Is there any tool available for fixing future-coercion-class-field
? We have a lot of those warnings -- 37 in Iris alone, probably hundreds over the repos we maintain -- so fixing them all manually is not really feasible.
IIRC the PR mentioned the possibility of another syntax that means "instance but not coercion"; if such a syntax existed then probably the line numbers printed by the warning would be enough to automate this reasonably well.
ah no that would not work, the line number is also wrong
Well, following your request, the warning is gone in 8.16. In 8.17 you'll just have to replace :>
with ::
(or probably rather silence the warning for a few version before doing that if you want to keep some backward compatibility).
it points at the Class
rather than the :>
and then there are multiple :>
in a class it only complains about the first
Pierre Roux said:
Well, following your request, the warning is gone in 8.16. In 8.17 you'll just have to replace
:>
with::
(or probably rather silence the warning for a few version before doing that if you want to keep some backward compatibility).
is it? I am seeing it in the rc1
Sure, that's when you (rightfully) complained, fixed in 8.16.0
oh okay I didnt realize that was after the rc
(I complained based on the PR where I got pinged, not based on seeing the warning in the wild)
okay I guess I'll have to wait for https://github.com/coq/opam-coq-archive/pull/2253 then before I can ensure on CI that we are warning-free with 8.16
thanks for reminding me that the warning is gone!
Indeed, the person who initially complained was @Paolo Giarrusso not you, sorry about that.
Last updated: Oct 13 2024 at 01:02 UTC