June 2009 - Wine-Devel - list.winehq.org

Re: Remove division from inner loop
by Henri Verbeet June 17, 2009

June 17, 2009

2009/6/16 Michael Abbott <michael(a)araneidae.co.uk>: > - /* The depth data is normalized, so needs to be > scaled, the stencil data isn't. */ Looks like the patch got wrapped. You'll probably want to just attach the output of git format-patch.

2 1

Re: comdlg32/tests: fix file dialog resize failures on NT4 and with tiny desk tops.
by Paul Vriens June 17, 2009

June 17, 2009

Rein Klazes wrote: > --- > dlls/comdlg32/tests/filedlg.c | 10 +++++++--- > dlls/comdlg32/tests/rsrc.rc | 16 ++++++++++++++++ > 2 files changed, 23 insertions(+), 3 deletions(-) > Thanks Rein, This fixes the failures on my boxes. -- Cheers, Paul.

1 0

RFC: dmime: Implement the DirectMusicTool8 class factory.
by Francois Gouget June 16, 2009

June 16, 2009

--- So all the code in dlls/dmime/tool.c was unused because nothing was calling DMUSIC_CreateDirectMusicTool8Impl(). Apparently the intent was to call it from a class factory so I looked around and came up with the following patch that implements a class factory for DirectMusicTool8. However I have never implemented a class factory before and don't know much about them. So does this patch make sense? Does the fact that we have an IID_IDirectMusicTool8 instead of a CLSID_XXX hint at something I'm missing? How come this was not done when the file was first added? That said IDirectMusicTool8Impl looks like one big stub. So if it makes more sense to remove it entirely I can do that too. dlls/dmime/dmime_main.c | 56 +++++++++++++++++++++++++++++++++++++++++++- dlls/dmime/dmime_private.h | 2 +- dlls/dmime/tool.c | 2 +- 3 files changed, 57 insertions(+), 3 deletions(-) diff --git a/dlls/dmime/dmime_main.c b/dlls/dmime/dmime_main.c index f4abddd..8bd95a9 100644 --- a/dlls/dmime/dmime_main.c +++ b/dlls/dmime/dmime_main.c @@ -729,6 +729,56 @@ static const IClassFactoryVtbl WaveTrackCF_Vtbl = { static IClassFactoryImpl WaveTrack_CF = {&WaveTrackCF_Vtbl}; /****************************************************************** + * DirectMusicTool8 ClassFactory + */ +static HRESULT WINAPI Tool8CF_QueryInterface(LPCLASSFACTORY iface,REFIID riid,LPVOID *ppobj) { + FIXME("- no interface\n\tIID:\t%s\n", debugstr_guid(riid)); + + if (ppobj == NULL) return E_POINTER; + + return E_NOINTERFACE; +} + +static ULONG WINAPI Tool8CF_AddRef(LPCLASSFACTORY iface) { + DMIME_LockModule(); + + return 2; /* non-heap based object */ +} + +static ULONG WINAPI Tool8CF_Release(LPCLASSFACTORY iface) { + DMIME_UnlockModule(); + + return 1; /* non-heap based object */ +} + +static HRESULT WINAPI Tool8CF_CreateInstance(LPCLASSFACTORY iface, LPUNKNOWN pOuter, REFIID riid, LPVOID *ppobj) { + TRACE ("(%p, %s, %p)\n", pOuter, debugstr_dmguid(riid), ppobj); + + return DMUSIC_CreateDirectMusicTool8Impl (riid, ppobj, pOuter); +} + +static HRESULT WINAPI Tool8CF_LockServer(LPCLASSFACTORY iface,BOOL dolock) { + TRACE("(%d)\n", dolock); + + if (dolock) + DMIME_LockModule(); + else + DMIME_UnlockModule(); + + return S_OK; +} + +static const IClassFactoryVtbl Tool8CF_Vtbl = { + Tool8CF_QueryInterface, + Tool8CF_AddRef, + Tool8CF_Release, + Tool8CF_CreateInstance, + Tool8CF_LockServer +}; + +static IClassFactoryImpl Tool8_CF = {&Tool8CF_Vtbl}; + +/****************************************************************** * DllMain * * @@ -821,7 +871,11 @@ HRESULT WINAPI DllGetClassObject(REFCLSID rclsid, REFIID riid, LPVOID *ppv) *ppv = &WaveTrack_CF; IClassFactory_AddRef((IClassFactory*)*ppv); return S_OK; - } + } else if (IsEqualCLSID (rclsid, &IID_IDirectMusicTool8) && IsEqualIID (riid, &IID_IClassFactory)) { + *ppv = &Tool8_CF; + IClassFactory_AddRef((IClassFactory*)*ppv); + return S_OK; + } WARN("(%s, %s, %p): no interface found.\n", debugstr_dmguid(rclsid), debugstr_dmguid(riid), ppv); return CLASS_E_CLASSNOTAVAILABLE; diff --git a/dlls/dmime/dmime_private.h b/dlls/dmime/dmime_private.h index 308c2c8..d21061d 100644 --- a/dlls/dmime/dmime_private.h +++ b/dlls/dmime/dmime_private.h @@ -72,7 +72,7 @@ extern HRESULT WINAPI DMUSIC_CreateDirectMusicSegmentImpl (LPCGUID lpcGUID, LPVO extern HRESULT WINAPI DMUSIC_CreateDirectMusicSegmentStateImpl (LPCGUID lpcGUID, LPVOID *ppobj, LPUNKNOWN pUnkOuter); extern HRESULT WINAPI DMUSIC_CreateDirectMusicGraphImpl (LPCGUID lpcGUID, LPVOID *ppobj, LPUNKNOWN pUnkOuter); extern HRESULT WINAPI DMUSIC_CreateDirectMusicAudioPathImpl (LPCGUID lpcGUID, LPVOID *ppobj, LPUNKNOWN pUnkOuter); -extern HRESULT WINAPI DMUSIC_CreateDirectMusicToolImpl (LPCGUID lpcGUID, LPVOID *ppobj, LPUNKNOWN pUnkOuter); +extern HRESULT WINAPI DMUSIC_CreateDirectMusicTool8Impl (LPCGUID lpcGUID, LPVOID *ppobj, LPUNKNOWN pUnkOuter); extern HRESULT WINAPI DMUSIC_CreateDirectMusicPatternTrackImpl (LPCGUID lpcGUID, LPVOID *ppobj, LPUNKNOWN pUnkOuter); extern HRESULT WINAPI DMUSIC_CreateDirectMusicLyricsTrack (LPCGUID lpcGUID, LPVOID* ppobj, LPUNKNOWN pUnkOuter); diff --git a/dlls/dmime/tool.c b/dlls/dmime/tool.c index 1654bb8..81f4bca 100644 --- a/dlls/dmime/tool.c +++ b/dlls/dmime/tool.c @@ -119,7 +119,7 @@ static const IDirectMusicTool8Vtbl DirectMusicTool8_Vtbl = { }; /* for ClassFactory */ -HRESULT WINAPI DMUSIC_CreateDirectMusicobjImpl (LPCGUID lpcGUID, LPVOID *ppobj, LPUNKNOWN pUnkOuter) { +HRESULT WINAPI DMUSIC_CreateDirectMusicTool8Impl(LPCGUID lpcGUID, LPVOID *ppobj, LPUNKNOWN pUnkOuter) { IDirectMusicTool8Impl* obj; obj = HeapAlloc (GetProcessHeap(), HEAP_ZERO_MEMORY, sizeof(IDirectMusicTool8Impl)); -- 1.6.3.1

2 3

Re: comdlg32: Delay setting the initial size of resizable file dialog until the CDN_INITDONE notification has been processed.
by Paul Vriens June 16, 2009

June 16, 2009

Rein Klazes wrote: > With tests. Fixes bug #18516 > --- > dlls/comdlg32/filedlg.c | 6 +- > dlls/comdlg32/tests/filedlg.c | 126 +++++++++++++++++++++++++++++++++++++++++ > 2 files changed, 130 insertions(+), 2 deletions(-) > Hi Rein, This is introducing test failures on at least NT4. (Just as I was down to zero :( ). Could you have a look? -- Cheers, Paul.

3 4

Request for Patchreview
by André Hentschel June 16, 2009

June 16, 2009

Can someone please review my patch http://www.winehq.org/pipermail/wine-patches/2009-June/074195.html. Its about the Installer-Button in Appwiz.cpl Thanks Best Regards André Hentschel

4 8

How to force use of DirectX8 instead of DirectX9?
by Joerg-Cyril.Hoehle＠t-systems.com June 16, 2009

June 16, 2009

Hi, What is the recommended way not to use Wine's DirectX9 and fall back to DirectX8 instead? Some applications list DirectX 8.1 as their minimal graphics requirement, yet they come with DirectX9 on the CD and presumably use Wine's DirectX9. The author of the screenshot at http://www.adventure-treff.de/images/bild.php?bild=/features/ankh_grafikver… shows how graphics differ a lot for Ankh, depending on whether DirectX 8 or 9 is being used in MS-Windows. Perhaps forcing use of DirectX8 can help make some applications run? (Or help locate bugs) Thanks for your help, Jörg Höhle

5 4

Re: wintrust: Implement WVTAsn1SpcSpOpusInfoDecode (resend)
by Alexandre Julliard June 16, 2009

June 16, 2009

Juan Lang <juan.lang(a)gmail.com> writes: > No difference, just a rebase. Please let me know if I'm missing > something obvious. Fixes bug 17696. It doesn't work here: ../../../tools/runtest -q -P wine -M wintrust.dll -T ../../.. -p wintrust_test.exe.so asn.c && touch asn.ok asn.c:888: Test failed: CryptDecodeObjectEx failed: c0000005 make[2]: *** [asn.ok] Error 1 -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: (resend) kernel32: Fix error returns from CreateFile (at least the simple ones).
by Alexandre Julliard June 16, 2009

June 16, 2009

Guy Albertelli <galberte(a)neo.rr.com> writes: > Changelog: > - CreateFile[AW] should return ERROR_PATH_NOT_FOUND when > STATUS_FILE_IS_A_DIRECTORY is returned from NtCreateFile and > the requested filename has a trailing \ This breaks the tests: ../../../tools/runtest -q -P wine -M kernel32.dll -T ../../.. -p kernel32_test.exe.so profile.c && touch profile.ok profile.c:968: Test failed: Expected ERROR_ACCESS_DENIED, got 3 make[2]: *** [profile.ok] Error 1 -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [1/4] [wined3d] move creation of NP2 fixup uniforms into separate loop
by Henri Verbeet June 16, 2009

June 16, 2009

2009/6/15 Tobias Jakobi <Liquid.Acid(a)gmx.net>: > Why do you need this to be in a separate loop? Also, could you please fix your mailer to set a proper Content-Type, instead of application/octet-stream?

3 6

Re: wininet: Always create a URL cache entry.
by Jacek Caban June 16, 2009

June 16, 2009

Hi Hans, Hans Leidekker wrote: > Fixes file downloads in ie7. > > -Hans > > diff --git a/dlls/wininet/http.c b/dlls/wininet/http.c > index 8f9f404..2721427 100644 > --- a/dlls/wininet/http.c > +++ b/dlls/wininet/http.c > @@ -3807,8 +3807,7 @@ BOOL WINAPI HTTP_HttpSendRequestW(LPWININETHTTPREQW lpwhr, LPCWSTR lpszHeaders, > } > while (loop_next); > > - /* FIXME: Better check, when we have to create the cache file */ > - if(bSuccess && (lpwhr->hdr.dwFlags & INTERNET_FLAG_NEED_FILE)) { > + if(bSuccess) { > I think we should check at least INTERNET_FLAG_NO_CACHE_WRITE flag here. Jacek

2 1

Re: [4/5] WineD3D: Manage SFLAG_CLIENT separately for RGB and SRGB textures
by Henri Verbeet June 16, 2009

June 16, 2009

2009/6/14 Stefan Dösinger <stefan(a)codeweavers.com>: > +static void surface_allocate_surface(IWineD3DSurfaceImpl *This, GLenum internal, GLsizei width, GLsizei height, GLenum format, GLenum type, > + DWORD client_store_flag) { I think that's ugly. More importantly, I don't see how client storage can be enabled for one texture but not the other.

2 5

Re: [2/4] [wined3d] add ps_np2fixup_info_t structure
by Stefan Dösinger June 16, 2009

June 16, 2009

Am Montag, 15. Juni 2009 20:23:05 schrieb Tobias Jakobi: >+ * This structure is shared between the GLSL and the ARB backend.*/ > +typedef struct ps_np2fixup_info_s { > + WORD idx[MAX_FRAGMENT_SAMPLERS]; /* indices to the real constant */ > + WORD active; /* bitfield indicating if we can apply the fixup */ > + WORD num_consts; > +} ps_np2fixup_info_t; Please don't use a typedef, just use "struct ps_np2_fixup_info" in the code. We stopped using these typedefs a while ago and are removing them as we go. Other than that, the backend specific compiled shader structs and per-shader private data make it easier to use backend specific structures which use the correct data types. I'm not sure if I remember correctly, but I think the WORD was not entirely correct for ARB, it would rather use a GLuint there(and GLSL is happy with a smaller unsigned char)

1 0

Re: wined3d: Reserve two varyings for ATI DX10 cards
by Henri Verbeet June 15, 2009

June 15, 2009

2009/6/15 Robert Key <hpestilence(a)gmail.com>: > Changelog: > * This reserves 2 varyings for gl_FrontColor and gl_FrontSecondaryColor > for ATI DX10 cards. This should fix bug #17437 once the new fglrx > drivers get released this week. > That doesn't look quite right. gl_FrontColor and gl_FrontSecondaryColor should only be used as a fallback when we're already out of regular varyings to use. What does the shader that triggers this look like?

2 3

Trouble cloning git via http
by Juan Lang June 15, 2009

June 15, 2009

Anyone else cloning git via http? I'm behind an annoying proxy server, and when I try to clone, I get: error: Unable to find afce86b4bc6198575bbe6046d12fd77efe01cd46 under http://source.winehq.org/git/wine.git Cannot obtain needed object afce86b4bc6198575bbe6046d12fd77efe01cd46 Just curious if anyone else has seen this or knows what I could try. Thanks, --Juan

2 2

Wine Direct Input Issue
by MD.IMAM HOSSAIN June 15, 2009

June 15, 2009

Dear Wine Developers, The latest WINE 1.1.23 has a small regression about Direct Input. The game HITMAN 3 CONTRACTS has a problem with mouse clicking. By WINE 1.1.23 and HITMAN 3 CONTRACTS mouse clicking does not work. It is worked fine with wine 1.1.11. Since the game is a third person shooter it is not playable anymore. One final thing, latest WINE has broadened Linux Gaming Field. It now supports many old games then before which did not run before. I have tested these games: 1. Knight Rider 2 [ fully working ]. 2. Turok 2 [ fully working ]. 3. Need For Speed 4 [ fully working ]. 4. Driver [ fully working ]. 5. London Racer 2 [ fully working ]. 6. Midtown Madness 2 [ fully working ]. Thank you very much for great work. Your sincerely, MD. IMAM HOSSAIN

2 1

re: Ensure that GetSystemMenu returns valid HMENU unless bRevert specified
by Dan Kegel June 15, 2009

June 15, 2009

Brent wrote: >This was causing some apps to crash with invalid read from address 0x4. Which apps? Is there a bug report for this? - Dan

2 1

Duplicates page: Permission denied
by Ken Sharp June 15, 2009

June 15, 2009

Since the upgrade of Bugzilla, the duplicates page reports permission denied. http://bugs.winehq.org/duplicates.cgi Can anyone with the relevant access check the permissions? This is http://bugs.winehq.org/show_bug.cgi?id=18665

2 1

Re: winapi: Fix a $blevel / $plevel mismatch.
by Francois Gouget June 15, 2009

June 15, 2009

Ignore this one, I patched the wrong place. On Mon, 15 Jun 2009, Francois Gouget wrote: [...] > } elsif(s/^\}//) { > $blevel--; > - if($blevel <= 0) { > + if($plevel <= 0) { > $self->_parse_c_error($_, $line, $column, "file", "} without {"); -- Francois Gouget <fgouget(a)free.fr> http://fgouget.free.fr/ Research is the transformation of money to knowledge. Innovation is the transformation of knowledge to money. -- Dr. Hans Meixner

1 0

Re: [1/5] WineD3D: Unroll loops in ARB if needed (try 2)
by Henri Verbeet June 15, 2009

June 15, 2009

2009/6/14 Stefan Dösinger <stefan(a)codeweavers.com>: > A few fixes from the first version: > > *) Copypaste index issues fixed > *) aL is incremeted at the end of the loop, not at the beginning > *) emulated c[aL + offset] indexing works > > The loop codepath is now tested as well, not just the rep stuff - I have a > test for aL constant indexing, its deeper down in my send queue > Since I replied to the original version of this patch... Why does get_int_const() (still) return a pointer to a static int array?

1 0

Request for comments on patch
by Erich Hoover June 15, 2009

June 15, 2009

In an effort to come up with a (hopefully palatable) solution to a couple of bugs I've put together the attached patch. Only the first of the two issues is currently handled, though the patch paves the way for the second issue. 1) Bug #15839: (not this title) Wine does not report the manufacturer driver (for example, "nv4_disp.dll" for nVidia users) 2) Wine does not always report the correct amount of video RAM (not sure if there is a bug number for this) The approach handles graphics requests like this: "manufacturer" driver --> "OS" driver (normally winex11.drv) Where the manufacturer driver (at present) forwards all requests to the OS driver. Later additions could provide driver-specific code for returning information such as the correct amount of video RAM. Also, the manufacturer driver is skipped (reverting to the current behavior) unless "HKEY_CURRENT_USER\Software\Wine\Drivers\ReportNativeVideoDriver" is set to "Y". Note: patch makes changes to configure.ac, so cover your bases: autoheader && autoconf && ./configure && make depend && make I'm particularly interested in comments related to how the driver is selected (__wine_get_video_driver in gdi32), but all feedback is welcome. Erich Hoover ehoover(a)mines.edu

5 16

Unused code in dmime.dll
by Francois Gouget June 14, 2009

June 14, 2009

So DMUSIC_CreateDirectMusicobjImpl() is unused. What's troubling about that is that, if my analysis is correct, it means that none of the code in dlls/dmime/tool.c is used. It seems like all that's missing is something that would let applications actually get at the IDirectMusicTool8 implementation. Could someone familiar with DCOM and/or dmime look into it? It would be a shame to strip all that unused code :-( -- Francois Gouget <fgouget(a)free.fr> http://fgouget.free.fr/ La terre est une bêta...

1 0

Command for creating git patches - possible discrepancy in Wine doc pages
by Frédéric Delanoy June 14, 2009

June 14, 2009

Hi, I'm about to generate my first wine patches and saw 2 pages in the Wine wiki giving two different commands : - on http://wiki.winehq.org/GitWine , section 3.6. Generating a patchset to submit, it says git format-patch --stdout --keep-subject origin - on http://www.winehq.org/site/sending_patches - Generating a patch for WineHQ with Git, it says git format-patch --stdout --keep-subject --attach origin (note the additional "--attach"). Being a fresh (couple of days) git user, I just wonder : which command should I use? Frédéric Delanoy

2 1

dotests script
by chris ahrendt June 14, 2009

June 14, 2009

Does anyone have the dotests script they could send me ... I moved to a new machine and lost the script. Went to the wiki and it fails to find it when I click on the link. I get an error stating that I am unable to add an attachment (even after logging in). Chris

3 3

Re: [PATCH] wined3d: mirror d3dresourcetype
by Stefan Dösinger June 14, 2009

June 14, 2009

Am Sonntag, 14. Juni 2009 14:26:53 schrieb Marcus Meissner: > Hi, > > gcc 4.5trunk warned that we do not handle case 7 (D3DRTYPE_INDEXBUFFER), > so add it here too. There's a reason why this doesn't exist - WineD3D has index and vertex buffers merged, so there is no such thing as a specialized index buffer. Where does this warning occur? The code that maps D3DRTYPE to WINED3DRTYPE or vice versa should be fixed to map INDEXBUFFER to BUFFER, and take care for the reverse conversion

4 5

Re: [5/5] WineD3D: Use flags instead of bools to control shader_hw_sample
by Stefan Dösinger June 14, 2009

June 14, 2009

Am Sonntag, 14. Juni 2009 18:34:13 schrieb Tobias Jakobi: > Why not use an enum instead of the defines? Because as far as I can see the other places that deal with flags use defines to define them, not an enum type. Specifically, I looked at the WINED3D_GLSL_SAMPLE_* flags, which have the equivalent job in glsl_shader.c Besides that, using an enum for OR-able flags seems not quite right to me. If you have something like enum flags { flag1 = 1; flag2 = 2; flag3 = 4; } "flag1 | flag3" gives a value(=5) that does not exist in the enum.

1 0