Add Autocorrect for Lint/UselessAssignment cop

[edbarbera]

Is your feature request related to a problem? Please describe.

The Lint/UselessAssignment cop often appears when linting tests. I would like to be able to run auto-correct to fix these linting errors.

Describe the solution you'd like

A method in the cop to auto-correct the useless assignment (happy to write it myself).

Example:

# Variable only assigned once, never referenced again
item = create(:item)

# Autocorrect to
create(:item)

Describe alternatives you've considered

I have searched around the repo to see if this has been mentioned in a prior Issue or PR, but couldn't find anything. I also don't see any reason to not allow an auto-correct method to automatically fix this error, but if there is please correct me.

Additional context

N/A

[vlad-pisanov]

In a lot of cases, you'd want the whole line gone, not just the assignment -- so you don't end up with things like:

item = 42

# autocorrect to 😬
42

[koic]

Removing a whole line is unsafe change when RHS as a method call with side effects is used instead of 42. So it would not be the cop's responsibility to remove the whole line I think.

[vlad-pisanov]

@koic yeah, that's what I was getting at -- in some cases you'd want the assignment removed, in others you'd want the whole line removed -- which is something only the developer can decide correctly, I think

unused = 42                       # should be safe to remove line
unused = some_useful_call         # dev should decide (keep `some_useful_call`)
unused = some_useless_call        # dev should decide (remove whole line)

EDIT: actually, no -- automatically removing the whole line is unsafe, as it could be the result of a method/block: 😨

def foo
  unused = 42
end

[dduugg]

In a lot of cases, you'd want the whole line gone, not just the assignment -- so you don't end up with things like:

item = 42

# autocorrect to 😬
42

…but this is handled by another cop (Lint/Void)

[edbarbera]

@vlad-pisanov definitely a good point that it would be unsafe to remove the entire line and that wasn't what I intended as I didn't think that the cop would need to handle a case such as (as @dduugg just mentioned):

unused = 42

In my head, the cop would only auto-correct when the RHS was a method call by using one of the inbuilt Rails methods? I haven't tested anything yet in code, so not sure how doable this approach is yet. However wanted to get people's opinions of whether or not this was a sensible auto-correction.

[r7kamura]

I have prepared the following pull requests for this issue. If you guys are interested, please take a look:

• Add autocorrection for Lint/UselessAssignment #11840
• Add autocorrection for Lint/Void #11848