wine-devel

wine-devel@winehq.org

10 participants
84459 discussions

Re: Serious Bug
by Boaz Harrosh 29 Oct '03

29 Oct '03

I'm Just exactly now fighting same problems. See my last post: "Help with shared library make files" Note 2 things A. Initialization order is opposite of msvc++ Last object on the linker gets to run first. (msvc first on the command line first to initialize) B. Hard, in gdb (kdevelop), to set a break point on pre-main code. use below file to do that. Put it as last object linked and after any library. In effect it is an "int 0x03" instruction to hard code a break point. C. All DLLs linked to the MyApp.dll.so should also link to MyApp.exe.so see my wine post: "wine msvcrt on Library initialization". These are the ones that get linked through winebuild. <Cpp_Globs_stop.cpp> #include "windows.h" #define _CrtDbgBreak() __asm__ ("\tint $0x3\n") class TestStaticClass { public: TestStaticClass() ; ~TestStaticClass() ; } ; TestStaticClass::TestStaticClass() { MessageBox(NULL ,"Constructor" ,"Cpp_Globs_stop" ,MB_OK) ; _CrtDbgBreak(); } TestStaticClass::~TestStaticClass() { MessageBox(NULL ,"Destructor" ,"Cpp_Globs_stop" ,MB_OK) ; } TestStaticClass g_StaticClass ; </Cpp_Globs_stop.cpp> <Makefile snippet> $(mfctest_dll_MODULE).so: $(mfctest_dll_MODULE).spec.o $(mfctest_dll_OBJS) Cpp_Globs_stop.o $(LDXXSHARED) $(LDDLLFLAGS) -o $@ \ $(mfctest_dll_MODULE).spec.o \ $(mfctest_dll_OBJS) \ $(mfctest_dll_LIBRARY_PATH) \ $(mfctest_dll_LIBRARIES:%=-l%) \ $(ALL_LIBRARIES) \ Cpp_Globs_stop.o \ </Makefile snippet> Now you can single step the initialization chain. flyker wrote: > Boaz Harrosh wrote: > >> use --wrap in winemaker. Read about this option in the documentation >> > Thanks. It is work :) > > > >

1 0

Re: Serious Bug
by flyker 29 Oct '03

29 Oct '03

Boaz Harrosh wrote: > use --wrap in winemaker. Read about this option in the documentation > Thanks. It is work :)

1 0

Serious Bug
by flyker 29 Oct '03

29 Oct '03

The simple test crashes wine. And i think many other commands in constructor A() may crash wine. test.cpp: #include <windows.h> #include <stdio.h> class A { public: A(); }; A::A() { LoadLibrary("user32.dll"); // all ok user32.dll is present } A a; int WinMain(HINSTANCE hInst, HINSTANCE, LPSTR szCmdLine, int nCmdShow) { return 0; }

2 2

Help with shared library make files
by Boaz Harrosh 29 Oct '03

29 Oct '03

I have the following setup mfctest/Makefile - Main make file mfctest.exe.so - is the wrapper for an mfc gui application mfctest.dll.so - is the mfc gui application (as generated by the msvc++ wizard and than winemaker) msvc/Makefile - make file for mfc & stuff missing from msvcrt.dll libmsvc.a - is all the cpp files archived by [ ar -rs $(ALL_OBJECT_FILES) ] libmsvc.so - same object files but linked with [ gcc --shared ] Now I link mfctest.dll.so with libmsvc.a. (-l) The spec.c I prepare by including all the object files of the msvc project since winebuild does not work with libraries. It all runs fine. Now I try to link with libmsvc.so ( exact same way as before with the -l option linux style, not as a DLL as usually done in wine ). At the beginning it would not link because of undefined symbols. ( before with the .a option not all files get pulled in by the linker). So I fix the undefined s. The spec.c file I have tried in to ways. 1) a libmsvc.spec.c for libmsvc.so and a small one for mfctest.dll.spec.c - and 2) A big mfctest.dll.spec.c that includes all objects same as above libmsvc.a . In both cases I had the same exact problem. It would not run. I get a SIGSEGV. On the (gdb) back-trace it is directly from "dlopen" (from BUILTIN32_dlopen) it SIGs 10 functions deep into dlopen. The last function that makes any sense is _dl_init in ld-linux.so.2 from than on I get un-logical symbols that should not be in the init path. Including undefined s like 0x00000094 (see complete back-trace below) the SIGSEGV itself is reported at "dllname" which is not a function but an array in dll.spec.c . So we can see that the debug info got totally screwed up. Big help!!! Any body got any ideas? Do you think I need to fix the start address of libmsvc.so some how? What is the command line option for that? The reason I need the libYYY.so option to work is because I have a big (1900 files) C++ project divided into 36 libraries. On the windows side it is dlls & static libs, all automatically exported by msvc++ (#define class class declspec(dllexport) ). Since we do not have a good C++ way for pure dlls in wine (yet) and the 1900 files on one gcc link will not work. ( truncated by the linker) .And I have more Applications in the future that use the same libraries. The .so option looks best. One more question: What do the gurus say? In theory, if I have a XXX.dll.so linked Linux style to libYYY.so both have calls into wine. what is best method (XXX.dll.so is only because of the wrapper no real exports here) A. XXX.dll.so has an XXX.spec.c && libYYY.so has a YYY.spec.c of its own each linked to their respective .spec.c . (Note that those are only imports since export symbols from libYYY.so are linked against - Linux style). B. one big XXX+YYY.spec.c linked against XXX.dll.so (very very slow, it looks like winebuild has an o(n^2) efficiency ) C. Both are good if no exports are made. (or exports only at the top lib) Thanks for any help or pointer. Please share past experience for good or for worse Free Life Boaz some info: <(gdb) backtrace> #0 0x40cb3a4d in dllname () from /home/wine/lib/wine/mfctest.dll.so #1 0x00000094 in ?? () #2 0x40ffec8e in std::basic_iostream<wchar_t, std::char_traits<wchar_t> >::~basic_iostream() () from /usr/lib/libstdc++.so.5 #3 0x40ffed0a in std::basic_iostream<wchar_t, std::char_traits<wchar_t> >::~basic_iostream() () from /usr/lib/libstdc++.so.5 #4 0x4104dbb7 in __cxa_dyn_string_eq () from /usr/lib/libstdc++.so.5 #5 0x40feea11 in std::collate<char>::id () from /usr/lib/libstdc++.so.5 #6 0x4000aa4c in _dl_init (main_map=0x3c02eaf0, argc=4, argv=0xbffff454, env=0xbffff468) at dl-init.c:68 #7 0x4030a51e in _dl_open () from /lib/i686/libc.so.6 #8 0x4000a828 in _dl_catch_error (objname=0x406d28e0, errstring=0x406d28e4, operate=0x4030a120 <_dl_open+432>, args=0x406d28e8) at dl-error.c:162 #9 0x40309fcf in _dl_open () from /lib/i686/libc.so.6 #10 0x40352f21 in dlopen () from /lib/libdl.so.2 #11 0x4000a828 in _dl_catch_error (objname=0x3c0017e0, errstring=0x3c0017e4, operate=0x40352ef0 <dlopen+80>, args=0x406d2a68) at dl-error.c:162 #12 0x40353294 in dlerror () from /lib/libdl.so.2 #13 0x40352ed3 in dlopen () from /lib/libdl.so.2 #14 0x40107abc in wine_dlopen (filename=0x3c02eaa0 "/home/wine/lib/wine/mfctest.dll.so", flag=2, error=0x406d2b08 " ¨<@´l\016@X,m@HÍ\a@", errorsize=256) at loader.c:577 #15 0x40107420 in dlopen_dll (name=0x406d2c2c "mfctest.dll", error=0x406d2b08 " ¨<@´l\016@X,m@HÍ\a@", errorsize=256, test_only=0) at loader.c:156 #16 0x4010795b in wine_dll_load (filename=0x406d2c2c "mfctest.dll", error=0x406d2b08 " ¨<@´l\016@X,m@HÍ\a@", errorsize=256) at loader.c:390 #17 0x4009b537 in BUILTIN32_dlopen (name=0x406d2c2c "mfctest.dll", handle=0x406d2c28) at ../../relay32/builtin32.c:63 #18 0x4009b94c in BUILTIN32_LoadLibraryExA (path=0x403ca710 "C:\\WINDOWS\\SYSTEM\\mfctest.dll", flags=0, pwm=0x406d2cc4) at ../../relay32/builtin32.c:171 #19 0x400b3c6d in load_dll (libname=0x403ca6f0 "mfctest.dll", flags=0, pwm=0x406d2cc4) at loader.c:1058 #20 0x400b3e34 in LdrLoadDll (path_name=0x0, flags=0, libname=0x406d2cfc, hModule=0x406d2cf8) at loader.c:1120 #21 0x4007e158 in LoadLibraryExA (libname=0x404b41a0 "mfctest.dll", hfile=0x0, flags=0) at ../../loader/module.c:655 #22 0x4007e28d in LoadLibraryA (libname=0x404b41a0 "mfctest.dll") at ../../loader/module.c:707 #23 0x404a3147 in WinMain (hInstance=0x404b0000, hPrevInstance=0x0, szCmdLine=0x403908f2 "", iCmdShow=1) at mfctest.exe_wrapper.c:127 #24 0x404b40ea in __wine_exe_main () at mfctest.exe.spec.c:140 #25 0x4009fced in start_process () at ../../scheduler/process.c:455 #26 0x400a402a in call_on_thread_stack (func=0x4009fb10) at ../../scheduler/sysdeps.c:113 </(gdb) backtrace> <(gdb) info sharedlibrary> >From To Syms Read Shared Object Library 0x40048000 0x400c6314 Yes /home/wine/lib/libntdll.dll.so 0x40105790 0x40107c14 Yes /home/wine/lib/libwine.so.1 0x4011cc80 0x4011f354 Yes /home/wine/lib/libwine_unicode.so.1 0x40212620 0x4030f328 Yes /lib/i686/libc.so.6 0x403334a0 0x4034abc4 Yes /lib/i686/libm.so.6 0x40352db0 0x40353d64 Yes /lib/libdl.so.2 0x40000a60 0x4000f49b Yes /lib/ld-linux.so.2 0x40382d10 0x4038b054 Yes /lib/libnss_files.so.2 0x404a3000 0x404b4184 Yes /home/wine/codeward/mfctest/mfctest.exe.so 0x406f3000 0x40732ef4 Yes /home/wine/lib/wine/kernel32.dll.so 0x407b6000 0x40848b14 Yes /home/wine/lib/wine/user32.dll.so 0x408cf000 0x40916c44 Yes /home/wine/lib/wine/gdi32.dll.so 0x40941000 0x4095e814 Yes /home/wine/lib/wine/advapi32.dll.so 0x40974000 0x4099def4 Yes /home/wine/lib/wine/msvcrt.dll.so 0x409b1850 0x409ee254 Yes /usr/lib/libfreetype.so.6 0x409f98e0 0x40a02194 Yes /lib/libz.so.1 0x40a13000 0x40a5e794 Yes /home/wine/lib/wine/x11drv.dll.so 0x404b8200 0x404bc8a4 Yes /usr/X11R6/lib/libSM.so.6 0x40a73540 0x40a81dd4 Yes /usr/X11R6/lib/libICE.so.6 0x40a89910 0x40a934a4 Yes /usr/X11R6/lib/libXext.so.6 0x40aa6380 0x40b17384 Yes /usr/X11R6/lib/libX11.so.6 0x40b99640 0x40bd91e4 Yes /usr/X11R6/lib/libGL.so.1 0x40beb110 0x40bf3c14 Yes /lib/i686/libpthread.so.0 0x40c39320 0x40c3e234 Yes /usr/X11R6/lib/libXcursor.so.1 0x40c41390 0x40c46324 Yes /usr/X11R6/lib/libXrender.so.1 0x40014860 0x40015794 Yes /usr/X11R6/lib/X11/locale/lib/common/xlcDef.so.2 0x40c74260 0x40c8ad84 Yes /usr/X11R6/lib/X11/locale/lib/common/ximcp.so.2 0x400185f0 0x40018f54 Yes /usr/lib/gconv/ISO8859-1.so 0x40ca0000 0x40cb2074 Yes /home/wine/lib/wine/mfctest.dll.so 0x40dcf920 0x40f2f8e4 Yes /home/wine/lib/libmsvc.so 0x40ff1070 0x4104dbc4 Yes /usr/lib/libstdc++.so.5 0x4001b540 0x40020534 Yes /lib/libgcc_s.so.1 </(gdb) info sharedlibrary>

1 0

Re: French keyboard layout with Euro symbol
by RemiAssailly＠aol.com 29 Oct '03

29 Oct '03

I sent a patch that looks like yours and it was simply refused ! I wasn't the fisrt...

4 10

Re: DINPUT device unacquire (mouse) fix
by Andreas Mohr 29 Oct '03

29 Oct '03

Hi, On Tue, Oct 28, 2003 at 02:53:44AM +0200, Hannu Valtonen wrote: > License: LGPL > > I stumbled upon this when looking at fixing mouse support for Crimson > lands. According to MSDN DirectInputDevice Unacquire is allowed even for > devices which haven't been acquired yet. > > Changelog: > - SysMouseAImpl_Unacquire should return DI_NOEFFECT if the device is not > in an acquired state. Ah, thanks!! Maybe this is why some game (I believe it was UT) broke... Darn, why didn't I discover that when I stumbled on that place years ago? But don't you think keyboard and joystick might need the same code?? (keyboard has it as well, but with completely different code --> mess) This could probably be cleaned up. Andreas Mohr

2 2

WinE-Bug? InternetCrackUrl incorrectly sets dwUrlPathLength
by Jan Sporbeck 29 Oct '03

29 Oct '03

Hi, During some application testing, we discovered that WinE does not correctly mimic the WinINet function InternetCrackUrl. The real WinAPI sets the dwUrlPathLength member of the URL_COMPONENTS structure to additionally hold the extra info of the given URL if the extra info is not requested as a separate component. Although this behaviour is not well documented by the Win SDK, it is for instance implemented on WinXP, Win2000 and WinME and therefore might be considered to be intended by Microsoft. The following lines of code demonstrate the problem: LPCTSTR lpszURL = _T("http://www.myserver.com/myscript.php?arg=1"); URL_COMPONENTS url; // call 1: crack out url path and extra info ::ZeroMemory(&url, sizeof(url)); url.dwStructSize = sizeof(url); url.dwHostNameLength = 1; url.dwUrlPathLength = 1; url.dwExtraInfoLength = 1; ::InternetCrackUrl(lpszURL, 0, 0, &url); // dwUrlPathLength is 13, holding "/myscript.php" // call 2: only crack out url path ::ZeroMemory(&url, sizeof(url)); url.dwStructSize = sizeof(url); url.dwHostNameLength = 1; url.dwUrlPathLength = 1; ::InternetCrackUrl(lpszURL, 0, 0, &url); // now, dwUrlPathLength should be 19, holding "/myscript.php?arg=1", // while WinE still sets it to 13 OK, it is of course no great deal to write code that correctly runs on real Win systems and on WinE (2-3 lines did it). But programers who rely on this behaviour will for example wonder why their applications don't transmit any HTTP GET parameters when they call HttpOpenRequest with lpszObjectName set to the URL path component returned by InternetCrackUrl... This applies to Wine 20031016, haven't found any recent report on this though. /jan.

1 0

spec files - syntax and generation
by "Martin Tröster" 29 Oct '03

29 Oct '03

Hi, once again I would like to get your advice on spec files. I now compiled a Windows application successful against winelib and can link it without problems. However, the application segfaults immediately when starting wine. I am tracking this back to .spec files (although I will need to debug in the next free time to be sure), since they are the weakest link I am seeing, since I wrote most of them by hand. So it would be great if you can answer the following questions: - Automatic generation: I composed the spec files by hand, based on Windows def files. Is there an automatic way to do that? I noticed a (somewhat hidden) tool called winedump, with nearly not existing documentation, but could not make much use of it. Any hints here? - Stack size and parameters in spec file: The def files are, as far as I know, very similar to the spec files: they contain the function name (either mangled or not mangled), as well as (most of the time) the stack space - format: functionname@stackspace. Is there anything wrong to translate an entry for "char* testfunc(char* str, int var1, long var2, long* ptr)" called "testfunc@16" into "@stdcall testfunc(ptr ptr ptr ptr)", without caring for the type of parameters? More explicitely asked, what use do I have from using str or long instead of ptr in the spec file? - Stdcall vs. cdecl How do I find out whether a function uses CDECL instead of STDCALL? I found the information (link lost unfortunately) that entries in the def file like testfunc1@0 are called via STDCALL, whereas in case of testfunc2 without any @ specifying the space is CDECL. Is this guess right? If not, how else do I find this out? Thanks again for any help! Cheers, Martin ______________________________________________________________________________ Bestes Testergebnis: Stiftung Warentest Doppelsieg fur WEB.DE FreeMail und WEB.DE Club. Nur fuer unsere Nutzer! http://f.web.de/?mc=021182

4 5

Re: PATCH: server/named_pipe.c (RESENT)
by Alexandre Julliard 28 Oct '03

28 Oct '03

Gerald Pfeifer <gerald(a)pfeifer.com> writes: > was there any problem with the following patch, or did it just fall > through the cracks? ;-) > > Gerald > > ChangeLog: > Remove unused variable pipe_client_fd_ops. Well, either the client side should really use pipe_client_fd_ops, or pipe_server_fd_ops should be renamed to show that it is used by both client and server. I'm not sure which one was intended, could the named pipe experts please check that? -- Alexandre Julliard julliard(a)winehq.com

1 0

Support for TLS in CrossOver
by Mike Hearn 28 Oct '03

28 Oct '03

Hi Alexandre, I noticed the ChangeLog for the new CXOffice release (congrats on that btw!) said it now works with all glibc 2.3 variants. Does that mean the TLS related named pipe breakage we saw earlier is now fixed? If so, do you know when the fix will be in WineHQ CVS? thanks -mike

3 3

WineHQ: PATCH to fix IPersistFile::Load() and associated functions
by Subhobroto Sinha 28 Oct '03

28 Oct '03

Hello all I finally produced the "official patch" after verifying that everything is OK. This patch has been inlined as well as attached as a ZIP file, so suit yourself. All reports regaring this patch is to be sent to <subhobrotosinha at yahoo.com> with a proper subject prefixed by "IPersistFile patch" in the subject header. Just don't simply hit the 'Reply' button ! This patch fixes: (1) Stream_LoadString: Conflict of unicode strings being taken as LPSTR and AGAIN being converted to Unicode due to wrong value of SCF_UNICODE .(I have really NOT changed the original code, but redirected this call to a fixed function which I named wineStream_LoadString. This should remain so until SCF_UNICODE value is corrected) (2) Stream_LoadLocation: Previously the GetPath() failed. It works great now. For the uninitiated, here's what has to be done: Please make sure that you have the 20031016 source (latest tarball) (1)Copy this .diff to the root folder of the Wine source (2)Run 'patch' on this diff file (3)It should patch "dlls/shell32/shelllink.c". Change to that directory (4)Type 'make' to rebuild 'shell32.dll.so' (5)'su' out and overwrite/symlink the current 'shell32.dll.so' over the installed one Now IPersistFile::Load() should work as intended. Try out McCormack's 'winemenubuilder' or my WineLib 'linkresolve' (NOT my C++ version) If the diff is not proper, etc, please alert me at the mentioned address PS: As I am not a member of 'wine-patches',can the moderator be kind enough... :=) I really want these functions to get fixed. That's all. Regards Subhobroto Sinha __________________________________ Do you Yahoo!? Exclusive Video Premiere - Britney Spears http://launch.yahoo.com/promos/britneyspears/ --- dlls/shell32/shelllink.c.original 2003-10-22 17:56:31.000000000 +0530 +++ dlls/shell32/shelllink.c 2003-10-26 16:30:57.000000000 +0530 @@ -401,9 +401,49 @@ return S_OK; } +static HRESULT wineStream_LoadString( IStream* stm, LPWSTR *pstr ) +{ + DWORD count=0; + USHORT len; + LPWSTR str=NULL; + HRESULT r; + /* + From Win98 upwards unicode==TRUE.(Atleast for shortcuts..) + However, the value of SCF_UNICODE does NOT seem to be 0x1000. + Until then, we assume it true. + Anybody differing mail to <subhobrotosinha at yahoo.com>. + */ + TRACE("%p\n", stm); + + r = IStream_Read(stm, &len, sizeof(len), &count); + if(FAILED(r)||(count != sizeof(len))) return E_FAIL; + + len *= sizeof(WCHAR); + + TRACE("reading %d\n", len); + str = HeapAlloc(GetProcessHeap(), 0, len+sizeof(WCHAR)); + if(!str) return E_OUTOFMEMORY; + + count = 0; + r = IStream_Read(stm, str, len, &count); + if( FAILED(r)||( count!=len) ) + { + HeapFree( GetProcessHeap(), 0, str ); + return E_FAIL; + } + + str[count/2]=0; + *pstr = str; + TRACE("read %s\n", debugstr_w(str)); + return S_OK; +} static HRESULT Stream_LoadString( IStream* stm, BOOL unicode, LPWSTR *pstr ) { +/*This is the original function short circuited to the above till we find out correct 'SCF_UNICODE'. +Now it's assumed that the data's unicode. (This assumption failes iff the OS was <=Win95*/ +return (wineStream_LoadString(stm,pstr)); +#if 0 DWORD count; USHORT len; LPVOID temp; @@ -453,6 +493,7 @@ *pstr = str; return S_OK; +#endif } static HRESULT Stream_LoadLocation( IStream* stm ) @@ -533,6 +574,14 @@ if( FAILED( r ) ) return r; } + ////////////////////////////////////////////////////////////////////////////////////// + char szTemp[MAX_PATH]={0}; + SHGetPathFromIDListA(This->pPidl,szTemp); + This->sPath=HeapAlloc( GetProcessHeap(), 0,(strlen(szTemp)+1)*sizeof(WCHAR)); + MultiByteToWideChar(CP_ACP,0,szTemp,(strlen(szTemp)+1),This->sPath,(strlen(szTemp)+1)); + //TRACE("%s\n",szTemp); + ////////////////////////////////////////////////////////////////////////////////////// + This->wHotKey = hdr.wHotKey; This->iIcoNdx = hdr.nIcon; FileTimeToSystemTime (&hdr.Time1, &This->time1);

3 3

Winesetuptk 0.7 released
by Ivan Leo Murray-Smith 28 Oct '03

28 Oct '03

Hi lists, Winesetuptk 0.7 is at the wine sourceforge page http://sourceforge.net/project/showfiles.php?group_id=6241 This new version contains an updated winedefault.reg, you don't need to run regedit winedefault.reg any more, and it's updated to configurate all the latest and greatest features of wine. This release was possible thanks to the help and contributions of Vincent Béron and Alex Pasadyn. Without them, winesetuptk 0.7 wouldn't exist. Download and enjoy!

1 0

Status of HeapAlloc Cleanup in shell32
by Steven Edwards 28 Oct '03

28 Oct '03

Is the HeapAlloc/ReAlloc stuff being tracked anywhere? I think it is the cause of a lot of problems I have with shell32 under Windows and ReactOS. Applications linked to wine shell32.dll under Windows2000 run fine but when you exit they get illigal op's where the memory could not be "read" kind of like the bug I had in uninstaller.exe that started the HeapAlloc cleanup. Thanks Steven __________________________________ Do you Yahoo!? Exclusive Video Premiere - Britney Spears http://launch.yahoo.com/promos/britneyspears/

2 1

Re: wine/ tools/winebuild/relay.c dlls/ntdll/signa ...
by Robert Shearman 28 Oct '03

28 Oct '03

This has fixed the problem slightly (see message titled "Bug in cxx_frame_handler") - the program now displays: err:seh:setup_exception stack overflow 592 bytes in thread 000e eip 401c425a esp 40640db0 stack 0x40540000-0x40750000 And then exits. I still believe there is a problem in dlls/msvcrt/cppexcept.c with the following lines (359-364): if (!exc_type) /* nested exception, fetch info from original exception */ { rec = msvcrt_get_thread_data()->exc_record; exc_type = (cxx_exception_type *)rec->ExceptionInformation[2]; } The problem is that msvcrt_get_thead_data()->exc_record returns NULL (probably because there is no nested exception?). Should there just be a check there for a NULL pointer or is it part of a bigger problem? Rob On Tuesday 28 October 2003 12:30 am, you wrote: > ChangeSet ID: 9891 > CVSROOT: /opt/cvs-commit > Module name: wine > Changes by: julliard(a)wine.codeweavers.com 2003/10/27 18:30:55 > > Modified files: > tools/winebuild: relay.c > dlls/ntdll : signal_i386.c > > Log message: > Switch back to the main process stack before calling exception > handlers. > > Patch: http://cvs.winehq.com/patch.py?id=9891 > > Old revision New revision Changes Path > 1.31 1.32 +27 -28 wine/tools/winebuild/relay.c > 1.73 1.74 +297 -165 wine/dlls/ntdll/signal_i386.c

1 0

Improper Focus handling
by Cogman 28 Oct '03

28 Oct '03

I have noticed this bug in wine for some time, (winex has corrected it btw). It seems that new windows, child windows, and message boxes are incorrectly handled in wine, when created they are often put at the bottem of the stack of wine windows and not given focus. This is bad expecialy in Install shield programs that bring the messages and windows up one after the other, but you cant see the others or move them because the first is covering them. Also Keyboard focus does not seem to be working properly all the time (sometimes a window gets keyboardfocus and sometimes it doesnt) One thing I would like to point out is the fact that if you click on a window in windows it gets both keyboard and mouse focus. A few examples of where the focus issue becomes a problem are in Cstike and Pharaoh. In Cstrike (or any HL game for that matter) if you loose focus on the main window (alt-tab or something else happens) you loose keyboardfocus, but not always mouse focus, So you endup stuck in a game with the keyboard focused on some other window (usually on a konsole so killall is the only way to get out other then changing virtual windows) In pharaoh the problem is when a new window pops up. The keyboard does not send messages to it, and you can click on it and everything but not be able to type anything (save game window). Now if you run games in desktop mode it will sometimes work. but I dont think that you should have to run games in desktop just to get proper focus handling. I really would love to see this bug smushed :)

1 0

wine msvcrt on Library initialization
by Boaz Harrosh 28 Oct '03

28 Oct '03

I have only managed to compile and link MFC with the msvcrt.dll the std way would not work. Now the MFC has lots of: <somefoo.cpp> void CleanFoo() { } int init = atexit( CleanFoo ) ; </somefoo.cpp> The call to atexit will in-turn call _lock if we look at _lock.c : <lock.c> ... /* If the lock doesn't exist yet, create it */ if( lock_table[ locknum ].bInit == FALSE ) { /* Lock while we're changing the lock table */ _lock( _LOCKTAB_LOCK ); .... </lock.c> Since we are calling on library initialization, (global variable) it might happen that the call is made before the initialization of the _LOCKTAB_LOCK (before a call to lock.c::msvcrt_init_mt_locks) and therefor I get an endless recursion. Now this is why we have the -wrap option with the delayed load of winelib apps. [Q}Should I submit a patch to protect that endless recursion in msvcrt. or should it be noted that all DLL s used by the winelib-application should be also linked to the wrapper (to have a chance to initialize) and submit a patch to winemaker? P.S Some similar problem occurs with native msvcrt.dll. That one is hard to pinpoint exactly. In any way making the wrapper link to msvcrt will solve the problem in both cases. Free Life Boaz

1 0

Proposal to resolve the issue of the broken wine-users gateway
by Jeremy White 28 Oct '03

28 Oct '03

Hi Folks, Say, Just chatting on IRC today, and ShadowHawk pointed out that a host of other sites (Franks corner, etc) have sprung up almost entirely because there are no user based forums on winehq.org. This is, in part, due to the fact that the gateway between the newsgroup and wine-users is busted. And it is that way, because the gateway we were using is a dead project, and Jer hasn't been able to wrestle it back into working (and doesn't really want to). Further, I think it may be time to recognize that cool looking forums are what users expect; that mailing lists and newsgroups are really becoming the 'old' way of doing things. Further, there is a PHP based forum system that Jer would be happy to put up on winehq.org. This would neatly side step the whole problem and perhaps provide a nice upgrade for winehq.org. What do others think? Cheers, Jeremy

5 6

Re: DMUSIC: fix in dmloader's behavior
by Jerry Jenkins 28 Oct '03

28 Oct '03

Rok Mandeljc wrote: > + /* OK, MSDN says that search order is the following: > + - current directory (DONE) > + - windows search path (FIXME: how do I get that?) Try this : DWORD GetEnvironmentVariable( LPCTSTR lpName, // environment variable name LPTSTR lpBuffer, // buffer for variable value DWORD nSize // size of buffer ); or NTSTATUS WINAPI RtlQueryEnvironmentVariable_U( PWSTR env, PUNICODE_STRING name, PUNICODE_STRING value ); > + - loader's search path (DONE) > + */

2 1

winelib documentation
by "Martin Tröster" 28 Oct '03

28 Oct '03

Hi, I want to propose a possible improvement for the wine (and especially winelib) documentation on the WineHQ website: I would like each chapter and section to be dated with the latest release date (like a CVS header etc.), so that the reader knows when the information has been composed. Furthermore, especially in case of development, it would be interesting for which version of Wine these changes are true. Another thing would be to change the FAQ into a Wiki-thing - although I usually do not like this concept too much, I think it is far better than having to refer to outdated documentation... I hope you guys see this as an improvement for the future - although it still shows that the current state of documentation needs some updates (especially in the case of winelib development, see my questions on .spec files that is soon to come :-)) Cheers, Martin ______________________________________________________________________________ Horoskop, Comics, VIPs, Wetter, Sport und Lotto im WEB.DE Screensaver1.2 Kostenlos downloaden: http://screensaver.web.de/?mc=021110

1 0

Re: Fix comctl32 conformance test
by Jakob Eriksson 28 Oct '03

28 Oct '03

Why is this not applied? Would be nice to have when cross compiling works. Hans Leidekker wrote: >Hi, > >It appears that the comctl32 conformance test calls functions by name >that are exported by ordinal only on Windows. Still the test runs (and >thus run-time links) on Wine. Is this wanted behavior? > >When I run it as a cross compiled test on my W2kSP4 machine or on Wine(!) >it bombs with linking errors. To fix this I made this patch that uses >GetProcAddress to get at the functions via their ordinals. > > -Hans > >Changelog: > Don't call functions by name. They are exported by ordinal only. > > >Index: dlls/comctl32/tests/dpa.c >=================================================================== >RCS file: /home/wine/wine/dlls/comctl32/tests/dpa.c,v >retrieving revision 1.2 >diff -u -r1.2 dpa.c >--- dlls/comctl32/tests/dpa.c 5 Sep 2003 23:08:42 -0000 1.2 >+++ dlls/comctl32/tests/dpa.c 19 Oct 2003 16:24:49 -0000 >@@ -29,6 +29,9 @@ > > #include "wine/test.h" > >+static HDPA (WINAPI *pDPA_Create)(INT)=NULL; >+static INT (WINAPI *pDPA_Search)(HDPA,LPVOID,INT,PFNDPACOMPARE,LPARAM,UINT)=NULL; >+ > static INT CALLBACK dpa_strcmp(LPVOID pvstr1, LPVOID pvstr2, LPARAM flags) > { > LPCSTR str1 = (LPCSTR)pvstr1; >@@ -42,14 +45,18 @@ > HDPA dpa_ret; > INT int_ret; > CHAR test_str0[]="test0"; >- >- dpa_ret = DPA_Create(0); >+ >+ HMODULE hcomctl32=LoadLibraryA("comctl32.dll"); >+ pDPA_Create=(void*)GetProcAddress(hcomctl32, (LPSTR)328); >+ pDPA_Search=(void*)GetProcAddress(hcomctl32, (LPSTR)339); >+ >+ dpa_ret = pDPA_Create(0); > ok((dpa_ret !=0), "DPA_Create failed"); >- int_ret = DPA_Search(dpa_ret,test_str0,0, dpa_strcmp,0, DPAS_SORTED); >+ int_ret = pDPA_Search(dpa_ret,test_str0,0, dpa_strcmp,0, DPAS_SORTED); > ok((int_ret == -1), "DPA_Search found invalid item"); >- int_ret = DPA_Search(dpa_ret,test_str0,0, dpa_strcmp,0, DPAS_SORTED|DPAS_INSERTBEFORE); >+ int_ret = pDPA_Search(dpa_ret,test_str0,0, dpa_strcmp,0, DPAS_SORTED|DPAS_INSERTBEFORE); > ok((int_ret == 0), "DPA_Search proposed bad item"); >- int_ret = DPA_Search(dpa_ret,test_str0,0, dpa_strcmp,0, DPAS_SORTED|DPAS_INSERTAFTER); >+ int_ret = pDPA_Search(dpa_ret,test_str0,0, dpa_strcmp,0, DPAS_SORTED|DPAS_INSERTAFTER); > ok((int_ret == 0), "DPA_Search proposed bad item"); > } > > > > >

2 2

Longhorn info
by Mike Hearn 28 Oct '03

28 Oct '03

Of course, as we don't actually implement the "new" XP APIs yet, this is all rather academic. Still, why not take a look at what we'll be reimplementing in 2010 ;) SDK: http://longhorn.msdn.microsoft.com/ Articles/Info: http://msdn.microsoft.com/longhorn thanks -mike

1 0

Segmentation fault
by flyker 28 Oct '03

28 Oct '03

Under Windows my program exe file is about 6Mb. Under Linux and Wine it is about 50Mb (.so file). It is normal ? And Wine makes Segmentation fault on initialising global variables of my program. #1 0x40bb0116 in __static_initialization_and_destruction_0 (__initialize_p=1, __priority=65535) at main.cpp:681 #2 0x40bb0176 in _GLOBAL__I_oldfCentury () at main.cpp:3809 #3 0x40bcc151 in __do_global_ctors_aux () from /home/mc/PRJ/main.exe.so #4 0x40872dc6 in _init () from /home/mc/PRJ/main.exe.so #5 0x4000c4b1 in _dl_init_internal () from /lib/ld-linux.so.2 #6 0x40154f42 in dl_open_worker () from /lib/libc.so.6 #7 0x4000c266 in _dl_catch_error_internal () from /lib/ld-linux.so.2 #8 0x401549af in _dl_open () from /lib/libc.so.6 #9 0x4017bf6b in dlopen_doit () from /lib/libdl.so.2 #10 0x4000c266 in _dl_catch_error_internal () from /lib/ld-linux.so.2 #11 0x4017c316 in _dlerror_run () from /lib/libdl.so.2 #12 0x4017bf14 in dlopen@@GLIBC_2.1 () from /lib/libdl.so.2 #13 0x4002ea32 in wine_dlopen (filename=0xbfffc5f8 "/home/mc/PRJ/main.exe.so", flag=2, error=0xbfffd628 "\200.", errorsize=1024) at loader.c:454 #14 0x4049e920 in __wine_kernel_init () at process.c:841 #15 0x401d8947 in __wine_process_init (argc=0, argv=0xbfffda84) at loader.c:1954 #16 0x4002e9fd in wine_init (argc=2, argv=0xbfffded4, error=0xbfffda84 "", error_size=1024) at loader.c:431 #17 0x3c001771 in main (argc=2, argv=0xbfffded4) at main.c:33 #18 0x40057917 in __libc_start_main () from /lib/libc.so.6 Any ideas ?

2 1

New compiler warnings
by Gerald Pfeifer 28 Oct '03

28 Oct '03

My nightly autobuilder on FreeBSD 5.1 noticed the following warning introduced yesterday: signal_i386.c:524: warning: `save_context' defined but not used signal_i386.c:602: warning: `restore_context' defined but not used Also, there are two OpenGL warnings left, after most have been fixed a few months ago: directx.c:852: warning: implicit declaration of function `glXGetProcAddressARB' drawprim.c:829: warning: implicit declaration of function `glMultiTexCoord4fARB' Could you please have a look? Thanks! Gerald -- Gerald Pfeifer (Jerry) gerald(a)pfeifer.com http://www.pfeifer.com/gerald/

1 0

Re: French keyboard layout with Euro symbol
by Sylvain Petreolle 28 Oct '03

28 Oct '03

Hi Jerome, It seems your patch includes a currency sign '¤' instead of the Euro : Altgr-E and Altgr-$ look the same (0xA4.) Please use the unified format when sending patches. ===== Sylvain Petreolle (spetreolle_at_users_dot_sourceforge_dot_net) ICQ #170597259 Say NO to software patents Dites NON aux brevets logiciels "What if tomorrow the War could be over ?" Morpheus, in "Reloaded". ___________________________________________________________ Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en français ! Yahoo! Mail : http://fr.mail.yahoo.com

1 0

Newbie questions: ddraw conformance tests
by Sami Aario 27 Oct '03

27 Oct '03

Hi, I'd like to write some conformance tests for ddraw, and I notice that there aren't any preexisting tests. What is the preferred way of setting up the "tests" subdirectory (makefiles)? Do I just do it by hand? BTW, why aren't there any preexisting patches for ddraw? Would my tests possibly be welcome as patches? -- Sami

1 1

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

wine-devel