November 2004 - Wine-Devel - list.winehq.org

Access Tokens and Threads
by Elad Lahav Nov. 3, 2004

Nov. 3, 2004

I've been looking into the annoying "Set ThreadImpersonationToken handle" messages. In order to fix this issue, I need to know a few things: 1. Does Wine maintain access tokens at all? It seems that most of the relevant functions (e.g., NtQueryInformationThread and NtSetInformationThread) simply ignore it. 2. It looks like the thread's token is a part of the so-called "security descriptor" member of the thread's SECURITY_ATTRIBUTES structure. Does anyone know what structure this descriptor points to? Elad

2 1

source level debugging with WineLib?
by Geoff K. Hart Nov. 3, 2004

Nov. 3, 2004

Hello, We just started using WineLib, and have very basic question: which -g??? option do you use to do source level debugging with winedbg? I searched the documentation and mailing list for any step-by-step about this, but found nothing. So either the answer is so obvious that I just can't see it - or it doesn't work? By default, winemaker seems configure things for no debug - so I edited one Makefile manually, and added -g. Winedbg can attach to the process, and can even show the stack - but didn't know much beyond that. I tried all the -g variants that I could find, but only -gstabs and -gstabs+ seemed to help. They allowed the debugger to find the correct place in the source, and "whatis" seemed to work - however, if I tried to dereference something, it complained about "no type ...", like this: Wine-dbg>info process pid threads parent executable (all id:s are in hex) 00000008 3 00000000 'Z:\home\s\src\SysSrv \DataDistributionService\DDSServer2\ddsserver2.exe' Wine-dbg>attach 8 In 32 bit mode. 0x546387a2: ret Wine-dbg>bt Backtrace: =>1 0x546387a2 (0x10252c50) 2 0x48714a7e NTDLL_wait_for_multiple_objects+0x12e in ntdll (0x10252cf8) 3 0x48712fec vm86_return_end+0xc55 in ntdll (0x10252d1c) 4 0x1000b808 (0x10352d64) 5 0x48714a7e NTDLL_wait_for_multiple_objects+0x12e in ntdll (0x10352e0c) 6 0x48714afc NtWaitForMultipleObjects+0x6c in ntdll (0x10352e30) 7 0x61e57a95 WaitForMultipleObjectsEx+0xb5 in kernel32 (0x10352f60) 8 0x61e5797b WaitForSingleObjectEx+0x3b in kernel32 (0x10352f80) 9 0x1002878afixme:dbghelp:elf_new_wine_thunks Duplicate in ddsserver2<elf>: _ZZ17BroadCastIdentityPvmmE22lRepCntrClientDeletion<10037720-00000004> lRepCntrClientDeletion<10037720-00000000> fixme:dbghelp:elf_new_wine_thunks Duplicate in ddsserver2<elf>: _ZZ17BroadCastIdentityPvmmE20lRepKeepAliveCounter<10037724-00000004> lRepKeepAliveCounter<10037724-00000000> fixme:dbghelp:elf_new_wine_thunks Duplicate in ddsserver2<elf>: _ZZ17BroadCastIdentityPvmmE17lRepCntrBroadcast<1003771c-00000004> lRepCntrBroadcast<1003771c-00000000> _Z13DDSServerMainPv+0x88(lpParameter=0x44c47098) [/home/s/src/SysSrv/ DataDistributionService/DDSServer2/AcceptSocket.cpp:220] in ddsserver2 (0x10353150) 10 0x61e5eeda (0x10353228) 11 0x48716523 (0x10353a78) 12 0x1000508d (0x10353b0c) 13 0x660d255a (0x00000000) Wine-dbg>up 8 220 dwRetVal = ::WaitForSingleObjectEx (pServer- >m_hStopEvent, INFINITE, TRUE); Wine-dbg>p pServer 0x44c47098 Wine-dbg>p *pServer No type or type mismatch Wine-dbg>whatis pServer struct DDSServer* Wine-dbg> Any help? Geoff

5 4

Re: InternetGetConnectedStateExA
by Alexandre Julliard Nov. 3, 2004

Nov. 3, 2004

Aric Stewart <aric(a)codeweavers.com> writes: > +BOOL WINAPI InternetGetConnectedStateExA(LPDWORD lpdwStatus, > + LPSTR lpszConnectionName, DWORD dwNameLen, DWORD dwReserved) > +{ > + INT len; > + LPWSTR lpwszConnectionName = NULL; > + BOOL rc; > + > + TRACE("(%p, %s, %ld, 0x%08lx)\n", lpdwStatus, debugstr_a(lpszConnectionName), dwNameLen, dwReserved); > + > + if (lpszConnectionName) > + { > + len = MultiByteToWideChar(CP_ACP, 0, lpszConnectionName, -1, NULL, 0); > + lpwszConnectionName= HeapAlloc(GetProcessHeap(), 0, len*sizeof(WCHAR)); > + MultiByteToWideChar(CP_ACP, 0, lpszConnectionName, -1, > + lpwszConnectionName, len); > + } > + > + rc = InternetGetConnectedStateExW(lpdwStatus,lpwszConnectionName, dwNameLen, lpszConnectionName is an output parameter, it should be converted on return from the W call. -- Alexandre Julliard julliard(a)winehq.org

2 1

Re: WINSOCK: Implement WSAAddressToStringA/W
by Alexandre Julliard Nov. 3, 2004

Nov. 3, 2004

Hans Leidekker <hans(a)it.vu.nl> writes: > + if (stringW) > + { > + ret = WSAAddressToStringW( sockaddr, len, NULL, stringW, &lenstrW ); > + *lenstr = lenstrW / sizeof(WCHAR); To be really correct you have to use WideCharToMultiByte to compute the ASCII length. Though in this case it would probably be easier to just duplicate the code and do a simple sprintf in the A function instead of trying to call the W function. -- Alexandre Julliard julliard(a)winehq.org

1 0

compiling wine with msvc
by James Hawkins Nov. 3, 2004

Nov. 3, 2004

I'm trying to get the wine tests to compile under msvc, so I ran msvcmaker in the wine source under cygwin (which was successful). Then I opened the winetest project and did a 'rebuild all'. When it tries to compile advapi32_test/security.c I get an warning and 102 errors: C:\Program Files\Microsoft Visual Studio\VC98\INCLUDE\rpc.h(89) : warning C4005: 'RPCRTAPI' : macro redefinition ..\..\..\include\rpc.h(45) : see previous definition of 'RPCRTAPI' :\Program Files\Microsoft Visual Studio\VC98\INCLUDE\rpcnsi.h(37) : error C2143: syntax error : missing ')' before 'type' C:\Program Files\Microsoft Visual Studio\VC98\INCLUDE\rpcnsi.h(37) : error C2491: 'RpcNsBindingExportA' : definition of dllimport function not allowed ......etc I'm pretty sure the problem is that the the msvc's rpc.h and rpcnsi.h files are being included when they shouldn't. Is that about right? If so, what could be a possible solution? BTW I'm compiling with MSVC++ 6.0. wine cvs from 11-01-2004. -- James Hawkins

3 6

Re: Fix for metafile test
by Dmitry Timoshkov Nov. 3, 2004

Nov. 3, 2004

"Jakob Eriksson" <jakov(a)vmlinux.org> wrote: > Seems the test fails in non-interactive mode without this patch. No, the problem seems to be related to a window's visibility. Windows does not play EMF records if the DC was created for a not visible window. A lot of other tests fail due to exactly that problem: user32/win.c and user32/msg.c are the examples. The question is what makes the tests create invisible windows? Does winetest run as a service on a not visible desktop? Anything else? -- Dmitry.

4 5

Re: The Return of IE6 Problems
by Mike Hearn Nov. 3, 2004

Nov. 3, 2004

> The common denominator for this situation is that these patches modified > some of the same files. Patch 13475 implemented itss.dll and modified > wine/dlls/Makefile.in, wine/configure.ac and wine/configure. A partial > implementation of advpack.RegInstall, Patch 13938 also modified > wine/dlls/Makefile.in, wine/configure.ac and wine/configure in addition > to adding /wine/dlls/advpack/reg.c. Although IE6 would still > initialize, patch 13937 added advpack to the repository. Hi Dan, Great work on the regression search! This is exactly the sort of information we need. ITSS was implemented by Mike - as it's a component of IE I guess it should at least work with it so I'm CCing him. It may be that he implemented it for some other app rather than IE and he doesn't have time to bring it up to scratch for IE right now. I'm not sure. The second patch suggests that we got RegInstall wrong. That needs to be fixed, but for now you can probably get IE back to a working state again by overriding both these DLLs as native (itss, advpack). If you look at the script I wrote back in January, quite a lot of what it does is simply overriding various incomplete DLLs. I no longer maintain that script (buy CrossOver! :) but if you want to take over maintenance by all means do so, I'd be happy to help you get up to speed. thanks -mike

2 1

RE: Test Drive failure
by Dagan Shai Nov. 3, 2004

Nov. 3, 2004

This results in a helloworld executable. When I execute it I get lots of output lines, here's a sample: fixme:ttydrv:TTYDRV_GetBitmapBits (0x3bc, 0x77cafa7c, 32): stub fixme:ttydrv:TTYDRV_GetBitmapBits (0x3b8, 0x77cafa9c, 32): stub err:imagelist:ImageList_ReplaceIcon no color! fixme:ttydrv:TTYDRV_DC_StretchBlt (0x94, 288, 0, 16, 16, 0x3c4, 0, 0, 16, 32, 13 369376): stub fixme:ttydrv:TTYDRV_DC_StretchBlt (0x98, 288, 0, 16, 16, 0x3c4, 0, 0, 16, 32, 13 369376): stub fixme:ttydrv:TTYDRV_GetBitmapBits (0x3cc, 0x77cafa7c, 128): stub fixme:ttydrv:TTYDRV_GetBitmapBits (0x3c8, 0x77cafafc, 128): stub err:imagelist:ImageList_ReplaceIcon no color! fixme:ttydrv:TTYDRV_DC_StretchBlt (0xb0, 576, 0, 32, 32, 0x3d4, 0, 0, 32, 64, 13 369376): stub fixme:ttydrv:TTYDRV_DC_StretchBlt (0xb4, 576, 0, 32, 32, 0x3d4, 0, 0, 32, 64, 13 369376): stub These are all functions that helloworld does not use, so I don't understand the problem here. Regards, Shai Dagan -----Original Message----- From: Dimitrie O. Paun [mailto:dpaun(a)rogers.com] Sent: Tuesday, November 02, 2004 6:57 PM To: Dagan Shai Cc: Dimitrie O. Paun; wine-devel(a)winehq.org Subject: Re: Test Drive failure On Tue, Nov 02, 2004 at 06:40:57PM +0200, Dagan Shai wrote: > Thanks. > > Here they are. Winemaker is buggy, the generated Makefile has a small problem. Here's the fix: --- Makefile.orig 2004-11-02 09:52:59.000000000 -0500 +++ Makefile 2004-11-02 11:52:57.451738280 -0500 @@ -58,7 +58,7 @@ ### Generic targets -all: $(SUBDIRS) $(DLLS:%=%.so) $(EXES:%=%.so) +all: $(SUBDIRS) $(DLLS:%=%.so) $(EXES) ### Build rules @@ -106,7 +106,7 @@ $(helloworld_exe_MODULE).dbg.c: $(helloworld_exe_C_SRCS) $(helloworld_exe_CXX_SRCS) $(WINEBUILD) -o $@ --debug -C$(SRCDIR) $(helloworld_exe_C_SRCS) $(helloworld_exe_CXX_SRCS) -$(helloworld_exe_MODULE).so: $(helloworld_exe_MODULE).dbg.o $(helloworld_exe_OBJS) +$(helloworld_exe_MODULE): $(helloworld_exe_MODULE).dbg.o $(helloworld_exe_OBJS) $(CXX) $(helloworld_exe_LDFLAGS) -o $@ $(helloworld_exe_OBJS) $(helloworld_exe_MODULE).dbg.o $(helloworld_exe_LIBRARY_PATH) $(LIBRARY_PATH) $(helloworld_exe_DLLS:%=-l%) $(helloworld_exe_LIBRARIES:%=-l%) -- Dimi. ______________________________________________________________________ This email message has been scanned by PineApp Mail-Secure and has been found clean.

1 0

Re: EnumDateFormatsA: Add ru_RU locale and default DATE format
by Dmitry Timoshkov Nov. 3, 2004

Nov. 3, 2004

"Vitaly Lipatov" <lav(a)etersoft.ru> wrote: > Changelog: > EnumDateFormatsA: > - Add ru_RU locale (0x419) That's a completely broken implementation, please do not extend it. It should retrieve the data from the kernel32 locale database (have a look at dlls/kernel/nls). > - Use DATE_SHORTDATE when dwFlags is 0 > - Ignore LOCALE_USE_CP_ACP bit in dwFlags (I can't find a difference in > Windows behavior with this flag) It would be nice to see a test case for that. Also please use a common indent 4 spaces as (almost) everywhere in Wine. -- Dmitry.

2 2

dinput - mouse
by Vitaliy Margolen Nov. 3, 2004

Nov. 3, 2004

Is there are anyone working on it? It seems like there are some strange problem with mouse input, and I can't find it! I trying to get Descent 3 running. So far so good. It installs, and runs. But mouse does not work. It's not like it doesn't work at all. It does, but it moves erratically and mostly left-right. Could it be that it can't switch to full screen mode and fails because of it? Dome could be downloaded from here: http://www.descent3.com/4_downloads.html#d3_demos P.S. I know there is Linux version of it. Too bad I don't have it. It cost $$ and have to by purchased separately. Vitaliy Here is what dinput log looks like: trace:dinput:dinput_mouse_hook msg 200 pt 320 240 (W=0) trace:dinput:dinput_mouse_hook (X: 0 - Y: 0 L: 00 M: 00 R: 00) trace:dinput:dinput_mouse_hook msg 200 pt 320 240 (W=0) trace:dinput:dinput_mouse_hook (X: 0 - Y: 0 L: 00 M: 00 R: 00) trace:dinput:dinput_mouse_hook queueing -2 at offset 0 (queue head 46 / size 64) trace:dinput:dinput_mouse_hook queueing 8 at offset 4 (queue head 47 / size 64) trace:dinput:dinput_mouse_hook msg 200 pt 318 248 (W=1) trace:dinput:dinput_mouse_hook (X: -2 - Y: 8 L: 00 M: 00 R: 00) trace:dinput:dinput_mouse_hook queueing -2 at offset 0 (queue head 48 / size 64) trace:dinput:dinput_mouse_hook queueing 6 at offset 4 (queue head 49 / size 64) trace:dinput:dinput_mouse_hook msg 200 pt 316 254 (W=1) trace:dinput:dinput_mouse_hook (X: -4 - Y: 14 L: 00 M: 00 R: 00) trace:dinput:dinput_mouse_hook queueing -2 at offset 0 (queue head 50 / size 64) trace:dinput:dinput_mouse_hook queueing 6 at offset 4 (queue head 51 / size 64) trace:dinput:dinput_mouse_hook msg 200 pt 314 260 (W=1) trace:dinput:dinput_mouse_hook (X: -6 - Y: 20 L: 00 M: 00 R: 00) trace:dinput:dinput_mouse_hook queueing -2 at offset 0 (queue head 52 / size 64) trace:dinput:dinput_mouse_hook queueing 8 at offset 4 (queue head 53 / size 64) trace:dinput:dinput_mouse_hook msg 200 pt 312 268 (W=1) trace:dinput:dinput_mouse_hook (X: -8 - Y: 28 L: 00 M: 00 R: 00) trace:msg:MSG_peek_message got type 7 msg 200 (WM_MOUSEMOVE) hwnd 0x10022 wp 0 lp 0 trace:dinput:SysMouseAImpl_GetDeviceData (0x4036bd48)->(dods=16,entries=64,fl=0x00000000) trace:dinput:SysMouseAImpl_GetDeviceData Application retrieving 12 event(s). trace:dinput:SysMouseAImpl_GetDeviceData Warping mouse to 320 - 240

1 0