[Biopython-dev] Online tools to track our test coverage & code quality

[Peter Cock]

On Fri, Nov 27, 2015 at 9:30 AM, Peter Cock <p.j.a.cock at googlemail.com> wrote:
> On Thu, Nov 26, 2015 at 5:06 PM, Peter Cock <p.j.a.cock at googlemail.com> wrote:
>> On Thu, Nov 26, 2015 at 4:01 PM, Peter Cock <p.j.a.cock at googlemail.com> wrote:
>>> Hi all,
>>>
>>> I've just been looking at https://codecov.io/ for logging unit test coverage,
>>> which can be used an add on on top of the GitHub and TravisCI with just
>>> a few small changes to the .travis.yml file ...
>>
>> I'd forgotten about another option in this area, https://coveralls.io
>>
>
> Tiago raised this on the list back in May,
> http://lists.open-bio.org/pipermail/biopython-dev/2015-May/020934.html
>
> We had a good discussion about what to do with the online tests and
> those will many external dependencies - which otherwise get poor
> test coverage scores in our current TravisCI configuration. On a
> related point, I've been meaning to try the Mac OS X support under
> TravisCI out.
>
> I started comparing coveralls.io and codecov.io yesterday, both using
> the same coverage data as input. It took a little tweaking to get them
> to report about the same overall coverage score:
>
> e.g. using my test branch and personal repository:
>
> https://codecov.io/github/peterjc/biopython?branch=codecov
> https://coveralls.io/github/peterjc/biopython?branch=codecov
>
> It seems that for now codecov.io does a much better job of combining
> all the TravisCI build targets to give overall coverage - while coveralls.io
> seems to list each file once per build target.
>
> My overall impression so far is a preference for the CodeCov.io
> interface - it also seems to have a more responsive website.
> We could of course setup both.
>
> Peter

No one has objected, and I found it useful at least, so I'm planing
to enable the CodeCov.io coverage integration in the next few
days. Currently I do not plan to enable Coveralls.io for the main
Biopython repository.

Any comments/objections/better ideas?

Peter

P.S. See also the code quality report from www.quantifiedcode.com
which Tiago has been exploring:

http://lists.open-bio.org/pipermail/biopython-dev/2015-December/021179.html