Re: Code coverage tests

From: Emil Dotchevski (emildotchevski_at_[hidden])
Date: 2020-04-13 18:36:40

I'd add that if someone contributes a high quality PR, it's better for the
community if the library maintainers accept it rather than reject it based
only on gaps in a coverage report.

On Mon, Apr 13, 2020 at 11:21 AM Edward Diener via Boost <
boost_at_[hidden]> wrote:
>
> On 4/13/2020 1:17 PM, Glen Fernandes via Boost wrote:
> > On Mon, Apr 13, 2020 at 12:44 PM Edward Diener via Boost
> > <boost_at_[hidden]> wrote:
> >>
> >> A number of libraries have code coverage/project tests in Github. I am
> >> not against such tests but I am against the idea that "failing" such a
> >> test, whatever that means, should prevent a PR from being merged into
> >> a Boost library.
> >
> > Edward, are you referring to Boost libraries where a maintainer has
> > specifically chosen to have Code Coverage part of the CI? Or are there
> > libraries where the maintainer has not made this decision but their
> > libraries CI automatically involve coverage because they use some
> > shared CI configuration (boost-ci)?
>
> I am just creating a valid PR for a Boost library. How that library uses
> code coverage is not my problem. I just do not want the validity of a PR
> being affected by how code coverage works for a library. I do realize
> that Github still allows a maintainer to merge a PR even when some CI
> testing fails. I just do not want code coverage, rather than correctness
> of code, determining whether a PR is merged or not. I am certainly not
> going to change a correct PR based on a code coverage report, unless
> that report shows me somehow that my change is logically incorrect, but
> I doubt that code coverage can do that.