March 2009 - Wine-devel - list.winehq.org

Re: [1/3]msctf: ThreadMgr sink framework
by Alexandre Julliard 24 Mar '09

24 Mar '09

Aric Stewart <aric(a)codeweavers.com> writes: > typedef struct tagACLMulti { > const ITfThreadMgrVtbl *ThreadMgrVtbl; > const ITfSourceVtbl *SourceVtbl; > LONG refCount; > > ITfDocumentMgr *focus; > + > + /* kept as separate lists to reduce unnecessary iterations */ > + struct list pActiveLanguageProfileNotifySink; > + struct list pDisplayAttributeNotifySink; > + struct list pKeyTraceEventSink; > + struct list pPreservedKeyNotifySink; > + struct list pThreadFocusSink; > + struct list pThreadMgrEventSink; Please pick better names for your variables. Hungarian notation is not a good idea in general, but incorrect Hungarian prefixes are even worse, these are not pointers. -- Alexandre Julliard julliard(a)winehq.org

2 1

Unicode error
by Robert Lunnon 24 Mar '09

24 Mar '09

I get this error under solaris building wine. Any clues would be welcome (Things have probably changed a lot since I last hacked on Wine) make[2]: Entering directory `/export/home/src/wine2004/wine/dlls/kernel32' ../../tools/wrc/wrc --nostdinc -I. -I. -I../../include -I../../include -D__WINESRC__ -D_KERNEL32_ -fokernel. res kernel.rc Source: ¤Q¤ë a4 51 a4 eb Unicode: 5341 6708 Back: ¢Ì¤ë a2 cc a4 eb nls/cht.nls:84:16: Error: String ¤Q¤ë does not convert identically to Unicode and back in codepage 950. Try us ing a Unicode string instead make[2]: *** [kernel.res] Error 1 make[2]: Leaving directory `/export/home/src/wine2004/wine/dlls/kernel32' make[1]: *** [kernel32] Error 2 make[1]: Leaving directory `/export/home/src/wine2004/wine/dlls' make: *** [dlls] Error 2

6 7

WineHQ.org Downtime
by Jeremy Newman 24 Mar '09

24 Mar '09

WineHQ.org (websites [www,appdb,bugzilla,forums,source,test], git, mailing lists) will be off-line tomorrow starting at 10:00 CDT and lasting at least 2 hours. We will be putting up a temp website that will display an off-line message during the duration of the downtime. This is the planned server hardware upgrade. -Newman

2 4

Re: [PATCH 2/2] secur32: Use indices to link the free handles.
by Henri Verbeet 24 Mar '09

24 Mar '09

2009/3/24 Florian Köberle <florian(a)fkoeberle.de>: > Pointers can get invalid when the table gets moved in memory. > This can happen when the table grows. Did you actually see this happen? The table is only supposed to grow when there are no free handles.

1 0

Re: user32/tests: Check effect of magic font height value 0x7fff in dialog template (try 2)
by Vitaliy Margolen 24 Mar '09

24 Mar '09

Koro wrote: > This is the second try for my testcase that shows a discrepancy between > Wine and Windows' handling of dialog templates in the specific case > where the font size member is set to 0x7fff. > > + static unsigned char dlgTemplate[] = > + { > + /* Dialog header */ > + 0x01,0x00, /* Version */ > + 0xff,0xff, /* Extended template marker */ > + 0x00,0x00,0x00,0x00, /* Context Help ID */ > + 0x00,0x00,0x00,0x00, /* Extended style */ > + 0xc0,0x00,0xc8,0x80, /* Style (WS_SYSMENU|WS_CAPTION|WS_POPUP|DS_SETFONT|DS_MODALFRAME) */ Why don't you add this to the resource file? > + hDlg = CreateDialogIndirectParamW(g_hinst, (LPCDLGTEMPLATE)dlgTemplate, NULL, messageBoxFontDlgWinProc, 0); > + > + if (!hDlg) > + { > + todo_wine win_skip("dialog wasn't created\n"); This is wrong. It shouldn't be todo_wine. If you properly put template into resources you can use CreateDialogIndirectParamA Vitaliy.

2 1

Office 2007 installer crash in msi - Backtrace on Solaris
by Robert Lunnon 24 Mar '09

24 Mar '09

The cause might be a Null pointer into strlenW as below ::stack libwine.so.1.0`strlenW+0xf(0, 7bdef160, 7fbf8612, 7bdeef28) kernel32.dll.so`lstrlenW+0x34(0, 0, 1, 0, 80, 0) 0x7da74480(6a8c3648, 0, 0, 7da801d4, 7f09f998, 3) 0x7da65e08(2, 0, 0) msi.dll.so`ACTION_CallDllFunction+0x21c(6a8c090c, 7ffc1598, 7bdef4a8, 7f02f044) msi.dll.so`DllThread+0x79(6a8c090c, 7ffc1598, 7bdef558, 7ffa0a85, 7f02f038, 6a8c090c) 0x7ffa09fe(7f02f038, 6a8c090c, 0, 7ffa0a10) ntdll.dll.so`call_thread_func+0x85(7f02f038, 6a8c090c, 7bdeffd8, 7ffa0c34, 0, 0) ntdll.dll.so`start_thread+0x12d(7da80fb8) libc.so.1`_thr_setup+0x4e(7e770a00) libc.so.1`_lwp_start(7e770a00, 0, 0, 7bdefff8, d0dc5ca0, 7e770a00 ::dis libwine.so.1.0`strlenW+0xf: cmpw $0x0,(%eax) but eax = 0 libwine.so.1.0`strlenW+0x13: je +0xa <libwine.so.1.0`strlenW+0x1d> libwine.so.1.0`strlenW+0x15: leal -0x4(%ebp),%eax libwine.so.1.0`strlenW+0x18: addl $0x2,(%eax) libwine.so.1.0`strlenW+0x1b: jmp -0xf <libwine.so.1.0`strlenW+0xc> libwine.so.1.0`strlenW+0x1d: movl 0x8(%ebp),%edx libwine.so.1.0`strlenW+0x20: movl -0x4(%ebp),%eax libwine.so.1.0`strlenW+0x23: subl %edx,%eax libwine.so.1.0`strlenW+0x25: sarl $0x1,%eax libwine.so.1.0`strlenW+0x27: leave libwine.so.1.0`strlenW+0x28: ret Bob

1 0

Re: How to enable font anti-aliasing in Wine?
by Rick Jones 24 Mar '09

24 Mar '09

> 2009/3/15 Roderick Colenbrander <thunderbird2k at gmx.net>: > > ... A string of registry changes enabled smoothed > fonts for me: > http://forum.winehq.org/viewtopic.php?p=20061&sid=6fbbcf362e44a66b310ef88f6… I was already using FontSmoothing=2 which was some improvement over nothing, but I didn't know these other settings. FontSmoothingType=2 made things much better, the others didn't seem to have any effect. However, I then discovered that with FontSmoothingType=2, a Java app I run in Wine crashes when it opens. :( The interesting item out of the Java stack trace is: Caused by: java.lang.IllegalArgumentException: 0 incompatible with Text-specific LCD contrast key Quite why Java is getting involved at this level I don't really know. Running on Windows with the same font smoothing settings it works fine. Puzzling ...

6 5

Re: [4] [wined3d] trigger constant reloading when RECT textures change
by Henri Verbeet 24 Mar '09

24 Mar '09

2009/3/23 Tobias Jakobi <Liquid.Acid(a)gmx.net>: > + if (!use_ps(stateblock) || (stateblock->wineD3DDevice->shader_backend != &glsl_shader_backend)) { Checking which backend is used is pretty ugly.

1 0

Re: [3] [wined3d] GLSL: implement texrect coord fixup
by Henri Verbeet 24 Mar '09

24 Mar '09

2009/3/23 Tobias Jakobi <Liquid.Acid(a)gmx.net>: > > + if (prog->ps_args.texrect_fixup) { > + for (i = 0; i < MAX_FRAGMENT_SAMPLERS; ++i) { Using something like the following would avoid checking the entire prog->rectFixup_location array when only the first few samplers are used: DWORD fixup = prog->ps_args.texrect_fixup; for (i = 0; fixup; fixup >>= 1, ++i) { if (!(fixup & 1)) continue; ... } > + /* Casting to IWineD3DTextureImpl* does work here, because rectFixup_location[i] > + * only gets initialized (with value != -1) when textures[i] is of type RECT. */ > + const IWineD3DTextureImpl* const tex = (const IWineD3DTextureImpl*)stateBlock->textures[i]; It works, but it's still unecessary. You can access the pow2Matrix through IWineD3DBaseTextureImpl just fine. Also note that * is part of the declarator rather than the base type in C. > + char texrect_fixup[28] = ""; Please don't do that, handle the fixup the same way bias is handled. > + sprintf(name, "PsamplerRectFixup%d", i); i is unsigned here.

1 0

Re: [2/2] quartz: Add a libavformat based MPEG demuxer.
by Chris Robinson 24 Mar '09

24 Mar '09

I wouldn't recommend this patch as being very safe. Having dealt with ffmpeg, my experience has been that it's ABI is not stable (the .so version number increments often, forcing recompilations). And when trying to update to new lib versions, things sometimes mysteriously break. Ultimately, FFmpeg is better off as a "built-in" code base. A project wishing to use it should pick a version and include it, and not rely on it externally.

4 4