Before the test we remmber which test is currently running.
During the test we collect all the calls to steps.
Beside that we persist the names of all steps within this shared
lib itself.
After the test(s) we write a corresponding json file. In fact we
write the file after each test, which is too often. But since we
don't know which test is the last test we can't do better.
The resulting file can be used later on for resolving the plugins
contributing the particular steps.
With that we are able to create a list of required plugins for each
step.
a42e727da2