succ, pred, inc and dec accept some integers instead of int

[ringabout]

supersedes #22457

[konsumlamm]

This is better than #22457, but I'd still prefer nim-lang/checksums#6. +/+= don't allow different argument types for good reason imo. This also breaks the JS backend, which relies on the second argument always being an int. This could be solved by only using the magic if the second argument is an int (i.e. adding an overload for SomeInteger, that doesn't use the magic).

[ringabout]

Unrelated CI failure.

[ringabout]

This could be solved by only using the magic if the second argument is an int (i.e. adding an overload for SomeInteger, that doesn't use the magic).

Feel free to improve that in the following PRs. I need this to unblock #22453

[konsumlamm]

Again, what's the problem with nim-lang/checksums#6?

[ringabout]

Well, not like I object to my own PR. Either one is fine to merge.

[Araq]

The signature is inc(Ordinal, int) etc, I don't know why this is debatable.

[konsumlamm]

The signature is inc(Ordinal, int) etc, I don't know why this is debatable.

That's how it originally was and currently is (without this PR or #22457), but checksums apparently relied on it also accepting uint64 as second argument, while the signature was inc[T: Ordinal, V: Ordinal](T, V).

[ringabout]

nim-lang/checksums#6 was merged