October 2001 - Wine-Devel - list.winehq.org

problems with dde, I think...

by Josh Duncan

Hi all, I'm a little new to the list, so please be forgiving if I'm asking old questions. I've tried to do some research on the net and list archives to see if my questions have been answered before. I'm trying to run a program called Logos, which is a reference system for Biblical texts and other various study materials. It has a two-(independent)-windowed type interface, with a list of reference materials in one window("library"), and a workspace-type window containing subwindows with the texts of individual books. I can get the "library" window to show up when running wine, and the initial "quick tip" window, but when I click the close button of the tip window wine encounters an exception. I believe the exception is in user32.dll, in the ddeuninitialize16 function. This is going from my memory of the one time I was able to attach to the wine process to catch the exception. I attempted copying the kernel32, user32 dll's from a win98 machine into my wine dir, but this did not help. after doing this, I could not run winedbg either, as it exits giving me the following error: err:win32:PE_fixup_imports No implementation for kernel32.dll.0(wine_get_unix_file_name) imported from winedbg, setting to 0xdeadbeef This error occurs even after removing the two dlls. If anyone can give me advice about how to fix this, I'll try to post up the winedbg results of running the program(backtrace, et al) I know that some people have mentioned dde being broken between apps, I don't know if this is the case for this app, because when I was able to do a winedbg "walk process" it only listed one w-process, with 4 w-threads running. I have attached my wine config file, any advice would be greatly appreciated. (Also, if this email should have gone to the wine-users list, please let me know) I'm using the wine-20011029.i386.rpm from LinuxEasy, linked off of the winehq.org download page. sincerely, Josh Duncan not that it matters, but a pic of the two windows I can produce are at http://www.msmd.net/logos.jpg I'd try and show a pic of the other window, but the system gets so unstable that it can hard-lockup the machine, I think this is a cpu-spinning-in-signal handler problem but I couldn't get you a log right now to prove it.

Oct. 30, 2001

Re: Colour problems with latest DIB fix

by Gerard Patel

At 12:53 PM 29/10/2001 -0800, you wrote: >> Anyway, the attached patch is displaying the failing picture correctly. >> What do you think ? > > If I apply your patch, then I see that the red and blue are exchanged >when I run vnc as: > vncserver :1 -geometry 800x600 -depth 16 -pixelformat bgr565 Could you explain a bit more what you are doing ? I would like to try the same but I don't see how to do it. Also, if you use vnc to test, did you consider the possibility that vnc could have a bug somewhere ? > Without the patch everything is fine. I also checked the formula and >I see nothing wrong with it (as long as you are on a little endian >computer, but you don't have a big endian x86, do you? ;-). Never heard that Amd is doing that :-) > So the question is why doesn't it work on your system. Can you run >The Sims (I assume it's with that the problem happens) with a +bitmap >trace? Then look for cases where either the bmp or the dib has >"depth=24". Then check the masks to see if they make sense. For example, >do they match what you know of your X server. You may also put a trace >in Convert_888_to_565_reverse and check that it is invoked at the right >time. Do they have a bgr888 dib (unsupported in windows AFAIK, but, >except for 1 or 2 shortcuts, supported in Wine)? Maybe it is the DIB >masks that are wrong. That's a bit involved for this evening :-/ > Also it would be great if you could run my test application and check >whether it displays fine. The line to watch for is the sixth line >(rgb888 dib to 565 rgb/bgr bmp and back). I had (of course) a number of problems with your app. First on my current computer - a Mandrake 8, with Xfree 4, your procedure ran fine until it tried to link the thing and then it reported that WinMain was missing. I changed in your code 'main' to 'WinMain' and all was well. I tried then to do the same on my other computer (Suse 6.1 with Xfree 3.3 and a hand-compiled glibc 2.1.3). No joy. Configure always failed with libwine.so failing to find symbols like dlclose and the like. I did not figure out what is happening, I just copied the binary from my other box and it worked. As for the results, what I see is 11 lines of 13 blocks of 4 colors : black, red, green, blue. I guess that when you speak of red and blue exchanged, you mean that in one of these blocks the red and blue lines are exchanged, right ? What is strange is that on both of my computers the display of your app is the same with or without my patch. No line exchanged that I can see. > As for The Sims, if you can also check the -debugmsg +bitmap output >and make sure it's coherent. I don't have The Sims... I use a little freeware app that is not available from the internet anymore, unfortunately :-/. And again, that's a bit complex for today. I'll try later. Gerard

Oct. 29, 2001

Re: problem with virtualdub.

by Roger Fujii

Nerijus Baliunas wrote: > On Mon, 29 Oct 2001 17:04:51 +0000 Gerard Patel <gerard.patel(a)nerim.net> wrote: > > GP> The web site says that this executable is compressed with a tool called UPX. This > GP> seems to be the reason for the problem, as the original program is compiled with > GP> Ms VC++ - not a very exotic tool. > You can unpack it with upx, even under Linux - http://upx.sourceforge.net/. well, this indeed seems to be part of the problem. Decompressing it makes it go further, but still dies. > GP> I tried a hack for that and virtualdub got further; however, I see that it is > GP> trying to load avicap32.dll and this is not going to work - even under Linux > I remember I was able to start Virtual Dub with wine (maybe a year ago), but don't > remember if it worked. BTW, just yesterday TAKESHIMA Hidenori added stub for avicap32. really? Doesn't seem to be in cvs yet... Looks like it's getting upset about not finding StdChannelMask, but this is just a guess. -r

Oct. 29, 2001

(no subject)

by Simon Britnell

I had an idea for using poll or select (rather than calculating expected play time and using a timeout) in audio.c The idea is to have a thread that sits on a select and dispatches a synthetic wine event (ie. we'd make up a WM_DSP_READY event) when the select returns. wodPlayer could then do an infinite wait and the select could be turned on or off when starting and ending playing. >From my relatively naive view a generic mechanism like this (ie. one that converts async IO selects into wine events) would be a good feature to have from a performance and code tidiness point of view. One caveat is that we'd need to ensure nothing evil happened if somebody sent the synthetic event from somewhere other than the select waiter. Comments? Am I missing something? __________________________________________________ Do You Yahoo!? Make a great connection at Yahoo! Personals. http://personals.yahoo.com

Oct. 29, 2001

Re[2]: problem with virtualdub.

by Nerijus Baliunas

On Mon, 29 Oct 2001 17:04:51 +0000 Gerard Patel <gerard.patel(a)nerim.net> wrote: GP> The web site says that this executable is compressed with a tool called UPX. GP> This GP> seems to be the reason for the problem, as the original program is compiled with GP> Ms VC++ - not a very exotic tool. You can unpack it with upx, even under Linux - http://upx.sourceforge.net/. GP> I tried a hack for that and virtualdub got further; however, I see that it is GP> trying to load avicap32.dll and this is not going to work - even under Linux I remember I was able to start Virtual Dub with wine (maybe a year ago), but don't remember if it worked. BTW, just yesterday TAKESHIMA Hidenori added stub for avicap32. Regards, Nerijus

Oct. 29, 2001

Re: Colour problems with latest DIB fix

by François Gouget

Gerard Patel wrote: > > At 02:18 PM 21/10/2001 -0700, you wrote: > > > >Johan Gill wrote: > >> > >> Since the big DIB fix, red and blue are exchanged in some cases, > >> for example in the install program for The Sims. > >> > >> This happens in both depth 16 and depth 24. > > > > Could you give me some more information about your X server (version, > >driver)? > > I see a very similar problem with my X server : 4.03 that comes with my > Mandrake 8, graphic card is a Sis 6326. > > I had tested your patch before it was committed BTW but I failed to notice > the problem :-/ > > Anyway, the attached patch is displaying the failing picture correctly. > What do you think ? If I apply your patch, then I see that the red and blue are exchanged when I run vnc as: vncserver :1 -geometry 800x600 -depth 16 -pixelformat bgr565 Without the patch everything is fine. I also checked the formula and I see nothing wrong with it (as long as you are on a little endian computer, but you don't have a big endian x86, do you? ;-). So the question is why doesn't it work on your system. Can you run The Sims (I assume it's with that the problem happens) with a +bitmap trace? Then look for cases where either the bmp or the dib has "depth=24". Then check the masks to see if they make sense. For example, do they match what you know of your X server. You may also put a trace in Convert_888_to_565_reverse and check that it is invoked at the right time. Do they have a bgr888 dib (unsupported in windows AFAIK, but, except for 1 or 2 shortcuts, supported in Wine)? Maybe it is the DIB masks that are wrong. Also it would be great if you could run my test application and check whether it displays fine. The line to watch for is the sixth line (rgb888 dib to 565 rgb/bgr bmp and back). I attached it to this email. Here is how to compile and run it: mkdir dib cp dib.c dib cd dib ~/wine/tools/winemaker . ./configure --with-wine=$HOME/wine make ./dib (note: don't run winemaker . in your home directory!) As for The Sims, if you can also check the -debugmsg +bitmap output and make sure it's coherent. -- François Gouget fgouget(a)codeweavers.com

Oct. 29, 2001

Re: problem with virtualdub.

by Gerard Patel

At 05:17 PM 26/10/2001 -0400, you wrote: > >there seems to be a funky problem with virtualdub >(http://virutaldub.sourceforge.net) Get differing >results between solaris and linux, but both seems to >fail before the app comes up. > >Under solaris, it seems like something is screwed up >with the loader - when it tries to load the binary, >it dies in dump_exports because what's at pe_exports >is just garbage. Not sure where it's dying under >linux (probably the same place) - don't have a devel >environment there (yet). > >Nothing unusual seems to happen before it gets there >(both solaris/linux binaries run sol.exe and other stuff just fine). >My suspicion is that the IMAGE_DIRECTORY_ENTRY_EXPORT data isn't pointing to what >wine thinks is there, but would like someone who knows more >about the loader to say something.... > >can anyone help shed more light onto this? > >-r > Yes, I have looked at this application. It loads fine under Windows - however, both dependancy walker and the Nt quick viewer are not happy with it (one displays that there is a format error in the exe, the other loops and finally crashes) The web site says that this executable is compressed with a tool called UPX. This seems to be the reason for the problem, as the original program is compiled with Ms VC++ - not a very exotic tool. I think that a fix could be to detect a corrupted header in dump_exports and to not try to display anything if numberoffunctions is not coherent with the export section size. I tried a hack for that and virtualdub got further; however, I see that it is trying to load avicap32.dll and this is not going to work - even under Linux ;:-) Gerard

Oct. 29, 2001

struct thread in server/thread.h

by Gerald Pfeifer

In server/thread.h we have a definition of struct thread. Unfortunately, though, this clashes with struct thread as defined in /usr/include/sys/thread.h on FreeBSD 5.x and probably other systems (Solaris) as well. How should this be addressed? Would any of you mind fixing this? (If it's okay to just rename struct thread to struct winethread, I'm volunteering to handle that, else it'd be great could one of the experts have a look!) Gerald -- Gerald "Jerry" pfeifer(a)dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

Oct. 29, 2001

Wine and Starcraft

by geschrader

I've been pecking away attempting to track down why the current wine build won't run Starcraft. When it starts up I get the effect that I playfully refer to as the TSOD (Technicolor Screen Of Death). What I've found is that if all of the references to gdi_display in X11DRV_DDHAL_SetPalEntries( ) are replaced with thread_display( ) then it actually starts up and looks pretty good. However, the keyboard and mouse don't work so this is at best a partial fix. I'm pretty sure that this is on the right track since X11DRV_XF86DGA2_CreatePalette( ) uses thread_display( ) to make the palette that is eventually given to X11DRV_DDHAL_SetPalEntries( ). The only other function in dlls/x11drv/x11ddraw.c that uses gdi_display is X11DRV_DDHAL_DestroyPalette( ). Should thread_display( ) be used there also? How is gdi_display related to the display returned by thread_display( )? Both are used in quite a few places but they're obviously not the same thing. Can anybody give me some hints on what might be happening with the keyboard and mouse? Thanks much, ges

Oct. 29, 2001

Re: Colour problems with latest DIB fix

by Gerard Patel

At 02:18 PM 21/10/2001 -0700, you wrote: > >Johan Gill wrote: >> >> Since the big DIB fix, red and blue are exchanged in some cases, >> for example in the install program for The Sims. >> >> This happens in both depth 16 and depth 24. > > Could you give me some more information about your X server (version, >driver)? I see a very similar problem with my X server : 4.03 that comes with my Mandrake 8, graphic card is a Sis 6326. I had tested your patch before it was committed BTW but I failed to notice the problem :-/ Anyway, the attached patch is displaying the failing picture correctly. What do you think ? Gerard

Oct. 28, 2001