[Piglit] Seperate NOTRUN and SKIP from fixes and regressions

[Marek Olšák]

I think there is a bug. It looks like subtest results of one of the
tests are also added to all the following tests in the "disabled"
page, making the page so huge that there are ~125000 test results. You
can reproduce the problem by generating a summary from these 2 result
files:

http://people.freedesktop.org/~mareko/result1.tar.gz
http://people.freedesktop.org/~mareko/result2.tar.gz

Marek

On Tue, Mar 18, 2014 at 9:34 PM, Dylan Baker <baker.dylan.c at gmail.com> wrote:
> Cool, I was mostly hoping that you and Daniel would look at them and see if
> they worked for you and didn't break your workflow.
>
> - Dylan
>
> On Tuesday, March 18, 2014 02:34:57 Marek Olšák wrote:
>> I cannot review this, because I'm not familiar with the code, but the
>> last patch looks good to me and I like the overall idea.
>>
>> Marek
>>
>> On Mon, Mar 17, 2014 at 8:06 PM, Dylan Baker <baker.dylan.c at gmail.com>
> wrote:
>> > This series separates the SKIP and NOTRUN status changes from fixes and
>> > regressions and puts them into separate categories, enabled and
>> > disabled.
>> >
>> > This makes two workflows possible where previously only one was:
>> > 1) creating baselines that are full tests, and then using small focused
>> >
>> >    runs to compare specific tests
>> >
>> > 2) running full comparisons, and being able to see when new tests are
>> >
>> >    enabled or tests are disabled
>> >
>> > Patches 1-3 add test to demonstrate current behavior or to model future
>> > behavior
>> >
>> > Patch 4 makes NOTRUN and SKIP a special subclass which when compared
>> > using numerics (not == or !=) will resolve False
>> >
>> > Patch 5 Plumbs it all in