…#11522)

Currently, the implicit-arrow-linebreak rule contains a lot of logic to determine how comments should be adjusted in code when an autofix is needed. The goal is to be able to autofix cases where there is a comment between an arrow token and the start of an arrow function body. Most other core rules simply decide not to fix cases when there is a comment interfering with the fix.

This logic accounts for a large fraction of the code in the rule, and seems to require a lot of different code for many individual cases. Unfortunately, bugs keep being reported identifying problems in the rule (e.g. #11268, #11521) and it's not clear that the fixes are moving us closer to making the rule correct in general, given that there are always more cases than we can explicitly account for.

To address those problems, this commit updates the implicit-arrow-linebreak rule to just skip autofixing when comments interfere, rather than trying to do an autofix anyway and find an alternate location for the comments. I'm reluctant to make this change given that a lot of time has been invested in the autofixing logic, but I think this is ultimately a better solution than trying to guess where the user wants their comments to go, and crashing/producing incorrect code if we get it wrong.