SyntaxNodes: implement the hash function for syntax nodes.

[nkcsgexi]

This hash function will concatenate all interesting pieces of information
of node definitions in a single string and call hash() on this string.

[akyrtzi]

I have a recommendation to improve how the hash is calculated. Instead of concatenating strings you can use hash's ability to hash a tuple and get the hash but "feeding" the values into the function, like this:

def hash_syntax_node(node):
	    # Hash into the syntax name and serialization code
	    result = hash((node.name, get_serialization_code(node.syntax_kind)))
	    for child in node.children:
	        # Hash into the expected child syntax
	        result = hash((result, child.syntax_kind))
	        # Hash into the child name
	        result = hash((result, child.name))
	        # Hash into whether the child is optional
	        result = hash((result, child.is_optional))
	    return result

[nkcsgexi]

@akyrtzi Done.

[akyrtzi]

👍

[nkcsgexi]

@swift-ci please smoke test

[akyrtzi]

Oops sorry, missed a nitpick, you don't have to convert the integer to a string in order to hash it.
Also no need for str(child.is_optional) and str(token.serialization_code) below.

[akyrtzi]

Recommend hash_syntax_node(node=node) -> hash_syntax_node(node), named arguments are not required in python, you usually put them for readability when there's a lot of them, but it's not offering any value here
Also remove named argument for hash_token_syntax(token=token) below.

[nkcsgexi]

Thanks! Addressed in #22469