[pytest-dev] can we remove the deprecated pytest.config variable assignment before 3.0

2016-07-20 Thread Ronny Pfannschmidt
Hi, while working on config.py i just noted we have that pytest.config assignment around in there for compatibility i'd like to remove it - and wonder if its something thats ok to slip into 3.0 or if we should go for deprecation in the 3 series and removal in the 4 series? -- Ronny -- Red Hat

Re: [pytest-dev] can we remove the deprecated pytest.config variable assignment before 3.0

2016-07-20 Thread Bruno Oliveira
Hi Ronny, AFAIK there’s no simple way to workaround it, correct? For example, I believe there might be code out there using pytest.config as a conditional in a pytest.mark.skipif decorator. In this case the change is not a simple find/replace. As a general guideline for deprecation vs removal, IM

Re: [pytest-dev] can we remove the deprecated pytest.config variable assignment before 3.0

2016-07-20 Thread holger krekel
On Wed, Jul 20, 2016 at 16:15 +, Bruno Oliveira wrote: > Hi Ronny, > > AFAIK there’s no simple way to workaround it, correct? For example, I > believe there might be code out there using pytest.config as a conditional > in a pytest.mark.skipif decorator. In this case the change is not a simple

Re: [pytest-dev] can we remove the deprecated pytest.config variable assignment before 3.0

2016-07-20 Thread Floris Bruynooghe
On 20 July 2016 at 21:50, holger krekel wrote: > On Wed, Jul 20, 2016 at 16:15 +, Bruno Oliveira wrote: >> Hi Ronny, >> >> AFAIK there’s no simple way to workaround it, correct? For example, I >> believe there might be code out there using pytest.config as a conditional >> in a pytest.mark.ski

Re: [pytest-dev] can we remove the deprecated pytest.config variable assignment before 3.0

2016-07-20 Thread Bruno Oliveira
On Wed, Jul 20, 2016 at 6:31 PM Floris Bruynooghe [email protected] wrote: On 20 July 2016 at 21:50, holger krekel wrote: > > On Wed, Jul 20, 2016 at 16:15 +, Bruno Oliveira wrote: > >> Hi Ronny, > >> > >> AFAIK there’s no simple way to workaround it, correct? For