-Dprism.dirtyopts

[Richard Bair]

>> Also worth noting is that prism.dirtyops is not API -- it really only existed to help people analyze whether graphical glitches with it turned on would disappear when turned off -- ironic that it is failing in the opposite manner :-).
> 
> I think he is saying that they appear with dirtyopts=true.
> They disappear *without* dirtyopts=false

Right, I think I said the same thing -- the normal operation is to turn dirty opts OFF and get things looking right (if there was a bug when they were turned on), ironic that in this case you have to turn them ON to get things looking right!