Caching of results often fails

[apaolillo]

Caching the results (with continuing=true) option often fails, and does not compose well with modern benchkit options, such as command wrapper and run hooks. We should find a more general way to support it. Perhaps forcing to enable the hierarchy of results would be a first step into this direction.

[aaronbog]

This is something I was struggling with as well, trying to find a way to label what has happened for command debugging.

Taking a look at the result folder and its structure might be in order.
Adding some metadata of sorts ,perhaps.

[apaolillo]

I think with the revamp of the internal API, with the execution engine, when things will be decoupled it will be easier to fix.