2018-02-05 22:33 GMT+01:00 Sijie Guo <guosi...@gmail.com>:

> On Mon, Feb 5, 2018 at 1:04 PM, Enrico Olivelli <eolive...@gmail.com>
> wrote:
>
> > Il lun 5 feb 2018, 18:11 David Rusek <d...@streaml.io> ha scritto:
> >
> > > It sounds like we didn't do anything with the info for a long time.
> > Enrico,
> > > I'm glad you're looking at it! Are you planning on filing some issues
> > > related to interpreting the coverage data and improving it?
> > >
> >
> >
> > It was long time ago when I started to experiment with bookkeeper
> codebase.
> > We had some problems and I had other priorities.
> > I will try to resume this thread on next weeks I think that the culprit
> of
> > our problems was the way we were performing BC tests.
> > I have not much time so I will go on one step at a time, if you have time
> > any help is appreciated.
> >
> > First step will be to test locally jacoco and then to restore the CI jobs
> >
>
> just one suggestion when you are trying to restore CI jobs, please start
> with a separate CI job and let the CI job run for a while to ensure it
> doesn't have any side efforts before enforcing it on the other jobs.
>
>

Create a new PR to upgrade Code Coverage configuration
https://github.com/apache/bookkeeper/pull/1129

This is an example of current master report:
https://coveralls.io/jobs/33538314

we are at 61 % (using default metrics)

Enrico


>
> >
> > Ideally I would like to have some automated way to keep an eye on BK and
> > maybe (not sure it is a big deal) to perform code coverage analysis even
> on
> > PRs.
> >
> > One big problem is that our corpus of tests is very heavy as most of the
> > tests start a new cluster.
> > Recently we started to use mockito in order to perform narrower unit
> > testing.
> >
> > Stay tuned
> >
> > Enrico
> >
> > Enrico
> >
> > >
> > > -Dave
>

Reply via email to