[Ecls-list] BUG: floating point underflow in TAN

[Juan Jose Garcia-Ripoll]

2007/5/20, Mark Hoemmen <mark.hoemmen at gmail.com>:
> On 5/19/07, Juan Jose Garcia-Ripoll <jjgarcia at users.sourceforge.net> wrote:
> > this is a bug in GLIBC 2.0, more precisely on the tanf() routine,
> > which causes this kind of error, but only on 64-bit platforms. You can
> > verify it writing the equivalent C program. I filed a bug report half
> > a year ago but it is not being taken very seriously.
>
> Could one work around it by disabling the underflow floating-point exception?

Not really. The function will not give you a good result IIRC. The
only reasonable workaround is to replace tanf() with tan().

Juanjo

--
Facultad de Fisicas, Universidad Complutense,
Ciudad Universitaria s/n Madrid 28040 (Spain)
http://juanjose.garciaripoll.googlepages.com