Avoid overflow in RNG

Instead of checking for overflow, get necessary information about
number of bits from type attributes.

This technically wasn't undefined behavior but with strict UBSAN
settings got tagged in tests.

diff --git a/src/gromacs/random/uniformrealdistribution.h b/src/gromacs/random/uniformrealdistribution.h
index bc7cb61405354569caf0207b387153f16db3722f..bdaf090018fab9fc571868e5897909f728b2bead 100644 (file)
--- a/src/gromacs/random/uniformrealdistribution.h
+++ b/src/gromacs/random/uniformrealdistribution.h
@@ -107,8 +107,7 @@ RealType generateCanonical(Rng& g)
     // No point in using more bits than fit in RealType
     const uint64_t digits   = std::numeric_limits<RealType>::digits;
     const uint64_t realBits = std::min(digits, static_cast<uint64_t>(Bits));
-    const uint64_t range    = Rng::max() - Rng::min() + uint64_t(1);
-    uint64_t       log2R    = (range == 0) ? std::numeric_limits<uint64_t>::digits : log2I(range);
+    const uint64_t log2R    = std::numeric_limits<typename Rng::result_type>::digits;
     uint64_t       k        = realBits / log2R + (realBits % log2R != 0) + (realBits == 0);
     // Note that Rng::max and Rng::min are typically an integer type.
     // Only unsigned integer types can express the range using the