August 2009 - Wine-Devel - list.winehq.org

Re: [2/4] ntdll: Add ProcessExecuteFlags handling to NtQueryInformationProcess
by Alexandre Julliard Aug. 14, 2009

Aug. 14, 2009

Michael Karcher <wine(a)mkarcher.dialup.fu-berlin.de> writes: > This patch adds todo_wine statements to the test case, as this patch > enables the test case to run on wine. It would be cleaner to store the state in process.c rather than adding yet another internal function to retrieve it. -- Alexandre Julliard julliard(a)winehq.org

1 0

d3d9 testsuite crash since 1.1.27 in DestroyContext
by Joerg-Cyril.Hoehle＠t-systems.com Aug. 14, 2009

Aug. 14, 2009

Hi, did anyone notice this recent bug? The d3d9 and d3d8 testsuites fail for me since 1.1.27. My crash backtrace is similar to that already visible in test.winehq.org "WinXP-AMD2600-1.1.27" (submitted by somebody else): http://test.winehq.org/data/b79639699766e7c8f3e236f3fef54e5daac0e5f9/index_… I'm using Ubuntu Intrepid with Intel i915 graphics. Is that worth regresion testing or is somebody already looking at this? Note that Austin English's machines show no failure in test.winehq.org, as well as Chris Ahrendt's! The bug is still present in git as of 2009-08-12. Regards, Jörg Höhle

2 1

[PATCH 3/4] kernel32: implement CancelIoEx (try 2)
by Mike Kaplinskiy Aug. 13, 2009

Aug. 13, 2009

1 0

Re: add skeleton for new attrib.exe application
by Juan Lang Aug. 13, 2009

Aug. 13, 2009

Hi EA, some comments on your tiny little app: +#define WIN32_LEAN_AND_MEAN + +#include "config.h" + +#include <windows.h> +#include <stdio.h> +#include <stdlib.h> You don't need to define WIN32_LEAN_AND_MEAN for compiling with Wine's headers, and you don't need to include <windows.h>, either. Also, since you never use any of the standard library functions, you don't need to include <stdio.h> or <stdlib.h>. A useful starting point when compiling with Wine's headers is: #include <stdarg.h> #include "windef.h" #include "winbase.h" +int main( int argc, char *argv[] ) +{ + exit(0); As Jeff pointed out in bug 18059, just return 0 rather than calling exit(0). I don't think you need to use wmain rather than main for an empty application, but if you ever intend to work with the command line, wmain is probably better. --Juan

1 0

Re: [trans] update Chinese translation for winecfg
by Henri Verbeet Aug. 13, 2009

Aug. 13, 2009

2009/8/13 Cheer Xiao <xiaqqaix(a)gmail.com>: > Henri Verbeet<hverbeet(a)gmail.com> told me that I should provide my > full name in the file as well as in my e-mail account name. Here is > the new patch. > Actually, that patch doesn't apply. Looks like it's against a version of Wine from before commit 4e046a6759b059b3c424480bea052af01cc85ebe on Aug. 5. You should rebase it against the current git version.

1 0

D3D shader assembler (round 2)
by Matteo Bruni Aug. 13, 2009

Aug. 13, 2009

2009/8/10 Stefan Dösinger <stefandoesinger(a)gmx.at>: > Hi, > A few comments - mostly things I haven't spotted earlier. Hi, I fixed the code, almost completely following your reviews (thank you, of course). I'm reporting the differences with your suggestions: > >> --- /dev/null >> +++ b/dlls/d3dx9_36/asmshader.h >> --- /dev/null >> +++ b/dlls/d3dx9_36/asmshader_private.h > I think it doesn't need two include files, especially since both of them are > d3dx9-private anyway. The question is, is it better to just merge them into > d3dx9_36_private.h, or have assembler things in a separate file? I tend > towards putting everything into d3dx9_36_private.h, but I don't have a strong > opinion about that. > I merged everything into d3dx9_36_private.h >> +%option prefix="asmshader_" >> + buffer = asmshader__scan_string(text, asm_ctx.yyscanner); > It seems that flex adds an underscore to the prefix anyway, so I guess its > fine to remove the one from the prefix option. > Removing the underscore from the flex prefix produces a (better) asmshader_scan_string function but also an ugly asmshaderlex_init. I decided to keep the underscore. >> + WINED3DSIO_NOP = 0, >> + WINED3DSIO_MOV = 1, >> + WINED3DSIO_ADD = 2, >> and others > Can you give them a different name prefix, to make it clear that they don't > have any relationship with wined3d any more? > I named them BWRITERxxxx (the same for wine_shader, now it is named bwriter_shader, enforcing the idea that that represents the bytecode writer interface). Is it ok? >>+DWORD SlWriteBytecode(const wine_shader *shader, int dxversion, DWORD > **result) { >> +static struct bc_writer *create_writer(DWORD version, DWORD dxversion) { > Does it still need the dxversion? > It's not needed, but I kept it in anticipation of D3D10 support. Can be removed anyway. > asmparser_srcreg_vs_3: >> + if(src->type != WINED3DSPR_TEMP && src->type != WINED3DSPR_INPUT && >>+ src->type != WINED3DSPR_CONST && src->type != WINED3DSPR_ADDR && >> + src->type != WINED3DSPR_CONSTBOOL && src->type != > WINED3DSPR_CONSTINT && >> + src->type != WINED3DSPR_LOOP && src->type !=WINED3DSPR_SAMPLER && >> + src->type != WINED3DSPR_LABEL && src->type != WINED3DSPR_PREDICATE) > A swich-case statement or a lookup table would probably be nicer. (A lookup > table has the disadvantage that you still have to check min and max > separately though) Looking better at this, I noticed that I wasn't checking the register number, while native has this check. So I reworked this part, and now the code is as the attached file. Regarding Henri's review, I followed everything, except keeping my_alloc/realloc/free (now asm_alloc and c.) functions, as it seems Stefan likes them very much :) Supposing my changes are ok, what should I do now? Wait for reviews of the rest? Or maybe send the updated code? In this case, should I send it again in wine-devel or have a try with wine-patches? Cheers, Matteo Bruni

3 2

Font can be changed in Dictation Box after today's git
by Susan Cragin Aug. 12, 2009

Aug. 12, 2009

Yippee!! A big step forward for usability. I'll mark the bug fixed tonight.

1 0

Re: [1/2] gdiplus: Implement GdipGetVisibleClipBounds
by Henri Verbeet Aug. 12, 2009

Aug. 12, 2009

2009/8/12 Andrew Eikum <aeikum(a)codeweavers.com>: > + GpStatus stat = Ok; The initialization is redundant, since you always initialize stat before using it. > + if((stat = GdipCreateRegionRect(&wnd_rectf, &wnd_rgn) != Ok)) > + return stat; I don't think that does what you want it to do.

3 2

Re: [2/2] gdiplus/tests: Add tests for GdipGetVisibleClipBounds
by Nikolay Sivov Aug. 12, 2009

Aug. 12, 2009

Andrew Eikum wrote: > --- > dlls/gdiplus/tests/graphics.c | 244 > +++++++++++++++++++++++++++++++++++++++++ > 1 files changed, 244 insertions(+), 0 deletions(-) I suggest to split this: --- + ok(rectf.X == exp.X && + rectf.Y == exp.Y && + rectf.Width == exp.Width && + rectf.Height == exp.Height, + "Expected clip bounds (%0.f, %0.f, %0.f, %0.f) to be the size of " + "the screen (%0.f, %0.f, %0.f, %0.f)\n", + rectf.X, rectf.Y, rectf.Width, rectf.Height, + exp.X, exp.Y, exp.Width, exp.Height); --- to four separate tests to simplify customization in possible failures in future (mean broken() cases on native).

1 0

re: winetricks, dotnet3 and winsxs
by Dan Kegel Aug. 12, 2009

Aug. 12, 2009

In http://www.winehq.org/pipermail/wine-devel/2009-August/077858.html Hin-Tak wrote >... >The last part (missing charset) seems to be some harmless thing winetrick does - instead of gtk it calls good old xlib message to pop up a message; I haven't looked at why. the code is warn() { echo "$@" case x"$GUI" in x1) xmessage -center " Winetricks error: $@ " ;; *) ;; esac } i.e. we haven't bothered to use the nicer dialog for errors. Think we should?

2 1