> On Apr 14, 2015, at 9:06 AM, Gaurav Aradhye <gaurav.arad...@clogeny.com> > wrote: > > Hi all, > > I feel that the test code in BVT and Regression tests (smoke and component > folders) should be more consistent. There are feature developers adding tests > into smoke and different test developers adding more tests into component > folder, it is evident that inconsistencies are bound to arise. But with time, > we should be striving to minimize them. > > I have listed few of the inconsistencies below. Feel free to add to the list > if you think of more. > > 1. List Methods: > > If we want to list accounts, some tests use Account.list method, while some > use list_accounts method. Both methods finally do the same thing, but there > are many list methods in common.py file corresponding base class methods in > base.py. This is code duplication. > > I feel that BaseClass.method() is more readable than list_xyz(), or should be > preferred. > > > 2. Checking empty/None lists: > > We already have validateList utility method which checks both empty lists and > None objects. No need to use “isinstance” method, or add additional None > checks. > > > 3. Pep8 issues: > Many files are not pep8 consistent. In the past I have tried to fix pep8 > issues in files and the mission is ON. Efforts need to be put into this. We > already have autopep8 tool which does most of the things. >
+1 that’s really a low hanging fruit and best practice. we should not be merging anything that breaks pep8 we should also add pylint tests checking for pep8 and python of all python code can be done at the high level using something like tox, kind of like the RAT tests > > I will be creating separate issues in JIRA for listed things. Everyone is > welcome to add pull requests for these tasks. > I personally will be putting more efforts into these in upcoming few months. > 2-3 months and we should see most of the tests consistent with each other. > > Regards, > Gaurav >
