March 2005 - Wine-devel - list.winehq.org

Config Problems ?
by Andrew Neil Ramage 06 Mar '05

06 Mar '05

I have downloaded and installed wine-20050211, but when I try to start a Windows program by clicking on it nothing happens. Starting it from the command line produces screens of output similar to the following: PatBlt (0x74, 0, 109, 472, 1, 15728673): stub fixme:ttydrv:TTYDRV_DC_PatBlt (0x74, 0, 110, 472, 2, 15728673): stub fixme:ttydrv:TTYDRV_DC_PatBlt (0x74, 0, 112, 472, 1, 15728673): stub fixme:ttydrv:TTYDRV_DC_PatBlt (0x74, 0, 113, 472, 1, 15728673): stub fixme:ttydrv:TTYDRV_DC_PatBlt (0x74, 0, 114, 472, 1, 15728673): stub fixme:ttydrv:TTYDRV_DC_PatBlt (0x74, 0, 115, 472, 2, 15728673): stub fixme:ttydrv:TTYDRV_DC_BitBlt (0x98c, 0, 0, 521, 130, 0x74, -24, 0, 13369376): stub fixme:ttydrv:TTYDRV_DC_StretchBlt (0x98c, 1, 0, 520, 86, 0x998, 0, 0, 520, 86, 8913094): stub fixme:ttydrv:TTYDRV_DC_StretchBlt (0x98c, 1, 0, 520, 86, 0x9a0, 0, 0, 520, 86, 15597702): stub fixme:ttydrv:TTYDRV_DC_BitBlt (0x74, -24, 0, 521, 130, 0x98c, 0, 0, 13369376): stub fixme:ttydrv:TTYDRV_DC_BitBlt (0x98c, 0, 0, 344, 40, 0x74, 64, 77, 13369376): stub fixme:ttydrv:TTYDRV_DC_BitBlt (0x74, 64, 77, 344, 40, 0 Anyone know what's wrong ? I am using Mandrake 10.0, Gnome 2.4 and wine-20050211. -- Andrew [hackcoughsniffle] Internal Virus Database corrupt

2 1

help wanted - testing "stuck Alt" patch
by Krzysztof Foltman 06 Mar '05

06 Mar '05

A few days ago I've sent this small patch to wine-patches: http://www.winehq.org/hypermail/wine-patches/2005/03/0091.html It's supposed to fix the long-existing (and annoying) problem where switching windows with Alt+TAB caused accelerator keys to work improperly (ie. pressing Ctrl+A was misrecognized as Ctrl+Alt+A). The patch seems to do its job (accelerator keys still work correctly after reactivating the app's window). However, I don't have many different configurations to test it. It would be nice if someone else applied it and tested if I haven't introduced a regression with this patch (it is possible, considering that it affects basic window management code used in all the GUI applications), and if it did the job. Any testing would be very helpful, regardless of whether the Alt+TAB problem existed on the particular machine or not. Although some people said that Alt+TAB problem doesn't exist in their configuration, I suspect that the problem is related to the particular usage patterns, and not configurations. Thanks in advance, Krzysztof

1 0

Running ZooTycoon 2 Demo
by Paul Vriens 06 Mar '05

06 Mar '05

Hi, I'm running current CVS with a fresh tools/wineinstall. Only changes to the config-file: "Drivers" = "winealsa.drv" "HardwareAcceleration" = "Emulation" ; Zoo Tycoon 2 Demo install [AppDefaults\\Zoo2Trial.exe\\Version] "Windows" = "winxp" ; Zoo Tycoon 2 Demo run [AppDefaults\\zt2demoretail.exe\\x11drv] "Desktop" = "800x600" installation of the app is fine (set version to winxp). When trying to run the demo it immediately fails: fixme:d3d9:IDirect3DDevice9Impl_CreateSurface (0x77e38ea8) call to IWineD3DDevice_CreateRenderTarget failed First chance exception: page fault on read access to 0x00000000 in 32-bit code (0x14a9481a). Fixed/circumvented that by: Index: dlls/wined3d/directx.c =================================================================== RCS file: /home/wine/wine/dlls/wined3d/directx.c,v retrieving revision 1.25 diff -u -p -r1.25 directx.c --- dlls/wined3d/directx.c 2 Mar 2005 13:44:58 -0000 1.25 +++ dlls/wined3d/directx.c 5 Mar 2005 16:11:45 -0000 @@ -1578,7 +1578,8 @@ HRESULT WINAPI IWineD3DImpl_CreateDevi object->render_ctx = object->glCtx; object->renderTarget = object->backBuffer; - IWineD3DSurface_AddRef((IWineD3DSurface *) object->renderTarget); + if (NULL != object->renderTarget) + IWineD3DSurface_AddRef((IWineD3DSurface *) object->renderTarget); /* TODO: Depth Stencil support object->stencilBufferTarget = object->depthStencilBuffer; if (NULL != object->stencilBufferTarget) { It comes a bit further now but fails again: trace:d3d:IWineD3DDeviceImpl_SetViewport glDepthRange call ok device.c / 1811 trace:d3d:IWineD3DDeviceImpl_SetViewport This->renderTarget : ((nil)) First chance exception: page fault on read access to 0x0000003c in 32-bit code (0x4658033a). The extra trace was added by me. Again this points to *->renderTarget being NULL. So there must be more to it. Any ideas? Oliver? Cheers, Paul.

2 3

Re: Picasa2- slideshow mode fixes [patch #1]
by Matthew Mastracci 06 Mar '05

06 Mar '05

Woohoo! With my patch from before and some extra hacking, I managed to get the slideshow working in Picasa by faking out some DirectDraw capabilities. Apparently, it requires the presence of the DDCAPS_OVERLAY and DDCAPS_OVERLAYSTRETCH to do its slideshow properly (and aborts silently if they aren't there). Adding the fake capabilities lets the slideshow run, though the images aren't stretched to the entire size of the window unless the slideshow overlay is showing at the bottom. It's hard to describe and a little jarring when you're interacting with the slideshow. Not sure if this is because of the lack of true overlay surface support, but this message seems to indicate so: trace:ddraw:Main_DirectDraw_CreateSurface App didn't request a valid surface type - assuming offscreenplain I don't know if I should submit this as part of a patch, since it doesn't implement the functionality. Here is the change, for those interested: Index: dlls/ddraw/ddraw/user.c =================================================================== RCS file: /home/wine/wine/dlls/ddraw/ddraw/user.c,v retrieving revision 1.22 diff -u -r1.22 user.c --- dlls/ddraw/ddraw/user.c 8 Sep 2004 01:50:37 -0000 1.22 +++ dlls/ddraw/ddraw/user.c 6 Mar 2005 07:04:29 -0000 @@ -143,7 +143,7 @@ #define BLIT_CAPS (DDCAPS_BLT | DDCAPS_BLTCOLORFILL | DDCAPS_BLTDEPTHFILL \ | DDCAPS_BLTSTRETCH | DDCAPS_CANBLTSYSMEM | DDCAPS_CANCLIP \ | DDCAPS_CANCLIPSTRETCHED | DDCAPS_COLORKEY \ - | DDCAPS_COLORKEYHWASSIST) + | DDCAPS_COLORKEYHWASSIST | DDCAPS_OVERLAY | DDCAPS_OVERLAYSTRETCH) #define CKEY_CAPS (DDCKEYCAPS_DESTBLT | DDCKEYCAPS_SRCBLT) #define FX_CAPS (DDFXCAPS_BLTALPHA | DDFXCAPS_BLTMIRRORLEFTRIGHT \ | DDFXCAPS_BLTMIRRORUPDOWN | DDFXCAPS_BLTROTATION90 \ Matthew Mastracci wrote: > I've been trying to get the slideshow mode in Picasa2 to work, but it > seems like it's relying on IDirectDrawSurface->Lock() to return > DDERR_SURFACEBUSY if the surface is already locked. If this call > doesn't return the error, it seems to keep hammering the function > until it does. It then proceeds to hit a stack overflow condition > somehow and dies. :) > > I think it's a bug in Picasa, but my ddraw-surface-lock.diff patch in > wine-patches fixes this behaviour (but not the feature, unfortunately). > > Note that now the slideshow mode now just terminates without error. > I'll keep on looking to see what's happening. > > Matt. >

2 2

Re: Re: [winedbg] Get expressions working again.
by Pouech Eric DMI AEI CAEN 06 Mar '05

06 Mar '05

> Ok, I'll take a look at this, but if where shouldn't > be using two pointer types we should be using > SymGetTypeInfo in all cases? the only (two things) you have to do are: - if one cannot find the pointer type in the pointee's module, then create a winedbg's type for the pointer type and store this type in a lookup table - in type_get_info, extend the types supported by also searching the lookup table for the pointers type. the lookup table should contain pointee type and module of pointee. You could also extend step 1 by not adding a pointer that already exists. A+

3 2

IP Address control broken
by Robert Reif 06 Mar '05

06 Mar '05

I am unable to edit the text in an IP Address control. I can select it and delete it with the mouse but typing into it causes the test to go to the window it was started in. Disabling the control does not change it's color. Is anyone working on this control?

3 7

Re: [winecfg] Audio device selection
by Robert Reif 06 Mar '05

06 Mar '05

Paul van Schayck wrote: >Changelog: >Let the audio device be changed from the audio tab > > OSS also supports multiple devices. When you have multiple audio devices in Windows, it provides the friendly names of all the devices and allows you to select one as the default. You can do this now in OSS.

1 0

re: [OLE #78] Run RPCs on a new thread client side
by Dan Kegel 06 Mar '05

06 Mar '05

Mike Hearn wrote: > Almost there .... Stay on target... [thinking of Skywalker's last bombing run] -- Trying to get a job as a c++ developer? See http://kegel.com/academy/getting-hired.html

1 0

Regression in painting.c
by Ron Jensen 06 Mar '05

06 Mar '05

dlls/user/painting.c is getting caught in an infinite loop in the function update_now() I added two TRACE statements to try to figure out what is going on. Uncommenting the SendMessageW() line allows the loop to exit. See the attached win traces for the working and broken cases. I don't know enough about what this function is trying to do to fix it... The bug can be reproduced using Irfan View's ( www.irfanview.com ) slideshow functions. --- painting.c.orig 2005-02-28 12:01:55.000000000 -0700 +++ painting.c 2005-03-04 08:18:52.000000000 -0700 @@ -340,6 +340,7 @@ else if (rdw_flags & RDW_ALLCHILDREN) flags |= UPDATE_ALLCHILDREN; if (!get_update_flags( hwnd, &child, &flags )) break; +TRACE( "flags = %d, child = %d, prev = %d\n", flags, child , prev ); if (!flags) break; /* nothing more to do */ if (child == prev) /* same window again, didn't get repainted properly */ @@ -351,8 +352,10 @@ TRACE( "%p not repainted properly, erasing\n", child ); if ((hrgn = send_ncpaint( child, NULL, &erase_flags ))) { + TRACE( "hrgn = %p, deleting\n", hrgn ); send_erase( child, erase_flags, hrgn, &rect, NULL ); DeleteObject( hrgn ); +// SendMessageW( child, WM_PAINT, 0, 0 ); } } else

2 1

Re: Better test for visible desktop, should be working on all platforms.
by Dmitry Timoshkov 06 Mar '05

06 Mar '05

"Jakob Eriksson" <jakov(a)vmlinux.org> wrote: > I took this test from the gdi32 conformance tests, since it fails > reliably when run under a non visible desktop, regardless of platform. > > > It's not pretty, but I don't have an NT4 to test with, so I didn't want > to change > anything. Later I can refactor it into exactly the failing part. Did you consider implementing the approach Robert Shearman has pointed out? -- Dmitry.

3 3