September 2006 - Wine-Devel

Page Fault when try to start Railroad Tycoon
by Roland Kaeser Sept. 10, 2006

Sept. 10, 2006

Hello I'm currently on testing my apps for 0.9.20 and getting the following error when try to run a cleanly installed RailRoad Tycoon 3 (on new wine dir). Can somebody help me to figure out the problem? I use a Matrox Parhelia with Tripple-Monitor Configuration. Thanks Roland Kaeser err:x11drv:X11DRV_CreateWindow invalid window width 1919249263 err:x11drv:X11DRV_CreateWindow invalid window height 1952540704 err:ole:CoInitializeEx Attempt to change threading model of this apartment from multi-threaded to apartment threaded fixme:d3d:IWineD3DDeviceImpl_CreateAdditionalSwapChain The app requests more than one back buffer, this can't be supported properly. Please configure the application to use double buffering(=1 back buffer) if possible err:x11settings:X11DRV_ChangeDisplaySettingsEx No matching mode found! (desktop) fixme:d3d:IWineD3DDeviceImpl_GetAvailableTextureMem (0x18b400) : stub, simulating 64MB for now, returning 64MB left fixme:d3d:IWineD3DDeviceImpl_ValidateDevice (0x18b400) : stub fixme:d3d:IWineD3DDeviceImpl_ValidateDevice (0x18b400) : stub wine: Unhandled page fault on read access to 0x00001874 at address 0x7e03d9c1 (thread 000e), starting debugger... WineDbg starting on pid 0xd Unhandled exception: page fault on read access to 0x00001874 in 32-bit code (0x7e03d9c1). Register dump: CS:0073 SS:007b DS:007b ES:007b FS:003b GS:0033 EIP:7e03d9c1 ESP:003100a8 EBP:00000001 EFLAGS:00210206( - 00 - RIP1) EAX:00001860 EBX:7e3b6efc ECX:7c29daf0 EDX:00000000 ESI:00000000 EDI:7c2957e0 Stack dump: 0x003100a8: 7df65d78 7c29da6c 7c29ee94 00000000 0x003100b8: 3f800000 7e04ac27 00310340 ffffffff 0x003100c8: 7e3b6efc 7c2bf3d0 7e40b7ac 00310224 0x003100d8: 7dfb58ba 7c2957e0 7c2c11ec 7e1f5fb5 0x003100e8: 00000001 00000000 ffffffff 7e3b6efc 0x003100f8: 00000000 7e3b6efc 00000000 00000000 fixme:ntdll:RtlNtStatusToDosErrorNoTeb no mapping for c0000119 Backtrace: =>1 0x7e03d9c1 in libgl.so.1 (+0x1199c1) (0x7e03d9c1) 2 0x00000000 (0x00000000) 0x7e03d9c1: movl 0x14(%eax),%edx Modules: Module Address Debug info Name (91 modules) PE 400000-da2000 Deferred rt3c PE 21100000-21164000 Deferred mss32 PE 30000000-30072000 Deferred binkw32 ELF 7bf00000-7bf03000 Deferred <wine-loader> ELF 7d543000-7d559000 Deferred midimap<elf> \-PE 7d550000-7d559000 \ midimap ELF 7d57f000-7d597000 Deferred msacm32<elf> \-PE 7d590000-7d597000 \ msacm32 ELF 7d597000-7d5d7000 Deferred wineoss<elf> \-PE 7d5a0000-7d5d7000 \ wineoss ELF 7d6ed000-7d6f7000 Deferred libnss_files.so.2 ELF 7d6f7000-7d81d000 Deferred libcrypto.so.0.9.8 ELF 7d81d000-7d85a000 Deferred libssl.so.0.9.8 ELF 7d85a000-7d878000 Deferred libcups.so.2 ELF 7d90d000-7d93f000 Deferred uxtheme<elf> \-PE 7d910000-7d93f000 \ uxtheme ELF 7d941000-7d95d000 Deferred imm32<elf> \-PE 7d950000-7d95d000 \ imm32 ELF 7d9a3000-7d9a6000 Deferred libxrandr.so.2 ELF 7d9a6000-7d9ae000 Deferred libxrender.so.1 ELF 7d9ae000-7d9b3000 Deferred libxfixes.so.3 ELF 7d9b3000-7d9bc000 Deferred libxcursor.so.1 ELF 7d9bc000-7d9d9000 Deferred ximcp.so.2 ELF 7d9d9000-7d9db000 Deferred xlcutf8load.so.2 ELF 7d9db000-7da6a000 Deferred winex11<elf> \-PE 7d9f0000-7da6a000 \ winex11 ELF 7dc36000-7dc55000 Deferred libexpat.so.1 ELF 7dc55000-7dc8e000 Deferred libfontconfig.so.1 ELF 7dc8e000-7dca0000 Deferred libz.so.1 ELF 7dca0000-7dd0d000 Deferred libfreetype.so.6 ELF 7dd0d000-7dd27000 Deferred wsock32<elf> \-PE 7dd10000-7dd27000 \ wsock32 ELF 7dd27000-7dd40000 Deferred version<elf> \-PE 7dd30000-7dd40000 \ version ELF 7dd40000-7dd6d000 Deferred ws2_32<elf> \-PE 7dd50000-7dd6d000 \ ws2_32 ELF 7dd6d000-7ddb9000 Deferred dsound<elf> \-PE 7dd80000-7ddb9000 \ dsound ELF 7ddb9000-7ddf0000 Deferred dinput<elf> \-PE 7ddc0000-7ddf0000 \ dinput ELF 7deae000-7df24000 Deferred libglu.so.1 ELF 7df24000-7e40c000 Export libgl.so.1 ELF 7e40c000-7e503000 Deferred libx11.so.6 ELF 7e503000-7e511000 Deferred libxext.so.6 ELF 7e511000-7e529000 Deferred libice.so.6 ELF 7e529000-7e532000 Deferred libsm.so.6 ELF 7e533000-7e547000 Deferred lz32<elf> \-PE 7e540000-7e547000 \ lz32 ELF 7e547000-7e55f000 Deferred dinput8<elf> \-PE 7e550000-7e55f000 \ dinput8 ELF 7e55f000-7e613000 Deferred wined3d<elf> \-PE 7e570000-7e613000 \ wined3d ELF 7e613000-7e638000 Deferred d3d8<elf> \-PE 7e620000-7e638000 \ d3d8 ELF 7e638000-7e6c0000 Deferred winmm<elf> \-PE 7e640000-7e6c0000 \ winmm ELF 7e6c0000-7e6f0000 Deferred winspool<elf> \-PE 7e6d0000-7e6f0000 \ winspool ELF 7e6f0000-7e7c0000 Deferred comctl32<elf> \-PE 7e700000-7e7c0000 \ comctl32 ELF 7e7c0000-7e7d3000 Deferred libresolv.so.2 ELF 7e7d3000-7e7f1000 Deferred iphlpapi<elf> \-PE 7e7e0000-7e7f1000 \ iphlpapi ELF 7e7f1000-7e841000 Deferred rpcrt4<elf> \-PE 7e800000-7e841000 \ rpcrt4 ELF 7e841000-7e8d6000 Deferred ole32<elf> \-PE 7e850000-7e8d6000 \ ole32 ELF 7e8d6000-7e92d000 Deferred shlwapi<elf> \-PE 7e8e0000-7e92d000 \ shlwapi ELF 7e92d000-7ea15000 Deferred shell32<elf> \-PE 7e940000-7ea15000 \ shell32 ELF 7ea15000-7eab2000 Deferred comdlg32<elf> \-PE 7ea20000-7eab2000 \ comdlg32 ELF 7eab2000-7eaf8000 Deferred advapi32<elf> \-PE 7eac0000-7eaf8000 \ advapi32 ELF 7eaf8000-7eb03000 Deferred libgcc_s.so.1 ELF 7ebe3000-7ec99000 Deferred gdi32<elf> \-PE 7ec00000-7ec99000 \ gdi32 ELF 7ec99000-7edda000 Deferred user32<elf> \-PE 7ecb0000-7edda000 \ user32 ELF 7ee1c000-7ef23000 Deferred kernel32<elf> \-PE 7ee40000-7ef23000 \ kernel32 ELF 7ef58000-7ef7d000 Deferred libm.so.6 ELF 7ef7d000-7f000000 Deferred ntdll<elf> \-PE 7ef90000-7f000000 \ ntdll ELF b7d71000-b7d76000 Deferred libxxf86vm.so.1 ELF b7d78000-b7d7c000 Deferred libdl.so.2 ELF b7d7c000-b7e9c000 Deferred libc.so.6 ELF b7e9c000-b7eb0000 Deferred libpthread.so.0 ELF b7eb0000-b7fc1000 Deferred libwine.so.1 ELF b7fc2000-b7fde000 Deferred ld-linux.so.2 Threads: process tid prio (all id:s are in hex) 0000000f 00000011 0 00000010 0 0000000d (D) C:\Programme\Railroad Tycoon 3\RT3c.exe 00000013 15 0000000e 0 <==

1 0

Can't compile wine at Linux-ppc
by Peter Lemenkov Sept. 10, 2006

Sept. 10, 2006

Hello, All! The last wine version which at least compiles fine is 0.9.17. Since then I can't even compile Wine. It fails with the next message ==================================== ../../tools/widl/widl -I. -I. -I../../include -I../../include -D__WINESRC__ -D_SHDOCVW_ -t -T shdocvw_v1.tlb shdocvw_v1.idl warning: create_msft_typeinfo: ignoring attr 46 warning: create_msft_typeinfo: ignoring attr 46 warning: add_func_desc: ignoring attr 5 error: encode_type: VT_USERDEFINED - unhandled type 17 make[2]: *** [shdocvw_v1.tlb] Error 2 make[2]: Leaving directory `/home/petro/work/projects/wine/dlls/shdocvw' make[1]: *** [shdocvw] Error 2 make[1]: Leaving directory `/home/petro/work/projects/wine/dlls' make: *** [dlls] Error 2 ==================================== I have no clue how to debug this situation, so I wold ask you: how to fix this strange situation? -- With best regards!

1 0

RE: Setupapi [0/6] : ReactOS setupapi patches
by Hervé Poussineau Sept. 9, 2006

Sept. 9, 2006

Hi, > As setupapi merge from ReactOS has been a recurrent demand > [1], I'm doing this set of patches. > > [...] > > The goal of this set of patches is to set a solid ground to > implement lots of missing functions in setupapi. > I've tried to cut them in smallest possible parts. > Additionnaly, this patchset implements SetupDiEnumDeviceInfo > and (mostly) SetupDiGetClassDevsExW > > [...] Still one week, and I got absolutly no reaction about my patches. Is something wrong with them? Hervé

3 2

Re: shell32[1/2]: document the shell32 "mini-COM" functions
by Frank Richter Sept. 9, 2006

Sept. 9, 2006

On 08.09.2006 21:11, Mikołaj Zalewski wrote: > This patch adds the explanation why many shell32 functions are similar > to ole32 ones - something I've been wondering for some time. It is not > copied from MSDN. Maybe you've seen it already, but this blog post talks about the various allocators: http://blogs.msdn.com/oldnewthing/archive/2004/07/05/173226.aspx -f.r.

2 1

Re: Socket resets randomly
by James Hawkins Sept. 9, 2006

Sept. 9, 2006

On 9/8/06, Ryan Miller <rmiller(a)auctionpay.com> wrote: > Index: ChangeLog > =================================================================== > RCS file: /home/wine/wine/ChangeLog,v > retrieving revision 1.120 > diff -u -r1.120 ChangeLog > --- ChangeLog 24 Aug 2006 14:43:25 -0000 1.120 > +++ ChangeLog 8 Sep 2006 22:49:55 -0000 > @@ -1,3 +1,8 @@ > +2006-09-08 Ryan Miller <rmiller(a)auctionpay.com> & Zach Gorman <gormanjz(a)hotmail.com> > + > + * dlls/ws2_32/socket.c, dlls/ws2_32/sock.c > + ws2_32: Fixed bug with sockets randomly reseting and tests. > + > 2006-08-24 Alexandre Julliard <julliard(a)winehq.org> > > * configure, configure.ac: > > Index: dlls/ws2_32/socket.c > As someone already said, no changes to Changelog. =================================================================== > RCS file: /home/wine/wine/dlls/ws2_32/socket.c,v > retrieving revision 1.4 > diff -u -r1.4 socket.c > --- dlls/ws2_32/socket.c 8 Sep 2006 10:16:19 -0000 1.4 > +++ dlls/ws2_32/socket.c 8 Sep 2006 22:48:11 -0000 > @@ -2660,6 +2660,11 @@ > TRACE("socket: %04x, level %d, name %d, ptr %p, len %d\n", > s, level, optname, optval, optlen); > > + /* In Windows, if setsockopt succeeds, GetLastError() always returns zero. > + The native wininet.dll expects this behavior, and does not always > + check the return value of setsockopt() first.*/ > + SetLastError(0); > + 0 is ERROR_SUCCESS, and you need to use that define. > /* SO_OPENTYPE does not require a valid socket handle. */ > if (level == WS_SOL_SOCKET && optname == WS_SO_OPENTYPE) > { > > > Index: dlls/ws2_32/tests/sock.c > =================================================================== > RCS file: /home/wine/wine/dlls/ws2_32/tests/sock.c,v > retrieving revision 1.2 > diff -u -r1.2 sock.c > --- dlls/ws2_32/tests/sock.c 23 May 2006 12:49:01 -0000 1.2 > +++ dlls/ws2_32/tests/sock.c 8 Sep 2006 22:48:12 -0000 > @@ -1410,6 +1410,27 @@ > > } > > +static void test_setsockopt_last_error() { > + SOCKET s; > + BOOL bKeepAlive=0; > + struct sockaddr_in addr; > + struct timeval select_timeout; > + > + select_timeout.tv_sec=5; > + select_timeout.tv_usec=0; > + addr.sin_family = AF_INET; > + addr.sin_addr.s_addr = inet_addr(SERVERIP); > + addr.sin_port = htons(SERVERPORT); > + > + s = socket(AF_INET, SOCK_STREAM, 0); > + do_bind(s, (struct sockaddr *)&addr, sizeof(addr)); > + > + SetLastError(1); > + setsockopt(s,SOL_SOCKET, SO_KEEPALIVE, (char *)&bKeepAlive, sizeof(bKeepAlive)); > + ok( GetLastError()==0, "GetLastError() was not cleared by setsockopt\n"); > + closesocket(s); > +} > + You need to SetLastError to some value that you know will never be set by windows, something like 0xdeadbeef. The reasoning is that theoretically it could alter the results of the test. > /**************** Main program ***************/ > > START_TEST( sock ) > @@ -1438,6 +1459,8 @@ > test_WSAStringToAddressW(); > > test_select(); > - > + > + test_setsockopt_last_error(); > + > Exit(); > } > > > > > -- James Hawkins

3 2

Re: GetTextMetrics returns wrong First/Last char index for Symbol fonts
by Dmitry Timoshkov Sept. 9, 2006

Sept. 9, 2006

"Jie Zhu" <jonathan(a)corvu.com.au> wrote: > + ptmA->tmFirstChar = ptmA->tmBreakChar - 2; > + if (ptmW->tmCharSet == SYMBOL_CHARSET) > + { > + ptmA->tmLastChar = ptmW->tmLastChar > 0x0F0FF ? 0x0FF : ptmW->tmLastChar & 0x0FF; > + } > + else > + { > + ptmA->tmLastChar = ptmW->tmLastChar > 255 ? 255 : ptmW->tmLastChar; > + } Probably simple WideCharToMultiByte with CP_SYMBOL and CP_ACP would do a proper thing for tmFirstChar, tmLastChar, tmBreakChar. You have too many fixes for different problems in your patch, that makes it not acceptable. Please send fixes one by one for the problems you have encountered. Also, if you could make your tests use FreeType APIs (like Wine GDI does) instead of home-made ones that would simplify them a lot. It would be nice to see a real Wine test suite for the above conversions. -- Dmitry.

1 0

Symbol demangling and the games crashing
by Vijay Kiran Kamuju Sept. 9, 2006

Sept. 9, 2006

Hi, In the bug reports I have seen that we are facing problems with symbol demangling of '_R' operators and 'Y' datatypes. err:msvcrt:symbol_demangle Unknown operator: _R and err:msvcrt:demangle_datatype Unknown type Y Escpecially for games StarTrek Armada and Gothic. If anyone is intersted in symbol demangling stuff(msvcrt), here is an excellent guide for definition of symbols, etc : http://sparcs.kaist.ac.kr/~tokigun/article/vcmangle.html I will try to come up with a basic patch for that. I dont mind, If someone beats me to it. Thanks, VJ

3 2

Re: *RESEND* Fix bug #5800
by Mike McCormack Sept. 9, 2006

Sept. 9, 2006

Daniel Skorka wrote: > + res = connect(s, ptr->ai_addr, (int)ptr->ai_addrlen); > + if(res) Why not use InternetConnect? If the connection needs to be made through a http:// proxy, connect() may not work, but InternetConnect should handle that case. Mike

1 0

Re: Socket resets randomly
by Tom Wickline Sept. 8, 2006

Sept. 8, 2006

On 9/8/06, Ryan Miller <rmiller(a)auctionpay.com> wrote: > Index: ChangeLog > You don't want to change the changelog, only AJ does the changelog once the patch is accepted and makes its way into main line Wine. You should just add a note of what changed Tom

1 0

Re: svrapi: realizes library svrapi.dll (3 of 20 functions).
by Detlef Riekenberg Sept. 8, 2006

Sept. 8, 2006

Konstantin Petrov wrote: Your Patch is much to large. > dlls/svrapi/Makefile, > dlls/svrapi/libsvrapi.def, This files are created automatic by the build-system. > +16 stdcall NetShareAdd (str long str long) WIN98_NetShareAdd > +17 stdcall NetShareDel(str str long) WIN98_NetShareDel > +18 stdcall NetShareEnum(str long ptr long ptr ptr) WIN98_NetShareEnum > +19 stub NetShareGetInfo > +20 stub NetShareSetInfo Are the ordinals needed / Which app import this Functions by Ordinal? > --- /dev/null 2006-07-03 10:36:18 +0400 > +++ dlls/svrapi/svrapi_main.c 2006-07-24 15:08:49 +0400 I suggest to send a Patch, that has only "DllMain" in this File to reduce the size. Add a stub with description in a separate Patch. > +//SHPWLEN is not known !!!!!! C++ - Comments are not allowed in wine (it's not portable) > +#define SHPWLEN LM20_PWLEN > +typedef struct _share_info_1 { > + char shi1_netname[LM20_NNLEN+1]; > + char shi1_pad1; > + unsigned short shi1_type; > + char* shi1_remark; > +} share_info_1; > +typedef struct _share_info_50 { > + char shi50_netname[LM20_NNLEN+1]; > + unsigned char shi50_type; > + unsigned short shi50_flags; > + char* shi50_remark; > + char* shi50_path; > + char shi50_rw_password[SHPWLEN+1]; > + char shi50_ro_password[SHPWLEN+1]; > +} share_info_50; This seems to be the wrong location here. (svrapi.h) Should be the first patch (only this include-file) > +BOOL WINAPI DllMain(HINSTANCE hinstDLL, DWORD fdwReason, LPVOID > fImpLoad) > +{ > + TRACE("%p 0x%lx %p\n", hinstDLL, fdwReason, fImpLoad); > + > + switch(fdwReason) { Do not forget "DLL_WINE_PREATTACH" > + > + FIXME("Stub (%s %d %p %d %p %p)\n", (pszServer ? > pszServer:"NULL"), sLevel, pbBuffer, > + cbBuffer, pcEntriesRead, pcTotalAvail); you need debugstr_a() > + if (pbBuffer != NULL) > + HeapFree(GetProcessHeap(), 0, pbBuffer); HeapFree() handles NULL; we removed similar unneeded "if" recently. > + if(pszServer != NULL) return NERR_NetNameNotFound; Many Functions in other dll's handle an empty Servername as an alias for the local Computer (the same way as an NULL-Parameter). Did you test this? > + // if ((sLevel == 50) && (cbBuffer == > sizeof(share_info_50))) //in real Why do you not use this code, when it reflects the windows-behavior? -- By By ... ... Detlef

5 8

Anything wrong with this patch? Comments....
by Vijay Kiran Kamuju Sept. 8, 2006

Sept. 8, 2006

Hi, Anything wrong with this patch, http://www.winehq.org/pipermail/wine-patches/2006-August/030201.html Please do send me comments, this fixes installation of icqlite. Thanks, VJ

1 0

Re: ntdll/file.c: arttr is never... ("attr"-typo in the Changelog)
by Detlef Riekenberg Sept. 8, 2006

Sept. 8, 2006

> Changelog: > - ntdll/file.c: arttr is never NULL inside NtCreateFile This "arttr" was a typo and should be "attr" The Patch is correct. Thanks -- By by ... Detlef

1 0

Re: include: Add new Vista SDK constants, fields and structures documented at http://ian.joeuser.com/index.asp?AID=738
by Dmitry Timoshkov Sept. 8, 2006

Sept. 8, 2006

"Francois Gouget" <fgouget(a)free.fr> wrote: > --- a/include/commctrl.h > +++ b/include/commctrl.h > @@ -119,7 +119,9 @@ #define NM_TOOLTIPSCREATED (NM_FIRS > #define NM_LDOWN (NM_FIRST-20) > #define NM_RDOWN (NM_FIRST-21) > #define NM_THEMECHANGED (NM_FIRST-22) > - > +#define NM_FONTCHANGED (NM_FIRST-23) > +#define NM_CUSTOMTEXT (NM_FIRST-24) > +#define NM_TVSTATEIMAGECHANGING (NM_FIRST-24) Do the two above really have the same values? -- Dmitry.

2 1

Framebuffer objects
by Ivan Gyurdiev Sept. 8, 2006

Sept. 8, 2006

Hi... I've been trying to get FBOs working, but I have extremely limited time to work on such things, so I thought I'd just post my current patch here, and see if people can help me figure out what's wrong with it :) FBOs (framebuffer objects) are a newer extension, intended to replace pbuffers to accomplish offscreen rendering - so instead of drawing to the screen, you get to draw to other objects, more importantly textures. We currently support pbuffers, but the code is sketchy, and disabled for that reason. SPEC: http://msi.unilim.fr/~porquet/glexts/GL_EXT_framebuffer_object.txt.htm Known problems with this patch include: (1) The D3D allows the user to setDepthStencilBuffer. I do no such thing - only a default depth buffer is supported. Neither is sharing this depth buffer between offscreen and onscreen [ requires copying of pixels via lock/unlock ]. Resolution: next revision - we don't support that onscreen in the first place :) (2) Attempting to bind a framebuffer using GL_RGB32F_ARB, and GL_RGBA32F_ARB formats will fail on my hardware (Nvidia GeForce 6800GS) with error GL_FRAMEBUFFER_UNSUPPORTED_EXT, which means the format combination was invalid for implementation-dependent reasons. However, this doesn't make sense, based on information online, which indicates this should work. See: http://www.mathematik.uni-dortmund.de/~goeddeke/gpgpu/tutorial.htm : "Using textures as render targets".... or google. This is a problem, since certain demos which make use of the R32F format won't work with FBOs - but oddly work with pbuffers. See SoftShadowsII: http://www.humus.ca/index.php?page=3D&&start=8 On the other hand I know R16F -> GL_RGB16F works fine, based on 3Dc demo. (3) Attempting to use GL_STENCIL_INDEX8_EXT in glRenderbufferStorageEXT will fail on my hardware with GL_INVALID_OPERATION. Why - this should work according to the GL spec. On the other hand this support list says otherwise: http://developer.nvidia.com/object/nv_ogl_texture_formats.html ========= Note that bugs or no bugs, the patch fixes the following (demos from humus.ca): 3Dc demo runs correctly TransparentShadowMapping demo runs correctly Sketch demo runs correctly Half-Life 2 startup screen runs correctly ---- Water demo offscreen content is moved offscreen SoftShadows2 - offscreen content moved offscreen, shows something a lot more correct PhongLighting - offscreen content moved offscreen ---- ...but doesn't quite match pbuffers, because: SoftShadows2 - works completely in pbuffers (even if slow), has errors in FBOs PhongLighting - has errors in FBOs SelfShadowBump - no shadow artifacts on the wall, but now most of the demo's in shadow. >From 98efb529e850dafb66c331afbc5168f90180650a Mon Sep 17 00:00:00 2001 From: Ivan Gyurdiev <ivg2(a)cornell.edu> Date: Thu, 7 Sep 2006 07:17:00 -0400 Subject: [PATCH] Add support for offscreen rendering via FBOs. This uses: GL_EXT_framebuffer_objects A select_offscreen_mode() function was added, where FBOs are given precedence over pbuffers - another way to do offscreen rendering, which is older, not as nice, and currently disabled at compile time. A shared depth buffer between offscreen and onscreen is not currently supported. A custom depth buffer (SetDepthStencil) is not supported (but onscreen doesn't support it either). --- dlls/wined3d/device.c | 201 +++++++++++++++++++++++++++++++++++----- dlls/wined3d/directx.c | 17 +++ dlls/wined3d/utils.c | 38 +++++++- dlls/wined3d/wined3d_private.h | 70 +++++++++----- include/wine/wined3d_gl.h | 1 5 files changed, 275 insertions(+), 52 deletions(-) diff --git a/dlls/wined3d/device.c b/dlls/wined3d/device.c index 463607f..aeed6b0 100644 --- a/dlls/wined3d/device.c +++ b/dlls/wined3d/device.c @@ -8,6 +8,7 @@ * Copyright 2005 Oliver Stieber * Copyright 2006 Stefan Dösinger for CodeWeavers * Copyright 2006 Henri Verbeet + * Copyright 2006 Ivan Gyurdiev * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public @@ -82,6 +83,8 @@ static void WINAPI IWineD3DDeviceImpl_Ad static void WINAPI IWineD3DDeviceImpl_ApplyTextureUnitState(IWineD3DDevice *iface, DWORD Stage, WINED3DTEXTURESTAGESTATETYPE Type); +static void device_render_to_texture(IWineD3DDeviceImpl* This, BOOL isTexture); + /* helper macros */ #define D3DMEMCHECK(object, ppResult) if(NULL == object) { *ppResult = NULL; WARN("Out of memory\n"); return WINED3DERR_OUTOFVIDEOMEMORY;} @@ -239,6 +242,128 @@ static void setup_light(IWineD3DDevice * } /********************************************************** + * FBO helper functions follow + **********************************************************/ + +static void device_fbo_attach_depthstencil( + IWineD3DDeviceImpl *This) { + + TRACE("Attaching depth stencil to offscreen FBO.\n"); + + /* Attach a default Depth renderbuffer */ + GL_EXTCALL(glFramebufferRenderbufferEXT(GL_FRAMEBUFFER_EXT, GL_DEPTH_ATTACHMENT_EXT, + GL_RENDERBUFFER_EXT, This->offscreenRBO[OFFSCREEN_RBO_DEPTH])); + checkGLcall("glFramebufferRenderbufferEXT(OFFSCREEN_RBO_DEPTH)"); + + /* Attach a default Stencil renderbuffer */ +// GL_EXTCALL(glFramebufferRenderbufferEXT(GL_FRAMEBUFFER_EXT, GL_STENCIL_ATTACHMENT_EXT, +// GL_RENDERBUFFER_EXT, This->offscreenRBO[OFFSCREEN_RBO_STENCIL])); +// checkGLcall("glFramebufferRenderbufferEXT(OFFSCREEN_RBO_STENCIL)"); +} + +static void device_fbo_attach_rendertarget( + IWineD3DDeviceImpl *This, + IWineD3DSurface* surface) { + + glDescriptor* surface_desc; + + TRACE("Attaching render surface %p to offscreen FBO.\n", surface); + + /* Pre-load does some neessary prep., such as filling out the gl descriptor properly */ + IWineD3DSurface_PreLoad(surface); + IWineD3DSurface_GetGlDesc(surface, &surface_desc); + + /* Now attach surface to the FBO */ + GL_EXTCALL(glFramebufferTexture2DEXT(GL_FRAMEBUFFER_EXT, GL_COLOR_ATTACHMENT0_EXT, + surface_desc->target, surface_desc->textureName, surface_desc->level)); + checkGLcall("glFramebufferTexture2DEXT(0)"); + + checkGLfbstatus(); + + /* TODO: MRTs */ +} + +static void device_fbo_render_offscreen( + IWineD3DDeviceImpl *This, + IWineD3DSurface* initialRendertarget, + IWineD3DSurface* initialDepthStencil) { + + UINT render_width, render_height; + WINED3DSURFACE_DESC render_desc; + memset(&render_desc, 0, sizeof(WINED3DSURFACE_DESC)); + render_desc.Width = &render_width; + render_desc.Height = &render_height; + + /* Grab the current rendertarget's width and height */ + IWineD3DSurfaceImpl_GetDesc(initialRendertarget, &render_desc); + + /* The first time offscreen rendering is requested, + * reserve one FBO, and two RBOs, and initialize storage for depth and stencil */ + + if (!This->offscreenFBO) { + GL_EXTCALL(glGenFramebuffersEXT(1, &This->offscreenFBO)); + checkGLcall("glGenFramebuffersEXT"); + + GL_EXTCALL(glGenRenderbuffersEXT(2, This->offscreenRBO)); + checkGLcall("glGenRenderbuffersEXT"); + + GL_EXTCALL(glBindRenderbufferEXT(GL_RENDERBUFFER_EXT, This->offscreenRBO[OFFSCREEN_RBO_DEPTH])); + checkGLcall("glBindRenderbufferEXT(OFFSCREEN_RBO_DEPTH)"); + GL_EXTCALL(glRenderbufferStorageEXT(GL_RENDERBUFFER_EXT, GL_DEPTH_COMPONENT24, + render_width, render_height)); + checkGLcall("glRenderbufferStorageEXT(OFFSCREEN_RBO_DEPTH)"); + +// GL_EXTCALL(glBindRenderbufferEXT(GL_RENDERBUFFER_EXT, This->offscreenRBO[OFFSCREEN_RBO_STENCIL])); +// checkGLcall("glBindRenderbufferEXT(OFFSCREEN_RBO_STENCIL)"); +// GL_EXTCALL(glRenderbufferStorageEXT(GL_RENDERBUFFER_EXT, GL_STENCIL_INDEX4_EXT, +// render_width, render_height)); +// checkGLcall("glRenderbufferStorageEXT(OFFSCREEN_RBO_STENCIL)"); + } + + /* TODO: lock depthstencil */ + + /* Switch to offscreen FBO */ + GL_EXTCALL(glBindFramebufferEXT(GL_FRAMEBUFFER_EXT, This->offscreenFBO)); + checkGLcall("glBindFramebufferEXT"); + + device_fbo_attach_rendertarget(This, initialRendertarget); + device_fbo_attach_depthstencil(This); + + /* TODO: unlock depthstencil */ +} + +static void device_fbo_render_onscreen( + IWineD3DDeviceImpl *This) { + + TRACE("Reverting to onscreen FBO.\n"); + + /* TODO: lock depthstencil */ + + /* Detach color */ + GL_EXTCALL(glFramebufferTexture2DEXT(GL_FRAMEBUFFER_EXT, + GL_COLOR_ATTACHMENT0_EXT, GL_TEXTURE_2D, 0, 0)); + checkGLcall("glFramebufferTexture2DEXT(0)"); + + /* Detach depth */ + GL_EXTCALL(glFramebufferRenderbufferEXT(GL_FRAMEBUFFER_EXT, + GL_DEPTH_ATTACHMENT_EXT, GL_RENDERBUFFER_EXT, 0)); + checkGLcall("glFramebufferRenderbufferEXT(OFFSCREEN_DEPTH_RBO)"); + + /* Detach stencil */ + GL_EXTCALL(glFramebufferRenderbufferEXT(GL_FRAMEBUFFER_EXT, + GL_STENCIL_ATTACHMENT_EXT, GL_RENDERBUFFER_EXT, 0)); + checkGLcall("glFramebufferRenderbufferEXT(OFFSCREEN_STENCIL_RBO)"); + + /* Switch to onscreen FBO */ + GL_EXTCALL(glBindFramebufferEXT(GL_FRAMEBUFFER_EXT, 0)); + checkGLcall("glBindFramebufferEXT"); + + /* TODO: unlock depthstencil */ + + checkGLfbstatus(); +} + +/********************************************************** * GLSL helper functions follow **********************************************************/ @@ -2141,6 +2266,19 @@ static HRESULT WINAPI IWineD3DDeviceImpl This->cursorTexture = 0; } + /* Delete the offscreen FBO (if present), and switch back on-screen */ + if (wined3d_settings.offscreen_selected_mode == OFFSCREEN_FBO && This->offscreenFBO) { + ENTER_GL(); + device_fbo_render_onscreen(This); + device_render_to_texture(This, FALSE); + GL_EXTCALL(glDeleteFramebuffersEXT(1, &This->offscreenFBO)); + checkGLcall("glDeleteFramebuffersEXT"); + GL_EXTCALL(glDeleteRenderbuffersEXT(2, This->offscreenRBO)); + checkGLcall("glDeleteRenderbuffersEXT"); + LEAVE_GL(); + This->offscreenFBO = 0; + } + for(sampler = 0; sampler < GL_LIMITS(sampler_stages); ++sampler) { IWineD3DDevice_SetTexture(iface, sampler, NULL); } @@ -5984,8 +6122,10 @@ static HRESULT WINAPI IWineD3DDeviceImpl /* If the container of the rendertarget is a texture then we need to save the data from the pbuffer */ IUnknown *targetContainer = NULL; - if (WINED3D_OK == IWineD3DSurface_GetContainer(This->renderTarget, &IID_IWineD3DBaseTexture, (void **)&targetContainer) - || WINED3D_OK == IWineD3DSurface_GetContainer(This->renderTarget, &IID_IWineD3DDevice, (void **)&targetContainer)) { + if (wined3d_settings.offscreen_selected_mode == OFFSCREEN_PBUFFER && + (WINED3D_OK == IWineD3DSurface_GetContainer(This->renderTarget, &IID_IWineD3DBaseTexture, (void **)&targetContainer) + || WINED3D_OK == IWineD3DSurface_GetContainer(This->renderTarget, &IID_IWineD3DDevice, (void **)&targetContainer))) { + TRACE("(%p) : Texture rendertarget %p\n", This ,This->renderTarget); /** always dirtify for now. we must find a better way to see that surface have been modified (Modifications should will only occur via draw-primitive, but we do need better locking @@ -7199,34 +7339,34 @@ static HRESULT WINAPI IWineD3DDeviceImpl static HRESULT WINAPI IWineD3DDeviceImpl_SetDepthStencilSurface(IWineD3DDevice *iface, IWineD3DSurface *pNewZStencil) { IWineD3DDeviceImpl *This = (IWineD3DDeviceImpl *)iface; - HRESULT hr = WINED3D_OK; IWineD3DSurface *tmp; TRACE("(%p) Swapping z-buffer\n",This); if (pNewZStencil == This->stencilBufferTarget) { TRACE("Trying to do a NOP SetRenderTarget operation\n"); - } else { - /** OpenGL doesn't support 'sharing' of the stencilBuffer so we may incure an extra memory overhead - * depending on the renter target implementation being used. - * A shared context implementation will share all buffers between all rendertargets (including swapchains), - * implementations that use separate pbuffers for different swapchains or rendertargets will have to duplicate the - * stencil buffer and incure an extra memory overhead - ******************************************************/ - - - tmp = This->stencilBufferTarget; - This->stencilBufferTarget = pNewZStencil; - /* should we be calling the parent or the wined3d surface? */ - if (NULL != This->stencilBufferTarget) IWineD3DSurface_AddRef(This->stencilBufferTarget); - if (NULL != tmp) IWineD3DSurface_Release(tmp); - hr = WINED3D_OK; - /** TODO: glEnable/glDisable on depth/stencil depending on - * pNewZStencil is NULL and the depth/stencil is enabled in d3d - **********************************************************/ + return WINED3D_OK; } - return hr; + /** OpenGL doesn't support 'sharing' of the stencilBuffer so we may incure an extra memory overhead + * depending on the renter target implementation being used. + * A shared context implementation will share all buffers between all rendertargets (including swapchains), + * implementations that use separate pbuffers for different swapchains or rendertargets will have to duplicate the + * stencil buffer and incure an extra memory overhead + ******************************************************/ + + tmp = This->stencilBufferTarget; + This->stencilBufferTarget = pNewZStencil; + /* should we be calling the parent or the wined3d surface? */ + if (NULL != This->stencilBufferTarget) IWineD3DSurface_AddRef(This->stencilBufferTarget); + if (NULL != tmp) IWineD3DSurface_Release(tmp); + /** TODO: glEnable/glDisable on depth/stencil depending on + * pNewZStencil is NULL and the depth/stencil is enabled in d3d + **********************************************************/ + + /* TODO: Unlock texture */ + + return WINED3D_OK; } @@ -7334,9 +7474,7 @@ static void device_reapply_stateblock(IW This->updateStateBlock = oldUpdateStateBlock; } -/* Set the device to render to a texture, or not. - * This involves changing renderUpsideDown */ - +/* Set the device to render to a texture */ static void device_render_to_texture(IWineD3DDeviceImpl* This, BOOL isTexture) { DWORD cullMode; @@ -7352,6 +7490,7 @@ static void device_render_to_texture(IWi /* Set upside-down rendering, and update the cull mode */ /* The surface must be rendered upside down to cancel the flip produced by glCopyTexImage */ This->renderUpsideDown = isTexture; + This->offscreenRendering = isTexture; This->last_was_rhw = FALSE; This->proj_valid = FALSE; IWineD3DDevice_GetRenderState((IWineD3DDevice*) This, WINED3DRS_CULLMODE, &cullMode); @@ -7540,12 +7679,22 @@ #if defined(GL_VERSION_1_3) /* Reapply the stateblock, and set the device not to render to texture */ device_reapply_stateblock(This); + + /* If the offscreen FBO is being used, switch to the onscreen FBO */ + if (wined3d_settings.offscreen_selected_mode == OFFSCREEN_FBO && This->offscreenRendering) + device_fbo_render_onscreen(This); device_render_to_texture(This, FALSE); } + /* Offscreen rendering: FBOs */ + } else if (wined3d_settings.offscreen_selected_mode == OFFSCREEN_FBO) { + + device_fbo_render_offscreen(This, RenderSurface, This->depthStencilBuffer); + device_render_to_texture(This, TRUE); + /* Offscreen rendering: PBuffers (currently disabled). * Also note that this path is never reached if FBOs are supported */ - } else if (pbuffer_support && + } else if (wined3d_settings.offscreen_selected_mode == OFFSCREEN_PBUFFER && (cfgs = device_find_fbconfigs(This, implicitSwapchainImpl, RenderSurface)) != NULL) { /** ******************************************************************** diff --git a/dlls/wined3d/directx.c b/dlls/wined3d/directx.c index d4a0e97..ef9cdfb 100644 --- a/dlls/wined3d/directx.c +++ b/dlls/wined3d/directx.c @@ -26,7 +26,6 @@ /* Uncomment this to force only a single display mode to be exposed: */ /*#define DEBUG_SINGLE_MODE*/ - #include "config.h" #include "wined3d_private.h" @@ -237,6 +236,20 @@ static void select_shader_mode( } } +/* Select the offscreen mode, based on the given capabilities, and flags */ + +static void select_offscreen_mode( + WineD3D_GL_Info *gl_info, + int* offscreen_selected) { + + if (gl_info->supported[EXT_FRAMEBUFFER_OBJECT]) + *offscreen_selected = OFFSCREEN_FBO; + else if (pbuffer_support) + *offscreen_selected = OFFSCREEN_PBUFFER; + else + *offscreen_selected = OFFSCREEN_NONE; +} + /** Select the number of report maximum shader constants based on the selected shader modes */ void select_shader_max_constants(WineD3D_GL_Info *gl_info) { @@ -1808,6 +1821,7 @@ static HRESULT WINAPI IWineD3DImpl_GetDe } select_shader_mode(&This->gl_info, DeviceType, &wined3d_settings.ps_selected_mode, &wined3d_settings.vs_selected_mode); + select_offscreen_mode(&This->gl_info, &wined3d_settings.offscreen_selected_mode); select_shader_max_constants(&This->gl_info); /* ------------------------------------------------ @@ -2398,6 +2412,7 @@ static HRESULT WINAPI IWineD3DImpl_Crea LEAVE_GL(); select_shader_mode(&This->gl_info, DeviceType, &wined3d_settings.ps_selected_mode, &wined3d_settings.vs_selected_mode); + select_offscreen_mode(&This->gl_info, &wined3d_settings.offscreen_selected_mode); select_shader_max_constants(&This->gl_info); temp_result = allocate_shader_constants(object->updateStateBlock); diff --git a/dlls/wined3d/utils.c b/dlls/wined3d/utils.c index e0778c0..40ae793 100644 --- a/dlls/wined3d/utils.c +++ b/dlls/wined3d/utils.c @@ -6,6 +6,7 @@ * Copyright 2004 Christian Costa * Copyright 2005 Oliver Stieber * Copyright 2006 Henri Verbeet + * Copyright 2006 Ivan Gyurdiev * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public @@ -47,7 +48,7 @@ static const PixelFormatDesc formats[] = {WINED3DFMT_G8R8_G8B8 ,0x0 ,0x0 ,0x0 ,0x0 ,1/*?*/ ,TRUE ,0 ,0 ,0 }, {WINED3DFMT_R8G8_B8G8 ,0x0 ,0x0 ,0x0 ,0x0 ,1/*?*/ ,TRUE ,0 ,0 ,0 }, /* IEEE formats */ - {WINED3DFMT_R32F ,0x0 ,0x0 ,0x0 ,0x0 ,4 ,FALSE ,GL_RGB32F_ARB ,GL_RED ,GL_FLOAT }, + {WINED3DFMT_R32F ,0x0 ,0x0 ,0x0 ,0x0 ,4 ,FALSE ,GL_RGB32F_ARB ,GL_RED ,GL_FLOAT }, {WINED3DFMT_G32R32F ,0x0 ,0x0 ,0x0 ,0x0 ,8 ,FALSE ,0 ,0 ,0 }, {WINED3DFMT_A32B32G32R32F,0x0 ,0x0 ,0x0 ,0x0 ,16 ,FALSE ,GL_RGBA32F_ARB ,GL_RGBA ,GL_FLOAT }, /* Hmm? */ @@ -646,6 +647,41 @@ #undef POOL_TO_STR } } +const char* debug_glerror(GLint err) { + switch (err) { +#define GLERR_TO_STR(p) case p: return #p; + GLERR_TO_STR(GL_NO_ERROR); + GLERR_TO_STR(GL_INVALID_ENUM); + GLERR_TO_STR(GL_INVALID_VALUE); + GLERR_TO_STR(GL_INVALID_OPERATION); + GLERR_TO_STR(GL_STACK_OVERFLOW); + GLERR_TO_STR(GL_STACK_UNDERFLOW); + GLERR_TO_STR(GL_OUT_OF_MEMORY); + GLERR_TO_STR(GL_INVALID_FRAMEBUFFER_OPERATION_EXT); +#undef GLERR_TO_STR + default: + FIXME("Unrecognized %u GL error!\n", err); + return "unrecognized"; + } +} + +const char* debug_fbstatus(GLenum status) { + switch (status) { +#define FBSTATUS_TO_STR(p) case p: return #p; + FBSTATUS_TO_STR(GL_FRAMEBUFFER_COMPLETE_EXT); + FBSTATUS_TO_STR(GL_FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT_EXT); + FBSTATUS_TO_STR(GL_FRAMEBUFFER_INCOMPLETE_DIMENSIONS_EXT); + FBSTATUS_TO_STR(GL_FRAMEBUFFER_INCOMPLETE_FORMATS_EXT); + FBSTATUS_TO_STR(GL_FRAMEBUFFER_INCOMPLETE_DRAW_BUFFER_EXT); + FBSTATUS_TO_STR(GL_FRAMEBUFFER_INCOMPLETE_READ_BUFFER_EXT); + FBSTATUS_TO_STR(GL_FRAMEBUFFER_UNSUPPORTED_EXT); +#undef FBSTATUS_TO_STR + default: + FIXME("Unrecognized %#x FB status!\n", status); + return "unrecognized"; + } +} + /***************************************************************************** * Useful functions mapping GL <-> D3D values */ diff --git a/dlls/wined3d/wined3d_private.h b/dlls/wined3d/wined3d_private.h index f103705..7d22c0f 100644 --- a/dlls/wined3d/wined3d_private.h +++ b/dlls/wined3d/wined3d_private.h @@ -5,6 +5,7 @@ * Copyright 2002-2003 Raphael Junqueira * Copyright 2004 Jason Edmeades * Copyright 2005 Oliver Stieber + * Copyright 2006 Ivan Gyurdiev * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public @@ -127,6 +128,10 @@ #define VS_SW 2 #define PS_NONE 0 #define PS_HW 1 +#define OFFSCREEN_NONE 0 +#define OFFSCREEN_PBUFFER 1 +#define OFFSCREEN_FBO 2 + #define VBO_NONE 0 #define VBO_HW 1 @@ -156,6 +161,7 @@ typedef struct wined3d_settings_s { BOOL glslRequested; int vs_selected_mode; int ps_selected_mode; + int offscreen_selected_mode; /* nonpower 2 function */ int nonpower2_mode; int rendertargetlock_mode; @@ -225,18 +231,28 @@ #define MAX_PALETTES 256 /* Checking of API calls */ /* --------------------- */ -#define checkGLcall(A) \ -{ \ - GLint err = glGetError(); \ - if (err == GL_NO_ERROR) { \ - TRACE("%s call ok %s / %d\n", A, __FILE__, __LINE__); \ - \ - } else do { \ - FIXME(">>>>>>>>>>>>>>>>> %x from %s @ %s / %d\n", \ - err, A, __FILE__, __LINE__); \ - err = glGetError(); \ - } while (err != GL_NO_ERROR); \ -} +const char* debug_glerror(GLint err); +#define checkGLcall(A) \ +{ \ + GLint err = glGetError(); \ + if (err == GL_NO_ERROR) { \ + TRACE("%s call ok %s / %d\n", A, __FILE__, __LINE__); \ + \ + } else do { \ + FIXME(">>>>>>>>>>>>>>>>> %s returned from %s @ %s / %d\n", \ + debug_glerror(err), A, __FILE__, __LINE__); \ + err = glGetError(); \ + } while (err != GL_NO_ERROR); \ +} + +const char* debug_fbstatus(GLenum status); +#define checkGLfbstatus() \ +{ \ + GLenum status = GL_EXTCALL(glCheckFramebufferStatusEXT(GL_FRAMEBUFFER_EXT)); \ + if (status != GL_FRAMEBUFFER_COMPLETE_EXT) \ + FIXME(">>>>>>>>>>>>>>>> %s: incomplete framebuffer status\n", \ + debug_fbstatus(status)); \ +} /* Trace routines / diagnostics */ /* ---------------------------- */ @@ -309,17 +325,17 @@ #endif /* Checking of per-vertex related GL calls */ /* --------------------- */ -#define vcheckGLcall(A) \ -{ \ - GLint err = glGetError(); \ - if (err == GL_NO_ERROR) { \ - VTRACE(("%s call ok %s / %d\n", A, __FILE__, __LINE__)); \ - \ - } else do { \ - FIXME(">>>>>>>>>>>>>>>>> %x from %s @ %s / %d\n", \ - err, A, __FILE__, __LINE__); \ - err = glGetError(); \ - } while (err != GL_NO_ERROR); \ +#define vcheckGLcall(A) \ +{ \ + GLint err = glGetError(); \ + if (err == GL_NO_ERROR) { \ + VTRACE(("%s call ok %s / %d\n", A, __FILE__, __LINE__)); \ + \ + } else do { \ + FIXME(">>>>>>>>>>>>>>>>> %s returned from %s @ %s / %d\n", \ + debug_glerror(err), A, __FILE__, __LINE__); \ + err = glGetError(); \ + } while (err != GL_NO_ERROR); \ } /* TODO: Confirm each of these works when wined3d move completed */ @@ -483,6 +499,7 @@ void dumpResources(ResourceList *resourc /***************************************************************************** * IWineD3DDevice implementation structure */ + typedef struct IWineD3DDeviceImpl { /* IUnknown fields */ @@ -562,8 +579,13 @@ #define NEEDS_DI HRESULT state; BOOL d3d_initialized; - /* Screen buffer resources */ + /* Screen buffer resources. */ glContext contextCache[CONTEXT_CACHE]; + GLuint offscreenFBO; + GLuint offscreenRBO[2]; +#define OFFSCREEN_RBO_DEPTH 0 +#define OFFSCREEN_RBO_STENCIL 1 + BOOL offscreenRendering; /* A flag to check if endscene has been called before changing the render tartet */ BOOL sceneEnded; diff --git a/include/wine/wined3d_gl.h b/include/wine/wined3d_gl.h index 1a589a5..d570d8c 100644 --- a/include/wine/wined3d_gl.h +++ b/include/wine/wined3d_gl.h @@ -1023,6 +1023,7 @@ #define GL_HILO8_NV #define GL_SIGNED_HILO8_NV 0x885F #define GL_FORCE_BLUE_TO_ONE_NV 0x8860 #endif + /* GL_ATI_texture_env_combine3 */ #ifndef GL_ATI_texture_env_combine3 #define GL_ATI_texture_env_combine3 1 -- 1.4.1

3 6

x11dvr: return support for 24bit - but handle them like 32bit
by Christoph Frick Sept. 8, 2006

Sept. 8, 2006

hiho the game ``NASCAR Racing Season 2003`` from Papyrus actually asks when using the OpenGL driver in the game for 24 bit modes. Otherwise it complains, that it can not find any display modes (it works with 16bit - but the game seem only to accept the modes the current desktop runs in). so what i did is adding the 24bit depth back into the list (last one) and replaced some magic ints and some unused code on my way through this. still this game now does no longer work with wine (it worked a year ago). no it stumbles when calling wglShareLists. License: LGPL ChangeLog: - also list 24bit modes; but still handle them later as 32bit -- cu diff --git a/dlls/winex11.drv/settings.c b/dlls/winex11.drv/settings.c index 5de07eb..f6e2807 100644 --- a/dlls/winex11.drv/settings.c +++ b/dlls/winex11.drv/settings.c @@ -39,7 +39,8 @@ static LPDDHALMODEINFO dd_modes = NULL; static unsigned int dd_mode_count = 0; static unsigned int dd_max_modes = 0; static int dd_mode_default = 0; -static const unsigned int depths[] = {8, 16, 32}; +static const unsigned int depths[] = {8, 16, 32, 24}; +static const unsigned int nr_depths = sizeof(depths) / sizeof(unsigned int); /* pointers to functions that actually do the hard stuff */ static int (*pGetCurrentMode)(void); @@ -62,7 +63,7 @@ LPDDHALMODEINFO X11DRV_Settings_SetHandl TRACE("Resolution settings now handled by: %s\n", name); if (reserve_depths) /* leave room for other depths */ - dd_max_modes = (3+1)*(nmodes); + dd_max_modes = nr_depths*nmodes; else dd_max_modes = nmodes; @@ -111,7 +112,7 @@ void X11DRV_Settings_AddDepthModes(void) int existing_modes = dd_mode_count; DWORD dwBpp = screen_depth; if (dwBpp == 24) dwBpp = 32; - for (j=0; j<3; j++) + for (j=0; j<nr_depths; j++) { if (depths[j] != dwBpp) { @@ -164,8 +165,6 @@ void X11DRV_Settings_Init(void) */ BOOL X11DRV_EnumDisplaySettingsEx( LPCWSTR name, DWORD n, LPDEVMODEW devmode, DWORD flags) { - DWORD dwBpp = screen_depth; - if (dwBpp == 24) dwBpp = 32; devmode->dmDisplayFlags = 0; devmode->dmDisplayFrequency = 0; devmode->dmSize = sizeof(DEVMODEW);

1 0

WGL patches cause problems with RBR and LFS
by Christoph Frick Sept. 8, 2006

Sept. 8, 2006

hiho i am currently running Mike's (?) mmbranch git version of wine and beside some games are working without problems (or at least no new problems) Richard Burns Rally and Live For Speed quit instantly with the more or less same message, that there are no valid display modes available. using git-bisect i can at least tell, that before Rodericks WGL patches everything was ok - then came at least one patch that broke compilation for me and at some pointer later it does compile but no longer work (so its hard to pinpoint the exact patch, that cause the regression). would someone - preferably Roderick - point me in a direction, what traces are needed to look into this? beside that for both games there are demos out there. i am running the games in desktop mode (non desktop does helps nothing); LFS with UseGLSL and RBR not. -- cu

2 4

Re: Add comments so we know why we are not self-registering some dlls, which most of the time is because it is not doing anything anyway.
by Paul Vriens Sept. 8, 2006

Sept. 8, 2006

On Thu, 2006-09-07 at 19:54 +0200, Francois Gouget wrote: > +;11,,wintrust.dll,1 FIXME: Just a stub Hi Francois, I'm pretty busy with the registering of wintrust currently, so why is this marked as a stub? Cheers, Paul.

1 0

Re: msi 2: Allow non-key columns to be used with the join query
by James Hawkins Sept. 7, 2006

Sept. 7, 2006

On 9/7/06, James Hawkins <truiken(a)gmail.com> wrote: > Hi, > > This fixes bug 4632. http://bugs.winehq.org/show_bug.cgi?id=4632 > > Changelog: > * Allow non-key columns to be used with the join query. > Ignore this for now. I forgot to remove the second column key check, and I'm going to add a test case for that explicitly. -- James Hawkins

1 0

Re : [DDRAW] Only detach surfaces when releasing a non-texture surface attached to a complex root.
by Elie Morisse Sept. 7, 2006

Sept. 7, 2006

Le 28.08.2006 18:10:37, Stefan Dösinger a écrit : > Do you mean the game releases the texture manually until it is destroyed??? > Urgh. Can you verify that this doesn't come from DllMain() in > dlls/ddraw/main.c? Nop I checked, and anyway can't be because this occurs also when switching resolutions. > The msdn suggests that destroying the root destroys all sublevels, and it > would be sane, because you create all surfaces with one call, so you should > destroy them with one call. > > If you are right, then every single ddraw application that works on wine now > would leak memory on windows. Well I don't understand, what leaks ? Complex attached surfaces are still destroyed when we fire the root. The new patch takes Moto Racer 2 memleak issues in hand, and Nomad Soul leaks may not be related to this — not a regression because exiting led to crash before ( and remaining surfaces are simple surfaces ). In fact i was a bit unfresh when i posted the patch :P .. > I suspect that something else, like a bad > capatiblity flag, causes weird behavior by the app, so it behaves differently > in wine then on windows, or that the real refcouning issue is that the front > buffer shouldn't be destroyed when the app messes with the back buffer. The > other thing could be that this application has an registry hack on windows: > Windows XP has ddraw compatiblity regitry entries in > HKEY_LOCAL_MACHINE\Software\Microsoft\DirectDraw\Compatiblitity. Can you > check if you find Nomad Soul there? I have no idea what these entries mean, > can you try to remove the whole compatiblity key and try nomad soul again? > (well, backup first :-) ) Hmm I'll try as soon as I procure it. But first tell me if i misunderstood something ;-) > It is hard to prove that a surface is really destroyed or not, but you can > write a regression test which accesses a back buffer which belongs to a > complex back buffer / front buffer compound after the front buffer has been > destroyed?

2 1

Help required to change function prototype
by Andrew Talbot Sept. 7, 2006

Sept. 7, 2006

Hi, I want to change UnmapViewOfFile() (in kernel) to take a LPCVOID, as per the SDK. To do that, I first need to constify one of the parameters of NtUnmapViewOfSection() (in ntdll), but changing it from PVOID to PCVOID in winternl.h breaks compilation (in the "make depend" stage), thus: In file included from ../../include/thread.h:29, from relay.c:30: ../../include/winternl.h:1933: error: expected declaration specifiers or ... before PCVOID make[2]: *** [relay.o] Error 1 make[1]: *** [winebuild] Error 2 make: *** [tools] Error 2 What am I doing wrong, please? The patch I want to submit is as follows: Changelog: ntdll: Constify a formal parameter. diff -urN a/dlls/ntdll/virtual.c b/dlls/ntdll/virtual.c --- a/dlls/ntdll/virtual.c 2006-07-31 17:30:58.000000000 +0100 +++ b/dlls/ntdll/virtual.c 2006-09-06 15:04:25.000000000 +0100 @@ -1925,7 +1925,7 @@ * NtUnmapViewOfSection (NTDLL.@) * ZwUnmapViewOfSection (NTDLL.@) */ -NTSTATUS WINAPI NtUnmapViewOfSection( HANDLE process, PVOID addr ) +NTSTATUS WINAPI NtUnmapViewOfSection( HANDLE process, PCVOID addr ) { FILE_VIEW *view; NTSTATUS status = STATUS_INVALID_PARAMETER; diff -urN a/include/winternl.h b/include/winternl.h --- a/include/winternl.h 2006-08-04 17:17:26.000000000 +0100 +++ b/include/winternl.h 2006-09-06 14:57:29.000000000 +0100 @@ -1930,7 +1930,7 @@ NTSTATUS WINAPI NtUnloadKeyEx(POBJECT_ATTRIBUTES,HANDLE); NTSTATUS WINAPI NtUnlockFile(HANDLE,PIO_STATUS_BLOCK,PLARGE_INTEGER,PLARGE_INTEGER,PULONG); NTSTATUS WINAPI NtUnlockVirtualMemory(HANDLE,PVOID*,SIZE_T*,ULONG); -NTSTATUS WINAPI NtUnmapViewOfSection(HANDLE,PVOID); +NTSTATUS WINAPI NtUnmapViewOfSection(HANDLE,PCVOID); NTSTATUS WINAPI NtVdmControl(ULONG,PVOID); NTSTATUS WINAPI NtWaitForSingleObject(HANDLE,BOOLEAN,const LARGE_INTEGER*); NTSTATUS WINAPI NtWaitForMultipleObjects(ULONG,const HANDLE*,BOOLEAN,BOOLEAN,const LARGE_INTEGER*); Thanks, -- Andy.

4 7

wine and dos support
by Vijay Kiran Kamuju Sept. 7, 2006

Sept. 7, 2006

Hi, A friend of mine is asking me to get dos based games to work using wine. Well, I generally tell to use dosbox for that. Is it possible make them work in wine. If yes, what effort should we put into that. I mean what should we need to implement. Is it possible to lift some parts of other LGPL dos projects. Thanks, VJ

1 0

Re: tools: Remove redundant NULL checks before free() (found by Smatch).
by Robert Shearman Sept. 7, 2006

Sept. 7, 2006

I think the checks are necessary to be compatible with libc's on other platforms. -- Rob Shearman

2 1

Should I give up?
by Damjan Jovanovic Sept. 7, 2006

Sept. 7, 2006

Hi My work on the still image system for wine has highlighted a bug in setupapi's SetupDiOpenClassRegKeyExW(). In short, the registry keys for device classes have the form HKEY_LOCAL_MACHINE\System\CurrentControlSet\ Class\{xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx} while wine incorrectly tries to open HKEY_LOCAL_MACHINE\System\CurrentControlSet\ Class\xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx ie. no curly braced around the guid. And in addition it incorrectly returns FALSE for a return type of HKEY (instead of INVALID_HANDLE_VALUE). I've submitted a test and patch to wine-patches 7 times now, and received very little feedback, which I've always followed. If this is how difficult it is to submit a patch, no wonder I hear people complaining about how wine doesn't have enough developers. How are you ever going to accept my still image code, which is 4000+ lines long and growing, when you can't accept a simple 182 line long patch? Normally I'm very patient - but this is ridiculous. I'm attaching my latest patch and would appreciate any comments. Thank you Damjan __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com --- a/dlls/setupapi/tests/Makefile.in 2006-07-10 18:01:08.000000000 +0200 +++ b/dlls/setupapi/tests/Makefile.in 2006-08-31 21:29:12.000000000 +0200 @@ -3,7 +3,7 @@ SRCDIR = @srcdir@ VPATH = @srcdir@ TESTDLL = setupapi.dll -IMPORTS = setupapi kernel32 +IMPORTS = setupapi kernel32 advapi32 rpcrt4 CTESTS = \ devinst.c \ --- a/dlls/setupapi/devinst.c 2006-08-28 20:35:44.000000000 +0200 +++ b/dlls/setupapi/devinst.c 2006-08-31 22:27:19.000000000 +0200 @@ -1413,6 +1413,7 @@ PVOID Reserved) { LPWSTR lpGuidString; + LPWSTR lpBracedGuidString; HKEY hClassesKey; HKEY hClassKey; LPCWSTR lpKeyName; @@ -1453,21 +1454,30 @@ if (UuidToStringW((UUID*)ClassGuid, &lpGuidString) != RPC_S_OK) { RegCloseKey(hClassesKey); - return FALSE; + return INVALID_HANDLE_VALUE; } + lpBracedGuidString = (LPWSTR) HeapAlloc(GetProcessHeap(), 0, + (1 + strlenW(lpGuidString) + 1 + 1) * sizeof(WCHAR)); + lpBracedGuidString[0] = (WCHAR) '{'; + memcpy(&lpBracedGuidString[1], lpGuidString, + strlenW(lpGuidString) * sizeof(WCHAR)); + lpBracedGuidString[1 + strlenW(lpGuidString)] = (WCHAR) '}'; + lpBracedGuidString[1 + strlenW(lpGuidString) + 1] = (WCHAR) 0; + RpcStringFreeW(&lpGuidString); + if (RegOpenKeyExW(hClassesKey, - lpGuidString, + lpBracedGuidString, 0, KEY_ALL_ACCESS, &hClassKey)) { - RpcStringFreeW(&lpGuidString); + RpcStringFreeW(&lpBracedGuidString); RegCloseKey(hClassesKey); - return FALSE; + return INVALID_HANDLE_VALUE; } - RpcStringFreeW(&lpGuidString); + RpcStringFreeW(&lpBracedGuidString); RegCloseKey(hClassesKey); return hClassKey; --- a/dlls/setupapi/tests/devinst.c 2006-08-28 20:35:53.000000000 +0200 +++ b/dlls/setupapi/tests/devinst.c 2006-08-31 22:14:36.000000000 +0200 @@ -26,6 +26,7 @@ #include "wingdi.h" #include "winuser.h" #include "winreg.h" +#include "rpc.h" #include "setupapi.h" #include "wine/test.h" @@ -34,6 +35,7 @@ static HMODULE hSetupAPI; static HDEVINFO (WINAPI *pSetupDiCreateDeviceInfoListExW)(GUID*,HWND,PCWSTR,PVOID); static BOOL (WINAPI *pSetupDiDestroyDeviceInfoList)(HDEVINFO); +static HKEY (WINAPI *pSetupDiOpenClassRegKeyExW)(GUID*,REGSAM,DWORD,PCWSTR,PVOID); static void init_function_pointers(void) { @@ -43,6 +45,7 @@ { pSetupDiCreateDeviceInfoListExW = (void *)GetProcAddress(hSetupAPI, "SetupDiCreateDeviceInfoListExW"); pSetupDiDestroyDeviceInfoList = (void *)GetProcAddress(hSetupAPI, "SetupDiDestroyDeviceInfoList"); + pSetupDiOpenClassRegKeyExW = (void *)GetProcAddress(hSetupAPI, "SetupDiOpenClassRegKeyExW"); } } @@ -79,6 +82,86 @@ ok(ret, "SetupDiDestroyDeviceInfoList failed : %ld\n", error); } +static void test_SetupDiOpenClassRegKeyExW() +{ + GUID guid; + RPC_STATUS rpcStatus; + HKEY hkey; + WCHAR *unbracedGuid; + WCHAR *bracedGuid; + + rpcStatus = UuidCreate(&guid); + if (rpcStatus != RPC_S_OK) + { + ok(FALSE, "failed to generate test guid, error %ld", rpcStatus); + return; + } + + /* Check return value for non-existant key */ + hkey = pSetupDiOpenClassRegKeyExW(&guid, KEY_ALL_ACCESS, + DIOCR_INSTALLER, NULL, NULL); + ok(hkey == INVALID_HANDLE_VALUE, + "invalid return value %p from SetupDiOpenClassRegKeyExW " + "for non-existant key, expected %p\n", hkey, INVALID_HANDLE_VALUE); + + rpcStatus = UuidToStringW(&guid, &unbracedGuid); + if (rpcStatus != RPC_S_OK) + { + ok(FALSE, "failed to get string form of guid, error %ld", rpcStatus); + return; + } + + bracedGuid = HeapAlloc(GetProcessHeap(), 0, + (lstrlenW(unbracedGuid) + 3) * sizeof(WCHAR)); + if (bracedGuid != NULL) + { + HKEY classesKey; + + bracedGuid[0] = (WCHAR) '{'; + memcpy(&bracedGuid[1], unbracedGuid, + lstrlenW(unbracedGuid) * sizeof(WCHAR)); + bracedGuid[1 + lstrlenW(unbracedGuid)] = (WCHAR) '}'; + bracedGuid[1 + lstrlenW(unbracedGuid) + 1] = (WCHAR) 0; + + SetLastError(0xdeadbeef); + classesKey = pSetupDiOpenClassRegKeyExW(NULL, KEY_ALL_ACCESS, + DIOCR_INSTALLER, NULL, NULL); + ok(classesKey != INVALID_HANDLE_VALUE, + "failed to open the device classes registry key: error %ld\n", + GetLastError()); + if (classesKey != INVALID_HANDLE_VALUE) + { + DWORD ret; + HKEY classKey; + ret = RegCreateKeyW(classesKey, bracedGuid, &classKey); + if (ret == ERROR_SUCCESS) + { + RegCloseKey(classKey); + SetLastError(0xdeadbeef); + classKey = pSetupDiOpenClassRegKeyExW(&guid, KEY_ALL_ACCESS, + DIOCR_INSTALLER, NULL, NULL); + ok(classKey != INVALID_HANDLE_VALUE, + "failed opening class key, error %ld\n", GetLastError()); + ret = RegCloseKey(classKey); + ok(ret == ERROR_SUCCESS, + "failed closing class key, error %ld\n", ret); + + RegDeleteKeyW(classesKey, bracedGuid); + } + else + ok(FALSE, "failed creating device key: error %ld\n", ret); + + RegCloseKey(classesKey); + } + + HeapFree(GetProcessHeap(), 0, bracedGuid); + } + else + ok(FALSE, "failed to allocate memory for braced guid\n"); + + RpcStringFreeW(&unbracedGuid); +} + START_TEST(devinst) { init_function_pointers(); @@ -88,5 +171,10 @@ if (pSetupDiCreateDeviceInfoListExW && pSetupDiDestroyDeviceInfoList) test_SetupDiCreateDeviceInfoListEx(); else - trace("Needed calls not all available, skipping tests.\n"); + trace("Needed calls for SetupDiCreateDeviceInfoListEx not all available, skipping test.\n"); + + if (pSetupDiOpenClassRegKeyExW) + test_SetupDiOpenClassRegKeyExW(); + else + trace("Needed call for SetupDiOpenClassRegKeyExW not available, skipping test.\n"); }

6 5

Re: user: test and fix SwitchToThisWindow [2nd try]
by James Hawkins Sept. 7, 2006

Sept. 7, 2006

On 9/7/06, Jan Zerebecki <jan.wine(a)zerebecki.de> wrote: > Bugreport about this patch: http://bugs.winehq.org/show_bug.cgi?id=5725 > I compiled-tested this and successfully ran the test on wine and > (thx kblin) on native, because the author said that he is not > able to compile wine. > > If this patch is rejected from inclusion, please tell me why, as > I would have to ask anyway. > > From: Neil Rashbrook <neil(a)parkwaycc.co.uk> > Changelog: > user: test and fix SwitchToThisWindow > > --- > > dlls/user/tests/Makefile.in | 1 > dlls/user/tests/win_sttw.c | 127 +++++++++++++++++++++++++++++++++++++++++++ > dlls/user/winpos.c | 15 +++++ > 3 files changed, 142 insertions(+), 1 deletions(-) > > diff --git a/dlls/user/tests/Makefile.in b/dlls/user/tests/Makefile.in > index 7a78a5b..4e10563 100644 > --- a/dlls/user/tests/Makefile.in > +++ b/dlls/user/tests/Makefile.in > @@ -23,6 +23,7 @@ CTESTS = \ > sysparams.c \ > text.c \ > win.c \ > + win_sttw.c \ > winstation.c \ > wsprintf.c > Why is this in a new file? It should be in win.c. -- James Hawkins

1 0

Google SOC projects Success/Failure
by Vijay Kiran Kamuju Sept. 6, 2006

Sept. 6, 2006

Hi, As the Google SOC concluded. We need to evaluate whether our SOC projects were successfull or not. Some of them got through like Oleview, Trash implementation, riched20 improvisation. I really dont know about the clamav integration, shell integration (there were some patches regarding this I think), improve mswscock.dll and ntlm using gensec. Wineconf is about to start, I think we need to re-think what might be good additions and how should we proceed for 1.0 release. I see about 50-60% success in the SOC - 2006. There were a lotta widl improvements, this thing should have been a SOC project ;) Please feel to comment on this. Thanks, VJ

2 1