[ODS] Quick fix

[amandatang]

Summary

Quick fix for a failing test in https://reviews.llvm.org/rG5267ed05bc4612e91409d63b4dbc4e01751acb75.

When properties is empty, the getProperties method in the OpAdaptor class returns an empty struct as opposed to a null pointer. This caused the null check on the properties of ConstShapeOp to fail. The fix involves performing the null check on the shape value within the Properties struct.

[jpienaar]

Please update the description to make it clear what is being fixed

[mehdi_amini]

Can you add a non-python test to exercise this please?

mlir/lib/Dialect/Shape/IR/Shape.cpp

void ConstShapeOp::getCanonicalizationPatterns(RewritePatternSet &patterns,
                                                MLIRContext *context) {
   patterns.add<TensorCastConstShape>(context);
 }
 
 LogicalResult mlir::shape::ConstShapeOp::inferReturnTypes(
     MLIRContext *context, std::optional<Location> location,
     ConstShapeOp::Adaptor adaptor, SmallVectorImpl<Type> &inferredReturnTypes) {
   Builder b(context);
-  const Properties *prop = &adaptor.getProperties();
+  const Properties prop = adaptor.getProperties();
   DenseIntElementsAttr shape;
   // TODO: this is only exercised by the Python bindings codepath which does not
   // support properties
-  if (prop)
-    shape = prop->shape;
-  else
-    shape = adaptor.getAttributes().getAs<DenseIntElementsAttr>("shape");
+  shape = prop.shape ? prop.shape : 
+            adaptor.getAttributes().getAs<DenseIntElementsAttr>("shape");
   if (!shape)
     return emitOptionalError(location, "missing shape attribute");
   inferredReturnTypes.assign({RankedTensorType::get(
       {static_cast<int64_t>(shape.size())}, b.getIndexType())});
   return success();
 }
 
 bool mlir::shape::ConstShapeOp::isCompatibleReturnTypes(TypeRange l,