Hi, Sorry for my ignorance about mingw-w64 - I am hoping for advice on submitting a patch.
When testing against the Python numpy library, we found a failing test for the mingw-w64 expm1 routine. Specifically, we found that: double i = -0.0; return expm1(i); returns 0 instead of -0, as I believe the standard requires: http://pubs.opengroup.org/onlinepubs/9699919799/functions/expm1.html http://lists.gnu.org/archive/html/bug-gnulib/2012-03/msg00029.html https://github.com/gagern/gnulib/blob/master/tests/test-expm1-ieee.h The following patch fixes this: diff -r -u mingw-w64-v3.1.0/mingw-w64-crt/math/expm1.def.h b/mingw-w64-crt/math/expm1.def.h --- mingw-w64-v3.1.0/mingw-w64-crt/math/expm1.def.h 2014-04-14 20:41:33.000000000 -0700 +++ b/mingw-w64-crt/math/expm1.def.h 2011-04-02 09:58:41.000000000 -0700 @@ -60,7 +60,7 @@ } else if (x_class == FP_ZERO) { - return __FLT_CST(0.0); + return (signbit (x) ? -__FLT_CST(0.0) : __FLT_CST(0.0) } if (__FLT_ABI (fabs) (x) < __FLT_LOGE2) { but I don't know how to submit tests for the change. What do y'all advise? Thanks a lot, Matthew ------------------------------------------------------------------------------ Learn Graph Databases - Download FREE O'Reilly Book "Graph Databases" is the definitive new guide to graph databases and their applications. Written by three acclaimed leaders in the field, this first edition is now available. Download your free book today! http://p.sf.net/sfu/NeoTech _______________________________________________ Mingw-w64-public mailing list Mingw-w64-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
