Avoid UB in maxIntN(64).

[jlebar]

Summary

Previously we were relying on 2's complement underflow in an int64_t.
Now we cast to a uint64_t so we explicitly get the behavior we want.

[majnemer]

include/llvm/Support/MathExtras.h

348 ↗

(On Diff #64237)

I'd just rewrite it as:

(UINT64_C(1)<<(N-1)) - 1

[jlebar]

Simplify the expression.

[jlebar]

include/llvm/Support/MathExtras.h

348 ↗

(On Diff #64237)

Oy. Done, thank you.

llvm/trunk/include/llvm/Support/MathExtras.h

inline int64_t minIntN(int64_t N) {
 
   return -(INT64_C(1)<<(N-1));
 }
 
 /// Gets the maximum value for a N-bit signed integer.
 inline int64_t maxIntN(int64_t N) {
   assert(N > 0 && N <= 64 && "integer width out of range");
 
-  return (INT64_C(1)<<(N-1)) - 1;
+  // This relies on two's complement wraparound when N == 64, so we convert to
+  // int64_t only at the very end to avoid UB.
+  return (UINT64_C(1) << (N - 1)) - 1;
 }
 
 /// isUIntN - Checks if an unsigned integer fits into the given (dynamic)
 /// bit width.
 inline bool isUIntN(unsigned N, uint64_t x) {
   return N >= 64 || x <= maxUIntN(N);
 }