On Tue, Aug 18, 2015 at 2:16 PM, David Fetter <david@fetter.org> wrote:
> I'm given to understand that this tight coupling is necessary for
> performance. Are you saying that it could be unwound, or that testing
> strategies mostly need to take it into account, or...?
I'm just saying that we shouldn't expect to find a magic bullet test
framework that solves all these problems. Without restructuring code,
which I don't think is really feasible, we won't be able to have good
unit test coverage for most existing code.
It might be more practical to start using such a new tool for new code
only. Then the new code could be structured in ways that allow the
environment to be mocked more easily and the results observed more
easily.
--
greg