Mercurial > octave-nkf
diff liboctave/oct-inttypes.cc @ 8856:ab4db66e286f
workaround gcc 4.3 explicit instantiation bug in octave_int_cmp_op
| author | Jaroslav Hajek <highegg@gmail.com> |
|---|---|
| date | Tue, 24 Feb 2009 11:15:05 +0100 |
| parents | ee3446ca6157 |
| children | 4d328b8979c8 |
line wrap: on
line diff
--- a/liboctave/oct-inttypes.cc Tue Feb 24 11:12:38 2009 +0100 +++ b/liboctave/oct-inttypes.cc Tue Feb 24 11:15:05 2009 +0100 @@ -56,7 +56,7 @@ template <class xop> OCTAVE_API bool -octave_int_cmp_op::mop (uint64_t x, double y) +octave_int_cmp_op::emulate_mop (uint64_t x, double y) { static const double xxup = std::numeric_limits<uint64_t>::max (); // This converts to the nearest double. Unless there's an equality, the @@ -76,7 +76,7 @@ template <class xop> OCTAVE_API bool -octave_int_cmp_op::mop (int64_t x, double y) +octave_int_cmp_op::emulate_mop (int64_t x, double y) { static const double xxup = std::numeric_limits<int64_t>::max (); static const double xxlo = std::numeric_limits<int64_t>::min (); @@ -123,7 +123,7 @@ template <class xop> OCTAVE_API bool -octave_int_cmp_op::mop (double x, uint64_t y) +octave_int_cmp_op::emulate_mop (double x, uint64_t y) { typedef typename rev_op<xop>::op rop; return mop<rop> (y, x); @@ -131,7 +131,7 @@ template <class xop> OCTAVE_API bool -octave_int_cmp_op::mop (double x, int64_t y) +octave_int_cmp_op::emulate_mop (double x, int64_t y) { typedef typename rev_op<xop>::op rop; return mop<rop> (y, x); @@ -499,8 +499,8 @@ } #define INSTANTIATE_INT64_DOUBLE_CMP_OP0(OP,T1,T2) \ - template OCTAVE_API bool \ - octave_int_cmp_op::mop<octave_int_cmp_op::OP> (T1 x, T2 y) + template bool \ + octave_int_cmp_op::emulate_mop<octave_int_cmp_op::OP> (T1 x, T2 y) #define INSTANTIATE_INT64_DOUBLE_CMP_OP(OP) \ INSTANTIATE_INT64_DOUBLE_CMP_OP0(OP, double, int64_t); \