[Mlir-commits] [mlir] [mlir][emitc] Arith to EmitC: handle floating-point<->integer conversions (PR #87614)

[Simon Camphausen]

================
@@ -201,6 +201,94 @@ class SelectOpConversion : public OpConversionPattern<arith::SelectOp> {
   }
 };
 
+// Floating-point to integer conversions.
+template <typename CastOp>
+class FtoICastOpConversion : public OpConversionPattern<CastOp> {
+public:
+  FtoICastOpConversion(const TypeConverter &typeConverter, MLIRContext *context)
+      : OpConversionPattern<CastOp>(typeConverter, context) {}
+
+  LogicalResult
+  matchAndRewrite(CastOp castOp, typename CastOp::Adaptor adaptor,
+                  ConversionPatternRewriter &rewriter) const override {
+
+    Type operandType = adaptor.getIn().getType();
+    if (!emitc::isSupportedFloatType(operandType))
+      return rewriter.notifyMatchFailure(castOp,
+                                         "unsupported cast source type");
+
+    Type dstType = this->getTypeConverter()->convertType(castOp.getType());
+    if (!dstType)
+      return rewriter.notifyMatchFailure(castOp, "type conversion failed");
+
+    if (!emitc::isSupportedIntegerType(dstType))
+      return rewriter.notifyMatchFailure(castOp,
+                                         "unsupported cast destination type");
+
----------------
simon-camp wrote:

Make i1 types illegal. 0.5 should be truncated to 0 for i1, but `(bool)0.5` would return true.

>Boolean conversion
A value of any scalar type can be implicitly converted to _Bool(until C23)bool(since C23). The values that compare equal to an integer constant expression of value zero(until C23)are a zero (for arithmetic types), null (for pointer types) or have a type of [nullptr_t](https://en.cppreference.com/w/c/types/nullptr_t)(since C23) are converted to ​0​(until C23)false(since C23), all other values are converted to 1(until C23)true(since C23).

>bool b1 = 0.5;              // b1 == 1 (0.5 converted to int would be zero)
bool b2 = 2.0*[_Imaginary_I](http://en.cppreference.com/w/c/numeric/complex/Imaginary_I); // b2 == 1 (but converted to int would be zero)
bool b3 = 0.0 + 3.0*I;      // b3 == 1 (but converted to int would be zero)
bool b4 = 0.0/0.0;          // b4 == 1 (NaN does not compare equal to zero)
bool b5 = nullptr;          // b5 == 0 (since C23: nullptr is converted to false)

https://github.com/llvm/llvm-project/pull/87614