Re: printf-frexp and the radix of floating point numbers

[Bruno Haible]

Paul Eggert wrote:
> A more up-to-date mainframe C compiler would be Systems/C
> <http://www.dignus.com/dcc/>.  Its most recent release is 1.85
> (released 2006-09-01).  It supports some C99 features but not all, and
> its documentation doesn't claim support for printf 'a' and 'A'.

OK, but nevertheless the practical problem is: Even if I wanted to adjust
the printf-frexp and the 'a'/'A' conversion code to avoid rounding errors
on such platforms, I don't have enough info about how to do it.

If FLT_RADIX = 16, multiplication of a number with 2.0 causes rounding
errors for 25% of the numbers, and multiplication with 0.5 causes rounding
errors for 75% of the numbers. (Btw, this makes it impossible to implement
a C99 compatible frexp() function, no?) I would at least need to know
whether the rounding problem when multiplying with 2.0 occurs in the
range [1.0,2.0) or [2.0,4.0) or [4.0,8.0) or [8.0,16.0).

Bruno