June 2002 - Wine-Devel - list.winehq.org

Re: FPU patch for bug 586
by Ove Kaaven June 11, 2002

June 11, 2002

On Tue, 11 Jun 2002, admiral coeyman wrote: > Ove Kaaven, > > > > On Mon, 10 Jun 2002, admiral coeyman wrote: > > > > > This patch adds the self-modifying code part of the FPU interrupts 34->3b and > > > moves int 3d into the winedos section. > > > > Oh, I've been meaning to reply, but hadn't got around to it yet... it > > makes no sense to use MapDosToLinear(MapRealToLinear()). It should be > > obvious from its name that MapRealToLinear already converted to a linear > > address, so that's all you need. And MapLinearToDos will probably *not* do > > what you want, you should change the real-mode pointer directly instead, > > as in general you just can't convert a linear address back to real mode. > > Otherwise, the code is probably OK... > > > I was unsure of the memory handling functions. Both of these functions report > that they return linear addresses, but I read one as a dos linear address and > the other as a wine linear address. It was a mistake on my part. I read it as > though I had to convert the seg:offset code to a dos linear address, then > convert it to a wine linear address. It may have made more sense to me if I > hadn't been worried about the addresses being in the wrong order. Well, the meanings of these address types are, if I remember right: Real = 16-bit real mode seg:ofs address Dos = 32-bit offset from DOSMEM_dosmem (aka DOSMEM_MemoryBase()) (unless it's above 1MB, then it's considered a linear pointer) (after the DOS restructure, DOSMEM_dosmem is always 0, which is probably why it worked for you) Linear = 32-bit linear pointer that you can dereference in Wine code I consider the "Dos" memory type pretty useless, there's almost never a reason to use it instead of linear memory or something... > What I read says that the stack's lowest address contains IP and the word > above that is the code segment for the return. IP:CS I'm not sure that I > could just point a segment:offset converting function at an address in that > format and get the linear address. Well, you can, as long as we stick to x86 (might be trickier if we were to write a CPU emulator, but I guess there would be lots of places to change then anyway). > I've done the reverse conversion be backing IP up 2 bytes. You have two versions of the Int3dHandler in your patch. One of them isn't changed... Anyway, if it helps, here's how I might have written your routine (without the error checking), if I thought that perhaps these interrupts are also used by win16 apps (otherwise the second CTX_SEG_OFF_TO_LIN(...) can be replaced with PTR_REAL_TO_LIN(stack[1], stack[0])) void FPU_ModifyCode(CONTEXT86 *context, BYTE Opcode) { WORD *stack = CTX_SEG_OFF_TO_LIN(context, context->SegSs, context->Esp) BYTE *code = CTX_SEG_OFF_TO_LIN(context, stack[1], stack[0]); code[-2] = 0x9b; code[-1] = Opcode; stack[0] -= 2; }

1 0

Re: GLU32 patch
by Marcus Meissner June 11, 2002

June 11, 2002

On Mon, Jun 10, 2002 at 11:52:34PM -0400, Vincent B?ron wrote: > The glu32 dll is now compliant with MSDN. It can be compiled with a > libGLU 1.1, but preferably by a libGLU 1.2. > > Changelog: > - Modified the glu32 prototypes to match MSDN. > - Made it work for libGLU 1.1 and 1.2. ... > -@ stdcall gluCheckExtension(str ptr) wine_gluCheckExtension Why is this function gone? And you will need to add checks for glu.h in configure.ac too. Ciao, Marcus

2 1

Problems converting DLL to unix SO with winelib
by List vmn June 11, 2002

June 11, 2002

Hello Wine-devel, I finaly got the expected results from my simple tests to get a winelib DLL and test application to work. I am pretty sure that I am still doing something wrong in these tests !!!!! Can somebody say if the below setup is correct and if not what would be the correct way to create a winelib "dll" and then call it from a winelib executable. I am using wine-20020509. =>I start with a simple DLL file called mytest.c and use winemaker to create a libmytest.so bash#cd ~/mytest.dll bash#ls mytest.c bash#cat mytest.c //mytest.c #include <windows.h> #include <stdio.h> BOOL APIENTRY DllMain(HINSTANCE hinstDLL , DWORD fdwReason, LPVOID lpvReserved ) { printf("DllMain called\n"); return TRUE; } void usefullFunction() { printf("Usefull function got called\n"); } //end of mytest.c bash#winemaker --dll --nomfc . <output suppressed> bash#cat libmytest.spec name mytest type win32 mode dll init DllMain import kernel32.dll import ntdll.dll bash# ./configure <output suppressed> bash# make && make install gcc -c -g -O2 -fPIC -D_REENTRANT -DWINELIB -I. -I/usr/local/include/wine -o mytest.o mytest.c ld -r mytest.o -o libmytest.tmp.o strip --strip-unneeded libmytest.tmp.o LD_LIBRARY_PATH="/usr/local/lib:$LD_LIBRARY_PATH" /usr/local/bin/winebuild -fPIC -L/usr/local/lib/wine -sym libmytest.tmp.o -o libmytest.spec.c -spec ./libmytest.spec gcc -c -g -O2 -fPIC -D_REENTRANT -DWINELIB -I. -I/usr/local/include/wine -o libmytest.spec.o libmytest.spec.c gcc -shared -Wl,-rpath,/usr/local/lib -Wl,-Bsymbolic -o libmytest.so mytest.o libmytest.spec.o -lwine -lwine_unicode -lwine_uuid -lm _list=""; for i in $_list; do (cd $i; make install) || exit 1; done _list=" "; for i in $_list; do install $i /usr/local/bin; done _list="libmytest.so"; for i in $_list; do install $i /usr/local/lib; done bash# #Now I have a /usr/local/lib/libmytest.so file \ #I think the next commands are wrong... bash# cd /usr/local/lib bash# cp libmytest.so wine/mytest.dll.so bash# ln -s wine/mytest.dll.so libmytest.dll.so bash# #This was done so that I can write \ # import mytest.dll \ # in a spec file and "register with wine" this so =>I now create a simple winelib application to call mytest called "callmytest" =>Call my test is defined in callmytest.c bash# cd ~/callmytest bash#ls callmytest.c bash# cat callmytest.c //start of callmytest.c #include "windows.h" extern usefullFunction(); int WINAPI WinMain(HINSTANCE a,HINSTANCE b,LPSTR c ,int d) { usefullFunction(); } //end of callmytest.c bash# winemaker --guiexe -imytest.dll -lmytest.dll . <output suppressed> bash# cat callmytest.spec name callmytest type win32 mode guiexe init WinMain import kernel32.dll import ntdll.dll import mytest.dll bash# ./configure <output suppressed> bash# make gcc -c -g -O2 -fPIC -D_REENTRANT -DWINELIB -I. -I/usr/local/include/wine -o callmytest.o callmytest.c ld -r callmytest.o -o callmytest.tmp.o strip --strip-unneeded callmytest.tmp.o LD_LIBRARY_PATH="/usr/local/lib:$LD_LIBRARY_PATH" /usr/local/bin/winebuild -fPIC -L/usr/local/lib/wine -sym callmytest.tmp.o -o callmytest.spec.c -spec ./callmytest.spec gcc -c -g -O2 -fPIC -D_REENTRANT -DWINELIB -I. -I/usr/local/include/wine -o callmytest.spec.o callmytest.spec.c gcc -shared -Wl,-rpath,/usr/local/lib -Wl,-Bsymbolic -o callmytest.so callmytest.o callmytest.spec.o -lmytest.dll -lwine -lwine_unicode -lwine_uuid -lm test -f callmytest || ln -s /usr/local/bin/wine callmytest bash# ./callmytest DllMain called Usefull function got called DllMain called QUESTION 1: I think that adding -imytest.dll -lmytest.dll to winemaker --guiexe -imytest.dll -lmytest.dll . is wrong. QUESTION 2: Is DllMain called when the DLL is loaded and then unloaded ? Can anyone help me on this? Thanks, Miguel Feitosa

1 0

Problems converting DLL to unix SO with winelib
by mfeitosa June 11, 2002

June 11, 2002

Hello Wine-devel, I finaly got the expected results from my simple tests to get a winelib DLL and test application to work. I am pretty sure that I am still doing something wrong in these tests !!!!! Can somebody say if the below setup is correct and if not what would be the correct way to create a winelib "dll" and then call it from a winelib executable. I am using wine-20020509. =>I start with a simple DLL file called mytest.c and use winemaker to create a libmytest.so bash#cd ~/mytest.dll bash#ls mytest.c bash#cat mytest.c //mytest.c #include <windows.h> #include <stdio.h> BOOL APIENTRY DllMain(HINSTANCE hinstDLL , DWORD fdwReason, LPVOID lpvReserved ) { printf("DllMain called\n"); return TRUE; } void usefullFunction() { printf("Usefull function got called\n"); } //end of mytest.c bash#winemaker --dll --nomfc . <output suppressed> bash#cat libmytest.spec name mytest type win32 mode dll init DllMain import kernel32.dll import ntdll.dll bash# ./configure <output suppressed> bash# make && make install gcc -c -g -O2 -fPIC -D_REENTRANT -DWINELIB -I. -I/usr/local/include/wine -o mytest.o mytest.c ld -r mytest.o -o libmytest.tmp.o strip --strip-unneeded libmytest.tmp.o LD_LIBRARY_PATH="/usr/local/lib:$LD_LIBRARY_PATH" /usr/local/bin/winebuild -fPIC -L/usr/local/lib/wine -sym libmytest.tmp.o -o libmytest.spec.c -spec ./libmytest.spec gcc -c -g -O2 -fPIC -D_REENTRANT -DWINELIB -I. -I/usr/local/include/wine -o libmytest.spec.o libmytest.spec.c gcc -shared -Wl,-rpath,/usr/local/lib -Wl,-Bsymbolic -o libmytest.so mytest.o libmytest.spec.o -lwine -lwine_unicode -lwine_uuid -lm _list=""; for i in $_list; do (cd $i; make install) || exit 1; done _list=" "; for i in $_list; do install $i /usr/local/bin; done _list="libmytest.so"; for i in $_list; do install $i /usr/local/lib; done bash# #Now I have a /usr/local/lib/libmytest.so file \ #I think the next commands are wrong... bash# cd /usr/local/lib bash# cp libmytest.so wine/mytest.dll.so bash# ln -s wine/mytest.dll.so libmytest.dll.so bash# #This was done so that I can write \ # import mytest.dll \ # in a spec file and "register with wine" this so =>I now create a simple winelib application to call mytest called "callmytest" =>Call my test is defined in callmytest.c bash# cd ~/callmytest bash#ls callmytest.c bash# cat callmytest.c //start of callmytest.c #include "windows.h" extern usefullFunction(); int WINAPI WinMain(HINSTANCE a,HINSTANCE b,LPSTR c ,int d) { usefullFunction(); } //end of callmytest.c bash# winemaker --guiexe -imytest.dll -lmytest.dll . <output suppressed> bash# cat callmytest.spec name callmytest type win32 mode guiexe init WinMain import kernel32.dll import ntdll.dll import mytest.dll bash# ./configure <output suppressed> bash# make gcc -c -g -O2 -fPIC -D_REENTRANT -DWINELIB -I. -I/usr/local/include/wine -o callmytest.o callmytest.c ld -r callmytest.o -o callmytest.tmp.o strip --strip-unneeded callmytest.tmp.o LD_LIBRARY_PATH="/usr/local/lib:$LD_LIBRARY_PATH" /usr/local/bin/winebuild -fPIC -L/usr/local/lib/wine -sym callmytest.tmp.o -o callmytest.spec.c -spec ./callmytest.spec gcc -c -g -O2 -fPIC -D_REENTRANT -DWINELIB -I. -I/usr/local/include/wine -o callmytest.spec.o callmytest.spec.c gcc -shared -Wl,-rpath,/usr/local/lib -Wl,-Bsymbolic -o callmytest.so callmytest.o callmytest.spec.o -lmytest.dll -lwine -lwine_unicode -lwine_uuid -lm test -f callmytest || ln -s /usr/local/bin/wine callmytest bash# ./callmytest DllMain called Usefull function got called DllMain called QUESTION 1: I think that adding -imytest.dll -lmytest.dll to winemaker --guiexe -imytest.dll -lmytest.dll . is wrong. QUESTION 2: Is DllMain called when the DLL is loaded and then unloaded ? Can anyone help me on this? Thanks, Miguel Feitosa

1 0

RE: Linux patch for supporting shortcuts and symlinks on VFAT
by Patrik Stridvall June 11, 2002

June 11, 2002

> On Tue, 11 Jun 2002, Patrik Stridvall wrote: > [...] > > However it would be nice be nice to handle .lnk files as symlinks > > somehow. > > I'm not sure this is very useful, Well, just a feeling, but I have always been irritated about that the .lnk files are not real symlinks. Wine is not only supposed to be as good as Windows but also better if possible. > especially since few > symlinks will be > supported in any case. For instance in his proposal he ignores: > * all symlinks to absolute paths. That would be pretty much all > symlinks created under windows Yes. However you could have a drive table like echo /windows-c-drive > /proc/windows/drive/c echo /windows-d-drive > /proc/windows/drive/d So the kernel could do the look up. BTW this would also means that Wine could read /proc/windows/drive and map the drives. > * all UNC symlinks except those to \\localhost. But don't these paths > have a drive letter too? Even worse they have the first part after \\localhost should be the name of the share so that means that we need to have a series of /proc/windows/share directories or something as well... > * I guess he would also ignore the command line options > specified in shortcuts to executables > * and also the current directory Hmm. Well, that is a problem as well. Perhaps you could link to some sort of autogenerated shell script in /proc/windows/links say you access /mnt/c/test.lnk and you get a shell script in /proc/windows/links/mnt/c/test.lnk or something that /mnt/c/test.lnk symlinks to. > * and the icon of course. Which then means tools like konqueror or > nautilus cannot extract the icon either unless they make use of the > proposed ioctl hack. I don't consider this a problem. However solving the other problem is probably to ugly eventhough they are solvable in some meaning. > And if the goal is to be able to create symlinks on a filesystem that > does not support them, I believe the right way to do so would > be to have > a module sit on top of the file system that would do things > ala umsdos. > This would for instance allow one to also get things like ownership, > file permissions, device files, socket files, etc. And if you > are going > to use a VFAT filesystem as a regular Unix filesystem, having > permissions at least while in Linux seems like it would be a very good > thing. Indeed. It would be better if it was generalized somehow. Hmm... > > > * on Windows if you open("foo.lnk") you get the .lnk > files, not the > > > file it 'points' to. On Linux you would get the file it points to > > > instead which is a different behavior. > > > > True. However we could check if the extension is .lnk in OpenFile > > as friends and then call special funcitons like readlink to read the > > file. But it probably would be a too ugly hack to be worth it. > > Yes that's the soluti^H^H^H^H ugly hack that was suggested as a > workaround. It would probably involve an ioctl of some sort > though here > we are talking about reading the content of a file, not just > a couple of > metadata bits, not sure how ioctls handle such things (maybe > that's not > a problem, I'm not very familiar with ioctls). Doing it is not the problem. Accepting the overhead and the ugly hack is the real problem... > The other issue is that this workaround is not implemented > yet so we may > get stuck in a bad place for a while. True. > [...] > > Anyway, the big question seem to be whether we should > oppose the patch > > or not. It might could trouble for Wine is such a feature > exists some > > user might active it so we have to handle that case. > > It is proposed as the default behavior. This means it would likely end > up as the default in most distributions, so many users will > be hit by it > even if they did not activate it. > > I definitely oppose it. I believe it is not only harmful to Wine, but > also the wrong thing for Linux. As the default behavior I also oppose it. Unfortunely I very much fear that I won't be possible to do it in an acceptable way at all...

1 0

RE: Linux patch for supporting shortcuts and symlinks on VFAT
by Patrik Stridvall June 11, 2002

June 11, 2002

> On Tue, 11 Jun 2002, Patrik Stridvall wrote: > > > I notice that I patch have been posted to the linux-kernel that > > might be of intrest to us. > > > > vfat patch for shortcut display as symlinks for 2.4.18 > > http://marc.theaimsgroup.com/?l=linux-kernel&m=102358795411222&w=2 > > > > It seem that at least some people are against it inclusion so I > > wonder whether we should support inclusion of the patch or not. > > > > I'm personally a little doubtful that it is a good idea... > > > I think it's a bad idea. The reason is that the VFAT driver > is the wrong > abstraction layer to support the '.lnk' files: Probably. The more I think about the more I worry about the consequences. However it would be nice be nice to handle .lnk files as symlinks somehow. > * on Windows if you open("foo.lnk") you get the .lnk files, not the > file it 'points' to. On Linux you would get the file it points to > instead which is a different behavior. True. However we could check if the extension is .lnk in OpenFile as friends and then call special funcitons like readlink to read the file. But it probably would be a too ugly hack to be worth it. > * Windows supports .lnk files on FAT, VFAT, NTFS, ISO9660, etc. So if > such support is added in the Linux kernel, it should be added > to all of > the above filesystems. And then there is no reason not to add it to > ext2, NFS, etc! True. The feature should probably be at file system independant layer, if anywhere. > This would also hurt Wine as: > > * it would prevent us from reading the information in the '.lnk' > file... at least for 'supported' '.lnk' files Well, as I said above, we can read them with readlink and friends. > * it's not entirely clear to me what they do with unsupported '.lnk' > files. Are they just dead symlinks or can one read their contents? In > the first case Wine is SOL again, and in the latter case we'll have to > play games to know which kind we got. As I said, you always can read a symlinks content with readlink. > * it was suggested to implement a hack to let Wine access the '.lnk' > data. Why implement a hack which is going to be Linux specific when > doing nothing works just fine and on any Unix system? True. > * making it a default option does not help Wine at all. Then > we have to > keep telling users that they have to modify their fstab if they want > Wine to work. Yes, that is a problem. See below. > The right level to implement symlink support is: > > * in Wine. Of course! :-) > * in an LD_PRELOAD library. Then they would work for all filesystems, > be selectable on a per-user or even per-process basis. Of course it > would most likely not work with Wine (ld_preload libraries seldom do). > > * as an option in the KDE/Gnome file browsers. That's the layer which > seems closest to the Windows shell 'layer'. True. Anyway, the big question seem to be whether we should oppose the patch or not. It might could trouble for Wine is such a feature exists some user might active it so we have to handle that case. BTW. I just though of another thing. Perhaps we should support the opposite. Translate existing symlinks to .lnk files? > This looks like it could be the next 'unhide' thing. hmmm, probably, > people not subscribed to linux-kernel missed the 'unhide' issue. See: > * Jeremy White - isofs unhide option: troubles with Wine > http://marc.theaimsgroup.com/?l=linux-kernel&m=102230123607864&w=2 I basicly agree with your opinion. The hidden attribute is metadata meant as advice for the application not something the operating system should concern itself.

2 1

Linux patch for supporting shortcuts and symlinks on VFAT
by Patrik Stridvall June 11, 2002

June 11, 2002

I notice that I patch have been posted to the linux-kernel that might be of intrest to us. vfat patch for shortcut display as symlinks for 2.4.18 http://marc.theaimsgroup.com/?l=linux-kernel&m=102358795411222&w=2 It seem that at least some people are against it inclusion so I wonder whether we should support inclusion of the patch or not. I'm personally a little doubtful that it is a good idea... Comments?

2 1

GLU32 status & update
by Vincent Béron June 11, 2002

June 11, 2002

After investigating bug #783 (Wine20020605: Problems with build of dlls/glu32), here are my conclusions so far (not finished yet, but I need a bit of advice on next direction to pursue): First, we need to not compile glu32 is no libGLU.so is found. The test for libGLU is already in configure, the only thing missing is some logic in the dlls/Makefile (or maybe dlls/glu32/Makefile?) to not build it. Next, the version of the GLU implementation of MS is 1.2. That's fine as long as the same (or more recent) version is found while building. I'm currently using Mesa 3.4.2 on RH 7.2, which is GLU 1.1 compliant. The GL/glu.h has the defines for 1.2 and 1.3, but the lib/libGLU.so doesn't have any implementation. Some more recent distributions (eg, RH 7.3) has the SGI GLU implementation which is 1.3. So we need to check this also. What would be the best thing to do if the native GLU is < 1.2? Not build dlls/glu32 at all? There's a way to map 1.1 call to 1.2 or beyond, but not the other way around (or I'm not aware of it). I finished correcting the prototypes (compliant with MSDN) in glu.c for a native GLU 1.2. Now I'd like to have some input as to what to do wrt a native GLU 1.1. If you have some advice, I'll appreciate them. Vincent

2 1

bugs.winehq.com upgraded
by Jeremy Newman June 11, 2002

June 11, 2002

I have *finally* gotten around to upgrading our Bugzilla. We are now running version 2.14.2. This build seems to solve some long overdue usability complaints. One of the big differences I noticed is that there is levels of "Resolved" now. If you notice any issues with the new bugzilla, please submit issues about it to the "Bugzilla" product. -- __ _WebGeek/NetAdmin CodeWeavers -= http://www.codeweavers.com | \| |_____ __ ___ __ __ _ _ _ -= http://www.dracowulf.com | .` / -_) | | / ' \/ _` | ' \ -= mailto:jnewman(a)codeweavers.com |_|\_\___|\_/\_/|_|_|_\__,_|_||_| -= ICQ: 1842980 Yahoo: laxdragon

2 3

tabctl32.ocx
by Medland, Bill June 10, 2002

June 10, 2002

Anyone any ideas? Our application basically results in running an OLE container which loads Visual Basic ocx's, some of which use the TABCTL32.OCX tab control. However the TABCTL32.OCX doesn't "load" in some way. (If I put a Debug MessageBox into the Visual Basic I get "Failed to load control 'SSTab' from TABCTL32.OCX." I have done a relay trace to see what happens between the end of the OCX's init routine (which succeeds) and the message and there is no obvious complaint. (This is with native ole32, oleaut and rpcrt4) Bill

1 0