builtins: do the TODO on compile() #9567
Conversation
| optimize: int = -1, | ||
| *, | ||
| _feature_version: int = -1, | ||
| ) -> _ast.AST: ... |
There was a problem hiding this comment.
We can potentially be more precise here: if mode is "eval", the node must be an _ast.expr.
There was a problem hiding this comment.
Would we have to duplicate all 8 overloads from ast.parse, or does compile only take "exec", "eval" or "single" for the mode parameter?
There was a problem hiding this comment.
Yes, we would. ast.parse just calls compile(). I think it's fine to stick with AST for simplicity.
There was a problem hiding this comment.
Though I think we'd only need four overloads because the second argument to compile() is required, unlike ast.parse.
There was a problem hiding this comment.
Hmm. Four overloads isn't that bad? But I don't have a strong opinion either way.
|
Diff from mypy_primer, showing the effect of this PR on open source code: sympy (https://github.com/sympy/sympy)
+ sympy/parsing/sympy_parser.py:1075: error: Incompatible types in assignment (expression has type "CodeType", variable has type "str") [assignment]
jinja (https://github.com/pallets/jinja)
+ src/jinja2/environment.py:704: error: Unused "type: ignore" comment
|
This is a side-effect of more precise types; previously mypy allowed the redefinition because compile() returned Any.
https://github.com/pallets/jinja/blob/main/src/jinja2/environment.py#L704 This fixes a false positive from mypy. |
No description provided.