xorg-prop-drivers-testers team mailing list archive

[Bryce Harrington <bryce@xxxxxxxxxxxxx>]

On Wed, Apr 21, 2010 at 06:59:34AM +0200, Sean Boran wrote:
> A few additional notes:
> - why was suspend and hibernate not tested too, problems with this were also
> relate to nvidia I believe?

Traditionally the kernel team has done extensive suspend/resume and
hibernate testing, so I did not feel it would be necessary to duplicate
their efforts.

> - when testing from one week to the next, it would be good to know what has
> actually changed in the driver (changelog). i.e. understand the expected
> evolution of the test results

Not a bad idea, although for testing purposes what we were mainly
looking for was not testing of the drivers themselves (presumably they
were quite well tested by NVIDIA and AMD before release), but rather
their integration with other components in Ubuntu (e.g. upstart, gdm,
plymouth, etc.)  History has shown that sometimes changes in these
secondary components can lead to problems unique to the proprietary
video drivers.  So, just knowing the driver changelog probably wouldn't
give you as strong an insight as you'd want.

Also, I'm not sure we have the manpower to write a weekly changes
highlights list - most people knowledgable enough to write such a thing
probably would be more effective focusing that time on reviewing and
following up on bug reports instead.  But if there were a
non-developer volunteer technical enough to follow the changes who could
write it, I think it could work out well.

> - linked could be provide to frequent issues, which are expected failure
> such as nvidia/xrandx, nvidia/hotplug and rotation support?

Agreed.  It would be handy if the isotracker software could highlight
"known issues" a bit more clearly for folks.  Also to differentiate
"serious regressions" from issues that aren't regressions but just
hadn't been filed before.

> - tell users to be prepared to spend 2-4 hrs and not 1h per week..

Thanks, this is good data to know for next time.

Bryce