Do not allow super() without parameters in generator expressions

[sobolevn]

This bug: https://bugs.python.org/issue46175

We need to require super(cls, self) in this case.

[ghost]

Hi ! I would like to work on this issue :)

[sobolevn]

@nasdev-cmd thank you!

[kitsiosvas]

Hello, is this issue still open?
If so, I would appreciate it greatly if I could provided with the information of where is the code linked to this behavior located.

[sobolevn]

@kitsiosvas I think that we can add this check here: https://github.com/wemake-services/wemake-python-styleguide/blob/master/wemake_python_styleguide/visitors/ast/classes.py

Are you familiar with python's AST?

[kitsiosvas]

@sobolevn I am somewhat familiar with AST indeed. Are there any other similar checks that I could consult?
I've been studying the code you included in your previous message to understand what exactly I will need to implement.
Thank you.

[sobolevn]

You can take a look at this:

wemake-python-styleguide/wemake_python_styleguide/visitors/ast/functions.py

Lines 95 to 102 in 61d4fa5

	def _check_wrong_function_called(self, node: ast.Call) -> None:
	function_name = functions.given_function_called(
	node, FUNCTIONS_BLACKLIST,
	)
	if function_name:
	self.add_violation(
	WrongFunctionCallViolation(node, text=function_name),
	)

It just checks for wrong functions. But, in this case you will also need the context. See https://github.com/wemake-services/wemake-python-styleguide/blob/master/wemake_python_styleguide/logic/walk.py

[kitsiosvas]

If I understand it correctly, I need to somehow check if super() exists as a node with parent a generator expression and it's kids
are not in the form you have pointed out initially (super(cls, self)).
I don't know if I am missing something. Any input is more than welcome.

[sobolevn]

Yes, sounds right!

If super call:

• is inside a gen expr
• and does not have two args

We raise a violation.

[kitsiosvas]

@sobolevn Thank you for your quick reply.
I am having trouble figuring out how do I check if a node is inside a generator expression.
Is there a tactic that I am missing?
Sorry for making this threat a long one.

Edit: I am thinking of something like using get_closest_parent(node, parents) method. As first argument there will be the super() call node, and as second argument I will have to include generator expressions (not sure on how that is implemented).
If it returns null then super() is not in a gen expr.

[sobolevn]

I would go this way:

1. Find ast.Call nodes, check name to be super
2. Check args count (it is fast and cheap)
3. Then get_closest_parent(call, ast.GeneratorExpr)

[kitsiosvas]

In order to get the node's name I would have to use given_function_called(node, to_check: Container[str]) method.
What would I use as Container in this case. I couldn't find anything clear on this. Or maybe there is another way of getting the name of the node?

[sobolevn]

given_function_called(node, {'super'})

[kitsiosvas]

I have implemented the feature. Should I issue a Pull Request for review?

[sobolevn]

Please! 👍

[kitsiosvas]

According to the "Raise a violation" part, I tried doing something similar to #2310 (comment) , but it raises errors apparently. I couldn't figure out what is the proper way of achieving this, so far. Will keep searching, but any help is amazing.
Thank you.

[akshaybhadange]

does this issue still exists , I want to contribute in this project , actually its my first opensource contribution , so any help regarding this really appreciated

[lensvol]

Can I take this issue over?

[sobolevn]

Please! Can you also check other comprehensions? And nested functions?

[lensvol]

I'll do what I can.