Updated to get correct gradient when reference variable is initialized with pointer dereference

[kaushal-malpure]

@parth-07
Referring to #734
According to my understanding when a reference variable is initialized with pointer dereference, the clad::gradient function creates a new adjoint variable for the given reference variable whereas it should share the same adjoint variable as the pointer. To solve this issue we can replace the reference variable with the original pointer. This might be a potential solution. I couldn't verify this because my system gives an error when I try to build the program.

[github-actions]

clang-tidy made some suggestions

[github-actions]

warning: do not use 'else' after 'return' [llvm-else-after-return]

Suggested change

	} else {
	return CladFunction<DerivedFnType, ExtractFunctorTraits_t<F>, true>(
	derivedFn /* will be replaced by gradient*/, code, f);
	}
	} return CladFunction<DerivedFnType, ExtractFunctorTraits_t<F>, true>(
	derivedFn /* will be replaced by gradient*/, code, f);

[PetroZarytskyi]

Hi @kaushal-malpure. Thank you for your interest in Clad. I think the solution to this issue should not be in the Clad's front end (like the gradient function) but rather in the back end. Perhaps the changes should be introduced to ReverseModeVisitor.cpp (this is where the gradient code is generated). I would suggest getting familiar with how Clang AST is generated there. You can start with the main function Derive. And yes, you need to be able to build Clad. Being able to take a look at the gradient code you produce is crucial for understanding and testing your changes.