ResourceScope attachment

[Maurizio Cimadamore]

On 29/03/2021 11:12, Maurizio Cimadamore wrote:
> Now, with ResourceScope we can probably set up a cleanup action which 
> checks some atomic counter (per library) and unloads the library when 
> the counter reaches zero. But it's not a slamdunk. 

Tl;dr - I think there is a tension between two cases:

a) I can buy the argument that: Maurizio, now we have a new and powerful 
way to express lifecycles, we do we have to bother with attachment 
objects? Can't we just use scopes when we previously would have started 
mocking with reachability?
b) how do we make some of the functionalities provided by the Foreign 
API "friendly" to use cases adopting implicit deallocation?

As for (b) we can now safely dereference segments that are implicitly 
allocated, and we can safely pass these segments to functions. But how 
do we deal with a _library_ whose scope is implicit? How do we make sure 
it is not deallocated prematurely?

I think in a lot of these questions is there is an underlying assumption 
that we don't care about (b), but we do. If implicit allocation only 
worked _some_ times, but not others, then I wouldn't see a lot of point 
for having it in the API in the first place?

Maurizio