The check incorrectly results in catan returning nan + inf i when real part is +/- 1 and imaginary part is 0. The same occurs for real 0.8 and imaginary 0.6. The change ends up matching glibc behaviour.
Signed-off-by: Yaakov Selkowitz <yselkowi@redhat.com>
libm/complex/ctan.c, libm/complex/ctanf.c: Use HUGE_VAL instead of MAXNUM.
* libm/Makefile.am: added complex functions documentation * libm/libm.texinfo: ditto * libm/complex/Makefile.am: Ditto * libm/complex/complex.tex: Ditto * libm/complex/(cabs.c, cacos.c, cacosh.c, carg.c, casin.c, casinh.c, catan.c, catanh.c, ccos.c, ccosh.c, cexp.c, cimag.c, clog.c, conj.c, cpow.c, cproj.c, creal.c, csin.c, csinh.c, csqrt.c, ctan.c, ctanh.c): Ditto * libm/Makefile.in: Regenerate * libm/complex/Makefile.in: Ditto
* libm/common/fdlibm.h: Added ifdef _COMPLEX_H. * libm/complex/*: New complex functions imported from NetBSD. * libm/Makefile.am: Added complex subdir. * libm/Makefile.in: Regenerate. * libm/configure.in: Added complex subdir. * libm/configure: Regenerate.
Joseph Faulls
Yaakov Selkowitz
Corinna Vinschen
Jeff Johnston