Hi Luiz,
Thank you again for the support on this issue. I have just provided a
patch series here:
https://patchwork.kernel.org/project/bluetooth/list/?series=390411
to include test coverage for the new APIs via mgmt-tester. In
addition, as this coverage helped me find a minor bug in returning
remaining adv data size in the MGMT response, I've submitted a fix in
the kernel patch series. Please let me know if there is anything
further I can provide.
Thanks!
Daniel
On Tue, Nov 3, 2020 at 1:25 PM Luiz Augusto von Dentz
<[email protected]> wrote:
>
> Hi Daniel,
>
> On Tue, Nov 3, 2020 at 9:42 AM Daniel Winkler <[email protected]> wrote:
> >
> > Hello Luiz,
> >
> > Thank you for the information. It is good to know that this tool is
> > actively used and that there is a way to skip existing flaky tests.
> > Just for clarification, is this a requirement to land the kernel
> > changes, i.e. should I prioritize adding these tests immediately to
> > move the process forward? Or can we land the changes based on the
> > testing I have already done and I'll work on these tests in parallel?
>
> We used to require updates to mgmt-tester but it seems some of recent
> command did not have a test yet, but if we intend to have the CI to
> tests the kernel changes properly I think we should start to requiring
> it some basic testing, obviously it will be hard to cover everything
> that is affected by a new command but the basic formatting, etc, we
> should be able to test, also tester supports the concept of 'not run'
> which we can probably use for experimental commands.
>
> > Thanks,
> > Daniel
> >
> > On Thu, Oct 29, 2020 at 5:04 PM Luiz Augusto von Dentz
> > <[email protected]> wrote:
> > >
> > > Hi Daniel,
> > >
> > > On Thu, Oct 29, 2020 at 3:25 PM Daniel Winkler <[email protected]> wrote:
> > > >
> > > > Hi Luiz,
> > > >
> > > > Thank you for the feedback regarding mgmt-tester. I intended to use
> > > > the tool, but found that it had a very high rate of test failure even
> > > > before I started adding new tests. If you have a strong preference for
> > > > its use, I can look into it again but it may take some time. These
> > > > changes were tested with manual and automated functional testing on
> > > > our end.
> > > >
> > > > Please let me know your thoughts.
> > >
> > > Total: 406, Passed: 358 (88.2%), Failed: 43, Not Run: 5
> > >
> > > Looks like there are some 43 tests failing, we will need to fix these
> > > but it should prevent us to add new ones as well, you can use -p to
> > > filter what tests to run if you want to avoid these for now.
>
>
>
> --
> Luiz Augusto von Dentz