Questions about bug reporting for DX10 games

[Das Letzte Einhorn]

In this forum post : http://forum.winehq.org/viewtopic.php?t=9260 , it has been suggested that we test DX10 games and that we report issues.

However, with only about 5% of the entire implementation done, we can expect a lot of things to not work properly. Hence, I was wondering how would it be best to proceed with reporting. Is it recommended to test with winetricks d3dx10, or clean wine? What should we be looking for in our logs?

I have 4-5 DX10 games which I can test, therefore having a clear method could help me and others to report what the devs need.

Thanks!

[James McKenzie]

Das Letzte Einhorn wrote:

In this forum post : http://forum.winehq.org/viewtopic.php?t=9260 , it has been suggested that we test DX10 games and that we report issues.

However, with only about 5% of the entire implementation done, we can expect a lot of things to not work properly. Hence, I was wondering how would it be best to proceed with reporting. Is it recommended to test with winetricks d3dx10, or clean wine? What should we be looking for in our logs?

I have 4-5 DX10 games which I can test, therefore having a clear method could help me and others to report what the devs need.

Since I'm the one that suggested this:

1. Check the changelog that comes with each release (in the ANNOUNCE
file).

2. If there is a fix for DirectX10 functionality, run your programs.
If something that worked, doesn't file a regression bug and expect to be
asked to do a regression test.

3. If the new function does NOT work, advise the developers on the
wine-devel list that the function does not work in program XYZ. There
may be more that needs to be done. Again, expect to be asked to file a
bug report with an ATTACHED terminal log file.

4. This should be one of the criteria for Wine 1.4. Making
productivity programs and games work well is one of this project's goals....

James McKenzie

[Das Letzte Einhorn]

Sounds good. For now, should we report each missing functions which the games require to start? (e.g. http://bugs.winehq.org/show_bug.cgi?id=24013 , that I filled this morning)

[James McKenzie]

Das Letzte Einhorn wrote:

Sounds good. For now, should we report each missing functions which the games require to start? (e.g. http://bugs.winehq.org/show_bug.cgi?id=24013 , that I filled this morning)

If using native DirectX10 fixes the missing feature problem, yes. Also,
please list any and all affected programs (since you have a few.) This
will make developers aware of what features will give the 'biggest bang
for the work'.

Also, for each program please create or update entries in the
Applications Database (I know that I'm 'preaching to the choir', but
others read these message threads as well.)

Hopefully, adding DirectX9/10 functions will reduce/eliminate the need
to use Microsoft's DirectX products and move us more towards a Microsoft
free world.

James McKenzie

[DanKegel]

http://bugs.winehq.org/show_bug.cgi?id=24013 has some good points:
it's not a direct duplicate of any other bug,
it mentions which game is affected,
it is linked to that game's appdb entry,
and it seems to be assigned to the right component.
So I'm glad to see you do testing like this, even if it might be a while
before Wine gets around to implementing these functions.

It could be better:
you could give the exact command used
you could give the winetricks command to work around the problem
you could give the sha1sum of the dvd (more useful with downloads)

But I'm not quite sure it's valid as written, since if you run
AssassinsCreed_Dx9.exe, that game *doesn't* really require
dx10. (Right?)

[r2rX]

Should a similar approach be given in regards to testing DX11? I'm not sure if WINE interfaces with OpenGL4 yet....

r2rX

[Das Letzte Einhorn]

http://bugs.winehq.org/show_bug.cgi?id=24013 has some good points:
it's not a direct duplicate of any other bug,
it mentions which game is affected,
it is linked to that game's appdb entry,
and it seems to be assigned to the right component.
So I'm glad to see you do testing like this, even if it might be a while
before Wine gets around to implementing these functions.

It could be better:
you could give the exact command used
you could give the winetricks command to work around the problem
you could give the sha1sum of the dvd (more useful with downloads)

But I'm not quite sure it's valid as written, since if you run
AssassinsCreed_Dx9.exe, that game *doesn't* really require
dx10. (Right?)

In the case of Assassin's Creed, the game has a DX9 and a DX10 mode. They are separate executables. I think that for testing purposes, it is good to use the DX10 executable.

I will add some of the comments you suggested.

[Das Letzte Einhorn]

New food for thought with World in Conflict

http://bugs.winehq.org/show_bug.cgi?id=24063

[James McKenzie]

r2rX wrote:

Should a similar approach be given in regards to testing DX11? I'm not sure if WINE interfaces with OpenGL4 yet....

It does not. MacOSX's native X11 does not support OpenGL4. Maybe the
MacPorts/XQuartz version does. I don't know about
FreeBSD/OpenBSD/Solaris either.

James McKenzie

[Charles Davis]

On 8/21/10 7:23 PM, James McKenzie wrote:

r2rX wrote:

Should a similar approach be given in regards to testing DX11? I'm not
sure if WINE interfaces with OpenGL4 yet....

It does not. MacOSX's native X11 does not support OpenGL4. Maybe the
MacPorts/XQuartz version does.

Nope. And the reason is that Mac OS X itself doesn't support OpenGL 4.
It doesn't even support OpenGL 3.x yet. We'll have to wait for at least
10.7 (and I know it's coming soon).

Chip