[ntuple] Fix wrong assert in Real32Quant

The assert in QuantizeReals() should not assume that all values are in
range, as we want to throw an exception at the end of the function if
that is the case.

Another mistake conspired to hide this bug from all non-windows builds:
LeadingZeroes() and TrailingZeroes() were incorrectly using the 64-bit
version of the intrinsics, where they should have used the 32-bit one.
This means they effectively never fired as they ranged from 0 to 63
rather than the intended 0 to 31.

Author: silverweed

tree/ntuple/src/RColumnElement.hxx

@@ -996,7 +996,7 @@ using Quantized_t = std::uint32_t;
       return static_cast<std::size_t>(31 - idx);
    return 32;
 #else
-   return static_cast<std::size_t>(__builtin_clzl(x));
+   return static_cast<std::size_t>(__builtin_clz(x));
 #endif
 }
 
@@ -1011,7 +1011,7 @@ using Quantized_t = std::uint32_t;
       return static_cast<std::size_t>(idx);
    return 32;
 #else
-   return static_cast<std::size_t>(__builtin_ctzl(x));
+   return static_cast<std::size_t>(__builtin_ctz(x));
 #endif
 }
 
@@ -1042,14 +1042,15 @@ int QuantizeReals(Quantized_t *dst, const T *src, std::size_t count, double min,
    for (std::size_t i = 0; i < count; ++i) {
       const T elem = src[i];
 
-      nOutOfRange += !(min <= elem && elem <= max);
+      bool outOfRange = !(min <= elem && elem <= max);
+      nOutOfRange += outOfRange;
 
       const double e = 0.5 + (elem - min) * scale;
       Quantized_t q = static_cast<Quantized_t>(e);
       ByteSwapIfNecessary(q);
 
       // double-check we actually used at most `nQuantBits`
-      assert(LeadingZeroes(q) >= unusedBits);
+      assert(outOfRange || LeadingZeroes(q) >= unusedBits);
 
       // we want to leave zeroes in the LSB, not the MSB, because we'll then drop the LSB
       // when bit packing.