[analyzer] Don't use makeIntVal to create a floating-point value.

SimpleSValBuilder processes a couple trivial identities, including 'x - x'
and 'x ^ x' (both 0). However, the former could appear with arguments of
floating-point type, and we weren't checking for that. This started
triggering an assert with r163069, which checks that a constant value is
actually going to be used as an integer or pointer.

llvm-svn: 163159

clang/lib/StaticAnalyzer/Core/SimpleSValBuilder.cpp

@@ -318,7 +318,9 @@ SVal SimpleSValBuilder::evalBinOpNN(ProgramStateRef state,
         return makeTruthVal(false, resultTy);
       case BO_Xor:
       case BO_Sub:
-        return makeIntVal(0, resultTy);
+        if (resultTy->isIntegralOrEnumerationType())
+          return makeIntVal(0, resultTy);
+        return evalCastFromNonLoc(makeIntVal(0, /*Unsigned=*/false), resultTy);
       case BO_Or:
       case BO_And:
         return evalCastFromNonLoc(lhs, resultTy);

clang/test/Analysis/idempotent-operations.c

@@ -234,3 +234,11 @@ void rdar8601243() {
   (void) start;
 }
 
+
+float testFloatCast(int i) {
+  float f = i;
+
+  // Don't crash when trying to create a "zero" float.
+  return f - f;
+}
+