- Wine-Devel - list.winehq.org

set_file_pointer and 64 bit file access
by Bill Medland June 18, 2001

June 18, 2001

I see we are in the middle of converting set_file_pointer towards 64 bit file systems. Currently the code is confused as to when to change from unsigned to signed int for the low 32 bits. Should it happen in SetFilePointer before sending the request (which means we end up changing the request format) or should it happen within set_file_pointer (after changing the argument back to a signed int; this would also require work in SetFilePointer to get it to set up the request correctly). Any views? Bill

1 0

RE: Use the libreadline in Winedbg, if available
by Ove Kaaven June 18, 2001

June 18, 2001

On Mon, 18 Jun 2001, Patrik Stridvall wrote: > A minor notice. Readline is GPL:ed (not LGPL:ed) which is not so good. > While it is without doubt legal for us to use and distribute, > it might cause trouble for other people. Not as long as it's optional (that patch even supported ./configure --without-readline, so no trouble, such people would just disable it).

1 0

RE: Use the libreadline in Winedbg, if available
by Patrik Stridvall June 18, 2001

June 18, 2001

> On Mon, Jun 18, 2001 at 04:10:14PM +0200, Uwe Bonnes wrote: > > Hallo, > > > > at present, wine uses unconditionally a self supplied readline > > function. Appended patch tries to find out if libreadline > is available > > and uses libreadline if available. I configured with and without > > libreadline and successfully compiled winedbg on my system. > > libreadline is a static library on my system and winedbg > grew 90 kByte > > stripped and 110 kByte compiled with -g. > > You are not allowed to link readline aganst non GPL code to > my knowledge. No, it is enough that it is GPL compatible which the Wine license is. Whether we should is another matter though. Just linking it to the debugger is rather harmless IMHO though.

2 1

Re: Use the libreadline in Winedbg, if available
by Marcus Meissner June 18, 2001

June 18, 2001

On Mon, Jun 18, 2001 at 04:10:14PM +0200, Uwe Bonnes wrote: > Hallo, > > at present, wine uses unconditionally a self supplied readline > function. Appended patch tries to find out if libreadline is available > and uses libreadline if available. I configured with and without > libreadline and successfully compiled winedbg on my system. > libreadline is a static library on my system and winedbg grew 90 kByte > stripped and 110 kByte compiled with -g. You are not allowed to link readline aganst non GPL code to my knowledge. Ciao, Marcus

1 0

Dosvm not initialized for Direct port access
by Uwe Bonnes June 18, 2001

June 18, 2001

Hallo, some license manager dll (lmgr327a) manages to read IOPORT 0x40 from 32-bit code. This results in an access to Dosvm.GetTimer() in line 297 in ioports.c, and as Dosvm isn't initialized this results in a crash. =>0 0x00000000 (ebp=404e4b38) 1 0x400cd595 (IO_inport+0x205(port=0x40, size=0x1) [ioports.c:297] in libntdll.so) (ebp=404e4b38, null call assumed) 2 0x400a7405 (INSTR_inport+0x25(port=0x40, size=0x1, context=0x404e4c50) [instr.c:328] in libntdll.so) (ebp=404e4b58) 3 0x400a888c (INSTR_EmulateInstruction+0x12cc(context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50, context=0x404e4c50) [instr.c:745] in libntdll.so) (ebp=404e4bc4) 4 0x400733c9 (do_segv+0x109(context=0x404e4c50, trap_code=0xd, cr2=0x0, err_code=0x0) [signal_i386.c:671] in libntdll.so) (ebp=404e4c2c) 5 0x4007368d (segv_handler+0x3d(__signal=0xb, __context=0x0) [signal_i386.c:787] in libntdll.so) (ebp=404e4d1c) 6 0x4029fb58 (__restore [sigaction.c] in libc.so.6) (ebp=405f57a0) 7 0x1c0b0004 (LMGR327A.DLL..reloc+0xbfe2004) (ebp=04007242) Bye -- Uwe Bonnes bon(a)elektron.ikp.physik.tu-darmstadt.de Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt --------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

1 0

16-bit and 32-bit separation
by Patrik Stridvall June 18, 2001

June 18, 2001

[Since the subject it up because of Dimitry's "Move 16-bit COMM stuff to USER, where it belongs" patch] My new tool winapi_fixup could without to much effort be extended to facillitate separation of 16-bit and 32-bit stuff by moving the 16-bit parts of a file (example: foo.c) to a new 16-bit only file (example: foo16.c). Of course because of dependecies, includes and such things it can only be partly automatic and it will in most cases require quite a lot editing afterwards, so the questions (primarily to Alexandre but other replies welcome to) are: 1. Will/Should Alexandre accept such patches at all? 2. Is it a useful effort to get the 16-bit stuff out of the way so we don't have to see it in the files when editing the 32-bit stuff? 3. Is it a useful effort to make it possible to have --disable-win16 configure option to compile out as much Win16 stuff as possible in order to save space in the binaries for say embedded stuff. 4. While this certainly it not a high or even medium priority effort, should it have low, very low or extremly low priority? PS. I have other stuff planed so I certainly do not lack things to do. However since the subject is up I might just as well ask.

1 0

SIGSEGV while trying to run Word97
by Heiko Nardmann June 17, 2001

June 17, 2001

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I am continously getting a dialog "Nicht genügend Arbeitsspeicher oder unzureichender Speicherplatz, um Word auszuführen" (German version). After trying to examine this problem with debugmsg all (without real success) I added an strace call to the whole call. This resulted in a log file which tells me that a SIGSEGV occurrs. Here is a part of the strace log file which contains the SIGSEGV message. After that the appropriate 'debugmsg all' part follows. What can be done to shed more light onto this problem? strace log: write(2, "08069280:trace:win32:_LeaveSysLe"..., 112) = 112 write(2, "08069280:Ret user32.CreateWindo"..., 68) = 68 write(2, "08069280:RET mso97.486: _MsoCre"..., 84) = 84 write(2, "08069280:trace:heap:HeapFree (40"..., 74) = 74 write(2, "08069280:Call kernel32.GetVersio"..., 49) = 49 write(2, "08069280:Ret kernel32.GetVersio"..., 65) = 65 write(2, "08069280:Call kernel32.GetVersio"..., 60) = 60 write(2, "08069280:Ret kernel32.GetVersio"..., 68) = 68 write(2, "08069280:Call kernel32.GlobalMem"..., 65) = 65 time(NULL) = 992531763 time(NULL) = 992531763 open("/proc/meminfo", O_RDONLY) = 11 fstat(11, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x4035c000 read(11, " total: used: free:"..., 1024) = 319 read(11, "", 1024) = 0 close(11) = 0 munmap(0x4035c000, 4096) = 0 write(2, "08069280:Ret kernel32.GlobalMem"..., 73) = 73 write(2, "08069280:CALL mso97.394: _MsoFCr"..., 147) = 147 write(2, ") ret=30011010\n", 15) = 15 write(2, "08069280:Call kernel32.VirtualAl"..., 86) = 86 write(2, "08069280:trace:virtual:VirtualAl"..., 68) = 68 mmap2(NULL, 131072, PROT_NONE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = -1 ENOSYS (Function not implemented) old_mmap(NULL, 131072, PROT_NONE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x40e90000 munmap(0x40ea0000, 65536) = 0 write(2, "View: 40e90000 - 40e9ffff (vallo"..., 35) = 35 write(2, " 40e90000 - 40e9ffff -----\n", 32) = 32 write(2, "08069280:Ret kernel32.VirtualAl"..., 67) = 67 write(2, "08069280:Call kernel32.VirtualAl"..., 86) = 86 write(2, "08069280:trace:virtual:VirtualAl"..., 68) = 68 write(2, "08069280:trace:virtual:VIRTUAL_S"..., 63) = 63 mprotect(0x40e90000, 4096, PROT_READ|PROT_WRITE) = 0 write(2, "View: 40e90000 - 40e9ffff (vallo"..., 35) = 35 write(2, " 40e90000 - 40e90fff c-rW-\n", 32) = 32 write(2, " 40e91000 - 40e9ffff -----\n", 32) = 32 write(2, "08069280:Ret kernel32.VirtualAl"..., 67) = 67 - --- SIGSEGV (Speicherzugriffsfehler) --- write(2, "08069280:RET mso97.394: _MsoFCr"..., 126) = 126 rt_sigprocmask(SIG_BLOCK, [HUP INT ALRM IO], [], 8) = 0 write(5, "F\0\0\0\0\0\34\1<\0336@\1\0\0\0\304\357\17@\0\0\0\0\0\0"..., 64) = 64 read(6, "F\0\0\0\0\0\34\1\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 64) = 64 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 write(2, "08069280:trace:seh:EXC_CallHandl"..., 87) = 87 write(2, "08069280:trace:seh:EXC_RtlUnwind"..., 55) = 55 write(2, "08069280:trace:seh:EXC_CallHandl"..., 87) = 87 write(2, "08069280:trace:seh:EXC_CallHandl"..., 54) = 54 - --- SIGSEGV (Speicherzugriffsfehler) --- write(2, "08069280:trace:seh:EXC_RtlRaiseE"..., 63) = 63 rt_sigprocmask(SIG_BLOCK, [HUP INT ALRM IO], [], 8) = 0 write(5, "F\0\0\0\0\0\34\1<\0336@\1\0\0\0\304\357\17@\0\0\0\0\0\0"..., 64) = 64 read(6, "F\0\0\0\0\0\34\1\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 64) = 64 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0 wine debug log: 08069280:trace:win32:_LeaveSysLevel (0x40753a74, level 2): thread 0x8069280 (fs 008f, pid 24063) count after 0 08069280:Ret user32.CreateWindowExW() retval=00000138 ret=306cc216 08069280:RET mso97.486: _MsoCreateMonitorWindow(a)0() retval = 00000000 ret=3000162f 08069280:trace:heap:HeapFree (40360000,00000002,403a4a80): returning TRUE 08069280:Call kernel32.GetVersion() ret=30001984 08069280:Ret kernel32.GetVersion() retval=05650004 ret=30001984 08069280:Call kernel32.GetVersionExA(405c686c) ret=300019b5 08069280:Ret kernel32.GetVersionExA() retval=00000001 ret=300019b5 08069280:Call kernel32.GlobalMemoryStatus(405c6900) ret=300019e2 08069280:Ret kernel32.GlobalMemoryStatus() retval=405c6900 ret=300019e2 08069280:CALL mso97.394: _MsoFCreateIPref(a)28(<unknown, check return>08069280:trace:heap:HeapAlloc (40360000,00000002,00000040): returning 403a4b98 ) ret=30011010 08069280:Call kernel32.VirtualAlloc(00000000,0000fffc,00002000,00000001) ret=3070b360 08069280:trace:virtual:VirtualAlloc 00000000 0000fffc 2000 00000001 View: 40e90000 - 40e9ffff (valloc) 40e90000 - 40e9ffff ----- 08069280:Ret kernel32.VirtualAlloc() retval=40e90000 ret=3070b360 08069280:Call kernel32.VirtualAlloc(40e90000,00001000,00001000,00000004) ret=3070b38b 08069280:trace:virtual:VirtualAlloc 40e90000 00001000 1000 00000004 08069280:trace:virtual:VIRTUAL_SetProt 40e90000-40e90fff c-rW- View: 40e90000 - 40e9ffff (valloc) 40e90000 - 40e90fff c-rW- 40e91000 - 40e9ffff ----- 08069280:Ret kernel32.VirtualAlloc() retval=40e90000 ret=3070b38b 08069280:RET mso97.394: _MsoFCreateIPref(a)28(30011020,00000000,08069280:trace:seh:EXC_RtlRaiseException code=c0000005 flags=0 08069280: queue_exception_event( first=1, record={context={flags=00000000,eax=00000000,ebx=400fefc4,ecx=4d535744,edx=00000000,esi=00000000,edi=00000007,ebp=405c6790,eip=400c8aca,esp=405c66cc,eflags=00010246,cs=0023,ds=002b,es=002b,fs=008f,gs=0000,dr0=00000000,dr1=00000000,dr2=00000000,dr3=00000000,dr6=00000000,dr7=00000000,float={00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000}},rec={code=c0000005,flags=0,rec=(nil),addr=0x400c8aca,params={0,4d535744}} ) 08069280: queue_exception_event() = 0 { handle=0 } 08069280:trace:seh:EXC_CallHandler calling handler at 0x4005b520 code=c0000005 flags=0 08069280:trace:seh:EXC_RtlUnwind code=c0000005 flags=2 08069280:trace:seh:EXC_CallHandler calling handler at 0x4005ad80 code=c0000005 flags=2 08069280:trace:seh:EXC_CallHandler handler returned 1 08069280:trace:seh:EXC_RtlRaiseException code=c0000005 flags=0 08069280: queue_exception_event( first=1, record={context={flags=00000000,eax=00000000,ebx=400fefc4,ecx=4d535744,edx=00000000,esi=00000000,edi=00000007,ebp=405c6790,eip=400c8baa,esp=405c66cc,eflags=00010246,cs=0023,ds=002b,es=002b,fs=008f,gs=0000,dr0=00000000,dr1=00000000,dr2=00000000,dr3=00000000,dr6=00000000,dr7=00000000,float={00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000,00000000}},rec={code=c0000005,flags=0,rec=(nil),addr=0x400c8baa,params={0,4d535744}} ) 08069280: queue_exception_event() = 0 { handle=0 } 08069280:trace:seh:EXC_CallHandler calling handler at 0x4005b520 code=c0000005 flags=0 08069280:trace:seh:EXC_RtlUnwind code=c0000005 flags=2 08069280:trace:seh:EXC_CallHandler calling handler at 0x4005ad80 code=c0000005 flags=2 08069280:trace:seh:EXC_CallHandler handler returned 1 - -- Heiko Nardmann (Dipl.-Ing.), h.nardmann(a)secunet.de, Software Development secunet Security Networks AG - Sicherheit in Netzwerken (www.secunet.de) Weidenauer Str. 223-225, D-57076 Siegen Tel. : +49 271 48950-13, Fax : +49 271 48950-50 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.4 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE7KcuPpm53PRScYygRAlfZAJ91r/8CwDzWpu0TU0vB+FRP18w14QCg4sCK GBnoWm0xkKgHgFucJcNrsno= =8hel -----END PGP SIGNATURE-----

5 6

CS_PARENTDC problems
by Malte.Starostik＠t-online.de June 17, 2001

June 17, 2001

Hi, we had some serious painting problems with dialog from an ActiveX control. All controls were invisible and only showed up sometimes on getting or losing focus. The cause is a custom class "_BitmapClass" that draws a logo and has the CS_PARENTDC class style (and no other class styles FWIW). Its WM_ERASEBKGND handler fills the complete dialog rect with that default Win95-Desktop-Background colour. According to the Win32 docs winelib behaves correctly wrt CS_PARENTDC and clipping. In this case it'd be Macromedia's fault as to how they implemented that window. However, I just found http://msdn.microsoft.com/library/techart/msdn_classy32.htm: <classy> If you use the CS_PARENTDC style for a class, each window in that class uses its parent window's DC. CS_PARENTDC is similar to CS_CLASSDC in that multiple windows share one DC. In CS_PARENTDC, however, the windows that share the DC need not belong to the same class. </classy> This contradicts the API docu at http://msdn.microsoft.com/library/psdk/winui/winclass_3b1v.htm: <api docs> CS_PARENTDC Sets the clipping rectangle of the child window to that of the parent window so that the child can draw on the parent. A window with the CS_PARENTDC style bit receives a regular device context from the system's cache of device contexts. It does not give the child the parent's device context or device context settings. Specifying CS_PARENTDC enhances an application's performance. <api docs> a little later, this is as near as the "classy" text comes to clipping: <classy> Another outcome of the CS_PARENTDC style is that child windows can draw anywhere in their parent's client area as well as drawing in their own client area. The CTL3D library, which creates 3-D effects around edit controls and list boxes, relies on this behavior. An application should not use CS_PARENTDC if it needs to adjust the mapping modes for different child windows. This will negate much of the benefit and could cause problems ensuring that each child window has the correct mapping mode set. </classy> Then I tried on Windows (2000), created a window with a child that has CS_PARENTDC set and * The paint rect of the PAINTSTRUCT as intialized by BeginPaint() equals the window's client rect - as expected. * but - surprise - so does the DC's clip box and trying to largen it to cover some or all areas of the parent fail and of course any drawing on the parent fails. * same for the DC returned by GetDC(hWnd) * only GetDCEx(hWnd, 0, DCX_PARENTCLIP) returns a DC that allows painting on the parent (no difference to a non-CS_PARENTDC window though) So, how is this painting on the parent with that style flag supposed to work? The things above make me think that it isn't possible at all or requires some additional steps. This would at least explain why the dialog described above looks busted in wine. If anyone could confirm my observations I'd come up with a patch. Thanks for any info, -Malte -- Wars don't determine who's right, only who's left.

1 0

Re: PATCH: wine-config
by Andreas Mohr June 17, 2001

June 17, 2001

On Sun, Jun 17, 2001 at 10:26:34AM +0200, Marcus Meissner wrote: > Hi, > > to help winelib using developers, this program exports flags and directories > to /usr/bin/. > > I am not sure whether exported cflags and libs are ok. > > Hmm, we probably should also tell how to build .so libs. And specs -> .cs. > (Hmm, supply Makefile snippets?) > > Ciao, Marcus > > Changelog: > Added a 'wine-config' for winelib developers. I don't think this is a good idea. That way it'll get confused with wineconf within fractions of a second... I'd much rather use a name like wlib-config or winelib-config instead. -- Andreas Mohr

1 0

wrc error
by Bang Jun-Young June 14, 2001

June 14, 2001

Hi, I got the following error when I tried to compile the latest CVS source: gmake[2]: Entering directory `/home/bjy/wine/dlls/user' LD_LIBRARY_PATH="../../unicode:$LD_LIBRARY_PATH" ../../tools/wrc/wrc -I. -I. -I../../include -I../../include -o resources/user32.res -r resources/user32.rc resources/user32_Ca.rc:2:1: Error: syntax error gmake[2]: *** [resources/user32.res] Error 1 After downgrading some files in wrc to earier revisions, Wine was successfully compiled again. I suspect that Francois Gouget's commit made on 2001-06-07 causes this. Jun-Young -- Bang Jun-Young <bjy(a)mogua.org>

2 1