Sat Aug 22 17:46:19 1998 Ulrich Weigand <weigand@informatik.uni-erlangen.de>
* [include/dosexe.h] [include/module.h] [include/pe_image.h]
[include/process.h] [include/windows.h] [loader/dos/module.c]
[loader/module.c] [loader/ne/module.c] [loader/pe_image.c]
[scheduler/process.c] [win32/process.c]
Partially implemented CreateProcess32.
* [win32/device.c] [relay32/kernel32.spec] [scheduler/k32obj.c]
[misc/registry.c] [win32/file.c]
Partially implemented VxDCall (VMM registry services).
* [files/dos_fs.c]
DOSFS_FindNext made thread-safe.
* [include/sig_context.h] [include/syslevel.h] [loader/signal.c]
[scheduler/syslevel.c] [tools/build.c]
Replaced CALLTO16_Current_fs by SYSLEVEL_Win16CurrentTeb.
* [win32/kernel32.c]
Bugfix: QT_Thunk/FT_Thunk should return 'long' in DX:AX, not EAX.
* [if1632/relay.c] [relay32/builtin32.c] [windows/msgbox.c]
[msdos/int20.c] [msdos/int21.c]
Use ExitProcess instead of TASK_KillCurrentTask.
* [include/task.h] [include/thread.h] [loader/task.c]
[scheduler/thread.c] [include/queue.h] [windows/message.c]
[windows/queue.c] [windows/win.c]
Prevent Win32 threads from entering the TASK_Reschedule loop.
(Note: Win32 messaging still doesn't work correctly; this patch
is just preventing the system from crashing when Win32 threads
call messaging functions. Messages will probably still get lost.)
* [scheduler/critsection.c]
Deactivated the use of SEM_UNDO for the SYSTEM_LOCK semaphore;
for some reason, this leads to problems after threads terminate...
Sat Aug 22 15:00:00 1998 Jürgen Schmied <juergen.schmied@metronet.de>
* [include/authors.h]
New file, includes all names of the developer (former shell.c)
* [Makefile.in][configure][configure.in][dlls/Makefile.in]
[dlls/shell32/Makefile.in][shres.rc]
Created dlls/shell32 and moved the shell32 stuff in it.
Started to create internal resources.
* [dlls/shell32/*]
Split the shell32 implementation into smaller files.
New classes: IContextMenu, IExtractIcon, IShellView.
Implemented Shell_GetImageList().
shell32 links to comctl32 now dynamically so it can use
internal/external implementations.
* [documentation/internal-dll] [documentation/shell32]
New, could anybody do a spellcheck?
* [include/commctrl.h]
Many new LV constants, structures, functions.
* [controls/comctl32undoc.c]
Rewrote the DSA* functions.
* [windows/winpos.c]
SetShellWindow32, GetShellWindow32.
Sat Aug 22 14:02:15 1998 Alexander Lukyanov <lav@long.yar.ru>
* [loader/resource.c]
Mark last accelerator as such in LoadAccelerators32W.
* [relay32/shell32.spec] [misc/shell.c]
Add stubs for SHGetSpecialFolderPath[AW].
Sat Aug 22 02:07:42 1998 Adrian Harvey <adrian@select.com.au>
* [include/file.h] [file/file.c] [msdos/int21.c] [msdos/vxd.c]
[misc/lzexpand.c] [win32/kernel32.c] [documentation/filehandles]
Fixed file handle handling. Created universal HFILE16 to HFILE32
translation macro from msdos/int21 code by Ove Kaaven.
Used macro in all Win16 functions so that win32 handles are translated
to avoid DOS/Win16 stdxx handles.
Removed handle translation from int21.c where Win16 functions are
called. Changed remaining calls to use new macro names.
Documented filehandle handling and differences between win 16 & 32.
Fri Aug 21 20:32:49 1998 Alexandre Julliard <julliard@lrc.epfl.ch>
* [server/process.c] [server/thread.c]
Implemented object wait queues and synchronization.
Fri Aug 21 18:40:02 1998 Huw D M Davies <daviesh@abacus.physics.ox.ac.uk>
* [graphics/psdrv/*]
DEVMODE dmPaper{Width|Length} fields are in 0.1mm.
Select a 100 pixel default font in CreateDC.
Thu Aug 20 22:47:39 1998 Uwe Bonnes <bon@elektron.ikp.physik.tu-darmstadt.de>
* [objects/bitmap.c]
Handle bits=32 in SetBitmapBits32 and GetBitmapBits32.
* [msdos/int21.c]
Add handling of Int21 0A and 37.
* [misc/commdlg.c]
Use MapHModuleLS and MapHModuleSL when translating HINSTANCE16 to
HINSTANCE32 and vice versa.
* [win32/file.c]
CreateFile32A: Abort if filename == NULL.
Thu Aug 20 12:28:31 1998 Marcus Meissner <marcus@jet.franken.de>
* [*/*]
Lots of missing prototypes added, some parameter types adapted to match
SDK.
* [debugger/stabs.c]
Don't loop forever if we don't find wine or one of the libxxx.so.
* [loader/ne/module.c]
Implemented MapHModuleLS,MapHModuleSL,MapHinstLS,MapHinstSL.
* [misc/network.c]
Implemented WNetGetUser32A.
* [misc/shellord.c]
Implemented ILRemoveLastID.
* [multimedia/dsound.c]
Fixed StarCraft memory leak.
* [graphics/ddraw.c]
Removed some unnecessary simple relaying functions, tried polishing
up the Xlib implementation (still doesn't work), temp. removed Xshm
code (to be remerged with working Xlib code).
Tue Aug 18 22:29:17 1998 Ove Kaaven <ovek@arcticnet.no>
* [multimedia/mmio.c] [multimedia/mmsystem.c]
Fixed most mmio bugs, fully implementing mmioSetBuffer
buffering, ability to read memory files, and the
sndPlaySound() SND_MEMORY flag. Most mmio-using programs
now work fine.
* [include/dosexe.h] [include/miscemu.h] [include/module.h]
[loader/module.c] [loader/task.c] [msdos/dosmem.c]
Improved DOS VM flexibility and portability somewhat. (Did
I get the #ifdefs right this time, BSD-ers?)
* [msdos/int21.c]
Made "Get Current PSP address" work as expected in a DOS VM.
* [loader/dos/*]
Began improving flexibility and portability somewhat. It
should be easier to add DPMI RMCB callbacks now. The
DOS VM no longer leaves big files lying around in /tmp
after a crash.
Tue Aug 18 12:38:31 1998 Turchanov Sergey <turchanov@usa.net>
* [relay32/winmm.spec]
This patch allows WinAmp to play WAV files (at least in PCM
format).
Sun Aug 16 05:34:13 1998 Pablo Saratxaga <srtxg@chanae.alphanet.ch>
* [windows/keyboard.c]
Corrected keyboard code to properly handle keys : ? ~ and "
on non US keyboards.
Sat Aug 15 18:47:14 1998 Brian Craft <bcboy@dorothy.wanglab.brandeis.edu>
* [windows/win.c]
Fixed severe bug in EnumChildWindwos().
Thu Aug 13 21:05:35 1998 Eric Kohl <ekohl@abo.rhein-zeitung.de>
* [controls/tooltips.c]
Fixed some bugs. Added subclassing support.
* [controls/toolbar.c]
Improved tooltip integration. Fixed some bugs.
* [controls/commctrl.c]
Changed control registration and added some documentation.
Fixed ShowHideMenuCtl.
* [controls/rebar.c][include/rebar.h][include/commctrl.h]
Improved rebar implementation (still no display).
* [controls/pager.c][include/pager.h][include/commctrl.h]
Improved pager implementation (still no display).
* [misc/imagelist.c]
Fixed a bug.
* [documentation/common_controls]
Updated.
Sun Aug 9 19:50:20 1998 James Juran <jrj120@psu.edu>
* [Makefile.in] [documentation/Makefile.in]
[programs/Makefile.in] [programs/*/Makefile.in]
Added uninstall rules, cleaned up install rules a little bit.
Sun Aug 9 13:21:35 1998 Andreas Mohr <100.30936@germany.net>
* [loader/ne/module.c] [if1632/kernel.spec]
Added the undocumented HIWORD of GetModuleHandle (hFirstModule).
* [loader/ne/segment.c]
Wine forgot to set some NE_SEGFLAGS_*.
Combined with another loader change, this fixed the
"BLINKER -- error in loading module" or ghost MessageBox problem
that about 1% of all Windows programs have.
Some BLINKER programs still don't work, though.
But I'm working on it, with great help from Blinkinc.
* [loader/task.c]
InitTask needs to decrement the SP register by two as Win95 does.
Sun Aug 9 02:41:28 1998 Ulrich Weigand <weigand@informatik.uni-erlangen.de>
* [if1632/kernel.spec] [relay32/kernel32.spec] [scheduler/syslevel.c]
[loader/main.c] [win32/ordinals.c] [include/syslevel.h]
[scheduler/Makefile.in]
Implemented Win95 'syslevel' routines (including Win16Lock).
* [if1632/relay.c] [if1632/thunk.c] [tools/build.c] [loader/task.c]
[loader/ne/segment.c] [win32/kernel32.c] [memory/selector.c]
[include/stackframe.h] [include/thread.h]
16-bit %fs handling revised. Use Win16Lock where appropriate.
* [include/thread.h] [scheduler/synchro.c] [windows/message.c]
[windows/queue.c] [win32/process.c]
Implemented MsgWaitForMultipleObjects.
* [files/change.c] [files/Makefile.in] [scheduler/k32obj.c]
[win32/newfns.c]
Implemented (dummy) file change notification objects.
* [debugger/dbg.y] [scheduler/process.c] [scheduler/thread.c]
[include/process.h] [include/thread.h]
Suspend all threads except current when hitting debugger break point.
* [objects/dib.c]
Bugfix for CreateDIBSection.
This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
This is the core of the Wine debugger. The reverse assember
was stolen from Mach more or less intact. It turns out that there are
two variables that are set differently if you are reverse assembling
16 bit code, and on the whole it seems to work.
NEWS:
The internal debugger has *tons* more capability than it did before.
I have enclosed some examples that show usage at the end of this file.
New features include:
1) Ability of debugger to read debug information from wine executable
*and* from Win32 executables. Local variable and line number information is
also read and processed.
2) The internal debugger is capable of 'stepping' to the next
line number, just like gdb. Examples of the commands are:
step
stepi
si
step 3
si 5
next
nexti
cont 4
finish
All of these should be exactly like how gdb does things.
3) The internal debugger now has a sense of what source file and line
number a given PC is at. New commands to support this are just like gdb,
and include:
list
dir
show dir
there are a variety of formats of arguments for the list command. All
permutations supported by gdb should also be supported.
4) The internal debugger knows about datatypes of various objects,
for both Win32 *and* the debugging information in the wine executable itself.
I have enclosed an example of how this works at the end.
5) There are more ways the 'b' command can be used to set breakpoints.
Examples are:
b *0x8190000
b 1100
b Usage
b
I don't think this covers all of the permutations that gdb accepts (this should
be cleaned up someday so that all possibilities are acceptable).
6) The 'print' and 'x' commands should behave more or less exactly
as they do under gdb. The difference is that the way the data is presented
will be slightly different, but the content should be fundamentally the same.
7) The internal debugger now supports conditional breakpoints, and
automatic display expressions. An example is at the end of this file. The
syntax and usage should be identical to that of gdb.
8) Type casts can be made from within the debugger, but they currently
don't work with typedef'ed types. They only work with builtin types and
named structures unions, etc. The problem is that internally we don't always
record the typedefed names of structures, so we have no guarantee that we
would know what each type is. This can be fixed, of course - it just takes
more memory. Note that in some cases, typedefed structures could be cast
using '(struct typedfname)' instead of '(typedfname)'. Technically this
isn't quite correct, but if and when the rest of this stuff gets fixed,
this would need to get corrected too.
NOTES:
If it weren't for the fact that gdb doesn't grok the Win32 debug
information, you could just use gdb. The internal debugger should be able
to read and use debugging information for both Win32 and also for the
Wine executable, making it possible to debug the combination of the two
together as if it were one large (very large) entity.
LIMITATIONS AND DIFFERENCES FROM GDB:
You cannot set a breakpoint by file and line number as you can
with gdb. Adding support for this wouldn't be all that tough, I guess, but
it would be a nuisance. You can set a breakpoint given a function and
line number, however. An example would be 'b main:2993'. It turns out
that the way the internal data structures are arranged it is a whole lot
easier to do things in this way than it would be to try and get the
source:line type of breakpoint working, but it would probably be worth it
to try.
Getting stack traces through Wine itself can be a bit tricky.
This is because by default the thing is built with optimization
enabled, and as a result sometimes functions don't get frames, and
lots of variables are optimized into registers. You can turn off
optimization for a few key source files if it will help you.
Memory consumption is getting to be a real problem. I think 32Mb is
no longer sufficient to debug wine - 48 or 64 is probably a whole lot better.
Unfortunately I cannot shut down X to save memory :-).
*************************************************************************
EXAMPLES:
Here is an example of how I tracked down a bug in Wine. The program
is something that just maps and dumps the contents of a Win32 executable.
It was dying for some reason.
Start the first time through.
bash$ ls -l dumpexe.exe
-rw-rw-r-- 1 eric devel 168448 Jan 4 13:51 dumpexe.exe
bash$ ./wine -debug './dumpexe.exe -symbol ./dumpexe.exe'
Warning: invalid dir 'e:\test' in path, deleting it.
Win32 task 'W32SXXXX': Breakpoint 1 at 0x081a3450
Loading symbols from ELF file ./wine...
Loading symbols from ELF file /usr/X11R6/lib/libXpm.so.4.6...
Loading symbols from ELF file /usr/X11R6/lib/libSM.so.6.0...
Loading symbols from ELF file /usr/X11R6/lib/libICE.so.6.0...
Loading symbols from ELF file /usr/X11R6/lib/libXext.so.6.0...
Loading symbols from ELF file /usr/X11R6/lib/libX11.so.6.0...
Loading symbols from ELF file /lib/libm.so.5.0.5...
Loading symbols from ELF file /lib/libc.so.5.2.18...
Loading symbols from ELF file /lib/ld-linux.so.1...
Loading symbols from Win32 file ./dumpexe.exe...
Stopped on breakpoint 1 at 0x081a3450 (_mainCRTStartup)
In 32 bit mode.
*** Invalid address 0x414c5ff8 (KERNEL32_NULL_THUNK_DATA+0x3930ee6c)
0x081a3450 (_mainCRTStartup): movl %fs:0,%eax
Wine-dbg>b DumpFile
Breakpoint 2 at 0x081a0078 (DumpFile+0x9 [dumpexe.c:2723])
Wine-dbg>c
Dump File: ./dumpexe.exe
Stopped on breakpoint 2 at 0x081a0078 (DumpFile+0x9 [dumpexe.c:2723])
Enter path to file dumpexe.c: ../de
2723 HANDLE hFile = NULL;
0x081a0078 (DumpFile+0x9 [dumpexe.c:2723]): movl $0x0,0xfffffff4(%ebp)
Wine-dbg>list
2723 HANDLE hFile = NULL;
2724 HANDLE hMap = NULL;
2725 PSTR lpMap = NULL;
2726 DWORD dwFileSize = 0;
2727 DWORD dwFileSizeHigh = 0;
2728
2729 PIMAGE_DOS_HEADER lpImageDOS = NULL;
2730 PIMAGE_FILE_HEADER lpImageFile = NULL;
2731 PIMAGE_NT_HEADERS lpImageNT = NULL;
2732
2733 /*
Wine-dbg>n 10
2747 dwFileSize = GetFileSize(hFile, &dwFileSizeHigh);
0x081a00ea (DumpFile+0x7b [dumpexe.c:2747]): leal 0xfffffff0(%ebp),%eax
Wine-dbg>n
2749 && (GetLastError() != NO_ERROR) )
0x081a00fb (DumpFile+0x8c [dumpexe.c:2749]): cmpl $-1,0xffffffe8(%ebp)
Wine-dbg>x/d dwFileSize
x/d dwFileSize
168448
Wine-dbg>n
2758 PAGE_READONLY, 0, 0, (LPSTR) NULL);
0x081a0124 (DumpFile+0xb5 [dumpexe.c:2758]): pushl $0x0
Wine-dbg>list 2750
list 2750
2750 {
2751 Fatal("Cannot get size of file %s", lpFileName);
2752 }
2753
2754 /*
2755 * map the file
2756 */
2757 hMap = CreateFileMapping(hFile, (LPSECURITY_ATTRIBUTES) NULL,
2758 PAGE_READONLY, 0, 0, (LPSTR) NULL);
2759 if( hMap == NULL )
2760 {
Wine-dbg>n
2759 if( hMap == NULL )
0x081a013b (DumpFile+0xcc [dumpexe.c:2759]): cmpl $0,0xfffffffc(%ebp)
Wine-dbg>x hMap
08e48c30
Wine-dbg>n
2767 lpMap = (LPSTR) MapViewOfFile(hMap, FILE_MAP_READ, 0, 0, 0);
0x081a0156 (DumpFile+0xe7 [dumpexe.c:2767]): pushl $0x0
Wine-dbg>n
2768 if( lpMap == NULL )
0x081a016b (DumpFile+0xfc [dumpexe.c:2768]): cmpl $0,0xffffffe0(%ebp)
Wine-dbg>print lpMap
0x414c5f40
Wine-dbg>x lpMap
40007000
Wine-dbg> x/10x 0x40007000
x/10x 0x40007000
0x40007000 (KERNEL32_NULL_THUNK_DATA+0x37e4fe74): *** Invalid address 0x40007000 (KERNEL32_NULL_THUNK_DATA+0x37e4fe74)
Wine-dbg>quit
$
*******************************************************************
The first time through, we find that MapViewOfFile isn't mapping the file
correctly into the virtual address space. Try running again, and step into
MapViewOfFile to figure out what went wrong.
*******************************************************************
bash$ ./wine -debug './dumpexe.exe -symbol ./dumpexe.exe'
Warning: invalid dir 'e:\test' in path, deleting it.
Win32 task 'W32SXXXX': Breakpoint 1 at 0x081a3450
Loading symbols from ELF file ./wine...
Loading symbols from ELF file /usr/X11R6/lib/libXpm.so.4.6...
Loading symbols from ELF file /usr/X11R6/lib/libSM.so.6.0...
Loading symbols from ELF file /usr/X11R6/lib/libICE.so.6.0...
Loading symbols from ELF file /usr/X11R6/lib/libXext.so.6.0...
Loading symbols from ELF file /usr/X11R6/lib/libX11.so.6.0...
Loading symbols from ELF file /lib/libm.so.5.0.5...
Loading symbols from ELF file /lib/libc.so.5.2.18...
Loading symbols from ELF file /lib/ld-linux.so.1...
Loading symbols from Win32 file ./dumpexe.exe...
Stopped on breakpoint 1 at 0x081a3450 (_mainCRTStartup)
In 32 bit mode.
*** Invalid address 0x414c5ff8 (KERNEL32_NULL_THUNK_DATA+0x3930ee6c)
0x081a3450 (_mainCRTStartup): movl %fs:0,%eax
Wine-dbg>b DumpFile:2767
Breakpoint 2 at 0x081a0156 (DumpFile+0xe7 [dumpexe.c:2767])
Wine-dbg>c
Dump File: ./dumpexe.exe
Stopped on breakpoint 2 at 0x081a0156 (DumpFile+0xe7 [dumpexe.c:2767])
Enter path to file dumpexe.c: ../de
2767 lpMap = (LPSTR) MapViewOfFile(hMap, FILE_MAP_READ, 0, 0, 0);
0x081a0156 (DumpFile+0xe7 [dumpexe.c:2767]): pushl $0x0
Wine-dbg>step
390 0385 stdcall MapViewOfFile(long long long long long) MapViewOfFile
0x080d793c (KERNEL32_385 [kernel32.spec:390]): pushl %ebp
Wine-dbg>step
223 if (!debugging_relay) return;
0x080c83dc (RELAY_DebugCallFrom32+0xc [relay.c:223]): cmpw $0,0x644a
Wine-dbg>
244 }
0x080c848e (RELAY_DebugCallFrom32+0xbe [relay.c:244]): leal 0xfffffff4(%ebp),%esp
Wine-dbg>
103 return MapViewOfFileEx(handle,access,offhi,offlo,size,0);
0x080911a4 (MapViewOfFile+0x14 [file.c:103]): pushl $0x0
Wine-dbg>
113 FILEMAP_OBJECT *fmap = (FILEMAP_OBJECT*)handle;
0x080911cf (MapViewOfFileEx+0xf [file.c:113]): movl 0x8(%ebp),%esi
Wine-dbg>n
115 if (!size) size = fmap->size;
0x080911d2 (MapViewOfFileEx+0x12 [file.c:115]): testl %ebx,%ebx
Wine-dbg>list
list
115 if (!size) size = fmap->size;
116 if (!size) size = 1;
117 return mmap ((caddr_t)st, size, fmap->prot,
118 MAP_ANON|MAP_PRIVATE,
119 FILE_GetUnixHandle(fmap->hfile),
120 offlo);
121 }
122
123 /***********************************************************************
124 * UnmapViewOfFile (KERNEL32.385)
125 */
Wine-dbg>x size
00000000
Wine-dbg>n
116 if (!size) size = 1;
0x080911d9 (MapViewOfFileEx+0x19 [file.c:116]): testl %ebx,%ebx
Wine-dbg>x size
00000000
Wine-dbg>n
117 return mmap ((caddr_t)st, size, fmap->prot,
0x080911e2 (MapViewOfFileEx+0x22 [file.c:117]): pushl %eax
Wine-dbg>x size
00000000
Wine-dbg>info local
MapViewOfFileEx:handle == 0x08e48c90
MapViewOfFileEx:access == 0x00000004
MapViewOfFileEx:offhi == 0x00000000
MapViewOfFileEx:offlo == 0x00000000
MapViewOfFileEx:size == 0x00000000
MapViewOfFileEx:st == 0x00000000
MapViewOfFileEx:offlo optimized into register $eax
MapViewOfFileEx:size optimized into register $ebx
MapViewOfFileEx:st optimized into register $edi
MapViewOfFileEx:fmap optimized into register $esi
Wine-dbg>print $ebx
0x0001
Wine-dbg>bt
bt
Backtrace:
=>0 0x080911e2 (MapViewOfFileEx+0x22 [file.c:117])
1 0x080911b0 (MapViewOfFile+0x20(handle=0x8e48c90, access=0x4, offhi=0x0, offlo=0x0, size=0x0) [file.c:104])
2 0x08104ab5 (CallFrom32_stdcall_5+0x25 [callfrom32.s])
3 0x081a0168 (DumpFile+0xf9(lpFileName=0x414c61ed) [dumpexe.c:2767])
4 0x081a0c35 (main+0x410(argc=0x3, argv=0x414c61cc) [dumpexe.c:3078])
5 0x081a3514 (_mainCRTStartup+0xc4)
6 0x0810549f (Code_Start+0x13 [callto32.s])
7 0x0802fdac (TASK_CallToStart+0x8c [task.c:373])
Wine-dbg>
*******************************************************************
Notice that you can step through the thunks into our own transfer
routines. You will notice that the source line displays as something
like:
390 0385 stdcall MapViewOfFile(long long long long long) MapViewOfFile
This is just the source line from the spec file that caused the transfer
routine to be generated. From this you can step again, and you step
into the relay logging code - keep stepping and you eventually step into
the actual function that does the dirty work.
At this point an examination of the source to the Win32 program
and an examination of the source to win32/file.s showed where the problem
was. When you specify 0 for the size of the object in CreateFileMapping,
it is supposed to use the entire size of the file as the size of the
object. Instead we were just blindly copying the number over.
*******************************************************************
Wine-dbg>b main
Breakpoint 1 at 0x080108c0 (main [dbgmain.c:213])
Wine-dbg>print breakpoints[1]
{addr={type=0x08043000, seg=0, off=134285504}, addrlen=' ', opcode='U', enabled=1, skipcount=0, in_use=1}
Wine-dbg> print breakpoints[1].enabled
1
Wine-dbg>set breakpoints[0].enabled = 0
Wine-dbg>print breakpoints[0].enabled
0
Wine-dbg>print type_hash_table[1]->type
STRUCT
Wine-dbg>print type_hash_table[1]
0x08072020
Wine-dbg>print *type_hash_table[1]
print *type_hash_table[1]
{type=STRUCT, next=0x00000000, name="LOGPALETTE", un={basic={basic_type=8, output_format="<22>V<08>M", basic_size=-128, b_signed=0}, bitfield={bitoff=8, nbits=0, basetype=0x081d56c0}, pointer={pointsto=0x00000008}, funct={rettype=0x00000008}, array={start=8, end=136140480, basictype=0x08043e80}, structure={size=8, members=0x081d56c0}, enumeration={members=0x00000008}}}
Wine-dbg>
*******************************************************************
This example shows how you can print out various data structures.
Note that enumerated types are displayed in the symbolic form, and strings
are displayed in the expected manner.
You can use the set command to set more or less anything. Note
however that you cannot use enumerated types on the RHS of the expression.
*******************************************************************
Wine-dbg>list
2986 if( argc <= 1 )
2987 {
2988 Usage(argv[0]);
2989 }
2990
2991 for( i = 1; i < argc; i++ )
2992 {
2993 if( strncmp(argv[i], "-dos", sizeof("-dos") - 1) == 0 )
2994 {
2995 DmpCtrl.bDumpDOSHeader = TRUE;
2996 }
Wine-dbg>b 2993
Breakpoint 3 at 0x081a8861 (main+0x3c [dumpexe.c:2993])
Wine-dbg>condition 3 i == 2
Wine-dbg>c
Stopped on breakpoint 3 at 0x081a8861 (main+0x3c [dumpexe.c:2993])
2993 if( strncmp(argv[i], "-dos", sizeof("-dos") - 1) == 0 )
0x081a8861 (main+0x3c [dumpexe.c:2993]): pushl $0x4
Wine-dbg>print i
2
Wine-dbg>print argv[i]
"./dumpexe.exe"
*******************************************************************
This example shows how to use conditional breakpoints.
Here is another one that demonstrates another cool feature
conditional breakpoints that involve a function call:
condition 3 strcmp(argv[i], "./dumpexe.exe") == 0
*******************************************************************
Wine-dbg>list
2986 if( argc <= 1 )
2987 {
2988 Usage(argv[0]);
2989 }
2990
2991 for( i = 1; i < argc; i++ )
2992 {
2993 if( strncmp(argv[i], "-dos", sizeof("-dos") - 1) == 0 )
2994 {
2995 DmpCtrl.bDumpDOSHeader = TRUE;
2996 }
Wine-dbg>b 2993
Breakpoint 3 at 0x081a8861 (main+0x3c [dumpexe.c:2993])
Wine-dbg>condition 3 strcmp(argv[i], "./dumpexe.exe") == 0
Wine-dbg>info break
Breakpoints:
1: y 0x081ab450 (_mainCRTStartup)
2: y 0x081a882e (main+0x9 [dumpexe.c:2986])
3: y 0x081a8861 (main+0x3c [dumpexe.c:2993])
stop when ( strcmp(( argv[i] ), "./dumpexe.exe") == 0 )
Wine-dbg>c
Stopped on breakpoint 3 at 0x081a8861 (main+0x3c [dumpexe.c:2993])
2993 if( strncmp(argv[i], "-dos", sizeof("-dos") - 1) == 0 )
0x081a8861 (main+0x3c [dumpexe.c:2993]): pushl $0x4
Wine-dbg>print i
2
Wine-dbg>print argv[i]
"./dumpexe.exe"
Wine-dbg>