Bug 38840 - Adding coverage analysis to unit tests

[John Smith]

On Fri, Aug 24, 2012 at 10:07 PM, Philipp Riemer <ruderphilipp at gmail.com> wrote:
>
> As far as I learned in my Softw. Eng. courses, the main intent of this
> type of code coverage is _not_ to show which parts are under test but
> primarily to point out which parts are not tested at all so far.
>
Yes, of course.


>
> And as one can (now fortunately) see there are still quite some lines
> of code that are not touched during the tests... But now we all have a
> much better overview what parts are exactly missing tests -- which at
> least from my perspective is much better than just guessing and gut
> feeling! Thank you very much for all your great work so far, John!
>
Im still not convinced 'my' coverage report shows this *exactly*.



>
> Of course, in future, every bug should get a unit test (at best even
> before starting to fix it) so that regressions are easier get caught
> ;-)
>
In the ideal world, yes, of course.
;)



> In addition, it would be also good, to have two reports: (1) with only
> the unit test coverage and (2) one where all test, including
> integration tests etc., were executed.
Huh ? what ?



Regards,

John Smith.