October 2005 - Wine-Devel - list.winehq.org

Re: version<unspecified>

by Paul Vriens

> On Fri, 28 Oct 2005 20:15:06 +0200, Paul Vriens <Paul.Vriens(a)xs4all.nl> > wrote: > >> On Fri, 2005-10-28 at 19:53 +0200, wino(a)piments.com wrote: >>> I have an app that tests the version number of certain dlls it uses on >>> startup. >>> >>> The first seems to be riched20.dll >>> >>> If I run it on 20050524 it starts and works. >>> >>> If I run from the same installation after installing wine-0.9 (or just >>> about any winecfg based version) it starts throwing errors like: >>> >>> riched20.dll version<unspecified> was found, this program requires at >>> least..... >>> >>> >>> I then have to pull in a native dll and tell wine to go native when in >>> fact the buildin functions work perfectly. >>> >>> bug or feature? >>> >>> I guess this was done intentionally or is a result of an intentional >>> change and seems related to winecfg becoming active. >>> >>> What is the best way to deal with this? It seems a shame to install >>> native >>> dlls when the wine code does the job. >>> >>> TIA >>> >>> >>> >> Hi, >> >> don't have time (yet) to do stuff in Wine, but have a look at >> http://www.winehq.org/pipermail/wine-patches/2005-September/020423.html >> >> this shows you how to add version stuff to a dll. >> >> cheers, >> >> Paul >> >> > Thanks for the info. > > I was not really wanting to get into recoding the wine dlls, I'm not sure > I'm upto it and I would not have the inside knowlege to know what version > would best represent each dll. > > I have an app. which is working pretty well under 20050524 but if I run > one of the winecfg releases the app. throws this version <unspecified> > bit. > > I would like to know what change brought this about. > > Is the text "unspecified" actually returned by wine now or is this the > app's interpretation. > > There seems to be a definate regression here. > > Do you have any more detail on this? > > TIA > Hi, could you send me a +ver,+loaddll,+module trace of both versions. I will have a look (if time permits). I hope I can see enough from the logs as I'm a but reluctant to go back to (or install somewhere else) the 20050524 version. Cheers, Paul.

Oct. 31, 2005

Re: server: Add object type to struct object

by Vitaliy Margolen

Sunday, October 30, 2005, 9:52:40 PM, Robert Shearman wrote: > Vitaliy Margolen wrote: >>This is by far not what native does. But we are not implementing the whole >>kernel here. So having just an object type ID will work just fine for what >>we need. >> >>Vitaliy Margolen >> >>changelog: >> server >> - Add object type to struct object >> > Why do you need this? It seems that having an object type field is > redundant since we can uniquely identify an object by the address of its > object_ops. Because the way I've implemented directory object it's not fully abstract. It needs to have a special case for symlinks lookups. Also I will need that for NtQueryDirectoryObject as it have to return object type name. I suppose I could compare obj->ops but that means they won't be static anymore and will have to be visible outside the file where they are defined. Vitaliy

Oct. 31, 2005

Re: server: Add object type to struct object

by Robert Shearman

Vitaliy Margolen wrote: >This is by far not what native does. But we are not implementing the whole >kernel here. So having just an object type ID will work just fine for what >we need. > >Vitaliy Margolen > >changelog: > server > - Add object type to struct object > Why do you need this? It seems that having an object type field is redundant since we can uniquely identify an object by the address of its object_ops. -- Rob Shearman

Oct. 31, 2005

Anarchy Online, Specular Lighting, Pixel Shaders....

by Evil

Ivan Gyurdiev wrote: > I see that the patch from Oliver which implements pixel shaders hasn't > been merged yet, which might explain why they aren't working... This probably explains my results with Anarchy Online tonight too, if I don't misunderstand about shaders. I got the AO patcher working, and the main game too after some tweaking (basically, I had to force some native DLLs back to the built-in ones after playing with Sidenet yesterday). It even runs at a respectable speed, except there are a lot of textures coming out completely black. I see the following errors repeating while it's running: err:ddraw:set_render_state Specular Lighting not supported yet. err:ddraw:GL_IDirect3DDeviceImpl_7_SetLight dvTheta not fully supported yet ! err:ddraw:GL_IDirect3DDeviceImpl_7_SetLight dvTheta not fully supported yet ! There didn't appear to be any existing entries in the bug database for AO, so I added this info there. I'll re-test as I see patches coming into CVS. -Jesse

Oct. 31, 2005

Re: tracking interprocess comms.

by wino＠piments.com

On Mon, 31 Oct 2005 02:23:52 +0100, <wino(a)piments.com> wrote: > Hi, > > I am trying to capture output that an app normally sends to other > processes under windows operation. > > I know that in the case of major M$ apps this will be my ole but I > believe that in the more general case this must be by simulating > keystrokes > > I tried WINEDEBUG="+relay" but the output file filled the 1.7GB !! of > free space I had before the app got as far at putting up the main window. > > could s.o. advice on a more suitable channel to watch. > > Many thanks. > > > PS I tried "+key" and I can see the real keystrokes going to the app but not the relayed text , maybe I need to look to WMCHAR or so. Thx

Oct. 31, 2005

tracking interprocess comms.

by wino＠piments.com

Hi, I am trying to capture output that an app normally sends to other processes under windows operation. I know that in the case of major M$ apps this will be my ole but I believe that in the more general case this must be by simulating keystrokes I tried WINEDEBUG="+relay" but the output file filled the 1.7GB !! of free space I had before the app got as far at putting up the main window. could s.o. advice on a more suitable channel to watch. Many thanks.

Oct. 31, 2005

Re: Half-Life 2 / Counter-Strike: Source under Wine

by Ivan Gyurdiev

> Try disabling pixel shaders in winecfg to get around that. They never > did work for me. I tried to use them yesterday with HL2 to get some > screenshots, and they froze my computer. If you're lucky enough to even > star a game using them, they mess things up anyway. > Alternatively, you could try filing bug reports for the problems you experience, rather than working around them. If no bug reports are filed, nothing will get fixed. Anyway, I see that the patch from Oliver which implements pixel shaders hasn't been merged yet, which might explain why they aren't working... >> Visually, I would say it looks better than I recall getting from >> Cedega. The textures on the monoliths were always plain and flat when I >> was running Cedega. I'm tempted to renew my Cedega subscription to make >> sure I'm remembering right. >> > Yeah, I know what you mean. I've never been a fan of Cedega, so I say to > hell with it. I would stick with free WINE. ;-) > ...and yet you have no problem making use of the Transgaming ActiveX control, which contains Microsoft DLL(s) within?

Oct. 31, 2005

Re: Half-Life 2 / Counter-Strike: Source under Wine

by Evil

That should have said "regsvr32 mozctlx.dll", BTW. That's what I get for writing from memory. :) Evil wrote: >James Liggett wrote: > > > >>BTW, what did you do to get transgaming's mozctl working? Thanks. >> >> >> >> > >All you need to do is go to the control directory >(~/.transgaming_global/mozcontrol, by default) and run "regsvr >mozctlx.dll" to register the DLL. Many thanks to Fabian Bieler, who >gave me that info. > >Since I think I've gotten CS:S and HL2 working as good as they can with >the current code, I think I'll see if I can get anywhere with Anarchy >Online (I hate the game, but would like to see WINE running it. ;) ). > >-Jesse > > > > > >

Oct. 30, 2005

Re: Half-Life 2 / Counter-Strike: Source under Wine

by Evil

James Liggett wrote: >BTW, what did you do to get transgaming's mozctl working? Thanks. > > All you need to do is go to the control directory (~/.transgaming_global/mozcontrol, by default) and run "regsvr mozctlx.dll" to register the DLL. Many thanks to Fabian Bieler, who gave me that info. Since I think I've gotten CS:S and HL2 working as good as they can with the current code, I think I'll see if I can get anywhere with Anarchy Online (I hate the game, but would like to see WINE running it. ;) ). -Jesse

Oct. 30, 2005

Re: Half-Life 2 / Counter-Strike: Source under Wine

by James Liggett

On Sun, 2005-10-30 at 11:34 -0600, Evil wrote: > Thanks for the tip, James! Disabling all the debugging had a much > bigger impact than I would have expected. I've included my observations > and comments below, in hope that they might be of use to the devs > working on DX. Yeah, you wouldn't think turning off debugging messages would do much of anything but it does. Don't ask me why though... :) > > The Counter-Strike: Source Video Stress Test is now giving me 15.74fps > in 1024x768 windowed mode (vs. 24 in Cedega and 44 in Win98). It's a > definite step in the right direction, but still a little too low to > prevent you from getting shot by any newb. :) I get the same speed at > lower resolutions, which indicates that I'm CPU-limited. I can believe > that, since it's pegged-out the entire time I run the test. Hmmm... I can usually get double that. My card isn't all that different from yours (I have a Radeon 9800 pro) Personally I think it has more to do with the speed of the code and yet to be implemented features. I can imagine that since Oliver is still trying to stabilize things and implement stuff, speed isn't yet the priority. > > I get pretty much the same performance whether I am windowing the game > by setting wine to emulate a desktop, or just passing -window to the > hl2.exe (15.2fps). > > The odd thing to me is: I get the same speed when running in fullscreen > mode. This is even the case if I used XRandR to set my desktop to the > same resolution before starting Steam, so that I don't end up > accidentally scrolling across my virtual desktop when I move the mouse > down or to the right. Doing that is also a help because it stops my > newsticker from drawing over the top of the game screen, but it seems to > have no impact on performance. > > Do OpenGL applications typically perform the same in windowed mode as > fullscreen mode? My experience programming with DirectX made me expect > fullscreen to run faster since the video card only has to flip the > backbuffer pointer instead of blitting from one chunk of memory to > another. Of course, that all being handled by the video card, I suppose > this too is probably just a side-effect of my CPU-limited speed. I wouldn't think running a game in a window versus full-screen would change much of anything. Then again, I'm no 3D expert either. > > Here are the parameters I'm feeding to CS:S by the way, in case anyone's > interested: "-heapsize 512000 -fullscreen -width 1024 -height 768". > I've tried passing "-dxlevel 70" to the hl2.exe, but don't see much if > any performance difference. It is interesting to not that if you don't > pass this parameter, you will see little white boxes indicating each > light source. These boxes aren't clipped by any object. > > Passing "-dxlevel 80" causes the menu screen to come up completely > white, making it unuseable. Forcing DirectX 8.1 or 9.0 has the same > white-screen effect. Try disabling pixel shaders in winecfg to get around that. They never did work for me. I tried to use them yesterday with HL2 to get some screenshots, and they froze my computer. If you're lucky enough to even star a game using them, they mess things up anyway. > > Restarting CS:S without exiting Steam completely can have flaky > results. It might be related to the amount of time it actually takes > the game to cleanup and exit. I've noticed that right after exiting the > game and then exiting Steam, it can take several minutes or longer (of > maxed out CPU utilization) before a command prompt returns. Maybe the Valve minidump program is intercepting a crash. I see that a lot with Source multiplayer games like CS:S and HL2:DM. I believe that it might be because of the web browser control used to display MOTDs. BTW, what did you do to get transgaming's mozctl working? Thanks. > > Visually, I would say it looks better than I recall getting from > Cedega. The textures on the monoliths were always plain and flat when I > was running Cedega. I'm tempted to renew my Cedega subscription to make > sure I'm remembering right. Yeah, I know what you mean. I've never been a fan of Cedega, so I say to hell with it. I would stick with free WINE. ;-) > > Anyway, it's looking very promising. Keep up the great work! I'm not really involved with development on the DX emulation. I just follow Oliver's work, test, cross my fingers, and hope it works. :) > > -Jesse > > > > James Liggett wrote: > > >I have single player HL2 running decent. Not quite excellent, but > >playable. Try running steam with this command: > > > >WINEDEBUG="-all" wine steam.exe > > > >And you should get anywhere from 10-30 FPS. It's not the best in the > >world, but at least you can play. :) I have started a new game, and I am > >now in the beginning of the car sequence, and aside from some textures > >being lit up and the flashlight not working, it's pretty nice. But, > >beware of problems with pixel shaders. The make CS and HL2 deathmatch > >totally useless (they give a white screen in menu. It's not frozen, but > >you can't see anything.) I don't see that behavior with single player > >because of the animated menus. Plus, pixel shaders screw things up big > >time anyway. It inverts colors and makes it look like NPCs are holding > >tree trunks. I'll put up some shots up later so you can see what I mean. > > > >James > > > > >

Oct. 30, 2005