Boost logo

Boost :

Subject: Re: [boost] marking regression test failures as expected
From: Andrey Semashev (andrey.semashev_at_[hidden])
Date: 2014-05-06 09:46:52


On Tuesday 06 May 2014 15:38:59 Andrzej Krzemienski wrote:
> 2014-04-30 12:05 GMT+02:00 Daniel James <dnljms_at_[hidden]>:
> > On 30 April 2014 10:57, Andrzej Krzemienski <akrzemi1_at_[hidden]> wrote:
> > > Hi,
> > > I noticed in the regression tests results taht some failing tests are
> > > marked as "fail*" in greyish color indicating that the failure is
> >
> > expected
> >
> > > along with the explanation why. E.g.,
> >
> > http://www.boost.org/development/tests/develop/developer/output/teeks99-03
> > a-Ubuntu12-04-64-boost-bin-v2-libs-optional-test-optional_test_ref-test-gc
> > c-4-4-debug.html>
> > > Question: how can I indicate that the new failure is expected in the
> > > same
> > > manner?
> >
> > It's specified in the super project at
> > 'status/explicit-failures-markup.xml'. I think you'll need to create a
> > pull request.
>
> I am trying to learn the syntax of this file and I cannot understand one
> thing. When I look at tests results here:
> http://www.boost.org/development/tests/master/developer/optional.html,
> one test case
> (optional_test_ref<https://github.com/boostorg/optional/blob/master/test/op
> tional_test_ref.cpp>) is marked as expected failure for toolset "gcc- 4.4",
> but in the file 'status/explicit-failures-markup.xml'
> I cannot find a corresponding entry for this toolset. How does the testing
> framework know, then?

I think, regression markup is used from develop, even if master branch is
tested.


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk