>> X is another story.  [...] 1024x768 and 1280x1024.  Incidentally,
>> those are the only resolutions where the width is the same as
>> linebytes.  Coincidence or not?  I'm still trying to find out
>> exactly how X pixels map to screen pixels at the broken resolutions.

> I finished testing X.  And as I suspected, it doesn't know abou the
> gaps that are required in video memory when linebytes is greater than
> the true width.

This sounds like fairly trivial brokenness in the DDX layer.  Check the
code which calls cfbSetupScreen and/or cfbFinishScreenInit; the third
and seventh arguments should not be identical.  (The third argument is
the screen width; the seventh is the screen stride.)

At least they're the third and seventh args as of X11-R6.4.  UTSL....

If the server isn't using cfb, (a) why not :-), and (b) then it's even
less excusable, since the ddx code must then itself be handling the
stuff cfb would normally handle.  Unfortunately it's likely to be
harder to fix in that case, as the assumption that stride==width may be
much more pervasive.  If you can send me, or point me at, the source
your server is built from, I can have a stab at finding the relevant
code.  (Note that I have no macppc machines running headed; I can't
actually try it myself.)

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B