From: Raffi Enficiaud (raffi.enficiaud_at_[hidden])
Date: 2020-04-13 18:38:35
On 13.04.20 17:52, Edward Diener via Boost 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. To me it is nonsense that a perfectly logical change to
> software should lead to some code coverage "failure", and that therefore
> the change should be deemed incorrect. I admit I do not know what the
> criteria are for code coverage but I do not see code coverage
> determining whether any change to a library is incorrect or not.
Even if the code coverage is set as a "required check" to have in order
to enable the merge, an admin of that repo can always bypass those (even
if the branch is protected).
A problem to me is that most of the repos that have code coverage today
did not really participate in having it: coverage tests were given
during the effort to have Travis based CI. How coverage is calculated is
unknown to me (and in my case why Boost.Test numbers are so bad).
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk