Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!rutgers!seismo!uunet!munnari!moncskermit!trldeity!trlamct!sunk!richb
From: richb@sunk.UUCP
Newsgroups: comp.sources.bugs
Subject: Re: binary display problem in calctool.
Message-ID: <231@sunk.oz>
Date: Wed, 2-Sep-87 02:05:46 EDT
Article-I.D.: sunk.231
Posted: Wed Sep  2 02:05:46 1987
Date-Received: Fri, 4-Sep-87 01:46:46 EDT
Organization: Sun Microsystems Australia
Lines: 46
Keywords: Take 2.

A little while ago Paul Czarnecki found a bug in calctool
when binary numbers near the capacity of the calculator
were being displayed incorrectly.

I put out a fix which partly fixed this problem but caused
the accuracy feature to work incorrectly.

Paul kindly mailed me his correction, and after we decided
that the calculator shouldn't try to grovel (the word sorry
was removed from the display message), we have a solution
that appears to work correctly.

This is for version 1.2. Again, I can't give diffs because
the version I'm currently working on it two releases later.

This is in the routine make_number in calctool.c:

  if ((ndig = ddig + accuracy) > cmax)
    {
      if (ddig > cmax) {
        STRCPY(display,"Overflow") ;
        display_result(display) ;
        error = 1 ;
        return ;
      } else {
        STRCPY(display, "Reducing precision") ;
        display_result(display) ;
        sleep(1) ;
        bzero(display);
        accuracy = cmax - ddig ;
        if (accuracy < 0) accuracy = 0;
        ndig = ddig + accuracy;
      }
    }

If this is all getting too confusing, then you might wish to
just wait for v1.4, which I've just posted to comp.sources.misc.

Sorry for the confusion, and thanks again Paul.

    Rich.

Rich Burridge,            JANET richb%sunk.oz@uk.ac.ucl.cs
ACSnet  richb@sunk.oz     UUCP {seismo,hplabs,ukc}!munnari!sunk.oz!richb
PHONE: +61 3 811 9927     ARPAnet richb%sunk.oz@seismo.css.gov
MAIL: Sun Microsystems, 123 Camberwell Rd, Hawthorn, VICTORIA 3122, AUST
D
D
D