Fix lua-style dictionary operator performance

[anvilfolk]

Fixes #68834.

I'm having the GDScript analyzer replace an IdentifierNode with a LiteralNode so the behavior of dict.someIndex becomes the same as dict["someIndex"].

I don't believe I can do this in the parser, because the parser does not know whether dict is a dictionary or not. That is something the analyzer infers.

Unfortunately, this does require us to go into GDScriptParser's singly-linked list list and do a search for the IdentifierNode so we can remove it from there and replace it with the LiteralNode. For large scripts, this is not very efficient at all. If GDScriptParser::Node had a prev as well as a next, then it would become a doubly-linked list and be O(1) instead of O(n), where n is the number of parse nodes in a script.

I will likely add some more documentation in the future, but technically, this is ready for review.

Quick note: I really only added if (base_type.kind == GDScriptParser::DataType::BUILTIN && base_type.builtin_type == Variant::DICTIONARY && p_subscript->attribute->type == GDScriptParser::Node::IDENTIFIER) { ... }. The rest of the lines are counting as added lines but only because they got indented inside the else.

[adamscott]

I would create a converter inside gdscript_parser.{h,cpp} definition of IdentifierNode instead of handling the conversion inside an if. It would make the if/else more readable.

[anvilfolk]

Decided to close since there's a nice refactor in #68086 which fixes this and makes things better to boot :)

[anvilfolk]

Closed again since @rune-scape is investigating allowing StringName indexed dictionaries, which would make this not particularly desirable :) See discussion on #68834 for benchmarking & conclusions :)