> On March 30, 2016, 2:38 a.m., Greg Mann wrote: > > BenM and I discussed this while I was working on the previous patches: > > since the code paths for the authenticated and unauthenticated cases are > > the same once the handler is reached, it shouldn't be necessary to > > explicitly test the full functioning of the endpoints in both cases. We > > test the full functionality of the endpoints with authentication disabled, > > then we test that requests that don't authenticate properly are refused > > when authentication is on. This should test all of the code paths through > > the handlers, since we don't actually do anything with the principal > > currently. What do you think?
Yes, I agree. The authenticated cases are already covered implicitly and the authenticator differs between "bad credentials" vs "missing credentials" -- which is already covered elsewhere. I'd then probably only remove the `badCredentials` as they're unused. - Jan ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/45434/#review125996 ----------------------------------------------------------- On March 29, 2016, 6:36 p.m., Jan Schlicht wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/45434/ > ----------------------------------------------------------- > > (Updated March 29, 2016, 6:36 p.m.) > > > Review request for mesos, Adam B and Greg Mann. > > > Repository: mesos > > > Description > ------- > > The current authentication tests of the `/files/` endpoint only test the case > of missing credentials. This commit adds the additional case of wrong > credentials. > > > Diffs > ----- > > src/tests/files_tests.cpp 2f0f84cd78e7f382b70b97c8711f71cea64ce0a5 > > Diff: https://reviews.apache.org/r/45434/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Jan Schlicht > >
