Always use expect_correction or expect_no_corrections when autocorrect is supported
#2052
Conversation
…rrect is supported
| @@ -329,6 +329,8 @@ | |||
| DESC | |||
| end | |||
| RUBY | |||
|
|
|||
| expect_no_corrections | |||
There was a problem hiding this comment.
Is this what we want, though? Does it discourage people from implementing autocorrect, or will it work fine as a “placeholder” until autocorrect is implemented.
There was a problem hiding this comment.
This method only useful for cops that implement autocorrect but not for all cases. If someone doesn't implement autocorrect to begin with, there will also be no offense. There's no cop to check for extend AutoCorrector (;
In this specific case, I think autocorrect could theoretically be implemented but hasn't because heredocs are hard and nobody actually writes tests with heredoc text descriptions.
There was a problem hiding this comment.
Yeah, some cops have autocorrect for most cases, but some cases are left to implement later. Maybe we should have an expect_no_corrections_yet alias for those cases 😆
No, it’s probably fine with the expect_no_corrections. I guess we’ll only use it for exactly those cases, where it’s theoretically possible to implement autocorrection.
In rubocop/rubocop#13183 I want to add a cop that enforces this automatically. This fixes offenses here in anticipation of that.