[PATCH 3/3] d3drm: Implement IDirect3DRMViewport*::Clear. (v3)

[Aaryaman Vasishta]

On Tue, Aug 9, 2016 at 9:23 PM, Henri Verbeet <hverbeet at gmail.com> wrote:

> On 9 August 2016 at 03:20, Aaryaman Vasishta <jem456.vasishta at gmail.com>
> wrote:
> > v2: Remove some redundancies and add more tests. Handling of Clear
> without depth surface is still a bit weird as wined3d_device_clear seems to
> return an error on a device with no depth surface since the start vs device
> whose depth surface was deleted after it was created, in which case it
> (oddly) returns successfully, even having access to and successfully
> clearing the depth surface.
> >
> What does "deleted" and "surface" mean in the above? If it's just
> releasing the references to the depth/stencil texture and view, note
> that the device state keeps a reference to view that wouldn't be
> released until you set a different view or NULL.
>
> By "deleted" I mean calling DeleteAttachedSurface on the render target
surface, telling it to delete the depth surface.
See the tests below the comment "/* Create device without depth surface
attached */".

What's happening is that Clear is returning WINED3DERR_INVALIDCALL if you
pass a device without a depth surface attached on creation, but if you pass
a device with the depth surface attached (and later on deleted via
DeleteAttachedSurface), then it returns WINED3D_OK. It's weird because
wined3d_device_set_depth_stencil_view should have set fb.depth_stencil =
NULL, but it's not reaching that part for some reason.

Sorry for not being clear about this earlier.

Cheers,
Aaryaman
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.winehq.org/pipermail/wine-devel/attachments/20160809/b14f1627/attachment.html>