September 2010 - Wine-devel

Re: Release plans
by Luis Carlos Busquets Pérez 13 Sep '10

13 Sep '10

As a user that only uses wine for playing games built for the Windows platform what I would like the most is d3d10, d3d10.1 and d3d11. It seems that these API's are more similar between themselves than what is the difference between d3d9 and d3d10.

1 0

Re: Release plans
by Dan Kegel 13 Sep '10

13 Sep '10

Looks like Alexandre must have fired up his time machine again :-) And as long as it's up and running, how about a look forward at the 1.4 release plans?

5 6

Re: d3d9/tests: Test NVDB (depth bounds test) behavior
by Henri Verbeet 13 Sep '10

13 Sep '10

This is fine, except for the 8 space indentation. I wonder though if perhaps it would be more interesting to start with a quad with a depth gradient from 0.0 to 1.0, bottom to top, then set the min/max depth bounds, and draw another fullscreen quad. That way you could easily verify the cutoffs are in roughly the right spots.

1 0

Re: wined3d: Add NVDB support
by Henri Verbeet 13 Sep '10

13 Sep '10

2010/9/12 Ričardas Barkauskas <miegalius(a)gmail.com>: > Looks reasonable in general, though there are a few minor issues: > + /* Depth bound test. To query if the card supports instancing CheckDeviceFormat with the special Instancing doesn't have anything to do with this. > + union { > + DWORD d; > + float f; > + } zmin, zmax; Four space indent is enough. You do this in the test patch as well. > + if(stateblock->renderState[WINED3DRS_ENABLEADAPTIVETESSELLATION]) > + FIXME("(WINED3DRS_ENABLEADAPTIVETESSELLATION,%d) not yet implemented\n", stateblock->renderState[WINED3DRS_ENABLEADAPTIVETESSELLATION]); Note that you can just call state_tessellation() here. > { STATE_RENDER(WINED3DRS_ADAPTIVETESS_W), { STATE_RENDER(WINED3DRS_ENABLEADAPTIVETESSELLATION), NULL }, WINED3D_GL_EXT_NONE }, > + { STATE_RENDER(WINED3DRS_ENABLEADAPTIVETESSELLATION), { STATE_RENDER(WINED3DRS_ENABLEADAPTIVETESSELLATION), state_tessellnvdb }, EXT_DEPTH_BOUNDS_TEST }, You have two extra spaces here.

1 0

re: [PATCH 2/7] d3dx9_36/tests: Added HLSL test suite
by Dan Kegel 13 Sep '10

13 Sep '10

Henri wrote: > You still have basic formatting errors like trailing spaces, mixing > tabs and spaces, inconsistent "*" placement, etc. > ... unnecessary cast... > wine-patches reviews are supposed to catch non-trivial interactions with > other parts of the code, not trivial whitespace errors. Looking at http://wiki.winehq.org/SubmittingPatches, I didn't see a nice, concise checklist of things to check for, so I added one at the top, and linked to Henri's post. Improvements welcome. - Dan

3 2

Suggestion for the Game Explorer implementation
by אלעד 13 Sep '10

13 Sep '10

I have a little suggestion: integrate the Game Explorer with the Games menu of the desktop, so that games i install in wine will be added to the Games menu automaticlly.

2 1

ptrace, pokerstars and Ubuntu 10.10
by Michael Fox 13 Sep '10

13 Sep '10

Hi all, There's a problem running wine under the new Ubuntu 10.10 (Maverick). They have enabled kernel security to prevent processes from using ptrace. This affects any wine programs which use ptrace which is surprisingly common. For example I found this bug because PokerStars wouldn't connect to the network. I have filed this as a bug with Ubuntu, but they might choose not to fix it depending on how committed they are to this new security measure. It is filed as: https://bugs.launchpad.net/ubuntu/+source/wine1.2/+bug/636278 There is a simple workaround and it is to add the ptrace capability to wineserver using the command: sudo setcap cap_sys_ptrace=eip /usr/bin/wineserver I tried to submit this as a bug for wine, but it was closed because the maintainer felt there was nothing the wine people could do about the way Ubuntu sets their security. I would like to make the following suggestions: 1. Add a winetricks trick to fix wine on Maverick by executing: "sudo apt-get install setcap; sudo setcap cap_sys_ptrace=eip /usr/bin/wineserver" 2. Add an error message in wine when ptrace fails with error EPERM. This message should suggest the winetrick so the user will find it quickly. In my case it took quite a while to figure out exactly what was going wrong because there is no more terminal output when ptrace fails than when it succeeds. -- - Michael

4 5

Re: [PATCH 2/3] wined3d: store amount of video memory in gpu_description_table
by Roderick Colenbrander 13 Sep '10

13 Sep '10

Hi Jesse, I think the 465 change should go in after the gpu table patch. The gpu table patch could be modified but then it would essentially contain two patches (makes regression testing more complicated, though I doubt that would be an issue here). It is trivial to submit a patch after this is in. If you want I can do that. Roderick On Sat, Sep 11, 2010 at 9:34 AM, Evil Jay <evil(a)eternaldusk.com> wrote: > Roderick, > > Would you add the GTX 465 (1GB) to this table? That way the patch I > submitted to add the 465 just 30 minutes earlier can be safely ignored. > > -Jesse > > >

2 1

Running GDI32 tests for Wine on MacOSX 10.6
by James McKenzie 13 Sep '10

13 Sep '10

On 9/7/10 11:37 AM, doh123 wrote: > James Mckenzie wrote: >> I agree with both Ryan and you, LD_LIBRARY_PATH is supposed to do absolutely NOTHING under MacOSX. However, Wine needs it for some strange reason. If I leave it out, the built dependency libraries are not found for freetype and fontconfig (and I know that fontconfig is included with MacOSX 10.6 now.) Otherwise, I get a series of freetype cannot be found, install freetype 2.0.5 or higher. I'll look at the Fink setup script to see how they did this. >> > This error can be fixed with DYLD_FALLBACK_LIBRARY_PATH... > > I've uploaded three logging files to pastebin.com: Part #1 of running the gdi32 tests with LD_LIBRARY_PATH set to: $HOME/darwine/wineEnv/usr/lib:/usr/X11R6/lib http://pastebin.com/NyrQxE4Q Part # 2 of the same test http://pastebin.com/dF1PwvHJ Log of running the same tests with DYLD_LIBRARY_PATH set to: $HOME/darwine/wineEnv/usr/lib:/usr/X11R6/lib and DYLD_FALLBACK_LIBRARY_PATH set to: /usr/lib:$HOME/darwine/wineEnv/usr/lib:/usr/X11R6/lib http://pastebin.com/i3Mfh7NB The $HOME/darwine/wineEnv/usr/lib directory contains unaltered builds of the dependencies for Wine (freetype, libgphoto2, etc.) that are separate from Fink builds. Any clue as to why the LD_LIBRARY_PATH is working and the DYLD paths are not? Thank you.

1 0

Re: kernel32: add a stub for GetSystemDEPPolicy()
by Marcus Meissner 12 Sep '10

12 Sep '10

On Sun, Sep 12, 2010 at 12:50:39PM -0500, Austin English wrote: > -- > -Austin > diff --git a/dlls/kernel32/kernel32.spec b/dlls/kernel32/kernel32.spec > index 8a33013..43e339b 100644 > --- a/dlls/kernel32/kernel32.spec > +++ b/dlls/kernel32/kernel32.spec > @@ -624,6 +624,7 @@ > @ stdcall GetSystemDefaultLCID() > @ stdcall GetSystemDefaultLangID() > @ stdcall GetSystemDefaultUILanguage() > +@ stdcall GetSystemDEPPolicy() > @ stdcall GetSystemDirectoryA(ptr long) > @ stdcall GetSystemDirectoryW(ptr long) > @ stdcall GetSystemInfo(ptr) > diff --git a/dlls/kernel32/process.c b/dlls/kernel32/process.c > index c1816a8..5990943 100644 > --- a/dlls/kernel32/process.c > +++ b/dlls/kernel32/process.c > @@ -3385,3 +3385,14 @@ DWORD WINAPI WTSGetActiveConsoleSessionId(void) > FIXME("stub\n"); > return 0; > } > + > +/********************************************************************** > + * GetSystemDEPPolicy (KERNEL32.@) > + */ > +DEP_SYSTEM_POLICY_TYPE WINAPI GetSystemDEPPolicy(void) > +{ > + FIXME("stub\n"); > + return 0; better return one of the Enum values here ... "AlwaysOff" probably, instead of 0 (AlwaysOn). Ciao, Marcus > +} > + > + > diff --git a/include/winbase.h b/include/winbase.h > index ea10696..d8dba1d 100644 > --- a/include/winbase.h > +++ b/include/winbase.h > @@ -1230,6 +1230,13 @@ typedef struct tagHW_PROFILE_INFOW { > DECL_WINELIB_TYPE_AW(HW_PROFILE_INFO) > DECL_WINELIB_TYPE_AW(LPHW_PROFILE_INFO) > > +typedef enum _DEP_SYSTEM_POLICY_TYPE { > + AlwaysOn, > + AlwaysOff, > + OptIn, > + OptOut > +} DEP_SYSTEM_POLICY_TYPE; > + > /* Event Logging */ > > #define EVENTLOG_FULL_INFO 0 >

3 2