- Wine-Devel - list.winehq.org

IP Address control broken
by Robert Reif March 6, 2005

March 6, 2005

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 March 6, 2005

March 6, 2005

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 March 6, 2005

March 6, 2005

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 March 6, 2005

March 6, 2005

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 March 6, 2005

March 6, 2005

"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

winecfg
by Chris Morgan March 6, 2005

March 6, 2005

I noticed that winecfg is still reporting that it "does not modify your configuration", and in winecfg.h WINE_KEY_ROOT isn't set to the configuration key. Is there some reason why we haven't enabled writing to the actual configuration yet? Are there some bugs or lack of features that need to be resolved before it goes active? Without a ~/.wine/config file it seems like winecfg is the only way to let users easily adjust their configuration. Chris

2 1

Regression in GetModuleHandle16
by Vitaliy Margolen March 6, 2005

March 6, 2005

This patch http://cvs.winehq.org/patch.py?id=16245 broke my wine (changes to "winbase.h"). It fails to start. Could someone take a look at it? I'm sorry, my brain failing me this weekend too, and I couldn't find the cause myself. I tried running without ~/.wine - didn't help. I'm running wine from the tree without installing it. Could this be a reason why no one else seeing the problem? 0009:trace:module:load_dll Found loaded module L"c:\\windows\\system\\kernel32.dll" for L"kernel32.dll" at 0x404c0000, count=13 0009:trace:heap:RtlFreeHeap (0x40380000,00000002,40380f78): returning TRUE 0009:trace:module:GetModuleHandle16 (krnl386.exe) 0009:trace:module:MODULE_LoadModule16 Trying built-in 'krnl386.exe' 0009:trace:global:GLOBAL_Alloc 56 flags=0000 0009:trace:heap:RtlAllocateHeap (0x40380000,00000002,00000040): returning 40380f78 0009:trace:module:NE_InitResourceHandler InitResourceHandler[1026] 0009:trace:module:GetModuleHandle16 (KERNEL) 0009:trace:module:GetProcAddress16 0000 'DefResourceHandler' 0009:trace:loaddll:MODULE_LoadModule16 Loaded module "krnl386.exe" : builtin 0009:trace:module:GetModuleHandle16 (krnl386.exe) 0009:err:module:MODULE_LoadModule16 Serious trouble. Just loaded module 'krnl386.exe' (hinst=0x1026), but can't get module handle. Filename too long ? -- Best regards, Vitaliy

1 0

dbghelp problem
by Robert Reif March 6, 2005

March 6, 2005

I can't get a stack trace with symbol information from an exe file and tracked the problem down to this. dbghelp tries to load the symbols for the executable file by looking up the file name without the .exe extension and fails because the file is not found. This patch adds the .exe back on to the file name and tries to look it up again before failing. I doubt this is the right solution but it works for me. --- pe_module.c 2005-03-01 06:20:19.000000000 -0500 +++ /mnt/floppy/pe_module.c 2005-03-04 11:06:30.000000000 -0500 @@ -332,6 +332,7 @@ HANDLE hMap; void* mapping; char loaded_name[MAX_PATH]; + TRACE("(%p %s %p %08lx %08lx)\n", pcs, name, hFile, base, size); loaded_name[0] = '\0'; if (!hFile) @@ -341,13 +342,22 @@ /* FIXME SetLastError */ return NULL; } - if ((hFile = FindExecutableImage(name, NULL, loaded_name)) == NULL) - return NULL; + if ((hFile = FindExecutableImage(name, NULL, loaded_name)) == NULL) { + char full_name[MAX_PATH]; + TRACE("file: %s not found\n", name); + strcpy(full_name, name); + strcat(full_name, ".exe"); + if ((hFile = FindExecutableImage(full_name, NULL, loaded_name)) == NULL) { + TRACE("file: %s not found\n", full_name); + return NULL; + } + } opened = TRUE; } else if (name) strcpy(loaded_name, name); else if (dbghelp_options & SYMOPT_DEFERRED_LOADS) FIXME("Trouble ahead (no module name passed in deferred mode)\n"); + TRACE("loaded_name = %s\n", loaded_name); if (!(module = module_find_by_name(pcs, loaded_name, DMT_PE)) && (hMap = CreateFileMappingA(hFile, NULL, PAGE_READONLY, 0, 0, NULL)) != NULL) {

2 1

Picasa2- slideshow mode fixes [patch #1]
by Matthew Mastracci March 6, 2005

March 6, 2005

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.

1 0

Re: More destinations for wineoss's mixer
by wino＠piments.com March 6, 2005

March 6, 2005

On 05 Mar 2005 12:34:26 -0500, Vincent Béron <vberon(a)mecano.gme.usherb.ca> wrote: > According to MSDN, DST_HEADPHONES can be the same as DST_SPEAKERS, > DST_LINE and DST_VOICEIN as DST_WAVEIN. > > Peter, does this one helps Dragon NS 7? > > Changelog: > Make wineoss's mixer know about DST_HEADPHONES, DST_LINE and > DST_VOICEIN. > Use proper constants instead of 0 and 1. > > Vincent one more thing to add: it seems your patch has done some good since I now notice that entering the audio setup now sets the mike to zero just before it gets stuck. I went back to try NS4 and I did now manage to go through the mike setup routine and sig/noise check. Although the mike setting was not waiting for me to speak before returning. Another brick is in place. 8) -- Opera 7 mail on Linux

1 0