Should ValueTask factor the token value into the GetHashCode result?

[IDisposable]

It looks like ValueTask and ValueTask<TResult> both correctly use the value of the _token value in the implementation of the Equals method and the ValueTask<TResult> implementation.

The value of _token is not used in the non-generic GetHashCode() implementation, which seems suboptimal. In the non-generic version, we just use the _obj?.GetHashCode() value (or zero if null). In the generic version, we just use the _obj field which will be null for any successful synchronous completion, which likely happens a lot.

When _obj references a IValueTaskSource, the token is actually meaningful (and used to call GetStatus() and GetResult(), so if multiple ValueTask objects have the same IValueTaskSource and different _token values then they are definitely distinct object (which the Equals implementation does get right).

The same comments apply to the generic ValueTask<TResult> usages.

The question is:
Do objects of ValueTask get placed in a collection or used in places where GetHashCode returning the same value for ValueTask/ValueType<TResult> objects that have the same IValueTaskSource is problematic? Is putting them into a single collision-heavy bucket an issue?

[teo-tsirpanis]

Storing value tasks long-term, especially those backed by an IValueTaskSource seems like a code smell. Value tasks' purpose is to be awaited immediately. I don't see any actual impact with their implementation of GetHashCode().