Pattern Catching uses let instead of const

[bbrk24]

ESLint doesn't like that

[edemaine]

Civet uses let all over the place for refs. Is this really the first time it's come up with a single assignment? I'd recommend disabling thay rule.

In this case, it's implementing catch's normal semantics (you can assign to e).

[bbrk24]

Is this really the first time it's come up with a single assignment?

Usually it's either multiple assignment (x := try f() catch g()) or assigned in expression position (if foo := bar()), so the error hasn't been applicable before now. It still has the ability to be meaningful, e.g. forbidding .= where := would suffice.

[edemaine]

Ah, OK, the rule is indeed useful for Civet if it doesn't apply to assignments in expression position. I guess the question here is whether to detect whether const would suffice (probably doable... but might need some scope work, to ignore assignments to shadowing variables of the same name), or somehow get eslint to ignore these (via a comment, or by doing the assignment in expression position, or...).

[bbrk24]

Actually, I just discovered another situation in which this error can trigger:

->
  return = 5
  foo()

The return ref is never used again so it "should" be a const. The annoying thing about both of these is that I can't even manually add // eslint-disable-next-line prefer-const -- in the return case it gets placed after the let ret; declaration, and in the catch case it's eaten entirely.