isreal(complex(1,%nan),0) returns %T instead of %F (Scilab 6.0.0 regression)
@sgougeon)
Reported by Samuel GOUGEON (BUG DESCRIPTION:
----------------
A Nan imaginary part has an infinity more chances to actually be non-zero than to be zero.
Up to Scilab 5.5.2, isreal(complex(1,%nan),0) returned %F, and that was right.
Starting from 6.0. and at least up to now with 6.1.1, this wrongly returns %T.
--> isreal(complex(1,%nan),0)
ans =
T
ERROR LOG:
----------
HOW TO REPRODUCE THE BUG:
-------------------------
isreal(complex(1,%nan),0)