[mythtv] Auto-detect UI painter

Lawrence Rust lvr at softsystem.co.uk
Thu Feb 17 09:26:49 UTC 2011


On Thu, 2011-02-17 at 08:25 +0800, Mark Kendall wrote:
> On 17 February 2011 02:33, Rob Smith <kormoc at gmail.com> wrote:
> > On Wed, Feb 16, 2011 at 4:12 AM, Lawrence Rust <lvr at softsystem.co.uk> wrote:
> >> Personally, I can't see what's wrong with keeping the current UI
> >> control, defaulting to auto, and allowing the user to force Qt
> >> permanently if they need to.
> >
> > We've talked about ditching the QT painter entirely so I'm not sure
> > allowing the user to force QT permanently and not reporting the bugs
> > with the D3D9 painter to us is a good solution.
> 
> Exactly. Fixing the d3d9 painter is the real issue here.

To put this in perspective I have tested the current MythTV master
branch on 5 Windows systems of various pedigrees - i7 thru AMD Duron
with Intel, NVidia and Radeon graphics on Windows 7, XP and 2K.  I don't
think this mix is unrepresentative of the h/w that users may have.  Only
the most recent Windows 7 system with Intel Ironlake graphics is close
to stable across LiveTV, MythVideo, MythMusic & MythGallery using D3D9
for both UI and video.

Iff the D3D9 implementation were stable across a range of systems and
applications then I would welcome this change.  However, removing the UI
control and Qt fallback now is asking for trouble.

Fixing the D3D9 painter and video renderer are very important and making
them the default is a good way to achieve that but if users can't
fallback to a workable situation when encountering problems then you're
going to lose support.

Stability can be achieved in 2 ways:
- Constrain the range of variables (h/w, OS, video adapters etc).
- Test and debug on a wider set of boundary conditions.

Making it more difficult for Windows users is not the way to getting
more bug reports - it's the way to put them off entirely.

-- Lawrence




More information about the mythtv-dev mailing list