Exchange use of ClassDB::is_parent_class() with Object::is_class() where possible

[DeeJayLSP]

Replaces instances of ClassDB::is_parent_class() where you have access to the object itself, like:

ClassDB::is_parent_class(object->get_class(), inherits);

ClassDB::is_parent_class(object->get_class_name(), inherits);

String class_name = object->get_class_name();
ClassDB::is_parent_class(class_name, inherits);

With:

object->is_class(inherits);

Functionally identical, except we skip ClassDB's lookup by name and directly check using the object's own GDType.

Hierarchy is constructed alongside the GDType itself (not during initialization) and tests pass, so I assume this is a safe change.

[DeeJayLSP]

@Ivorforce this may be a reasonable follow-up of #118582

[dsnopek]

Could we get an existing StringName rather than parsing a char *? For example:

Suggested change

	const bool is_script = p_resource->is_class("Script");
	const bool is_script = p_resource->is_class(Script::get_class_static());

[KoBeWi]

The point of is_class() is to avoid including the class.

And if the class is already included, it's better to use Object::cast_to<>() or derives_from(). Using is_class() with get_class_static() does not make much sense.

Though for static strings, you can use SNAME at least.

[Ivorforce]

StringName lookup by key is about 3x faster than String allocation for the same key (i tested this once). ClassDB's lock + has lookup makes it even worse. So while using SNAME would be even faster, most (90+%?) of the performance gap is already addressed in this PR, even without SNAME use.

[DeeJayLSP]

Applied the SNAME suggestion to script_text_editor.cpp as I found it reasonable + fixed the commit title.

[Ivorforce]

Makes sense to me.