[GiNaC-devel] Floating complex numbers

[Alexei Sheplyakov]

Hi,

On Fri, Jan 17, 2014 at 10:50 AM, Stefan Weinzierl
<stefanw at thep.physik.uni-mainz.de> wrote:

> I would also propose the attached patch for the tutorial to make it clear to
> all users, what GiNaC and CLN do with floating point complex numbers.

I think the issue is a bit more generic. That is GiNaC (and CLN) converts
exact numbers into inexact ones only upon the user request (i.e. one need
to call evalf()). Likewise inexact numbers are never converted into exact ones.
This behavior is already documented.  Do we really need to document it
once more?

 Best regards,
        Alexei