June 2004 - Wine-Devel - list.winehq.org

Re: winegcc and winemaker- a few minor bugs

by Francois Gouget

On Thu, 17 Jun 2004, Steven Edwards wrote: > Hola, > I noticed two minor bugs when porting some sample DirectX code. > > 1. MSVC and Borland both allow for > > WINAPI WinMain(HINSTANCE hInstance, HINSTANCE, LPSTR, int) > > where under winegcc you will get the following error: "winmain.cpp:187: > error: ISO C++ forbids declaration of `WinMain' with no type". If you > change it to: > > int APIENTRY WINAPI WinMain(HINSTANCE hInstance, HINSTANCE, LPSTR, int) > > Then it will compile fine. So should we work around the incorrectness > of code allowed for by other compilers? Or document the changes that > you will run in to porting C++ from other compilers? Detecting this kind of problem would require us to have a C++ parser which is more than intended for winemaker. So we could document it though, imho, the g++ error message is documentation enough. > 2. Winemaker does not import all of the import libs from the MS_VC > makefile. winemaker does not use the Visual C++ makefile at all. That's why you get this error. -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ I haven't lost my mind, it's backed up on tape around here somewhere...

June 21, 2004

Re: Donations - RFC

by Brian Vincent

> >I would say sponsoring Wineconf 2005 would be a better idea (so we could go > >somewhere exotic (not that St Paul in winter was not exotic in a kinda > >perverted way for most of us :-) )). > > > > > I have volunteered to organize it in Cyprus. I've already started > lobbying for that location. After last year's last minute "hey, it's > more comfortable in the US" change, I decided that long term preparation > for votes, like you do in any self-disrespecting parliament, is the only > way to go. I just a did a quick search and found that airfare to Cyprus is about twice that of anywhere in western Europe. That's about what I would expect too. The cheapest flights from North America are generally into London, Paris, Amsterdam, or Frankfurt. Sometimes Madrid and Rome are reasonable too. Cyprus also adds a fair amount of travel time. Don't get me wrong - Cyprus would be a cool destination and I would love to travel there. It's also wonderful someone is will to try to organize a location other than North America. However, I'm willing to bet a Cyprus location will be such an inconvenient place to get to that we'd keep a lot of people away.

June 21, 2004

Game Testing Update - 20040615

by Mike Kost

Here a continuation of my game testing efforts. Same rules still apply in my efforts: All games tested are freely available as demos or F/OSS games on the Internet and they don't work right. Since my last report: - No bugs have closed - 2 more bugs filed - - #2299: Static in Dominions II Demo Audio - - #2300: Galactic Civilizations demo dies with unhandled exception - Finally started testing Galactic Civilizations Galactic Civilization Summary ----------------------------- Galactic Civilization is a bit grumpy and requires both a DCOM9x and an IE install to work correctly. It installed, but died with an unhandled exception when the game was started. Filed bug #2300 http://www.tashcorp.net/Wine/GalCiv/galciv-20040615.html Austerlitz Summary ------------------ There were no noticable changes over 20040505 testing. http://www.tashcorp.net/Wine/Austerlitz/austerlitz-20040615.html Dominions II ------------ There were no changes over 20040505, but I turned on the speakers and noticed that there was static in the audio at the startup screen. This was filed as bug #2299. http://www.tashcorp.net/Wine/DominionsII/dominionsII-20040615.html Remote Assault -------------- There were no noticable changes over 20040505 testing. http://www.tashcorp.net/Wine/RemoteAssault/remoteassault-20040615.html

June 21, 2004

MSVCRT - __unDNameEx

by Andrei Barbu

Hey, I encountered an app that heavily relies on __unDNameEx which is a stub, and I couldn't find any MSDN info on it. The stub did have a hint as to an implementation of it, said that it's been implemented pretty well in tools/winebuild/msmangle.c I looked for that, couldn't find it, but I did find it in the CVS attic in wine/tools/speckmaker http://cvs.winehq.com/cvsweb/~checkout~/wine/tools/specmaker/Attic/msmangle… Is there a reason this isn't used there? or can I clean it up a little and include it? Yes it does have some fixme's, but it's better than ignoring it. Thanks, Andrei

June 21, 2004

Re: MSVCRT - __unDNameEx

by Rein Klazes

On Mon, 21 Jun 2004 11:33:24 -0400, you wrote: > Hey, > > I encountered an app that heavily relies on __unDNameEx which is a stub, > and I couldn't find any MSDN info on it. > > The stub did have a hint as to an implementation of it, said that it's > been implemented pretty well in tools/winebuild/msmangle.c I guess that is a typo, and it should be tools/winedump/msmangle.c Rein. -- Rein Klazes rklazes(a)xs4all.nl

June 21, 2004

Re: MSVCRT - __unDNameEx

by Dimitrie O. Paun

On Mon, Jun 21, 2004 at 11:33:24AM -0400, Andrei Barbu wrote: > Is there a reason this isn't used there? or can I clean it up a little > and include it? Yes it does have some fixme's, but it's better than > ignoring it. I can't see why not. -- Dimi.

June 21, 2004

Remote operations -- need for advice

by Alexander Yaworsky

I'm thinking how to implement CreateRemoteThread and besides fix memory management functions. The complete (afaik) list includes: RtlCreateUserThread NtAllocateVirtualMemory NtFreeVirtualMemory NtProtectVirtualMemory NtQueryVirtualMemory NtLockVirtualMemory (do nothing?) NtUnlockVirtualMemory (do nothing?) NtFlushVirtualMemory NtMapViewOfSection NtUnmapViewOfSection Suggested implementation. Let's add two groups of handlers to the wineserver, something like remote_operation_xxxx and remote_operation_xxxx_complete. remote_operation should do the following: 1) suspend_for_ptrace(); 2) inject a piece of code into required process and start its execution; 3) resume_after_ptrace(); 4) place calling thread into suspended state (or into some wait state?) remote_operation_complete should prepare reply and resume thread suspended by remote_operation. Injected code should call required function and then remote_operation_complete in context of required process. The question is: how to correctly get address of function? imho possible solutions are: 1) assume ntdll loaded at the same address for all processes -- unreliable; 2) get dll base address from per-process dll list and parse ELF by hand -- too complicated (?); 3) pass relative offsets and add them later to the ntdll's base address -- unreliable a bit: ntdll may be replaced, although, it is unlikely; 4) extend the struct process (server/process.h) and the request init_process with pointers to required functions -- most reliable and simple but looks ugly. What do you think?

June 21, 2004

Re: Donations - RFC

by Tom

Shachar Shemesh wrote: > I have volunteered to organize it in Cyprus. I've already started > lobbying for that location. After last year's last minute "hey, it's > more comfortable in the US" change, I decided that long term > preparation for votes, like you do in any self-disrespecting > parliament, is the only way to go. > I vote for France :-) Tom

June 21, 2004

Re: Incoming Message

by Twickline

June 21, 2004

winedbg crashes

by Jon Griffiths

Hi, I'm having some trouble with the debugger and a large C++ application I'm trying to debug. This seems unrelated to the recent changes that were made (the old debugger used to have issues as well). When my app dies i get the following: wine: Unhandled exception (thread 0009), starting debugger... fixme:console:SetConsoleCtrlHandler (0x405db800,1) - no error checking or testing yet WineDbg starting on pid 0x8 fixme:dbghelp:SymLoadModule should no longer happen fixme:dbghelp:SymLoadModule should no longer happen Unhandled exception: page fault on read access to 0x00000001 in 32-bit code (0x40a272a2). In 32 bit mode. err:dbghelp_stabs:stabs_pts_read_type_def Forward declaration name mismatch _Obj <> __default_alloc_template<true,0> err:dbghelp_stabs:stabs_pts_read_type_def Forward declaration name mismatch _Impl <> locale err:dbghelp_stabs:stabs_pts_read_type_def Forward declaration name mismatch id <> locale wine-pthread: stabs.c:169: stabs_add_include: Assertion `idx >= 0' failed. wine: Unhandled exception (thread 000b), starting debugger... fixme:console:SetConsoleCtrlHandler (0x405db800,1) - no error checking or testing yet WineDbg starting on pid 0xa Unhandled exception: assertion failed in 32-bit code (0xffffe410). In 32 bit mode. 0xffffe410: popl %ebp Wine-dbg> The 'read access to 0x00000001' exception is my app, but the the debugger fails the above assertion and dies. I have stripped the application and .so's for this app (its winelib based), but this doesn't help the crash. I suspect maybe the template brackets may be screwing up the stabs parsing (since template names can contain commas). Has anyone seen something like this before? I'm happy to help debug the problem from this end if that helps. Cheers, Jon ===== "Don't wait for the seas to part, or messiahs to come; Don't you sit around and waste this chance..." - Live jon_p_griffiths(a)yahoo.com __________________________________ Do you Yahoo!? Yahoo! Mail - You care about security. So do we. http://promotions.yahoo.com/new_mail

June 21, 2004