March 2009 - Wine-devel - list.winehq.org

how/where to implement case-insensitive fourcc check?
by Joerg-Cyril.Hoehle＠t-systems.com 26 Mar '09

26 Mar '09

Hi, I'm looking for advice about how to properly implement the case-insensitive fourcc check of video formats required by the conformance test msvfw32, i.e. how to eliminate the 2 todo_wine in http://test.winehq.org/data/tests/msvfw32:msvfw.html This would allow the AppDB rating of the application "The Alien Nations / Die Völker" to jump from garbage to silver. Cf. http://bugs.winehq.org/show_bug.cgi?id=15915 http://source.winehq.org/git/wine.git/?a=blob;f=dlls/msvfw32/tests/msvfw.c;… The quick and dirty hack (to make the game run) is to change - if (icinfo && icinfo->fccType != ICTYPE_VIDEO) into + if (icinfo && icinfo->fccType != ICTYPE_VIDEO && icinfo->fccType != mmioFOURCC('V', 'I', 'D', 'C')) in files dlls/iccvid/iccvid.c, likewise in msvidc32/msvideo1.c and msrle32/msrle32.c But I wouldn't even think of submiting that to wine-patches. Somehow a case-insensitive check is required. For instance, msvfw32/msvideo_main.c:ICOpen() uses the case-insensitive helper function compare_fourcc(). But the above may not be the correct place for this check. Any suggestions? Thanks for your help, Jörg Höhle

1 0

google summer of code applicant
by wwinthan＠purdue.edu 26 Mar '09

26 Mar '09

hi i'm Win Than Aung from Purdue university. I have found your project listed on google summer of code website. I would like to ask for a chance to present myself. would you please tell me to whom should i contact for further information? PS: i have been full-time interning at the local companies at Purdue Research Park almost 1 year as software development engineer-intern. PS: some of the projects I have done so far can be found at the following links http://keshunli.byethost15.com/wwinthan/projects.php http://keshunli.byethost15.com/searchwebserver/(php web server) http://keshunli.byethost15.com/wwinthan (simple facebook-like add friends features) thanks winthan

1 0

Re: programs/dxdiag: Create dummy application
by Austin English 25 Mar '09

25 Mar '09

On Wed, Mar 25, 2009 at 5:09 PM, Allen Hair <allen.hair(a)gmail.com> wrote: Howdy Allen Glad to see y'all are finally trying to get this committed. Couple minor things: diff --git a/programs/dxdiag/En.rc b/programs/dxdiag/En.rc new file mode 100644 index 0000000..76e93ff --- /dev/null +++ b/programs/dxdiag/En.rc @@ -0,0 +1,39 @@ +/* + * Notepad (English resources) This isn't notepad ;-). + * + * Copyright 2009 Allen Hair <allen.hair(a)gmail.com> No need to put your e-mail. git log will show that. -- -Austin

1 0

Re: programs/dxdiag: Create dummy application
by Austin English 25 Mar '09

25 Mar '09

On Wed, Mar 25, 2009 at 5:09 PM, Allen Hair <allen.hair(a)gmail.com> wrote: Howdy Allen Glad to see y'all are finally trying to get this committed. Couple minor things: diff --git a/programs/dxdiag/En.rc b/programs/dxdiag/En.rc new file mode 100644 index 0000000..76e93ff --- /dev/null +++ b/programs/dxdiag/En.rc @@ -0,0 +1,39 @@ +/* + * Notepad (English resources) This isn't notepad ;-). + * + * Copyright 2009 Allen Hair <allen.hair(a)gmail.com> No need to put your e-mail. git log will show that. -- -Austin

1 0

Re: [kernel32] Fix inf file parsing
by James Hawkins 25 Mar '09

25 Mar '09

On Wed, Mar 25, 2009 at 8:25 AM, Paul Vriens <paul.vriens.wine(a)gmail.com> wrote: > Hi, > > This fixes bug 17843 but I'm not sure it's a 100% correct. James didn't > change > this just for the fun of it. > If you're unsure of the correct fix, you should write a test case that fails without your patch and succeeds with your patch. -- James Hawkins

2 4

Re: msi: avoid negative index in TABLE_insert_row
by James Hawkins 25 Mar '09

25 Mar '09

On Wed, Mar 25, 2009 at 10:39 AM, Aric Stewart <aric(a)codeweavers.com> wrote: > > fixes bug Bug 17600 > --- > dlls/msi/table.c | 3 ++- > 1 files changed, 2 insertions(+), 1 deletions(-) > Please add a test case that demonstrates what you're trying to fix. If an internal msi function is sending in a negative index, then there's a bug somewhere else. -- James Hawkins

2 5

Re: [4/4] [wined3d] trigger constant reloading when RECT textures change
by Stefan Dösinger 25 Mar '09

25 Mar '09

Am Mittwoch, 25. März 2009 14:52:15 schrieb Tobias Jakobi: A few comments on all the patches. I am a bit busy at university right now, so I didn't have a chance to answer earlier: >From patch 4/4: > + if (!use_ps(stateblock)) { > + TRACE("Non power two matrix multiply fixup\n"); > + glMultMatrixf(((IWineD3DTextureImpl *) stateblock->textures[texUnit])->baseTexture.pow2Matrix); > + } I think a comment would be helpful why this isn't done when pixel shaders are used, as this can't be seen from the context of this code. > + if (GL_TEXTURE_RECTANGLE_ARB == IWineD3DBaseTexture_GetTextureDimensions(stateblock->textures[sampler]) && > + use_ps(stateblock)) { Maybe it would be helpful to add a flag that says if the fixup matrix is identity. WineD3D can emulate np2 textures with padded GL_TEXTURE_2D textures as well. Now I can't think of any hardware that has shaders, but not GL_ARB_texture_rectangle, but checking for GL_TEXTURE_RECTANGLE_ARB to find out if we need a fixup isn't that nice(We might do it in other places too, though) >From patch 1: > + /* Bitmap for texture rect coord fixups (32 samplers max currently). > + * Increase size when MAX_FRAGMENT_SAMPLERS grows. */ > + UINT texrect_fixup; I think there's no point in supporting more than 16 samplers. That's the limit in d3d9 and will never change, and d3d10 makes unconditional NP2 support mandatory. The ps_compile_args structure is memcmp'ed every time the shader is changed, so the size matters. I also recommend to move it to the bottom of the structure, since it will be 0x0000 in most cases, so it prolongs the number of bytes the memcmp has to check before it finds a difference.(ok: In which direction does memcmp search?) General: I also think that you could merge the 4 patches into one or two. I appreciate the effort to make the patches small, but that shouldn't be a self-sufficient purpose. Patches 1 and 2 can certainly be merged, as patch 1 alone doesn't do anything. Patch 4 could be merged into patch 3 because it is required for the code in patch 3 to work correctly. The patches leave out one corner case(Subject for a separate patch though): The use of a vertex shader with fixed function fragment processing. In this case the vertex shader has to take care of the fixup. In GLSL that can conveniently be done in the pixelshader<->vertexshader glue function. It could read the same uniform and multiply the written texture coordinate. Once all the cases work in GLSL, we could experiment with the R and RG float formats from GL_NV_float_buffer, since they're supported on RECT textures only, but may come handy on pre-geforce 8 hardware which doesn't support the ARB_RG extension.

1 0

Re: comctl32: CCS_VERT flips toolbar separators' orientation
by Rob Shearman 25 Mar '09

25 Mar '09

2009/3/17 Igor Tarasov <tarasov.igor(a)gmail.com>: > Currently wine flips orientation of toolbar separators on > BTNS_DROPDOWN, which is odd, since even horizontal toolbar separators > may have this flag turned on. Tests with control spy show that native > comctl flips orientation on CCS_VERT toolbar style, which is logical. > ... > @@ -855,7 +855,7 @@ TOOLBAR_DrawButton (HWND hwnd, TBUTTON_INFO *btnPtr, HDC hdc, DWORD dwBaseCus > /* empirical tests show that iBitmap can/will be non-zero */ > /* when drawing the vertical bar... */ > if ((dwStyle & TBSTYLE_FLAT) /* && (btnPtr->iBitmap == 0) */) { > - if (btnPtr->fsStyle & BTNS_DROPDOWN) > + if (dwStyle & CCS_VERT) > TOOLBAR_DrawDDFlatSeparator (&rc, hdc, infoPtr); > else > TOOLBAR_DrawFlatSeparator (&rc, hdc, infoPtr); You should fix the name of the TOOLBAR_DrawDDFlatSeparator function and the comments above it if you think it should be called as a result of the CCS_VERT window style and not the BTNS_DROPDOWN button style. -- Rob Shearman

2 2

Re: [PATCH 3/6] kernel32: Handle writing empty sections and keys in WritePrivateProfileString.
by Paul Vriens 25 Mar '09

25 Mar '09

James Hawkins wrote: > --- > dlls/kernel32/profile.c | 32 ++++++++++++++++++++------------ > dlls/kernel32/tests/profile.c | 5 +---- > 2 files changed, 21 insertions(+), 16 deletions(-) > > > ------------------------------------------------------------------------ > > Hi James, Sorry if you've already picked the problem with this patch up (bug 17843). I tried looking into this myself but can't come up with a simple answer/fix. On a clean .wine the registry shows a lot of wrong entries (just look under HKCR, HKCU or HKLM). Any idea? -- Cheers, Paul.

1 0

Re: advapi32: stub TreeResetNamedSecurityInfo for the IE8 installer
by Alexandre Julliard 25 Mar '09

25 Mar '09

Aric Stewart <aric(a)codeweavers.com> writes: > diff --git a/dlls/advapi32/security.c b/dlls/advapi32/security.c > index b6d51d8..4013171 100644 > --- a/dlls/advapi32/security.c > +++ b/dlls/advapi32/security.c > @@ -58,6 +58,9 @@ static BOOL ParseStringSecurityDescriptorToSecurityDescriptor( > LPDWORD cBytes); > static DWORD ParseAclStringFlags(LPCWSTR* StringAcl); > > +typedef VOID (*FN_PROGRESS) ( LPWSTR pObjectName, DWORD Status, > + OUT PPROG_INVOKE_SETTING pInvokeSetting, PVOID Args, BOOL SecuritySet); > + Please put this in the appropriate header. -- Alexandre Julliard julliard(a)winehq.org

2 1