Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!lll-crg!lll-lcc!qantel!ihnp4!inuxc!pur-ee!uiucdcs!convex!infoswx!bert
From: bert@infoswx.UUCP
Newsgroups: net.lang.c
Subject: Re: Help with BDSC float
Message-ID: <5000001@infoswx>
Date: Tue, 16-Sep-86 11:08:00 EDT
Article-I.D.: infoswx.5000001
Posted: Tue Sep 16 11:08:00 1986
Date-Received: Sat, 20-Sep-86 20:57:45 EDT
References: <3211@mit-eddie.MIT.EDU>
Lines: 35
Nf-ID: #R:mit-eddie.MIT.EDU:3211:infoswx:5000001:000:1229
Nf-From: infoswx.UUCP!bert    Sep 16 10:08:00 1986


About float in BDSC. (I'm talking about the CPM version that I have used)

Basically, BDSC does not have float.  There is a library of float
functions available that will allow you to do some floating point
operations with some difficulty.  To use them, you must declare a 5 char
array for each float variable.  You assign it a value with a function.
You play with it with other functions.  Then you can print it with
a special printf routine that you must link in specifically to handle
the %e and %f control strings.  But don't try to use fprintf if you
link the float printf stuff in, it doesn't work on the version I have.

examples:

char float1[5], float2[5];
buf[30];

atof(float1, "10.0");        /* set float to 10.0 */

fpadd(float1, float1, atof(float2, "5.0");     /* add 5 to it */

fpsub(float1, float1, atof(float2, "2.0");     /* subtract 2 from it */

fpmult(float1, float1, float2);			/* multiply it by 5 */

fpdiv(float1, float1, float2);			/* divide it by 5 */

ftoa(buf, float1);				/* make ascii (%e format) */

printf("%f %s\n", float1, buf);				/* print it (%f, %e) */

Note that the result is left in the first arg on all the arith operations.
Also, they all return char * to the result.

Bert Campbell.