Alejandro Serrano

> The downside of the context-parameters approach is my desire to provide a scope with both `Raise` and `Raise` and the contextual ambiguity that ensues. I'd need to double check,...

Bringing this one up again, is this ready to merge?

Thanks for the reproducer. It turns out that this is a [reported bug in the Kotlin compiler](https://youtrack.jetbrains.com/issue/KT-52706/Bad-signature-for-generic-value-classes-with-substituted-type-parameter). It seems that if you use `NonEmptyList`, the compiler still "leaves" an undeclared...

Unfortunately the only option until the fix is released is to use `List` instead and perform validation manually.

Update: the [bug](https://youtrack.jetbrains.com/issue/KT-52706/Bad-signature-for-generic-value-classes-with-substituted-type-parameter) is fixed in the compiler, but will be released in 2.2.20. I'll close this issue once it's released and we check it works.

@dcampogiani @hugomd could any of you test whether 2.2.20-Beta1 fixes this issue for you? I think it does, but I'd like to double check before closing this issue.

@dcampogiani it's deployed as usual, so you can just change your Kotlin version 2.2.20-Beta1 as with any other

@revonateB0T I've reported this as an [issue in YouTrack](https://youtrack.jetbrains.com/issue/KT-76389/Provide-update-functions-for-common-atomics).

We've decided to keep our implementation of atomics for the time being (until the new ones are no longer experimental). The `update` family of functions is showing in 2.2.20, so...

⚠️ The proposal mentions some [**non**-goals for the KEEP](https://github.com/Kotlin/KEEP/blob/improved-resolution-expected-type/proposals/context-sensitive-resolution.md#the-issue-with-overloading), in particular relating to improving resolution in calls. Please read that section carefully.