build_client() now accepts a needs_diff= argument, indicating if the

test needs to perform diff functionality.
By default, this is False, and attempting to call diff() will result

in an error.
By setting to True, diff tool calculations will be performed and

cached, preventing the tests from having to account for any diff checks

in any spies. A missing compatible tool will cause the test to be

skipped.

Unit tests pass.
Tested with a client failing a diff tool check. Saw the test was skipped.