May 2005 - Wine-Devel - list.winehq.org

Regression in Half life
by Stefan Dösinger May 6, 2005

May 6, 2005

Hello, These two patches cause a regression in Half-Life (Version 1.1.1.0): (1)http://www.winehq.org/hypermail/wine-cvs/2005/04/0307.html (2)http://www.winehq.org/hypermail/wine-cvs/2005/04/0308.html Half-life reports "ChoosePixelFormat failed", followed by "The Selected OpenGL mode is not supported by your card" when starting the engine. Wine prints the following error: err:opengl:X11DRV_ChoosePixelFormat glXChooseFBConfig returns NULL (glError: 0) Reversing the above 2 patches fixes the problem. Reversing patch 1 only causes a crash, if I reverse only patch 2 the problem still exists. Cheers, Stefan Here's a +opengl trace trace:opengl:X11DRV_OpenGL_Init GLX is up and running error_base = 77 fixme:ddraw:Main_DirectDraw_SetCooperativeLevel (0x77c71a58)->(00000000,00000008) fixme:ddraw:Main_DirectDraw_SetCooperativeLevel (0x77c71a58)->(00000000,00000013) fixme:mmtime:timeBeginPeriod Stub; we set our timer resolution at minimum fixme:mmtime:timeEndPeriod Stub; we set our timer resolution at minimum fixme:ddraw:Main_DirectDraw_SetCooperativeLevel (0x77c71a58)->(00000000,00000008) fixme:ddraw:Main_DirectDraw_SetCooperativeLevel (0x77c71a58)->(00010028,00000013) fixme:win:LockWindowUpdate (0x10028), partial stub! fixme:mmtime:timeBeginPeriod Stub; we set our timer resolution at minimum fixme:win:LockWindowUpdate ((nil)), partial stub! fixme:ddraw:Main_DirectDraw_SetCooperativeLevel (0x77c71a58)->(00000000,00000008) fixme:ddraw:Main_DirectDraw_SetCooperativeLevel (0x77c71a58)->(00010028,00000013) fixme:x11drv:X11DRV_desktop_SetCurrentMode Cannot change screen BPP from 32 to 16 fixme:x11drv:X11DRV_desktop_SetCurrentMode Cannot change screen BPP from 32 to 16 fixme:x11drv:X11DRV_desktop_SetCurrentMode Cannot change screen BPP from 32 to 16 fixme:win:LockWindowUpdate (0x10028), partial stub! fixme:win:LockWindowUpdate ((nil)), partial stub! fixme:win:LockWindowUpdate (0x10028), partial stub! fixme:win:LockWindowUpdate ((nil)), partial stub! fixme:x11drv:X11DRV_desktop_SetCurrentMode Cannot change screen BPP from 32 to 16 trace:opengl:wgl_ext_initialize_extensions GL version : "1.3.5010 (X4.3.0-8.12.10)". trace:opengl:wgl_ext_initialize_extensions GL exts : "GL_ARB_multitexture GL_EXT_texture_env_add GL_EXT_compiled_vertex_array GL_S3_s3tc GL_ARB_occlusion_query GL_ARB_point_parameters GL_ARB_texture_border_clamp GL_ARB_texture_compression GL_ARB_texture_cube_map GL_ARB_texture_env_add GL_ARB_texture_env_combine GL_ARB_texture_env_crossbar GL_ARB_texture".... trace:opengl:wgl_ext_initialize_extensions GLX exts : "GLX_EXT_visual_info GLX_EXT_visual_rating GLX_EXT_import_context GLX_ARB_get_proc_address GLX_ARB_multisample ". trace:opengl:wgl_ext_initialize_extensions Server GLX exts : "GLX_ARB_multisample GLX_EXT_visual_info GLX_EXT_visual_rating GLX_EXT_import_context GLX_OML_swap_method GLX_SGI_make_current_read GLX_SGIS_multisample GLX_SGIX_fbconfig GLX_ARB_get_proc_address ". trace:opengl:wgl_ext_initialize_extensions Client GLX exts : "GLX_EXT_visual_info GLX_EXT_visual_rating GLX_EXT_import_context GLX_ARB_get_proc_address GLX_ARB_multisample GLX_ATI_pixel_format_float GLX_ATI_render_texture ". fixme:opengl:query_function_pbuffer gl_version is: "1.3.5010 (X4.3.0-8.12.10)" fixme:opengl:query_function_pbuffer glx_exts is: "GLX_EXT_visual_info GLX_EXT_visual_rating GLX_EXT_import_context GLX_ARB_get_proc_address GLX_ARB_multisample" fixme:opengl:query_function_pbuffer gl_version is: "1.3.5010 (X4.3.0-8.12.10)" fixme:opengl:query_function_pbuffer glx_exts is: "GLX_EXT_visual_info GLX_EXT_visual_rating GLX_EXT_import_context GLX_ARB_get_proc_address GLX_ARB_multisample" trace:opengl:wgl_ext_initialize_extensions Supporting following WGL extensions : "WGL_ARB_extensions_string WGL_EXT_extensions_string WGL_ARB_make_current_readWGL_ARB_multisample WGL_ARB_pbuffer WGL_ARB_pixel_format WGL_EXT_swap_control". trace:opengl:X11DRV_ChoosePixelFormat (0x77cc5448,0x1df4964) - size / version : 40 / 1 - dwFlags : PFD_DOUBLEBUFFER PFD_DRAW_TO_WINDOW PFD_SUPPORT_OPENGL - iPixelType : PFD_TYPE_RGBA - Color : 24 - Red : 0 - Green : 0 - Blue : 0 - Alpha : 0 - Accum : 0 - Depth : 32 - Stencil : 0 - Aux : 0 - iLayerType : PFD_MAIN_PLANE err:opengl:X11DRV_ChoosePixelFormat glXChooseFBConfig returns NULL (glError: 0) trace:opengl:wglMakeCurrent ((nil),(nil)) trace:opengl:wglMakeCurrent returning True fixme:win:LockWindowUpdate (0x10028), partial stub! fixme:win:LockWindowUpdate ((nil)), partial stub! fixme:x11drv:X11DRV_desktop_SetCurrentMode Cannot change screen BPP from 32 to 16 fixme:x11drv:X11DRV_desktop_SetCurrentMode Cannot change screen BPP from 32 to 16 fixme:mmtime:timeEndPeriod Stub; we set our timer resolution at minimum

2 2

Google Web Accelerator
by Matthew Mastracci May 6, 2005

May 6, 2005

To indulge my habit of trying the newest Google software releases under Wine, I decided to give the new Google Web Accelerator a shot with the latest CVS. The first issue that I ran into (patches sent to wine-devel!) was the lack of StgCreateStorageEx and StgOpenStorageEx. This seems to be used by Google to store their extra web cache data. Once I replaced the stubs with proper functions, the application started up and started caching data. From what I can tell, it even contacts the Google server for the compressed updates. It doesn't seem to "save time" in any way - I think this might have something to do with the inefficiency within the structured storage implementation (100% CPU for a short period of time as you request pages). My theory is that the application picks the request that returns first - Google's optimized version or the web page's response. It seems to die after a few minutes of browsing with what appears to be an async IO error. The strange thing is that the error code looks like a memory location rather than a winsock error: fixme:winsock:NtStatusToWSAError Status code 77ce16c0 converted to DOS error code 77ce16c0 fixme:winsock:WS2_async_send status: c0000120 [matthew(a)epsilon temp]$ cd /tmp/wine/dlls/winsock I tried to trace through what happens in an asynchronous send operation, but I couldn't figure out how the status of the response was translated back to the "->Internal" field of the OVERLAPPED structure. Any hints? Thanks, Matt.

2 1

New program: getsffile
by Vincent Béron May 6, 2005

May 6, 2005

Sent to wine-devel to get feedback on initial idea and implementation. New program in programs/getsffile (Get Sourceforge File). It downloads a file from Sourceforge download section, finding a suitable mirror by itself. This should solve part of the mirroring problem we have for the Mozilla ActiveX control, to make it easy to install on demand, and can most probably be used by other similar things as well. You can also specify a preferred mirror in the registry (Software\\Wine\\getsffile, key SFPreferredMirror. If the file is not available from that mirror (mirror down, not yet synced, etc.), it'll try another one (up to 3 times). The list of mirrors is got from sf.net, it's not maintained in getsffile. Basically, to use it to get the Mozilla ActiveX control, what needs to be done is to actually put the control on sf (along with a zip containing 2 needed dlls, one of them is msvcp60.dll), and change the download logic in shdocvw to call getsffile for both files rather than do the download itself. The installation part is almost the same (it'd be nice if the zip would be installed at the same time, even nicer if it wouldn't be needed but that's Mozilla's part). Note that it's still not 100% proof, some mirrors give a 404 back and it's not properly detected. Is it possible to detect such things while using URLDownloadToFile? The whole dialog is ripped from shdocvw (Mike McCormack's work), the rest is my own development. Vincent

4 4

Re: Commercial support
by Peter Kovacs May 5, 2005

May 5, 2005

>Rather than set threshholds on capacity, there might be a tiered >arrangement >whereby anybody can get a class D listing for nothing. Class C, B and >A >listings would cost $200, $1000, and $10000. The page would then be >ranked by >listing class, and within listing class by geography. How do you want to determine the different classes. If I pay 10.000$ I get listed on the Premium list? Or is it more that If I have more employees working on wine i pay more for being a official supporter? I am not sure if this is what the winhq should aim at. On the market it could work I do not know because I do not know how great the exsisting demand is. But we should be careful with lists. Every list is a official recommandation. And the wine Project should take care that these People do honestly contribute to wine. A general fee for all is better IMHO. We could make a fee 0f 200$ link the List to a profile where the Companies stats is listed. There we could make a Rubrik like the Company donated over X $ to the project. That would state the closeness and the support. You move up in the list if you collect enough points by producing code, patches and donation. Wich you can check on the profile. How bout that? >3) a link back to winehq.org from there site and not twenty pages into >there site. We could make a Button like Offical Wine support for the Commercial supporter. This button can be placed on the homepage. >7) if a banned party wants re-instatement they must pay a fine of >$25,000 and post a written apology to the community for there actions. It would be better if a official Wine Support Company signs a contract where it agrees to an penalty payment if violating the contract. The violation fee could be differ to the severity of the violation. This sounds serious to me. Because this is a two way road. We promise the company to treat them right and they ensure us to be honest on the project. Of course a contract is a bit more demanding then the simple list but they become the OFFICIAL Partner for wine. I think that is more worth then a fee. Greetings Peter

2 1

Re: Winelib's role in converting Windows applications
by Steven Edwards May 5, 2005

May 5, 2005

Hi Ira, --- Ira Krakow <ikrakow_1999(a)yahoo.com> wrote: > For example, if the application requires PAM > authentication, or a Linux-based help system, these > modules would be separated out and encapsulated as > Winelib objects. I was thinking of using PAM > authentication as a good example, since it works for > any authentication scheme that the application > requires. > > This is the approach I plan to take. I welcome all > feedback. Another good example is from the UI point of view where you don't want users to have to deal with drive letters in a Unix env. Using winelib it would be possible to directly call GTK/GDK or KDE and use browse dialogs. Of course some work seems to be going on to add support for this directly to shell32 so someone else might know suggest a better way. Thanks Steven __________________________________ Yahoo! Mail Mobile Take Yahoo! Mail with you! Check email on your mobile phone. http://mobile.yahoo.com/learn/mail

1 0

Re: PATCH: add test for non existing directory during createfile
by Troy Rollo May 5, 2005

May 5, 2005

On Fri, 6 May 2005 02:59, Marcus Meissner wrote: > I would like to have someone test this test on windows... > > I have the suspicion that Windows might create non-existing directories > in the middle. Testing on Win2k I get ERROR_PATH_NOT_FOUND.

1 0

Re: Regression in Half life
by fenix＠club-internet.fr May 5, 2005

May 5, 2005

2 1

Re: Re: Regression in Half life
by fenix＠club-internet.fr May 5, 2005

May 5, 2005

3 2

DCOM patch
by Mike Hearn May 5, 2005

May 5, 2005

Hi Rob, I've attached the patch I was working on at WineConf. The test cases pass now but I haven't checked it against InstallShield yet: can you look it over? Sorry about the extraneous changes. thanks -mike

1 0

Re: Missing HAVE_V4L2 check
by Maarten Lankhorst May 5, 2005

May 5, 2005

Francois Gouget wrote: > > Maarten, could you fix the following winapi_check warnings: > > dlls/avicap32/avicap32_main.c: HAVE_V4L2 is not declared as a conditional > dlls/avicap32/avicap32_main.c: HAVE_V4L2 is not declared as a conditional > > > The problem is that dlls/avicap32/avicap32_main.c checks for HAVE_V4L2 > but we lack the appropriate check in configure.ac and the macro > declaration in include/config.h. > > HAVE_V4L2 is not declared in wine itself, but by /usr/include/linux/videodev.h, which includes videodev2.h and defines HAVE_V4L2 if V4L2 is available, so in my opinion it doesn't need fixing..

3 3