On 02/17/2014 11:54 AM, Chris Evich wrote: > Since the test-runner pseudo-mocks <autotest + control + run_tests()> > the practical fix-point is in run_tests and the test runner. Both of > them are able to determine the super-classes required, and therefor are > able to pass those references down (where their base-class interface can > be used).
...err looking at the runner and virt.py more, there is quite a lot of duplicate functionality between: virt.py class virt(test.test) and virttest/standalone_test.py class Test(object) Within these classes, references are held to both the env-stuff and the TP-stuff. So I see two approaches (maybe there are others): A: 1) De-duplicate Test & Virt 2) Add shared behavior whereby env_process calls are given a reference to the main subject-subclass to use (i.e. libvirt_vm.VM, qemu_vm.VM, or whatever). -or- 2) Take over responsibility for instantiating the main test subject and handing it off / storing it in Env. 3) env_process goes about it's merry-way, relying on the BaseVM (or BaseWhatever) interface on whatever super-class (or instance) it receives. B: Similar to A2, except we duplicate the functionality of 2-3 in both virt() and Test() (or bring a shared implementation in by some other method). -- Chris Evich, RHCA, RHCE, RHCDS, RHCSS Quality Assurance Engineer e-mail: cevich + `@' + redhat.com o: 1-888-RED-HAT1 x44214 _______________________________________________ Virt-test-devel mailing list [email protected] https://www.redhat.com/mailman/listinfo/virt-test-devel
