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

RtlGetNtVersionNumbers
by Joshua Thielen 24 Jun '02

24 Jun '02

Hello everybody, I was searching Google for info about RtlGetNtVersionNumbers and I came across this message, http://www.winehq.com/hypermail/wine-users/2001/12/0488.html, posted on wine-users. Unfortunately, the attachment is not there. Does anyone know what happened to this patch? It doesn't appear to have been sent to wine-patches nor did it make it into the CVS. Perhaps the patch was incorrect as he said his implementation might be off. If this is the case, does anyone have more info about RtlGetNtVersionNumbers? Thanks, Josh

2 1

WINE Performance tweaking and known bottlenecks
by Adam Luchjenbroers 24 Jun '02

24 Jun '02

What options/tricks/techniques can I use to enhance performance under WINE. I am trying to get Total Annihilation to run smoothly on a K6-2 233MHz machine, it runs and is getting close to playable. The only unusual thing is that the framerate is the same at 1024x768 as in 640x480. What is likely to be the bottleneck and what can I do about it? I'm willing to go into the code if neccessary. I have limited programming ability but I am an exceptionally fast learner.

1 0

Endian-ness?
by Ann and Jason Edmeades 23 Jun '02

23 Jun '02

Hi, I appear to have hit a little endian-ness problem. Can anyone point out what I am doing wrong or how to fix it please? olefont.c #includes wtypes.h This defines a structure as #ifdef BIG_ENDIAN. We hit this case and the parameters are defined in the wrong order as they are passed in from a windows dll. We hit is because endian.h #defines BIG_ENDIAN if __USE_BSD is defined, and features.h #defines _BSD_SOURCE and subsequently __USE_BSD. This is on Mandrake 8.2 if it makes a difference. Regards, Jason

1 0

(PATCH) SPI_GETDESKWALLPAPER Implementation
by Philipp Wollermann 23 Jun '02

23 Jun '02

Hello, this is my first real patch for Wine. It implements the SystemParametersInfo SPI_GETDESKWALLPAPER action and fixes SPI_SETDESKWALLPAPER so it updates the Wallpaper entry in the registry. (Wine Bug-Tasklet #461) I've also included a simple test for these two actions. Please tell me if something is wrong so I can fix it. :-) ChangeLog: * windows/sysparams.c: Implemented SPI_GETDESKWALLPAPER action, fixed SPI_SETDESKWALLPAPER * dlls/user/tests/sysparams.c Added test for wallpaper actions License: X11 / LGPL Philipp Wollermann

1 0

Infinite loop in winedebug
by Duane Clark 23 Jun '02

23 Jun '02

During an installation, wine crashed and went into the debugger, which proceeded to go into an infinite loop. This is with current CVS. Oddly, it only goes into the infinite loop if I have: "UseXTerm"=dword:00000000 Allowing winedebug to use its own xterm, the debugger works fine. Here is one iteration of the loop. 0x00485cf0 (_INS0432._MP..text+0x84cf0 in C:\temp\_INS0432._MP): divl 0x18(%ebp),%eax Wine-dbg>parse error First chance exception: divide by zero in 32-bit code (0x00485cf0). Register dump: CS:0023 SS:002b DS:002b ES:002b FS:008f GS:0000 EIP:00485cf0 ESP:40476a00 EBP:40476a58 EFLAGS:00010246( R- 00 I Z- -P1 ) EAX:00000000 EBX:4055bc8c ECX:00000000 EDX:00000000 ESI:00000000 EDI:00020029 Stack dump: 0x40476a00 (NTDLL.DLL.sqrt+0x2790dc): 00020029 00000000 4055bc8c 40110e80 0x40476a10 (NTDLL.DLL.sqrt+0x2790ec): 405fcea4 405fcd20 0000135a 40476a34 0x40476a20 (NTDLL.DLL.sqrt+0x2790fc): 405d4c86 405fc324 405fcea4 402239c6 0x40476a30 (NTDLL.DLL.sqrt+0x27910c): 0000135a 40476a4c 405c86ff 00000000 0x40476a40 (NTDLL.DLL.sqrt+0x27911c): 00000000 00000000 402239c6 40476a68 0x40476a50 (NTDLL.DLL.sqrt+0x27912c): 405d4f8f 0000135a 40476ad8 00484a8e 0x40476a60 (NTDLL.DLL.sqrt+0x27913c): 0011: sel=008f base=40110e80 limit=00000fff 32-bit rw- Backtrace: =>0 0x00485cf0 (_INS0432._MP..text+0x84cf0 in C:\temp\_INS0432._MP) (ebp=40476a58) 1 0x00484a8e (_INS0432._MP..text+0x83a8e in C:\temp\_INS0432._MP) (ebp=40476ad8) 2 0x00482d25 (_INS0432._MP..text+0x81d25 in C:\temp\_INS0432._MP) (ebp=40476b5c) 3 0x0046c40e (_INS0432._MP..text+0x6b40e in C:\temp\_INS0432._MP) (ebp=40476bb0) 4 0x0046c358 (_INS0432._MP..text+0x6b358 in C:\temp\_INS0432._MP) (ebp=40476bfc) 5 0x00406d06 (_INS0432._MP..text+0x5d06 in C:\temp\_INS0432._MP) (ebp=40476c24) 6 0x004058fa (_INS0432._MP..text+0x48fa in C:\temp\_INS0432._MP) (ebp=40476c8c) 7 0x00401c0e (_INS0432._MP..text+0xc0e in C:\temp\_INS0432._MP) (ebp=40476cb4) 8 0x4050ccb7 (WINPROC_wrapper+0x17 in user32.dll.so) (ebp=40476cd8) 9 0x4050cd42 (WINPROC_CallWndProc+0x82(proc=0x401914, hwnd=0x20029, msg=0xf, wParam=0x0, lParam=0x0) [winproc.c:183] in user32.dll.so) (ebp=40476d08) 10 0x4051308f (CallWindowProcA+0xab(func=0x40bf0640, hwnd=0x20029, msg=0xf, wParam=0x0, lParam=0x0) [winproc.c:2795] in user32.dll.so) (ebp=40476d3c) 11 0x404f55d6 (DispatchMessageA+0x11e(msg=0x40476d9c) [message.c:1063] in user32.dll.so) (ebp=40476d80) 12 0x004011af (_INS0432._MP..text+0x1af in C:\temp\_INS0432._MP) (ebp=40476dc0) 13 0x00401824 (_INS0432._MP..text+0x824 in C:\temp\_INS0432._MP) (ebp=40476df8) 14 0x004873db (_INS0432._MP.EntryPoint+0x14b in C:\temp\_INS0432._MP) (ebp=40476e8c) 15 0x400bc1ec (start_process+0x264 [process.c:530] in libntdll.dll.so) (ebp=40476f38) 16 0x400c009b (call_on_thread_stack+0x23(func=0x400bbf88) [sysdeps.c:105] in libntdll.dll.so) (ebp=40476ff4) 17 0x400c0234 (SYSDEPS_CallOnStack+0x14 in libntdll.dll.so) (ebp=00000000) 0x00485cf0 (_INS0432._MP..text+0x84cf0 in C:\temp\_INS0432._MP): divl 0x18(%ebp),%eax Wine-dbg>parse error

2 1

Re: Updated patch for PROFILE_GetString
by Dmitry Timoshkov 23 Jun '02

23 Jun '02

"Joshua Thielen" <thielen(a)netprince.net> wrote: > This update to my earlier patch handles the case when entry is NULL and > no keys are found in the section. Joshua, very simple question: how about Profile APIs regression test? -- Dmitry.

2 2

[Request] Another view of http://cvs.winehq.com/patch.py?id=XXX
by Uwe Bonnes 22 Jun '02

22 Jun '02

Hallo, to my knowledge, viewing the patch contend can at the moment only be done by webpages with an address like http://cvs.winehq.com/patch.py?id=XXX, where XXX can only be known from the postings from wine-cvs(a)winehq.com. I would propose that we create another page, where all the patches are listed with their subject ( like "wine/server thread.c" for "http://cvs.winehq.com/patch.py?id=1024708836163791355983452"), the subject being a link to the real patch and the newest patches added to the top of the patch. Thanks -- Uwe Bonnes bon(a)elektron.ikp.physik.tu-darmstadt.de Free Software: If you contribute nothing, expect nothing --

2 2

Re: PATCH: documentation/samples/config
by Andreas Mohr 22 Jun '02

22 Jun '02

On Fri, Jun 21, 2002 at 12:14:37PM +0200, Johan Gill wrote: > Changelog: 48 was plain wrong as value for HELmargin. Removed the sample > EmulDriver appdefault since N is the default now, and an appdefault for > Y is not useful. Huh ?? I'm sure there are plenty of entries in there that are default, but a sample config file is supposed to be a *sample* file... How do you expect people to know about EmulDriver if you get plain rid of it ? EmulDriver is very important, BTW. Several games need a different setting for it. -- "My attitude is, everybody should try competing with Microsoft once in their life. Once." - Marc Andreessen, former Netscape lead employee, in a "browser wars" interview

2 1

Re: Patch 2/2 for SmartForms Installer: GetModuleFileNameW called by 16-bit task.
by Dmitry Timoshkov 22 Jun '02

22 Jun '02

"Joshua Thielen" <thielen(a)netprince.net> wrote: > - if ((wm = MODULE32_LookupHMODULE( hModule ))) > + if(!hModule && !(NtCurrentTeb()->tibflags &= TEBF_WIN32)) { ^^ Apparently '&=' above must be replaced by simple '&' -- Dmitry.

2 1

Re: Patch 1/2 for SmartForms Installer: PROFILE_GetString incorrect return value
by Andreas Mohr 22 Jun '02

22 Jun '02

On Sat, Jun 22, 2002 at 12:18:07AM -0400, Joshua Thielen wrote: > This patch fixes an error encountered in the Smart Forms installer > (somewhat related to bug 575). > > When calling GetPrivateProfileString with the second parameter as NULL, > all of the keys under the specified section are copied to the buffer. > The return value should be the number of characters copied minus the > last null terminator. Since the keys are copied as null-terminated > strings with the last string terminated with two null characters, strlen > cannot be used to determine the number of characters copied. This patch > corrects PROFILE_GetString for this case. Hmm, hmm. Could this happen to fix bug #724 also ? Word97 etc. broke when I handled the "default" case in PROFILE_GetString(). As the log files received didn't show relay lines, only profile lines, I haven't been able to fix it. But this sounds like a reasonable cause for it... -- Andreas Mohr Stauferstr. 6, D-71272 Renningen, Germany Tel. +49 7159 800604 http://mohr.de.tt

2 1