http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47716
Tobias Burnus <burnus at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |FIXED --- Comment #7 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-02-15 23:17:41 UTC --- (In reply to comment #6) > warning: warning: reference to compatibility cabsf() That seems to be a warning generated by NetBSD's libm, cf. http://mirror.leaseweb.com/NetBSD/NetBSD-release-5-0/src/lib/libm/compat/compat_cabsf.c That file contains: float cabsf __P((struct complex)); __warn_references(cabsf, "warning: reference to compatibility cabsf()"); float cabsf(struct complex z) { return hypotf(z.x, z.y); } Which causes the warning. The __warn_references is defined in <sys/cdefs.h>. Thus, you could also try to silence the warning. hypot(f) should be give an equivalent result to cabs(f) - thus I do not understand why there is a warning at the first place. (Depending, of course, on the quality of the hypot(f) implementation.) As the warning is due to NetBSD itself and as the quadmath_weak.h issue is solved, I close this PR. It would be nice if one could silence the warning - and again build with Fortran support.