wine-devel

wine-devel@winehq.org

84463 discussions

Re: Strange crash
by gerard patel 20 Apr '01

20 Apr '01

At 10:30 AM 20/04/2001 +0200, you wrote: <snip> >The lower 6 bits are the new floating point exception masks. Mine are >all set and yours are cleared, therefore the crash. > >Why that is the case, I haven't got a clue. I'm not sure I followed everything, but it could come from a problem that occurred a lot of time before, no? The bits could be already cleared on entry of the subroutine. If the exception handling bits are cleared and the app don't do anything with floating point, nothing will break, AFAICT. If this is the case a way could be to patch the code to print the mask with every +relay operation. BTW, I wonder if Dmitry has done some classic tricks like - getting a fresh tree from Winehq - compiling everything with -O0 - trying the same app on another computer with a (as much as possible) different configuration. Gerard

2 1

problem with native dll's.
by Rein Klazes 20 Apr '01

20 Apr '01

Hi, Since the cvs commits early this week or the last (I could be more precise but that will take a while) I have noticed a problem with some native dlls. If I start one app that uses such a dll and then another using the same dll may crash. If this second app is started as the first process it runs fine. As an example eudora, using native commctrl/comctl32 crashes when it loads these dlls: | Unhandled exception: priviledged instruction | in 32-bit code (0x406e12b8). | In 32-bit mode. | 0x406e12b8 (GetTextColor+0x34 [dcvalues.c:177]): hlt | 177 DC_GET_VAL( COLORREF, GetTextColor, textColor ) | Wine-dbg>bt | Backtrace: | =>0 0x406e12b8 (GetTextColor+0x34(hdc=0x0) [dcvalues.c:177]) (ebp=405b6dc4) | 1 0xbfebbd8d (COMCTL32.DLL.341+0x74) (ebp=405b6dd8) | 2 0xbfebbceb (COMCTL32.DLL.EntryPoint+0x36) (ebp=405b6e04) | 3 0x4008abe3 (MODULE_InitDLL+0x97(wm=0x40383d84, type=0x1, lpReserved=0x1) [module.c:115]) (ebp=405b6e24) | 4 0x4008acf1 (MODULE_DllProcessAttach+0xd1(wm=0x40383d84, lpReserved=0x1) [module.c:182]) (ebp=405b6e48) | 5 0x4008accd (MODULE_DllProcessAttach+0xad(wm=0x403838f4, lpReserved=0x1) [module.c:177]) (ebp=405b6e68) | 6 0x4008accd (MODULE_DllProcessAttach+0xad(wm=0x4038302c, lpReserved=0x1) [module.c:177]) (ebp=405b6e88) | 7 0x4008accd (MODULE_DllProcessAttach+0xad(wm=0x403826f8, lpReserved=0x1) [module.c:177]) (ebp=405b6ea8) | 8 0x4008accd (MODULE_DllProcessAttach+0xad(wm=0x0, lpReserved=0x1) [module.c:177]) (ebp=405b6ec8) | 9 0x400d288f (start_process+0x133 [process.c:373]) (ebp=405b6f30) | 10 0x400d5f99 (SYSDEPS_DoCallOnStack+0x25(func=0x400d275c, arg=0x0) [sysdeps.c:173]) (ebp=405b6ff0) | 11 0x400d603f (SYSDEPS_CallOnStack+0x2f) (ebp=bffff8d4) | 12 0x400d60e2 (SYSDEPS_SwitchToThreadStack+0x8e(func=0x400d275c) [sysdeps.c:236]) (ebp=bffff908) | 13 0x400d2c5d (PROCESS_InitWine+0x1a1(argc=0x5, argv=0xbffff9a4, win16_exe_name=0x804d5c0, win16_exe_file=0x804d6c4) [process.c:524]) (ebp=bffff924) | 14 0x0804c164 (Letext [main.c]) (ebp=bffff93c) | 15 0x4028116b (NTDLL.DLL.sqrt+0x3bc6b) (ebp=bffff978) | 16 0x08049021 (_start+0x21) (ebp=00000000) The ip register is pointing to some instruction byte, making it an bogus opcode, "hlt" in this case: | Wine-dbg>disass $eip-12,$eip+10 | 0x406e129c (GetTextColor+0x18 [dcvalues.c:177]): call 0x406bf664 (_init+0x67c) | 0x406e12a1 (GetTextColor+0x1d [dcvalues.c:177]): addl $4,%esp | 0x406e12a4 (GetTextColor+0x20 [dcvalues.c:177]): testl %eax,%eax | 0x406e12a6 (GetTextColor+0x22 [dcvalues.c:177]): jz 0x406e12b4 (GetTextColor+0x30 [dcvalues.c:177]) | 0x406e12a8 (GetTextColor+0x24 [dcvalues.c:177]): movl 0x88(%eax),%edi | 0x406e12ae (GetTextColor+0x2a [dcvalues.c:177]): pushl %esi | 0x406e12af (GetTextColor+0x2b [dcvalues.c:177]): call 0x406bfee4 (_init+0xefc) | 0x406e12b4 (GetTextColor+0x30 [dcvalues.c:177]): movl %edi,%eax | 0x406e12b6 (GetTextColor+0x32 [dcvalues.c:177]): leal 0xfffffff4(%ebp),%esp | 0x406e12b9 (GetTextColor+0x35 [dcvalues.c:177]): popl %ebx | 0x406e12ba (GetTextColor+0x36 [dcvalues.c:177]): popl %esi | 0x406e12bb (GetTextColor+0x37 [dcvalues.c:177]): popl %edi | 0x406e12bc (GetTextColor+0x38 [dcvalues.c:177]): leave | 0x406e12bd (GetTextColor+0x39 [dcvalues.c:177]): ret $0x4 | 0x406e12c0 (GetViewportExt16 [dcvalues.c:182]): pushl %ebp Redoing the experiment (by starting and stopping all apps's) will give crashes at slightly different places. Has any one a suggestion? Rein. -- Rein Klazes rklazes(a)xs4all.nl

1 0

Re: Strange crash
by Alexandre Julliard 20 Apr '01

20 Apr '01

"Dmitry Timoshkov" <dmitry(a)sloboda.ru> writes: > Unhandled exception: c000008f > [...] > > Commenting lines 36-37 > physDev->pen.width = GDI_ROUND((FLOAT)pen->logpen.lopnWidth.x * > dc->xformWorld2Vport.eM11 * 0.5); > in X11DRV_PEN_SelectObject (graphics/x11drv/pen.c) cures the problem. > > Why call to floor() leads to crash? Exception c000008f is EXCEPTION_FLT_INEXACT_RESULT. Probably one of the floating point values is invalid, or something similar. -- Alexandre Julliard julliard(a)winehq.com

4 10

Regedit from Win2k
by Dominik Strasser 19 Apr '01

19 Apr '01

Hi, trying to run regedit from Win2k I encounter the following problems: err:win32:fixup_imports No implementation for shlwapi.dll.0(StrRetToBufA) imported from shell32.dll, setting to 0xdeadbeef err:win32:fixup_imports No implementation for shlwapi.dll.0(StrRetToBufW) imported from shell32.dll, setting to 0xdeadbeef Hmm, but those are implemented. When I try to add a loadorder for shlwapi to get the builtin one, this doesn't work. "shlwapi" = "builtin, native" Still the native version is used. If I change the overall DefaultLoadOrder, to "builtin, native" it works. After this, the tree items in the Treeview vanished but I don't care. I assume this is the cause: fixme:listview:LISTVIEW_Create ANSI notify format is NOT used Trying to save the registry I get a crash wich I tracked down to this line in filedlg95.c: 453: AllocInArgWtoA(ofn->lpTemplateName, lpTemplateName); But ofn->lpTemplateName is not a pointer but seems to be a handle: 0x00000064. If I wrap the call into a if(!IsBadReadPtr(ofn->lpTemplateName, 4)), the filedialog comes up. Any clues anybody ? Regards Dominik -- Dominik Strasser | Phone: +49 89 636-43691 SIEMENS AG | Fax: +49 89 636-42284 CT SE 4 | E-Mail:Dominik.Strasser@mchp.siemens.de Otto-Hahn-Ring 6 | D-81739 Muenchen | Room: 53-263

1 0

Re: Store each glyph name only once in PS driver
by Alexandre Julliard 18 Apr '01

18 Apr '01

Ian Pilcher <ian.pilcher(a)home.com> writes: > Separating the string data from their indices will allow them to be > declared as const, so that should get them into the ".rodata" section of > the resulting assembly; the Adobe Glyph List vector is already there. > Is this what you mean by "read only memory"? Yes, exactly that; separating the strings and the indices seems the right thing to do. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: Store each glyph name only once in PS driver
by Alexandre Julliard 18 Apr '01

18 Apr '01

Ian Pilcher <ian.pilcher(a)home.com> writes: > typedef struct { > + INT index; > + CHAR sz[0]; > +} GLYPHNAME; > + > +typedef struct { > + LONG UV; > + GLYPHNAME *name; > +} UNICODEGLYPH; > + > +typedef struct { > + INT size; > + UNICODEGLYPH glyphs[0]; > +} UNICODEVECTOR; Zero-size arrays are not portable, and the standard method of using a size of 1 doesn't work well with initialized data (lots of warnings). So you probably need to change your data structures a bit; also it would be nice to store all the preinitialized stuff, or at least the strings, in read-only memory to save a bit of swap space. -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: SEH logging
by Alexandre Julliard 17 Apr '01

17 Apr '01

Francois Gouget <fgouget(a)free.fr> writes: > In this implementation exception.c reuses format_exception_msg() > which is normally used to display the dialog box asking whether to debug > the exception or not. Since this is all in ntdll it works. Actually win32/except.c is part of kernel, it's only linked in ntdll now because separation is not finished. So you cannot call format_exception_msg from ntdll. And anyway this would print "unhandled xxx" for all exceptions, which is misleading since they might be handled later on. Also I don't think the core exception routines should know about specific exception codes like EXCEPTION_WINE_STUB. If you want a message on stubs, it should be done in the .spec.c file before the call to RtlRaiseException. -- Alexandre Julliard julliard(a)winehq.com

2 1

IStorage file format.
by Psyon 17 Apr '01

17 Apr '01

I am interested in making an HTMLHelp (.CHM) viewer for Linux. I am told the files are actually IStorage files. I have been scouring the internet looking for information about this format with no avail. I have also been looking over the WINE implementation, to see if that would shed some light on the format, but the code dives into a bunch of cryptic functions that I cant seem to follow. Is there any documents describing the format of these files? even just a brief description would be nice at this moment. If you have any information please replay directly to me at x-odus(a)iname.com Thanks in advance --Don Becker

4 3

RH 7.1 gcc fixes compiler bug
by James Juran 17 Apr '01

17 Apr '01

As discovered about two weeks ago by Francois Jacques and Ulrich Weigand, Red Hat's gcc-2.96-69 package has a bug which shows up in the HOOK_CallHook function in windows/hook.c (and possibly other places). Red Hat 7.1 contains gcc-2.96-81, which fixes this bug. Users of Red Hat 7.0 should install the gcc and cpp packages from Red Hat 7.1 to fix this problem. -- James Juran jamesjuran(a)alumni.psu.edu

1 0

Re: strange msvcrt crash
by gerard patel 17 Apr '01

17 Apr '01

At 02:47 PM 16/04/2001 -0700, you wrote: > > Also I had lost of trouble running mspaint when using the builtin >msvcrt. None of the following works: >$ wine /mnt/win98/Program\ Files/Accessories/mspaint.exe >$ wine "/mnt/win98/Program Files/Accessories/mspaint.exe" >$ wine "c:/Program Files/Accessories/mspaint.exe" >$ wine "c:\\Program Files\\Accessories\\mspaint.exe" > > They all tell me that 'H:\wine\Files\Accessories\mspaint.exe was not >found.'. If I use native msvcrt they all work (and wine is the real >thing, not a script). Msvcrt includes command line parsing to give apps argc and argv - the current builtin handling in dlls/msvcrt/data.c is probably the cause of your problem : while (1) { if ((cmdline[end]==' ') || (cmdline[end]=='\0')) { if (cmdline[end]=='\0') last_arg=1; else cmdline[end]='\0'; Gerard

1 0

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