Re: [PATCH 3/5] ddraw/tests: Add a IDirect3DDevice2 GetCaps test (try 2)

[Saulius Krasuckas]

* On Tue, 6 Dec 2011, Stefan Dösinger wrote:
> try 2: Add infrastructure to rerun every test with new d3d objects.
>
> dlls/ddraw/tests/d3d.c |  211 ++++++++++++++++++++++++++++++++++++++++++++++++
>  1 files changed, 211 insertions(+), 0 deletions(-)

Hi Stefan,

this is 
http://source.winehq.org/git/wine.git/commitdiff/7ae81ba378c1633e574f130461b...

The patch introduced one failure on 64-bit XP running on virtual machine 
with VMware SVGA II" driver v11.6.0.35:
http://test.winehq.org/data/95f81d10c9c852b72c6513eca0a2d0e066457f9e/xp_wtb-...
http://test.winehq.org/data/1fa1ab54376bace57f78d27ac13b7229caa56a2e/xp_wtb-...

d3d.c:5075: Test failed: CreateDevice failed: 88760091.

Platform is i386 WOW64 -- probably 64-bit XP running 32-bit Winetest:
http://test.winehq.org/data/1fa1ab54376bace57f78d27ac13b7229caa56a2e/xp_wtb-...

Would mind having a look and fixing the failure:)?

S.

[PATCH 3/5] ddraw/tests: Add a IDirect3DDevice2 GetCaps test (try 2)

[Stefan Dösinger]

Am Samstag, 14. Januar 2012, 11:21:15 schrieb Saulius Krasuckas:
> http://test.winehq.org/data/1fa1ab54376bace57f78d27ac13b7229caa56a2e/xp_wtb
> -wxpx64-32/ddraw:d3d.html
>
> d3d.c:5075: Test failed: CreateDevice failed: 88760091.

This is DDERR_INVALIDPIXELFORMAT. The code doesn't provide a pixelformat for 
the offscreen surface, which means the screen's format should be used. What 
color depth is this machine runnning at?

My generic guess is that adding a broken(hr == DDERR_INVALIDPIXELFORMAT) is 
the best thing we can do here.

[PATCH 3/5] ddraw/tests: Add a IDirect3DDevice2 GetCaps test (try 2)

[Saulius Krasuckas]

* On Sat, 14 Jan 2012, Stefan Dösinger wrote:
> Am Samstag, 14. Januar 2012, 11:21:15 schrieb Saulius Krasuckas:
> > http://test.winehq.org/data/1fa1ab54376bace57f78d27ac13b7229caa56a2e/xp_wtb
> > -wxpx64-32/ddraw:d3d.html
> >
> > d3d.c:5075: Test failed: CreateDevice failed: 88760091.
>
> This is DDERR_INVALIDPIXELFORMAT. The code doesn't provide a pixelformat for 
> the offscreen surface, which means the screen's format should be used. What 
> color depth is this machine runnning at?

I know nothing about this machine as it belongs to Wine Test Bot.

Maybe we do need additional trace() for it (like it is done in 
d3d9:visual, opengl32:opengl, dsound:propset or winmm:wave to find out 
model/name of an appropriate hardware specifics).

Stefan, would GetDeviceCaps( GetDC(NULL), BITSPIXEL ) do the trick?
> My generic guess is that adding a broken(hr == DDERR_INVALIDPIXELFORMAT) 
> is the best thing we can do here.

I am all for broken() as it seems to be no other guy able to investigate 
this in more depth.

Even more -- I am against testing 3D related things inside virtual 
machines at all:)

S.