On Mon, May 03, 2010 at 01:30:34AM -0700, Niko Tyni wrote: > I'm seeing 5.12.0 test failures on the Debian powerpc architecture > with -Duselongdouble -Duse64bitint (but not with just -Duse64bitint.) > I assume the 128-bit long doubles on this platform are actually a sum of > two 64-bit doubles, which "does not conform to the IEEE floating-point > standard" [1] and "has serious arithmetic anomalies" [2]. That seems a reasonable assumption, given that sparc64 Debian with 16 byte long doubles does work. > If this something that can not be supported, perhaps Configure should > detect such a platform and disable the longdouble setting or error out? That seems reasonable. Not sure how to do it. > (This is also http://bugs.debian.org/578295 ) > > [1] http://en.wikipedia.org/wiki/Long_double > [2] http://gcc.gnu.org/ml/gcc-help/2006-10/msg00138.html Nicholas ClarkThread Previous