32 bpp cursors?

[Dan Kegel]

http://bugs.winehq.org/show_bug.cgi?id=4273 points to a
patch set that implements 32 bit per pixel cursors
and a bunch of other cursor stuff.

Looks like the patch got dropped by the author, though,
and since it makes server changes, it's going to be hard
to get past Alexandre.

Is anyone interested in picking this up?
- Dan

[Roderick Colenbrander]

> http://bugs.winehq.org/show_bug.cgi?id=4273 points to a
> patch set that implements 32 bit per pixel cursors
> and a bunch of other cursor stuff.
>
> Looks like the patch got dropped by the author, though,
> and since it makes server changes, it's going to be hard
> to get past Alexandre.
>
> Is anyone interested in picking this up?
> - Dan

I believe the problem is not that Henri has dropped the patches. The problem was that Alexandre didn't have time for reviewing the patches. The patches are very, very, very big and chance big parts of Wine. For this reason Alexandre is reluctant in applying it.

Roderick

[Dan Kegel]

On 2/16/07, H. Verbeet wrote:
> > http://bugs.winehq.org/show_bug.cgi?id=4273 points to a
> > patch set that implements 32 bit per pixel cursors
> > and a bunch of other cursor stuff.  [What's up with it?]
>
> I believe the problem is not that Henri has dropped the patches.
> The problem was that Alexandre didn't have time for reviewing the patches.
> The patches are very, very, very big and change big parts of Wine.
> For this reason Alexandre is reluctant in applying it.

Is it neccessary to move cursors into the server first?
Could you break off some of the features and
do them first, e.g. 32 bpp cursors?
- Dan

[H. Verbeet]

On 17/02/07, Dan Kegel  wrote:
> Could you break off some of the features and
> do them first, e.g. 32 bpp cursors?
> - Dan

My plan was to redo the set for local cursors only, but there's also
wined3d that needs attention :-)

[Duane Clark]

Dan Kegel wrote:
> http://bugs.winehq.org/show_bug.cgi?id=4273 points to a
> patch set that implements 32 bit per pixel cursors
> and a bunch of other cursor stuff.
>
> Looks like the patch got dropped by the author, though,
> and since it makes server changes, it's going to be hard
> to get past Alexandre.
>
> Is anyone interested in picking this up?

Well, no server changes are needed just to implement 32 bits per pixel. 
There are rather simple changes needed in the file 
dlls/winex11.drv/mouse.c in create_cursor(), and the places that need to 
be changed are cleverly marked by "switch (ptr->bBitsPerPixel)" ;)

I tried installing Google Sketchup, but could not get that to install. 
That said, it looks like 32 bit pixels use 8 bits each of RGB, plus an 
extra 8 bits of "alpha" (which apparently controls transparency). So 
probably you could duplicate the 24 bit case, read out the extra 8 bits 
and just throw them away for now. Standard X cursors don't support 
alpha, AFAIK.

[H. Verbeet]

On 17/02/07, Duane Clark  wrote:
> Well, no server changes are needed just to implement 32 bits per pixel.
> There are rather simple changes needed in the file
> dlls/winex11.drv/mouse.c in create_cursor(), and the places that need to
> be changed are cleverly marked by "switch (ptr->bBitsPerPixel)" ;)
>
> I tried installing Google Sketchup, but could not get that to install.
> That said, it looks like 32 bit pixels use 8 bits each of RGB, plus an
> extra 8 bits of "alpha" (which apparently controls transparency). So
> probably you could duplicate the 24 bit case, read out the extra 8 bits
> and just throw them away for now. Standard X cursors don't support
> alpha, AFAIK.

The only thing that really needs cursors to be in the server is
passing cursor/icon handles around between processes.