[igt-dev] ✓ Fi.CI.IGT: success for igt/drm_read: Exercise waking up the next waiter

[Chris Wilson]

Quoting Antonio Argenziano (2019-02-27 21:18:10)
> 
> 
> On 27/02/19 03:48, Patchwork wrote:
> > == Series Details ==
> > 
> > Series: igt/drm_read: Exercise waking up the next waiter
> > URL   : https://patchwork.freedesktop.org/series/57285/
> > State : success
> > 
> > == Summary ==
> > 
> > CI Bug Log - changes from CI_DRM_5664_full -> IGTPW_2531_full
> > ====================================================
> > 
> > Summary
> > -------
> > 
> >    **SUCCESS**
> > 
> >    No regressions found.
> > 
> >    External URL: https://patchwork.freedesktop.org/api/1.0/series/57285/revisions/1/
> > 
> > Possible new issues
> > -------------------
> > 
> >    Here are the unknown changes that may have been introduced in IGTPW_2531_full:
> > 
> > ### IGT changes ###
> > 
> > #### Possible regressions ####
> > 
> >    * {igt at drm_read@short-buffer-wakeup} (NEW):
> >      - shard-apl:          NOTRUN -> FAIL
> >      - shard-snb:          NOTRUN -> FAIL
> >      - shard-kbl:          NOTRUN -> FAIL
> >      - shard-hsw:          NOTRUN -> FAIL
> >      - shard-glk:          NOTRUN -> FAIL
> 
> Is it normal to have the run claim success if a new test is added and fails?

It's expected, let's call it a feature. New tests start as automatically
suppressed, so it falls out of that logic.

Now that Martin has the "its a brand spanking new test" bit worked out,
the next feature would be indeed to identify that an IGT run with a new
test is probably interested in that test result; and indeed such a
series should ideally not be committed until the failure has been
resolved.

It would also be useful if the NOTRUN -> {FAIL, PASS} be reported in the
issues.html so that we didn't have to go searching the full view for
them.
-Chris