GDScript: Improve error message for static/non-static function conflicts

[Lidang-Jiang]

Fixes #117912.

When a child class defines a function with the same name as a parent's static function (or vice versa), the error messages were not specific enough about the static/non-static mismatch.

Changes:

1. In check_class_member_name_conflict: When a member name conflicts with a parent class's static function, the error now says The member "X" already exists in parent class Y as a static function. instead of the generic The member "X" already exists in parent class Y.

2. In resolve_function_signature: Split the validation into two cases:

   • Static mismatch: Shows a specific error like The function "X" is not static, but the parent function of the same name is static.
   • Signature mismatch: Shows the existing signature comparison error (unchanged behavior).

Before

# Case 1: Child non-static overrides parent static
ERROR at line 8: The function signature doesn't match the parent. Parent signature is "my_function() -> void".

# Case 2: Variable overloads parent static function
ERROR at line 8: The member "overload_me" already exists in parent class A.

After

# Case 1: Child non-static overrides parent static
ERROR at line 8: The function "my_function" is not static, but the parent function of the same name is static. A static function cannot be overridden with a non-static function.

# Case 2: Child static overrides parent non-static
ERROR at line 8: The function "my_function" is static, but the parent function of the same name is not static. A non-static function cannot be overridden with a static function.

# Case 3: Variable overloads parent static function
ERROR at line 8: The member "overload_me" already exists in parent class A as a static function.

Tests added:

• override_static_function_with_non_static.gd/.out
• override_non_static_function_with_static.gd/.out
• variable_overloads_superclass_static_function.gd/.out

---

AI disclosure: AI tooling (Claude Code) was used to assist with writing and reviewing the code. All modifications were reviewed and tested by the author.

[dalexeev]

This is overcomplicated. You don't need to add code to check_class_member_name_conflict(), and you don't necessarily need to increase the indentation level in the if (!valid) branch (even though it means the function signature line is being generated in vain). Also, I have a PR #117360 on unifying signature handling, so I could add a more precise error message there.

[Lidang-Jiang]

Thanks for the review! You're right, this is overcomplicated. Since you already have #117360 for unifying signature handling and could add a more precise error message there, I'm happy to close this PR and defer to your approach. Let me know if you'd prefer that.

[dalexeev]

Please avoid creating multiple small files to test related features/bugs/warnings. Analyzer and runtime tests support multiple errors. Try to create comprehensive tests.

[Lidang-Jiang]

Good point, will consolidate into a single comprehensive test file if this PR continues.