August 2002 - Wine-devel - list.winehq.org

Question about unix path request in 2 of the wine dlls
by Steven Edwards 26 Aug '02

26 Aug '02

in urlmon umon.c calls cabextract for by requesting the unix path via wine_get_unix_file_name and in shell32 shelllink.c get_unix_file_name is used in a few locations. Does anyone have a suggestion on a clean simple way to handle both of these on non-nix platforms? Can this be moved to libwine? Thanks Steven __________________________________________________ Do You Yahoo!? HotJobs - Search Thousands of New Jobs http://www.hotjobs.com

2 2

--dll buggy behaviour ?
by Raul Dias 26 Aug '02

26 Aug '02

Hi, I think that I spot a bug in the way --dll command line behaves. When trying to install internet explorer 5.01 sp2 in a empty tree/fake windows (protected with WINEPREFIX), the IE 5 installer launchs: trace:process:CreateProcessA app (null) cmdline "acmsetup /T nashbase.stf /S C:\\windows\\msdownld.tmp\\AS00AA6F.tmp\\ /QT /g \"C:\\WINDOWS\\IE Setup Log.Txt\"" trace:process:find_exe_file looking for "acmsetup" trace:process:open_exe_file Trying native exe "C:\\windows\\msdownld.tmp\\AS00AA6F.tmp\\acmsetup.exe" trace:process:CreateProcessA starting "C:\\windows\\msdownld.tmp\\AS00AA6F.tmp\\acmsetup.exe" as Windows binary trace:process:PROCESS_InitWine starting process name="C:\\windows\\msdownld.tmp\\AS00AA6F.tmp\\acmsetup.exe" file=4 argv[0]="acmsetup" trace:process:PROCESS_InitWine starting Win32 binary "C:\\windows\\msdownld.tmp\\AS00AA6F.tmp\\acmsetup.exe" trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\windows\system\kernel32.dll' : builtin trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\windows\msdownld.tmp\AS00AA6F.tmp\advapi32.dll' : native trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\windows\system\user32.dll' : builtin trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\windows\system\rpcrt4.dll' : native trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\windows\system\ole32.dll' : builtin trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\windows\system\mpr.dll' : builtin trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\windows\system\lz32.dll' : builtin trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\windows\system\version.dll' : builtin trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\windows\msdownld.tmp\AS00AA6F.tmp\mssetup.dll' : native trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\WINDOWS\SYSTEM\shlwapi.dll' : builtin trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\windows\msdownld.tmp\AS00AA6F.tmp\comctl32.dll' : native trace:loaddll:MODULE_LoadLibraryExA Loaded module 'C:\WINDOWS\SYSTEM\SHELL32.dll' : builtin err:thunk:_loadthunk (commctrl.dll, Cctl1632_ThunkData16, comctl32.dll): Unable to load 'commctrl.dll', error 2 Note that tm comctl32.dll is native. However I have launched the instller with: wine --dll comctl32,commctrl=b ie5setup.exe In my config file however, I have the load order as "native, builtin" for both dlls. Is it possible that the --dll command line argument is not valid for subprocess/forks from the lauched process? Or is it a feature ? []'s Raul Dias

5 12

Problems compiling wine with Intel Compiler
by mkyprianou＠ICTV.com 26 Aug '02

26 Aug '02

I'm getting compiler errors when I try to compile wine with the Intel C++ 6.0 Compiler. This works fine with the g++ compiler. $ icc -c tstwine.cpp -I/usr/include/wine tstwine.cpp /usr/include/wine/basetsd.h(27): error: invalid combination of type specifiers typedef char __int8; ^ /usr/include/wine/basetsd.h(30): error: invalid combination of type specifiers typedef short __int16; ^ /usr/include/wine/basetsd.h(33): error: invalid combination of type specifiers typedef int __int32; ^ /usr/include/wine/basetsd.h(36): error: invalid combination of type specifiers typedef long long __int64; ^ /usr/include/wine/basetsd.h(36): error: invalid combination of type specifiers typedef long long __int64; ^ /usr/include/wine/winnt.h(59): catastrophic error: #error directive: You need gcc >= 2.7 to build Wine on a 386 # error You need gcc >= 2.7 to build Wine on a 386 ^ compilation aborted for tstwine.cpp (code 4) It looks like the wine headers need a level of the gcc version for the compiler, but since I'm not using gcc, there is no gcc version level set. Has anyone built wine with the Intel Compiler? If so, what was done to get this to work? Is any work being done to allow the usage of the Intel compiler with wine? thanx

2 1

How do I grab the screen contents referred to by an hDC?
by Bill Medland 26 Aug '02

26 Aug '02

Can anyone tell me how I can do this; I am trying to debug a problem with user32.dll under the native NT comctl32.? When drawing a ListView --debugmsg +relay shows three consecutive calls; SetRect (&rect...) FillRect (00000b04,&rect,00000082 (/* I reckon that is GetStockObject(BLACK_BRUSH)*/) DrawIcon (00000b04,0x40,0x60,hicon,...) Now as it stands I get random rubbish in the background of the icons (in the region covered by the mask). If I insert a hack into DrawIcon so that it does a FillRect (b04,the rectangle, GetStockObject (WHITE_BRUSH)) then things work out OK. (Also, the code seems to draw the icon to four hDCs and I can do the same hack on the second hDC to get the same effect but not the third and fourth). Firstly I don't understand how it has a chance of working given that it fills the rectangle with a black brush. So what I want is a way of seeing what the contents of the rectangular region is at the exit from the FillRect and the entrance to the DrawIconEx, based on the hDC. (Something like "give me the pixels of the 32x32 rectangle at this location of this hDC") Also it seems to me that these two function calls are consecutive, so if anything is changing I guess it would be as a result of code within comctl32 itself. Is there any other way that control can come back into the Wine code between a relay exit and entrance? Bill

1 0

Re: Add BiDi infrastructure
by Dimitrie O. Paun 26 Aug '02

26 Aug '02

On August 15, 2002 12:08 pm, Shachar Shemesh wrote: > configure.ac: > Added a check for the existance of the fribidi include file. No > check for the library, as we do not statically link with it. Like any programmer, I usually hate copying stuff around, but in this case isn't it better to simply have a copy of the fribidi heades around? I mean, we do the runtime load and checks anyway, why do we need to have the header around? If we don't use much out of the header, I would say simply include the prototype of the functions we're calling into the dlls/gdi/fribidi.c file directly, and be done with it. No configure, no macros, no ifdefs, no mess. -- Dimi.

4 12

RE: Problems compiling wine with Intel Compiler
by Patrik Stridvall 26 Aug '02

26 Aug '02

> On Mon, Aug 26, 2002 at 05:40:01PM +0200, Patrik Stridvall wrote: > > The Intel compiler like the Microsoft compiler probably > already supports > > __int{8,16,32,64}. > > > > This will have to be changed to > > > > #if !defined(_MSC_VER) && !defined(XXX) > > Or maybe a HAVE_xxx configure check? In general this is the way is should be done. However no rule without an exception. :-) Short explaination: In the header files it can't be done that way because the same header files must support multiple compilers and thus they can't have a shared config.h file.

1 0

RE: Problems compiling wine with Intel Compiler
by Patrik Stridvall 26 Aug '02

26 Aug '02

> I'm getting compiler errors when I try to compile wine with the Intel > C++ 6.0 Compiler. This works fine with the g++ compiler. > > $ icc -c tstwine.cpp -I/usr/include/wine > tstwine.cpp > /usr/include/wine/basetsd.h(27): error: invalid combination of type > specifiers > typedef char __int8; > ^ > > /usr/include/wine/basetsd.h(30): error: invalid combination of type > specifiers > typedef short __int16; > ^ > > /usr/include/wine/basetsd.h(33): error: invalid combination of type > specifiers > typedef int __int32; > ^ > > /usr/include/wine/basetsd.h(36): error: invalid combination of type > specifiers > typedef long long __int64; > ^ > > /usr/include/wine/basetsd.h(36): error: invalid combination of type > specifiers > typedef long long __int64; > ^ The Intel compiler like the Microsoft compiler probably already supports __int{8,16,32,64}. The typedef are protected by a #ifndef _MSC_VER This will have to be changed to #if !defined(_MSC_VER) && !defined(XXX) where XXX is some that indicate that the Intel compiler is used. Check the manual. > /usr/include/wine/winnt.h(59): catastrophic error: #error directive: > You need gcc >= 2.7 to build Wine on a 386 > # error You need gcc >= 2.7 to build Wine on a 386 > ^ > > compilation aborted for tstwine.cpp (code 4) > > It looks like the wine headers need a level of the gcc version for the > compiler, but since I'm not using gcc, there is no gcc version level > set. Presumably the Intel compiler like the Microsoft compiler already defined __cdecl and __stdcall. In that case similar to first problem above the line # elif defined(_MSC_VER) will have to be change to # elif defined(_MSC_VER) || defined(XXX) where XXX is the same as above. > Has anyone built wine with the Intel Compiler? If so, what was > done to get this to work? > > Is any work being done to allow the usage of the Intel > compiler with wine? Not that I know of.

2 1

Try for Thai input
by Supphachoke Suntiwichaya 25 Aug '02

25 Aug '02

hi,all I try modified keyboard.c for thai keymap static const char main_key_th[MAIN_LEN][4] = { "`~_%", "1!ๅ+", "2@/๑", "3#-๒", "4$ภ๓", "5%ถ๔", "6^ุู", "7&ึ฿", "8*ค๕", "9(ต๖", "0)จ๗", "-_ข๘", "=+ช๙", "\\|ฃฅ", "qQๆ๐", "wWไ\"", "eEำฎ", "rRพฑ", "tTะธ", "yYัํ", "uUี๊", "iIรณ", "oOนฯ", "pPยญ", "[{บฐ", "]}ล,", "aAฟฤ", "sSหฆ", "dDกฏ", "fFดโ", "gGเฌ", "hH้็", "jJ่๋", "kKาษ", "lLสศ", ";:วซ", "'\"ง.", "zZผ(", "xXป)", "cCแฉ", "vVอฮ", "bBิฺ", "nNื์", "mMท?", ",<มฒ", ".>ใฬ", "/?ฝฦ" }; but I still can't input Thai char.. err:keyboard:X11DRV_ToUnicode Please report: no char for keysym 0DB4 (Thai_dodek) : err:keyboard:X11DRV_ToUnicode (virtKey=47,scanCode=22,keycode=29,state=2010) err:keyboard:X11DRV_ToUnicode Please report: no char for keysym 0DCA (Thai_sosua) : err:keyboard:X11DRV_ToUnicode (virtKey=BA,scanCode=27,keycode=2E,state=2010) Help Me plz. mrchoke :::: Windows use cp874 for Thai :::: Linux use tis620 for Thai (Std) -- ========================================== Supphachoke Suntiwichaya WWW:http://jedi.links.nectec.or.th/~mrchoke UIN:10509387 TEL:+66-1-5408959 Love Me Love Linux.... ===========================================

1 0

IDA stopped working
by Shachar Shemesh 25 Aug '02

25 Aug '02

Hi all, somewhere back (havn't checked when, yet), some change in WINE made IDA (The Interactive Disassembler) stop working. I am talking about the bought version, have not checked the free one. Initial analysis (using IDA) suggest some heavy anti-disassembler techniques were used in this executable. One thing that is immediatly visible, however, is that the base address (as well as the address IDA is loading under windows) is different than the on in WINE. Wine: Execution starts at 0x006fb000 Windows, as well as static base address: Execution starts at 0x00599000 I believe this may be a hint, together with the fact it is employing some wierd arithmetics on the PC to stop static analysis using tools such as IDA ;-). Before I go through the tiring process of CVSing back and finding the patch that killed it, anyone happens to know who's using 00599000 and causing the conflict? Is there any simple way to check this? Shachar

2 2

extremely simple borland C builder app hangs
by Grzegorz Prokopski 25 Aug '02

25 Aug '02

Hi! This is still the same app, that used previously for testing freereport. How the app works. 0. show main window 1. call simple and meaningless freereport funtions (works now) 2. show [OK] dialog saying what above function returned 3. allow to press [OK] (of course ;-) 4. show main windows until it's closed. The app hangs after or around [2] it seems. Frames of main window and of the dialog show up, hovever their content is not painted and I cannot click/Enter/Esc/space anything. Final part of +relay log is this: 08072028:Ret x11drv.SetWindowPos() retval=00000001 ret=408c83ef 08072028:Ret x11drv.ShowWindow() retval=00000001 ret=408c7690 08072028:Call x11drv.MsgWaitForMultipleObjectsEx(00000000,00000000,00000000,00000000,00000000) ret=408deb6f 08072028:Ret x11drv.MsgWaitForMultipleObjectsEx() retval=00000102 ret=408deb6f 08072028:Call window proc 0x418a0fef (hwnd=00010021,msg=WM_USER+ac00,wp=00000000,lp=00000000) 08072028:Ret window proc 0x418a0fef (hwnd=00010021,msg=WM_USER+ac00,wp=00000000,lp=00000000) retval=00000000 08072028:Call x11drv.MsgWaitForMultipleObjectsEx(00000000,00000000,00000000,00000000,00000000) ret=408deb6f 08072028:Ret x11drv.MsgWaitForMultipleObjectsEx() retval=00000102 ret=408deb6f 08072028:Call window proc 0x418a0fef (hwnd=00010021,msg=WM_ENTERIDLE,wp=00000000,lp=0001002a) 08072028:Call user32.DefWindowProcA(00010021,00000121,00000000,0001002a) ret=40620e03 08072028:Ret user32.DefWindowProcA() retval=00000000 ret=40620e03 08072028:Ret window proc 0x418a0fef (hwnd=00010021,msg=WM_ENTERIDLE,wp=00000000,lp=0001002a) retval=00000000 08072028:Call x11drv.MsgWaitForMultipleObjectsEx(00000000,00000000,00000000,00000000,00000000) ret=408deb6f 08072028:Ret x11drv.MsgWaitForMultipleObjectsEx() retval=00000102 ret=408deb6f 08072028:Call x11drv.MsgWaitForMultipleObjectsEx(00000001,403c53e8,ffffffff,00000000,00000000) ret=408dee6d So it seems that MsgWaitForMultipleObjectsEx doesn't return if it has anything to do (first parameter is number of objects to wait for). It is almost exactly the same, as the "big app", which hangs at this: 08072028:Ret window proc 0x422408ed (hwnd=0001004d,msg=WM_PAINT,wp=00000000,lp=00000000) retval=0000000008072028:Call x11drv.MsgWaitForMultipleObjectsEx(00000000,00000000,00000000,00000000,00000000) ret=408deb6f 08072028:Ret x11drv.MsgWaitForMultipleObjectsEx() retval=00000102 ret=408deb6f 08072028:Call window proc 0x42240fef (hwnd=00010021,msg=WM_ENTERIDLE,wp=00000000,lp=00010052) 08072028:Call user32.DefWindowProcA(00010021,00000121,00000000,00010052) ret=40620e03 08072028:Ret user32.DefWindowProcA() retval=00000000 ret=40620e03 08072028:Ret window proc 0x42240fef (hwnd=00010021,msg=WM_ENTERIDLE,wp=00000000,lp=00010052) retval=00000000 08072028:Call x11drv.MsgWaitForMultipleObjectsEx(00000000,00000000,00000000,00000000,00000000) ret=408deb6f 08072028:Ret x11drv.MsgWaitForMultipleObjectsEx() retval=00000102 ret=408deb6f 08072028:Call x11drv.MsgWaitForMultipleObjectsEx(00000001,403c7eb8,ffffffff,00000000,00000000) ret=408dee6d It seems, that dll/x11drv/event.c:X11DRV_MsgWaitForMultipleObjectsEx is being called which in turn is calling scheduler/synchro.c:WaitForMultipleObjectsEx (I think). I looked at the code, but ATM it's still black magic for me - too many unknown wine interlans ivolved :-/ Is this known problem? What should I look for while debugging? Any pointers/hints? (I don't mean general wine-developers docs pointers) Regards Grzegorz B. Prokopski PS: I compiled wine from CVS, but not today's version (as it had some problems with compilation) but 2 weeks old.

4 5