norm@rocksanne.uucp (Norm Zeck) (10/07/88)
With all the noted problems of significant zeros in floating point
numbers, does anyone have a work around using cps? Looks like limiting
the precision to four decmial places in pprintf may work. But
has anyone fixed this?
I have also tried using the "fixed" type of 16.16 (16 bit integer,
16 bit fraction represented as a long, but pprintf gets a memory
violation in the code for parsing %e (fixed) numbers.
--------------
Norm Zeck
Xerox Corp.
800 Phillips Rd. 128-29E
Webster N. Y., 14580
ARPA: NZeck.Wbst@Xerox.com
UUCP: {decvax, allegra, seismo, cmcl2}!rochester!rocksanne!norm
Phone: [USA] (716) 422 6246
--------------greg@gergle.UUCP (10/14/88)
> I have also tried using the "fixed" type of 16.16 (16 bit integer, > 16 bit fraction represented as a long, but pprintf gets a memory > violation in the code for parsing %e (fixed) numbers. I found this one a few months ago, but forgot to post the fix. I don't have the checked fix on this machine, but it went something like this; in function pprintf file ctops.c look for case 'e': at the bottom of the while loop you will see n = *np++; change this to if(cnt > 0) n = *np++; -greg.