[clang] Use current rounding mode for float inc/dec (llvm#73770)

Increment and decrement are equivalent to adding or subtracting 1. For
the floating-point values these operations depend on the current
rounding mode. Teach constant evaluator to perform ++ and -- according
to the current floating-point environment.

Pull request: llvm#73770

clang/lib/AST/ExprConstant.cpp

@@ -4623,11 +4623,13 @@ struct IncDecSubobjectHandler {
     if (Old) *Old = APValue(Value);
 
     APFloat One(Value.getSemantics(), 1);
+    llvm::RoundingMode RM = getActiveRoundingMode(Info, E);
+    APFloat::opStatus St;
     if (AccessKind == AK_Increment)
-      Value.add(One, APFloat::rmNearestTiesToEven);
+      St = Value.add(One, RM);
     else
-      Value.subtract(One, APFloat::rmNearestTiesToEven);
-    return true;
+      St = Value.subtract(One, RM);
+    return checkFloatingPointResult(Info, E, St);
   }
   bool foundPointer(APValue &Subobj, QualType SubobjType) {
     if (!checkConst(SubobjType))

clang/test/SemaCXX/rounding-math.cpp

@@ -77,3 +77,24 @@ struct S1d {
   int f;
 };
 static_assert(sizeof(S1d) == sizeof(int), "");
+
+constexpr float incr_down(float k) {
+  float x = k;
+  ++x;
+  return x;
+}
+
+// 0x1.0p23 = 8388608.0, inc(8388608.0) = 8388609.0
+static_assert(incr_down(0x1.0p23F) == 0x1.000002p23F, "");
+// 0x1.0p24 = 16777216.0, inc(16777216.0) = 16777217.0 -> round down -> 16777216.0
+static_assert(incr_down(0x1.0p24F) == 0x1.0p24F, "");
+
+#pragma STDC FENV_ROUND FE_UPWARD
+constexpr float incr_up(float k) {
+  float x = k;
+  ++x;
+  return x;
+}
+static_assert(incr_up(0x1.0p23F) == 0x1.000002p23F, "");
+// 0x1.0p24 = 16777216.0, inc(16777216.0) = 16777217.0 -> round up -> 16777218.0
+static_assert(incr_up(0x1.0p24F) == 0x1.000002p24F, "");