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

Re: Resend : Implements MIIM_BITMAP state in menus
by Alexandre Julliard March 7, 2005

March 7, 2005

Maxime Bellengé <maxime.bellenge(a)wanadoo.fr> writes: > + * drawhbmbitmap : True to draw the hbmbitmap(MIIM_BITMAP)/False to draw the MF_BITMAP > */ > -static void MENU_DrawBitmapItem( HDC hdc, MENUITEM *lpitem, const RECT *rect, BOOL menuBar ) > +static void MENU_DrawBitmapItem( HDC hdc, MENUITEM *lpitem, const RECT *rect, BOOL menuBar, BOOL drawhbmbitmap ) What's the point of passing the drawhbmbitmap flag since you always set it to TRUE? > @@ -762,14 +773,15 @@ > int h = rect->bottom - rect->top; > int bmp_xoffset = 0; > int left, top; > + HBITMAP swtch = (drawhbmbitmap)?lpitem->hbmpItem:(HBITMAP)lpitem->text; That's a strange variable name for a bitmap... > + /* New style MIIM_BITMAP */ > + if (lpitem->hbmpItem) > + { > + if (lpitem->hbmpItem == HBMMENU_CALLBACK) > + { > + MEASUREITEMSTRUCT measItem; > + measItem.CtlType = ODT_MENU; > + measItem.CtlID = 0; > + measItem.itemID = lpitem->wID; > + measItem.itemWidth = lpitem->rect.right - lpitem->rect.left; > + measItem.itemHeight = lpitem->rect.bottom - lpitem->rect.top; > + measItem.itemData = lpitem->dwItemData; > + > + SendMessageW( hwndOwner, WM_MEASUREITEM, lpitem->wID, (LPARAM)&measItem); > + > + /* Keep the size of the bitmap in callback mode to be able to draw it correctly */ > + lppop->bmpSize.cx = max(lppop->bmpSize.cx, measItem.itemWidth - (lpitem->rect.right - lpitem->rect.left)); > + lppop->bmpSize.cy = max(lppop->bmpSize.cy, measItem.itemHeight - (lpitem->rect.bottom - lpitem->rect.top)); Why do you store the bitmap size in the parent popup? What if there are multiple items with bitmaps? -- Alexandre Julliard julliard(a)winehq.org

2 2

Trace from Page Plus 5
by Andrew Neil Ramage March 7, 2005

March 7, 2005

I tried to start Page Plus 5, a DTP program from Serif. I was presented with a dialog asking for me to select a printer. When I OK'ed that, with the splash screen still showing, wine crashed. Here is the trace. [andrew(a)zeus Program]$ wine pp.exe fixme:commdlg:PRINTDLG_OpenDefaultPrinter Could not open printer EPSONStylusC42?! fixme:commdlg:PRINTDLG_OpenDefaultPrinter Could not open printer EPSONStylusC42?! wine: Unhandled exception (thread 0009), starting debugger... WineDbg starting on pid 0x8 Unhandled exception: page fault on read access to 0x00000000 in 32-bit code (0x00447977). In 32 bit mode. Register dump: CS:0073 SS:007b DS:007b ES:007b FS:003b GS:0033 EIP:00447977 ESP:406aeadc EBP:406aeb00 EFLAGS:00210206( - 00 - RIP1) EAX:00000000 EBX:4056c508 ECX:00000002 EDX:406ae940 ESI:40361bfe EDI:00631c67 Stack dump: 0x406aeadc: 00000000 406aec5c 00001012 00000000 0x406aeaec: 00000000 406aec58 00631c67 40361bfe 0x406aeafc: 4056c508 406aecd4 0044988c 00000000 0x406aeb0c: 406aec5c 00000003 406aec58 00631c67 0x406aeb1c: 40361bfe 4056c508 406aeb40 406aeb84 0x406aeb2c: 00000040 406aec34 40045bd3 4020b864 Backtrace: =>1 0x00447977 in pp (+0x47977) (0x406aeb00) 2 0x0044988c in pp (+0x4988c) (0x406aecd4) 3 0x004c804b in pp (+0xc804b) (0x406af370) 4 0x0048a5c6 in pp (+0x8a5c6) (0x406af390) 5 0x0048a2f5 in pp (+0x8a2f5) (0x406af7c4) 6 0x0043acff in pp (+0x3acff) (0x406afe54) 7 0x0043b32f in pp (+0x3b32f) (0x406afe98) 8 0x00631d89 EntryPoint+0x122 in pp (0x406aff2c) 9 0x404d7db8 start_process+0xc8(arg=0x0) [/home/andrew/Download/Software/wine-20050211/dlls/kernel/process.c:1046] in kernel32 (0x406afff4) 10 0x40027dcd wine_pthread_abort_thread+0x2d(status=0x0) [port.c:85] in libwine.so.1 (0x00000000) 0x00447977: movl 0x0(%eax),%eax Wine-dbg> -- Andrew [hackcoughsniffle] Internal Virus Database corrupt

1 0

Re: fix wrc to parse some Borland dialog resources
by Filip Navara March 7, 2005

March 7, 2005

Michael Lin wrote: > * change wrc to parse '&' and '~' when specifying style I believe the attached patch would be better solution...you can use arbitrary expressions in the style definitions with Microsoft Resource Compiler... Regards, Filip

2 1

Wine logo (the glass)
by Dimitrie O. Paun March 7, 2005

March 7, 2005

Folks, do we have an .svg version of the logo? If so, where can I get it from? -- Dimi.

2 1

winebuild changes
by Dimitrie O. Paun March 7, 2005

March 7, 2005

Hi Alexandre, Ivan found that on Windows .exe's can export functions, just like DLLs. He needs this for his work on ntoskrnl.exe. To support this, we need a bit of an interface change to winebuild. I've put together a simple patch to see if you are OK with such a change. If so, I'll finish it, and modify winegcc accordingly. So, thing will change as follows: OLD: winebuild --dll=mystuff.spec NEW: winebuild --dll -E mystuff.spec OLD: winebuild --def=mystuff.spec NEW: winebuild --def -E mystuff.spec OLD: winebuild --exe=myprog.exe NEW: winebuild --exe -F myprog.exe to build an .exe with arbitrary exports, we would do: winebuild --exe -F myprog.exe -E mystuff.spec Index: tools/winebuild/main.c =================================================================== RCS file: /var/cvs/wine/tools/winebuild/main.c,v retrieving revision 1.59 diff -u -r1.59 main.c --- tools/winebuild/main.c 7 Dec 2004 17:19:54 -0000 1.59 +++ tools/winebuild/main.c 7 Mar 2005 01:50:44 -0000 @@ -144,6 +144,7 @@ " -C --source-dir=DIR Look for source files in DIR\n" " -d --delay-lib=LIB Import the specified library in delayed mode\n" " -D SYM Ignored for C flags compatibility\n" +" -E --export=FILE Export the symbols defined in the .spec or .def file\n" " -e --entry=FUNC Set the DLL entry point function (default: DllMain)\n" " -f FLAGS Compiler flags (only -fPIC is supported)\n" " -F --filename=DLLFILE Set the DLL filename (default: from input file name)\n" @@ -165,9 +166,9 @@ " --version Print the version and exit\n" " -w --warnings Turn on warnings\n" "\nMode options:\n" -" --dll=FILE Build a .c file from a .spec or .def file\n" -" --def=FILE.SPEC Build a .def file from a spec file\n" -" --exe=NAME Build a .c file for the named executable\n" +" --dll Build a .c file from a .spec or .def file\n" +" --def Build a .def file from a spec file\n" +" --exe Build a .c file for the named executable\n" " --debug [FILES] Build a .c file with the debug channels declarations\n" " --relay16 Build the 16-bit relay assembly routines\n" " --relay32 Build the 32-bit relay assembly routines\n\n" @@ -204,6 +205,7 @@ /* aliases for short options */ { "source-dir", 1, 0, 'C' }, { "delay-lib", 1, 0, 'd' }, + { "export", 1, 0, 'E' }, { "entry", 1, 0, 'e' }, { "filename", 1, 0, 'F' }, { "help", 0, 0, 'h' }, @@ -248,6 +250,10 @@ case 'D': /* ignored */ break; + case 'E': + spec_file_name = xstrdup( optarg ); + set_dll_file_name( optarg, spec ); + break; case 'F': spec->file_name = xstrdup( optarg ); break; @@ -324,21 +330,12 @@ break; case LONG_OPT_DLL: set_exec_mode( MODE_DLL ); - spec_file_name = xstrdup( optarg ); - set_dll_file_name( optarg, spec ); break; case LONG_OPT_DEF: set_exec_mode( MODE_DEF ); - spec_file_name = xstrdup( optarg ); - set_dll_file_name( optarg, spec ); break; case LONG_OPT_EXE: set_exec_mode( MODE_EXE ); - if ((p = strrchr( optarg, '/' ))) p++; - else p = optarg; - spec->file_name = xmalloc( strlen(p) + 5 ); - strcpy( spec->file_name, p ); - if (!strchr( spec->file_name, '.' )) strcat( spec->file_name, ".exe" ); if (!spec->subsystem) spec->subsystem = IMAGE_SUBSYSTEM_WINDOWS_GUI; break; case LONG_OPT_DEBUG: -- Dimi.

2 1

Where are Wine's webserver logs?
by Scott Ritchie March 7, 2005

March 7, 2005

I'd like to do some datamining around the Wine server logs to see where people are linked to from offsite the most, where they deadend, and such. But, I don't have any logs to mine. Can someone make the past few months logs available? Thanks, Scott Ritchie

2 1

Re: Re: [x11drv] Fix X11DRV_set_wm_hints
by Maxime BELLENGE March 7, 2005

March 7, 2005

> Message du 07/03/05 06:01 > De : "Dmitry Timoshkov" <dmitry(a)baikal.ru> > A : "Maxime Bellengé" <maxime.bellenge(a)wanadoo.fr> > Copie à : wine-devel(a)winehq.com > Objet : Re: [x11drv] Fix X11DRV_set_wm_hints > > "Maxime Bellengé" <maxime.bellenge(a)wanadoo.fr> wrote: > > > I don't agree with you because, if the application don't call this with > > the HTCAPTION parameter > > > > PostMessage( WM_NCLBUTTONDOWN, HTCAPTION, MAKELPARAM( point.x, > > point.y)); > > > > but with HTLEFT, HTRIGHT ... the WM must resize the window I think. > > > > I wondered first if MWM_FUNC_RESIZE should be added but I think it is > > correct. > > If WS_THICKFRAME style is not set, then a window is not supposed to be > resized by WM, that's the point. SetWindowPos still will be able to resize > a window. > > -- > Dmitry. > Ok, I've re-read the docs and you're right. Sorry to have insisted wrongly and thanks for making me notice the problem. I will resend a correct patch. Maxime

1 0

FlashFXP v 3.01 and wine + Odin
by Dietrich Teickner March 7, 2005

March 7, 2005

Currently I have tested FlashFXP v3.01 (ftp-client with ssl) with wine, after I found it chrashs in Odin with stack-overflow. And I found FlashFXP stops after the same actiom in wine. I think that is the same message-loop, that I found in odin. with SysListView 68000034 LVM_SETITEMSTATE in comctl32 and WM_NOTIFY in user32 t01 ( 4): (code=-150) t01 ( 4): USER32: GetWindowLongW 68000034 fffffff4 t01 ( 4): USER32: GetWindowLongW returned 68000034 t01 ( 4): USER32: SendMessageW 68000033 4e 68000034 12f25c t01 ( 5): SendMessageW WM_NOTIFY for 68000033 68000034 12f25c t01 ( 5): USER32: GetWindowThreadProcessId 68000034 12ee9c t01 ( 5): USER32: GetWindowThreadProcessId returned 27470001 t01 ( 5): KERNEL32: GetCurrentProcessId t01 ( 5): KERNEL32: GetCurrentProcessId returned 2747 t01 ( 5): USER32: GetPropA 68000034 f22b t01 ( 5): USER32: GetPropA returned 20035a04 t01 ( 5): USER32: SendMessageA 68000034 102b 0 12eb68 t01 ( 6): SendMessageA User Message 102B (WM_USER+3115) for 68000034 0 12eb68 t01 ( 7): USER32: GetWindowLongW 68000034 0 t01 ( 7): USER32: GetWindowLongW returned 2d45570 t01 ( 7): SysListView 68000034 LVM_SETITEMSTATE 0 12eb68 t01 ( 7): wvsnprintfA : 168D8038 256 "{iItem=%d, iSubItem=%d, " t01 ( 7): {iItem=0, iSubItem=0, t01 ( 7): wvsnprintfA : 168D804E 234 "state=%x, stateMask=%x, " t01 ( 7): state=0, stateMask=2, t01 ( 7): lvItem={iItem=0, iSubItem=0, state=0, stateMask=2} t01 ( 7): wvsnprintfA : 168D8138 256 "{iItem=%d, iSubItem=%d, " t01 ( 7): {iItem=0, iSubItem=0, t01 ( 7): wvsnprintfA : 168D814E 234 "state=%x, stateMask=%x, " t01 ( 7): state=0, stateMask=2, t01 ( 7): (lpLVItem={iItem=0, iSubItem=0, state=0, stateMask=2}, isW=1) t01 ( 7): () t01 ( 7): wvsnprintfA : 168D8238 256 "{iItem=%d, iSubItem=%d, " t01 ( 7): {iItem=0, iSubItem=0, t01 ( 7): wvsnprintfA : 168D824E 234 "state=%x, stateMask=%x, " t01 ( 7): state=0, stateMask=ffffffff, t01 ( 7): wvsnprintfA : 168D826B 205 "lParam=%lx, " t01 ( 7): lParam=0, t01 ( 7): (lpLVItem={iItem=0, iSubItem=0, state=0, stateMask=ffffffff, lParam=0}, isW=1) t01 ( 7): wvsnprintfA : 168D8338 256 "{iItem=%d, iSubItem=%d, " t01 ( 7): {iItem=0, iSubItem=0, t01 ( 7): wvsnprintfA : 168D834E 234 "state=%x, stateMask=%x, " t01 ( 7): state=0, stateMask=f00, t01 ( 7): pdi->item={iItem=0, iSubItem=0, state=0, stateMask=f00} t01 ( 7): (code=-150) the child ask the parent, this ask the child, this ask the parrennt, this..... never end, only if the stack at end. Dietrich

2 1

Re: [x11drv] Fix X11DRV_set_wm_hints
by Dmitry Timoshkov March 7, 2005

March 7, 2005

"Maxime Bellengé" <maxime.bellenge(a)wanadoo.fr> wrote: > This patch adds MWM_FUNC_MOVE | MWM_FUNC_RESIZE when WS_EX_APPWINDOW. > Now the windows are movable. I think that MWM_FUNC_RESIZE should be added only for windows with WS_THICKFRAME style set. -- Dmitry.

2 2

AUTO3STATE with BS_OWNERDRAW bug?
by Michael Lin March 6, 2005

March 6, 2005

Hi all, I am having problem with AUTO3STATE control that have BS_OWNERDRAW style set. Dialogs with this control will not show at all. For example, I have a control AUTO3STATE "Text", 103, 10, 9, 149, 47, BS_OWNERDRAW | WS_CHILD | WS_VISIBLE | WS_TABSTOP in my dialog resource and the dialog won't show when it is called. But when I change it to CONTROL "Text", 103, "BUTTON", BS_OWNERDRAW | WS_CHILD | WS_VISIBLE | WS_TABSTOP, 10, 9, 149, 47 the dialog works fine. This seems like a bug to me, does anyone else have seen this problem? Thanks, Michael

1 0

Config Problems ?
by Andrew Neil Ramage March 6, 2005

March 6, 2005

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 March 6, 2005

March 6, 2005

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 March 6, 2005

March 6, 2005

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 March 6, 2005

March 6, 2005

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 March 6, 2005

March 6, 2005

> 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 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