March 2008 - Wine-devel - list.winehq.org

Re: Implementing a new Direct3D function
by Jérôme Gardou 25 Mar '08

25 Mar '08

Le Sunday 09 March 2008 18:35:16 tony.wasserka(a)freenet.de, vous avez écrit : > D3DXGetImageInfoFromFile takes the D3DXIMAGE_INFO structure as parameter, > which has changed between d3dx8 an dd3dx9_36 though. > This means you can't forward the D3DXGetImageInfoFromFile calls from d3dx9 > to d3dx8, (even if it is done like that atm, you should fix that an you > patch) so you'll have to write two versions of each. > Our plan for implementing d3dx was to keep any d3dx code away from wined3d, > so you don't have to change anything about that. > So basically all you have to do is > (1) implement the d3dx8 version of each function > (2) add them to the d3dx8 specfile > (3) implement the d3dx9_36 version of each function > (3) correct the d3dx9_36 specfile (i.e. remove the forwarding to d3dx8) > > The d3dx9_24 to 35 dlls forward their calls to d3dx9_36, so they are okay, > too. > > > Tony > > > > > Unbegrenzter Speicher, Top-Spamschutz, 120 SMS und eigene E-MailDomain > inkl. > http://office.freenet.de/dienste/emailoffice/produktuebersicht/power/mail/i >ndex.html OK, then I'll do this that way. Also, d3dx9 provides support for multiple file format (PNG, TGA, BMP...). I assume I'll have to add those dependencies in the build process. ___________________________________________________________________________ Yahoo! Mail réinvente le mail ! Découvrez le nouveau Yahoo! Mail et son interface révolutionnaire. http://fr.mail.yahoo.com

4 3

New valgrind errors in GdiRealizationInfo?!
by Dan Kegel 25 Mar '08

25 Mar '08

http://kegel.com/wine/valgrind/logs-2008-03-24/vg-gdi32_font-diff.txt shows a number of new warnings, e.g. -fixme:font:GdiRealizationInfo (0x2e8, 0x7f22fc6c): stub! -font.c:1774: Tests skipped: skipping GdiRealizationInfo with truetype font ... + Use of uninitialised value of size 4 + at WineEngGetGlyphOutline (freetype.c:4152) + by GetGlyphOutlineW (font.c:2352) + by GetGlyphOutlineA (font.c:2330) + by enum_truetype_font_proc (font.c:1568) + by FONT_EnumInstance (font.c:589) + by WineEngEnumFonts (freetype.c:3762) + by FONT_EnumFontFamiliesEx (font.c:626) + by EnumFontFamiliesExA (font.c:664) + by func_font (font.c:1714) + by run_test (test.h:406) + by main (test.h:455)

1 0

Patch : fixme:msvcrt:MSVCRT__sopen : pmode 0x01a0 ignored
by Erik de Castro Lopo 25 Mar '08

25 Mar '08

Hi all, My interest in Wine is mainly because it allows me run the test suite for my projects when I use the Linux -> windows MinGW compiler. Currently, the programs in my test suite spit out a lot of fixme warnings like: fixme:msvcrt:MSVCRT__sopen : pmode 0x01a0 ignored I've tracked this down in the code to the function in dlls/msvcrt/file.c. The patch below removes this fixme warning. I beleive that simply removing the warning is valid in this case because the variable being warned about, pmode, is not used after the warning. In fact, the only use of this variable is the FIXME or WARN messages. Cheers, Erik --- dlls/msvcrt/file.c~ 2008-03-25 23:42:47.000000000 +1100 +++ dlls/msvcrt/file.c 2008-03-25 23:44:14.000000000 +1100 @@ -1442,11 +1442,6 @@ pmode = va_arg(ap, int); va_end(ap); - if(pmode & ~(MSVCRT__S_IREAD | MSVCRT__S_IWRITE)) - FIXME(": pmode 0x%04x ignored\n", pmode); - else - WARN(": pmode 0x%04x ignored\n", pmode); - if (oflags & MSVCRT__O_EXCL) creation = CREATE_NEW; else if (oflags & MSVCRT__O_TRUNC) -- ----------------------------------------------------------------- Erik de Castro Lopo ----------------------------------------------------------------- Only people who suck have a problem with elitism.

2 1

Re: Create stub dosx.exe
by Alexandre Julliard 25 Mar '08

25 Mar '08

Justin Chevrier <jchevrier(a)nexicom.net> writes: > Changelog: > > Create stub dosx.exe Why do you need this? -- Alexandre Julliard julliard(a)winehq.org

2 4

Re: Create stub dosx.exe
by Dmitry Timoshkov 25 Mar '08

25 Mar '08

"Justin Chevrier" <jchevrier(a)nexicom.net> wrote: > +#define WIN32_LEAN_AND_MEAN > + > +#include <windows.h> What's the reason to include anything in an empty stub? > +int main(int argc, char *argv[]) > +{ > + return 0; > +} -- Dmitry.

2 1

Re: (resend)imm: implement ImmGetIMEFileNameW and ImmGetIMEFileNameA correct registry lookups
by Alexandre Julliard 25 Mar '08

25 Mar '08

Aric Stewart <aric(a)codeweavers.com> writes: > + if (rc > 0) > + { > + if (uBufLen && lpszFileName) > + rc = WideCharToMultiByte(CP_ACP, 0, bufW, uBufLen, lpszFileName, > + uBufLen, NULL, NULL); > + else /* get the length */ > + rc = WideCharToMultiByte(CP_ACP, 0, bufW, wBufLen, NULL, 0, NULL, > + NULL); The source length should be -1, you don't want to convert the whole buffer. > + static CHAR KeyboardKey[] = "System\\CurrentControlSet\\Control\\Keyboard Layouts"; > + static WCHAR szImeFileW[] = {'I','m','e',' ','F','i','l','e',0}; > + > + HKEY keyboardKey; > + DWORD length; > + DWORD rc; > + CHAR regKey[sizeof(KeyboardKey) + 11]; > + > + wsprintfA( regKey, "%s\\%08x", KeyboardKey, (unsigned)hKL ); > + rc = RegOpenKeyA( HKEY_LOCAL_MACHINE, regKey, &keyboardKey); This should use Unicode like the rest of the function. -- Alexandre Julliard julliard(a)winehq.org

2 1

Re: WineD3D: Always report D16 (fixes bug 12130) [attempt 2]
by Stefan Dösinger 25 Mar '08

25 Mar '08

> > 3dmark2000 broke recently because we aren't offering D16 anymore if a > > card doesn't expose it using GLX. On a Geforce 8800 it isn't offered by > > GLX (because the hardware is 24bit) but D3D9 on Windows offers it, so > > offer it too. Second D3D <= 7 requires it too as at that time there was > > no 24bit z-buffer. The bug in 3Dmark2000 is a bug in ddraw, we do not report 24/32 bit formats in a way the game likes. I have a patch for that deep down in my sending queue that fixes the depth and stencil masks. This patch is still perfectly valid, just pointing out that in case of 3DMark2k it fixes the wrong bug. Other older apps are likely do demand D16, just 3Dmark2000 doesn't.

1 0

Re: [PATCH 7/9] wined3d: Make device palettes dynamically allocated.
by Alexandre Julliard 25 Mar '08

25 Mar '08

Alexander Dorofeyev <alexd4(a)inbox.lv> writes: > + if (PaletteNumber >= This->NumberOfPalettes) { > + NewSize = This->NumberOfPalettes; > + do { > + NewSize *= 2; > + } while(PaletteNumber >= NewSize); > + palettes = HeapReAlloc(GetProcessHeap(), 0, This->palettes, sizeof(PALETTEENTRY[256]) * NewSize); You should be using an array of pointers, not a single array for all palettes, so that you don't need to move around the whole data on each realloc. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 9/9] d3d9/tests: Fix precision problems in shademode_test.
by Stefan Dösinger 25 Mar '08

25 Mar '08

Am Dienstag, 25. März 2008 01:04:10 schrieb Alexander Dorofeyev: > --- > dlls/d3d9/tests/visual.c | 26 +++++++++++++++++++++----- > 1 files changed, 21 insertions(+), 5 deletions(-) The drawback of this is that if a driver behaves differently in D3DSHADEMODE_PHONG this will slip through as imprecision most likely. That is something we have to live with I am afraid

1 0

[Repost] wineasio source now on SourceForge
by Peter L Jones 24 Mar '08

24 Mar '08

Hi all, wineasio, the ASIO to JACK driver for WINE, has been under some development over recent months by various interested parties. Co-ordination has been on a best endeavours basis but really needed something more organised. The project really needed a proper home, too. So I've gathered the various threads of source and committed them into a SourceForge SVN project. Robert Reif's wineasio-0.0.x and the JackLab work on wineasio-0.x have their own branches. The current work on wineasio-0.7.x forms the SVN trunk. wineasio-0.7.x releases (i.e. 0, 1, 2 and 3) are tagged. HEAD currently has two changes pending release (a fix for Nuendo and a possible fix for some xruns being experienced by a 64studio user). wineasio developers and packagers are encouraged to consider using the SourceForge facilities - let me know if you want to be added as a committer and please use the trackers for bugs, feature request, etc. Right now, feedback would be appreciated on the current code in HEAD and around the xrun issued mentioned, more details here: http://forum.jacklab.net/viewtopic.php?f=2&t=726 I spent oh... seconds... considering a name for the SourceForge project. Fortunately "wineasio" was available. You can find the project on the end of this url: http://sf.net/projects/wineasio wineasio would not exist without Robert Reif. Ralf Beck undertook much work for JackLab. William Steidtmann and Nedko Arnaudov had provided bug fixes. I provided the config file handling and tried to combine Ralf and Robert's work (thus kicking off 0.7.x). Many thanks, -- Peter

1 0