Hello Is there a status update on this patch? I am curious to know if it will be merged, and if not, what correction is needed to make it mergeable.
Thanks Jeff On Mon, Sep 3, 2012 at 1:51 PM, Adam Jakubek <ajaku...@gmail.com> wrote: > On Mon, Sep 3, 2012 at 6:29 PM, Stefan Dösinger <stefandoesin...@gmx.at> > wrote: >> >> I skimmed bug #18490, and I'm wondering about two things: Does the app >> create >> a D3DDEVTYPE_REF or D3DDEVTYPE_NULLREF device? Once the device is created, >> what does it to with it? A +d3d log would provide some answers. > > > Ok, this needs a little correction on my part. The game actually creates > D3DEVTYPE_HAL (I somehow previously concluded it wouldn't work, since on my > test machine it required software vertex processing). > Here is the relevant log line: > trace:d3d:wined3d_device_create wined3d 0x235ce558, adapter_idx 0, > device_type 0x1, focus_window 0x20026, flags 0x22, device_parent 0x2284cc0c, > device 0x2284cc14. > > Focus window is the desktop handle. Presentation parameters specify windowed > mode and hDeviceWindow=NULL, so in the end it uses the desktop window. > > It seems to be some kind of short-lived auxiliary device for texture > operations. It's not used for rendering at all. > The only operations on that device I could find is a series of CreateTexture > calls: > trace:d3d9:d3d9_device_CreateTexture iface 0x2284cc08, width 2048, height > 2048, levels 1, usage 0, format 0x15, pool 0x2, texture 0x33be40, > shared_handle (nil). > > They are followed by several LockRect/UnlockRect calls, after which the > textures are released. > The device is released immediately after that. > > Originally I suspected this might be used to generate mip-maps or load > textures with d3dx, but the presence of D3DPOOL_SYSTEMMEM excludes both. > > I admit that it's a very non-standard way of creating a D3D device, since it > can't render anything. However it seems to work for this particular use > case. > > Regards, > Adam Jakubek > > > >