Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!uunet!samsung!usc!pollux.usc.edu!kjh
From: kjh@pollux.usc.edu (Kenneth J. Hendrickson)
Newsgroups: comp.os.minix
Subject: floating point in Minix C
Message-ID: <26232@usc.edu>
Date: 30 Jul 90 01:57:38 GMT
Sender: news@usc.edu
Organization: EE-Systems, USC, Los Angeles
Lines: 22

Is there any simple flag to use for cc so that the libfp.a library will
be used before libc.a?  Is the only way to compile with -S, and then
explicitly execute asld?

Or maybe I have it all wrong?  Maybe I need to put the routines from the
libfp.a that I have into libc.a (replacing those routines that don't
support floating point like printf(), and adding missing routines)?

Where is the source code for these routines (in libfp.a)?

Why do I get the "Warning: dummy floating point constant(s)" message
from the C compiler every time I compile any program with any float
variables?

Please help.  I want to get floating point working.  (My next step will
be to introduce code to support the 80x87 math co-processor chips in the
kernel, and to get math co-processor code into the libraries also.)
Who knows?  Maybe we can turn Minix into a real tool.

Not having floating point is a worse limitation than 64k I+D.

Ken Hendrickson N8DGN/6      kjh@usc.edu      ...!uunet!usc!pollux!kjh