------- Comment #3 from kargl at gcc dot gnu dot org 2005-10-11 18:48 ------- I don't think this is a glibc problem. gfortran will use a GCC builtin function. See gcc/fortran/mathbuiltins.def To confirm this, I get
troutmask:kargl[210] ./z ( 0.000000 , -1.000000 ) (-0.7071068 , 0.7071068 ) ( 1.1754944E-38, -1.000000 ) ( 0.7071068 ,-0.7071068 ) (-1.1754944E-38, -1.000000 ) ( 0.7071068 ,-0.7071068 ) on amd64-*-freebsd, which does not have a csqrtf in libm. Note, I've replaced Mick's subnormal number with TINY(1.). -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24313
