nanogui: Thread: Re: GrMoveWindow() inplementation

: I wrote the included program to start testing ideas for a
: nano-windowmanager, but I quickly ran into some bug/features in the
: nano-x implementation.

Dan - I'm moving finally towards fixing the problems with your
demo program.  I've included comments FYI.

: 1.  If you move a window in nano-X, it's children don't move.  They
:     stay in the same position relative to the root window.  This is
:     sub-optimal.

Yes, this does indeed just need the for loop for children like Microwindows
has.  And I'll change Microwindows and Nano-X to work for all grandchildren
as well.


:
: 2.  If you move window A under window B, the server draws on B.  You
:     can demonstrate this by dragging the two windows in the test
:     program over each other.  The behavior is decidedly different,
:     depending on which one is on top.

Yes, I fixed this in the Microwindows api.  The clipping calculation code needed
to be rewritten for Microwindows, and I'm going to have to rewrite it
for Nano-X.  The problem is that the traversal is not quite correct
when setting the global clipping region for the current drawable window.
Now is maybe the time to combine these two sections of code for
Microwindows and Nano-X.  There's a slightly different WINDOW structure
for the two api's, but maybe they should be made the same, so that the clipping
calculation code can be shared between the two api's.  Currently, the
clipping "clip" code is shared, in devclip.c, but the Microwindows clip
code is in winclip.c, and Nano-X in nanox/srvutil.c.



:
: 3.  I haven't tested this, but from looking at winuser.c, it looks
:     like bug #1, is partly in microwindows also.  Unless I miss my
:     guess, MoveWindow() will update the positions of child windows,
:     but will *not* update the position of grandchildren.

Yep

Regards,

Greg