GDScript source text into a flat token list.

Indentation is significant in GDScript, so the lexer is line-oriented: it tracks an indent stack and emits :indent / :dedent tokens, the same shape a Python lexer produces. Everything downstream then works on a flat stream with explicit block delimiters.

A token is {type, value, line}. Types:

• :ident :int :float :string — literals and names
• :keyword — a reserved word (func, var, if, …)
• :op — an operator or delimiter
• :newline :indent :dedent :eof — structure

Blank lines and comment-only lines produce no tokens at all, and newlines inside brackets are swallowed so a call can wrap across lines.