On Oct 29, 2009, at 4:35 PM, Robert Kern wrote: > On 2009-10-29 18:22 PM, Lisandro Dalcin wrote: >> Sorry for the late reply.. >> >> On Wed, Oct 28, 2009 at 4:13 AM, Stefan >> Behnel<[email protected]> wrote: >>> Hi, >>> >>> most of the 'complex' tests currently fail in Py3.1 because >>> complex.__repr__() became more correct. It now prints (-0-2j) for - >>> (2j). >> >> Are you sure this is "more correct" ? What's the point of printing a >> "0" for the real part? I really do not know what's going on there... > > +0 and -0 are different in IEEE floating point representation. It is > information > that should be preserved particularly since signed zeros are more > important with > complex numbers than reals because of branch cuts.
That is a very good point that I never thought about. - Robert _______________________________________________ Cython-dev mailing list [email protected] http://codespeak.net/mailman/listinfo/cython-dev
