endian byte ordering.