Mail Archives: djgpp/1997/06/17/04:10:28
On Mon, 16 Jun 1997, Phil Galbiati wrote:
> If somebody who has access to these standards could comment on the
> subject, I would appreciate it.
Amazingly enough, ANSI standard doesn't seem to say anything about
this, at least in the context of `printf' and friends. The
description of <math.h> functions does say that they should NOT
generate any exceptions but set `errno' to EDOMAIN or ERANGE (which
apparently is NOT the case with math functions in libc.a), but it
keeps silence about float conversions in `printf'.
From various related sources, it seems that the most relevant place to
look is the IEEE standard, but I don't have access to it. Anybody?
- Raw text -