Path: utzoo!attcan!utgpu!news-server.csri.toronto.edu!mailrus!cs.utexas.edu!uunet!decwrl!sgi!shinobu!odin!joker.sgi.com!myoung
From: myoung@joker.sgi.com (Mark Young)
Newsgroups: comp.sys.mac.programmer
Subject: TC4.0 - problems with sscanf("%f") conversions with 881 enabled
Message-ID: <11679@odin.corp.sgi.com>
Date: 11 Aug 90 19:36:51 GMT
Sender: news@odin.corp.sgi.com
Reply-To: myoung@joker.sgi.com (Mark Young)
Followup-To: comp.sys.mac.programmer
Organization: Silicon Graphics, Inc., Mountain View, CA
Lines: 24

I have a piece of code that is similar to:

    {
        float a = 1.0;
        float b = 2.0;
        float c = 3.0;
        char *s = "6.02";

        sscanf(s,"%f",&b);
    }

when I compile the code with the 68881 compiler option DISabled, the code
works as you would expect.  with the 68881 option ENabled, the result in
the variable 'b' is ok, but the value in 'a' is getting munged by the call
to sscanf().

is there a patch for this?  I looked at the code to verify that the
absence of a length modifier in the format implies single precision and 
found that I was correct.  I really don't want to have to waste the time
tracking this down if it's already been fixed.

any leads are welcome.

	...myoung@joker.asd.sgi.com