September 2003 - Wine-devel

20 Sep '03

Hello, There are a couple slots in the FAQ that need a little TLC I would like to start on Memory useage and Compile times. I would like to get some compile times to add to *our* faq. 1) 3Ghz box with a 1GB+ of RAM (range) 2) 2Ghz box 512 to 1GB of RAM (range) 3) 1Ghz box with 128 to 512MB RAM (range) So, I would like to ask everyone one who can.. to post there "compile" times so I can add them to our most beautiful FAQ. Tom

1 0

implement interlocked* functions on Alpha AXP
by Steven Edwards 20 Sep '03

20 Sep '03

Hello, I know nothing about this except what I have learned in the past 4 hours so......... After looking at the interlocked code and reading how PEACE and Mono implement these functions I think this will work. This is assuming that the SUN code for implementing this is compatible with pthreads. I only get one warrning about a cast in interlocked_xchg_ptr so I assume I am on the right track with this. I have not done any work at this low level before so is there was something wrong with using pthreads here? I am not worried about breaking compatiblity with exsting Win32/64 AXP apps as there are few left. Changlog: Implement interlocked* function for the Alpha CPU Index: wine/libs/port/interlocked.c =================================================================== RCS file: /home/wine/wine/libs/port/interlocked.c,v retrieving revision 1.3 diff -u -r1.3 interlocked.c --- wine/libs/port/interlocked.c 2 Jul 2003 04:29:33 -0000 1.3 +++ wine/libs/port/interlocked.c 20 Sep 2003 02:07:04 -0000 @@ -244,6 +244,65 @@ _lwp_mutex_unlock( &interlocked_mutex ); return retv; } + +#elif defined(__ALPHA__) || defined(__AXP__) +/* This code was adapted from the Sparc port. After checking + * the way Mono and PEACE implement interlocked functions on + * pthreads I think it will work. + */ +#include <pthread.h> + +static pthread_mutex_t interlocked_mutex = PTHREAD_MUTEX_INITIALIZER; + +long interlocked_cmpxchg( long *dest, long xchg, long compare ) +{ + pthread_mutex_lock( &interlocked_mutex ); + if (*dest == compare) *dest = xchg; + else compare = *dest; + pthread_mutex_unlock( &interlocked_mutex ); + return compare; +} + +void *interlocked_cmpxchg_ptr( void **dest, void *xchg, void *compare ) +{ + pthread_mutex_lock( &interlocked_mutex ); + if (*dest == compare) *dest = xchg; + else compare = *dest; + pthread_mutex_unlock( &interlocked_mutex ); + return compare; +} + +long interlocked_xchg( long *dest, long val ) +{ + long retv; + pthread_mutex_lock( &interlocked_mutex ); + retv = *dest; + *dest = val; + pthread_mutex_unlock( &interlocked_mutex ); + return retv; +} + +void *interlocked_xchg_ptr( void **dest, void *val ) +{ + long retv; + pthread_mutex_lock( &interlocked_mutex ); + retv = *dest; + *dest = val; + pthread_mutex_unlock( &interlocked_mutex ); + return retv; +} + +long interlocked_xchg_add( long *dest, long incr ) +{ + long retv; + pthread_mutex_lock( &interlocked_mutex ); + retv = *dest; + *dest += incr; + pthread_mutex_unlock( &interlocked_mutex ); + return retv; +} + #else # error You must implement the interlocked* functions for your CPU #endif + __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com

1 0

Patch to fix build without optimization
by Pavel Roskin 19 Sep '03

19 Sep '03

Hello! Before I send it to wine-patches I'd like developers to look at it because it's an important issue and I'm not 100% sure that it's the best fix. I'm trying to compile CVS with without optimization and it fails: make[1]: Entering directory `/usr/local/src/wine/miscemu' gcc -o wine -Wl,--section-start,.interp=0x3c000100 main.o -L../dlls -lntdll.dll -L../libs/wine -lwine -L../libs/unicode -lwine_unicode -L../libs/port -lwine_port ../dlls/libntdll.dll.so: undefined reference to `InterlockedCompareExchange' InterlockedCompareExchange() is defined as extern inline in include/winbase.h for i386 and PowerPC. It's defined extern for other architectures. It's implemented in dlls/kernel/sync.c for all architectures, either in assembly code or as a call to interlocked_cmpxchg(). The implementation is linked into kernel32.dll. interlocked_cmpxchg() is declared in include/wine/port.h and implemented in libs/port/interlocked.c. It's exported by libwine_port.a. The unresolved call to InterlockedCompareExchange comes from wine/win32/device.c, which is linked into ntdll.dll. Actually, device.c calls InterlockedCompareExchangePointer(), which is a static inline function that calls InterlockedCompareExchange(). If optimization if off, InterlockedCompareExchangePointer() is used as a static function in wine/win32/device.c, and it calls external function InterlockedCompareExchange(). When the wine executable is linked, it's linked against ntdll.dll, but not against kernel32.dll. Therefore, the reference to InterlockedCompareExchange() in ntdll.dll is not satisfied. I see two solutions. One is to link wine executable against kernel32.dll. Another is to use the low-level function interlocked_cmpxchg_ptr(), which is what other code linked with "wine" does. I prefer the second solution for its simplicity and consistency. But I don't know, maybe dependency of ntdll.dll and wine executable on kernel32.dll is OK. Then we could eliminate low-level interlocked_cmpxchg() and use Win32 API functions for interlocking. Anyway, here's the patch. It works for me (Linux 2.4.23-pre4, gcc 3.2.2 from Red Hat 9, AMD Athlon). Tested with winemine, notepad and Windows Commander under Wine compiled with CFLAGS=-g ================================= --- win32/device.c +++ win32/device.c @@ -611,7 +611,7 @@ static HKEY create_special_root_hkey( HK if (NtCreateKey( &hkey, access, &attr, 0, NULL, 0, NULL )) return 0; } - if (!(ret = InterlockedCompareExchangePointer( (PVOID) &special_root_keys[idx], hkey, 0 ))) + if (!(ret = interlocked_cmpxchg_ptr( (PVOID) &special_root_keys[idx], hkey, 0 ))) ret = hkey; else NtClose( hkey ); /* somebody beat us to it */ ================================= -- Regards, Pavel Roskin

2 2

winedbg and loading symbols
by Michael Ost 19 Sep '03

19 Sep '03

Howdy, list. I can't figure out how to get winedbg (6/18/03 build) to load my application's symbol table. I am writing an ELF application which uses the wine libraries, and trying to debug it with winedbg. I can run that program fine, but when it loads the shared library built by winebuild (rm-host.exe.so) it says that there is no debug information in the ELF shared library. The file is built with -g, and 'nm' shows symbols galore. I tried using the 'symbolfile' command. The docs suggest that it would help, but I can't figure out what sort of 'path' argument it wants. Here's what I typed: Wine-dbg>symbolfile /home/most/src/rm-host/x86/rm-host.exe.so Parse error Wine-dbg>symbolfile F:\src\rm-host\x86\rm-host.exe.so Parse error Any suggestions? By the way has anyone got a GUI front end to winedbg in the works (ala kdbg for gdb)? I'd love to ge my hands on it. - mo PS: Sorry if this isn't the right forum, perhaps it's not a developer topic. But I can't find any other useful resources.

2 3

Trying to install ChurchWindows, but it requires IE4?
by Dan Kegel 19 Sep '03

19 Sep '03

My church is setting up a technology committee to decide what to do in the face of recent (possibly virus-related) problems they've been having. Naturally, Linux and Wine come to mind. The one app they run that I know of no Linux equivalent for is called ChurchWindows (from http://churchwindows.com, natch). A freely downloadable demo is available at their web site. Sadly, when I try to install it with 20030911, I get the error "Certain Microsoft files can only be provided by installing Internet Explorer version 4.01 Service Pack 2 or later. It appears that your computer does not curently have this version (or later) installed. Please install a newer version of Internet Explorer before running this install." I then tried --debugmsg +reg to see what it was sensing, but then the familiar old race condition reared its ugly head: "Error Number: 0x80040706 Description: Object reference not set Setup will now terminate" Can someone with more Wine juju than me have a look at http://209.190.80.19/trial/CW2003demo.exe and see how far off Wine is from supporting it? Thanks! - Dan

4 9

Re: WineHQ: Janitorial update
by Eric Pouech 19 Sep '03

19 Sep '03

Dimitrie O. Paun wrote: > ChangeLog > Add new task to track the removal of non-standard dirs. > The QueryDosDevice W->A crosscall was fixed. > The non-standard headers snoop.h and syslevel.h were removed. > to keep it uptodate, a couple of fixes in dlls/winmm (heap usage and W->A calls) have been fixed too A+ -- Eric Pouech

2 2

Re: [DXT3] menu in Tribunal now vanishes (regression?)
by MichaelGuennewig＠gmx.de 19 Sep '03

19 Sep '03

"Ann and Jason Edmeades" <us(a)the-edmeades.demon.co.uk> writes: > Thanks! I'll take a look. It'll probably be the weekend before I get a > chance to dig in, so bear with me. From your last post I think I understand > what you mean, and it should be easy to resolve. I'll get back to you Thanks, the patches dx76 up to dx79 have fixed the menu issue and als the most textures in the game are now back. But in the game the textures for the body of the persons are still missing as before, see screenshot at http://ls6-www.cs.uni-dortmund.de/~guennewi/Morrowind-Juib.png Don't know why because it doesn't give any error. Pherhaps these textures are only present with DXT2/4 support which isn't available. Michael

1 0

Re: Fix SetMenu after CreateWindow for invisible window
by Dmitry Timoshkov 19 Sep '03

19 Sep '03

"Michal Janusz Miroslaw" <M.Miroslaw(a)elka.pw.edu.pl> wrote: > +/********************************************************************** > + * SetMenu (USER32.@) > + */ > +BOOL WINAPI SetMenu( HWND hWnd, HMENU hMenu ) > +{ > + TRACE("(%p, %p);\n", hWnd, hMenu); > + > + if (GetWindowLongA( hWnd, GWL_STYLE ) & WS_CHILD) return FALSE; A better fix is to remove the above check altogether. -- Dmitry.

2 1

Re: dsound primary buffer volume/pan
by Oleg Prokhorov 19 Sep '03

19 Sep '03

Hello Robert, Thursday, September 18, 2003, 8:34:04 AM, you wrote: RR> Adds volume and pan support to primary buffers. RR> Adds more property set support. This patch creates with all the rest (excluding wineoss) drivers, which don't have IDsDriverBufferImpl_SetVolumePan implemented. Dsound is not working on them due to PrimaryCreate / PrimaryOpen fails. -- Best regards, Oleg mailto:xolegpro(a)rbcmail.ru

1 0

Re: wine/ tools/winapi/msvcmaker ./configure.ac ./ ...
by Dimitrie O. Paun 19 Sep '03

19 Sep '03

On September 19, 2003 12:36 am, Alexandre Julliard wrote: > Log message: > Get rid of relay32 directory. I still can't believe this is gone... I thought we'd never get rid of relay32, if1632, etc. They were a constant eye sore for me :) Thank you! -- Dimi. P.S. And now a little dance around the fire to celebrate...

2 1