September 2017 - Wine-devel

Re: [PATCH v2 1/5] d3d11.idl: Add missing D3D11_DEVICE_CONTEXT_TYPE enums and D3D11_FEATURE_DATA_* declaration
by Henri Verbeet 18 Sep '17

18 Sep '17

On 17 September 2017 at 07:39, Ihsan Akmal <ihsanapps(a)gmail.com> wrote: > +typedef enum D3D11_CONSERVATIVE_RASTERIZATION_TIER { > + D3D11_CONSERVATIVE_RASTERIZATION_NOT_SUPPORTED = 0, > + D3D11_CONSERVATIVE_RASTERIZATION_TIER_1 = 1, > + D3D11_CONSERVATIVE_RASTERIZATION_TIER_2 = 2, > + D3D11_CONSERVATIVE_RASTERIZATION_TIER_3 = 3, > +} D3D11_CONSERVATIVE_RASTERIZATION_TIER; > + > + > +typedef struct D3D11_FEATURE_DATA_D3D11_OPTIONS2 > +{ > + BOOL PSSpecifiedStencilRefSupported; > + BOOL TypedUAVLoadAdditionalFormats; > + BOOL ROVsSupported; > + D3D11_CONSERVATIVE_RASTERIZATION_TIER ConservativeRasterizationTier; > + D3D11_TILED_RESOURCES_TIER TiledResourcesTier; > + BOOL MapOnDefaultTextures; > + BOOL StandardSwizzle; > + BOOL UnifiedMemoryArchitecture; > +} D3D11_FEATURE_DATA_D3D11_OPTIONS2; > + > + The formatting is inconsistent here. > +typedef enum D3D11_SHADER_CACHE_SUPPORT_FLAGS > +{ > + D3D11_SHADER_CACHE_SUPPORT_NONE = 0x0, > + D3D11_SHADER_CACHE_SUPPORT_AUTOMATIC_INPROC_CACHE = 0x1, > + D3D11_SHADER_CACHE_SUPPORT_AUTOMATIC_DISK_CACHE = 0x2, > +} D3D11_SHADER_CACHE_SUPPORT_FLAGS; > + > +typedef struct D3D11_FEATURE_DATA_SHADER_CACHE > +{ > + UINT SupportFlags; > +} D3D11_FEATURE_DATA_SHADER_CACHE; Should "SupportFlags" be of type D3D11_SHADER_CACHE_SUPPORT_FLAGS?

2 3

Re: [PATCH v2 3/5] include: Add dxgi1_3.idl
by Henri Verbeet 17 Sep '17

17 Sep '17

On 17 September 2017 at 07:39, Ihsan Akmal <ihsanapps(a)gmail.com> wrote: > +typedef struct DXGI_MATRIX_3X2_F > +{ > + FLOAT _11; > + FLOAT _12; > + FLOAT _21; > + FLOAT _22; > + FLOAT _31; > + FLOAT _32; > +} DXGI_MATRIX_3X2_F; I'd prefer these as "float", although it doesn't really matter. > +typedef enum DXGI_FRAME_PRESENTATION_MODE > +{ > + DXGI_FRAME_PRESENTATION_MODE_COMPOSED»····· = 0, > + DXGI_FRAME_PRESENTATION_MODE_OVERLAY»······ = 1, > + DXGI_FRAME_PRESENTATION_MODE_NONE»· = 2, > + DXGI_FRAME_PRESENTATION_MODE_COMPOSITION_FAILURE»··= 3 > +} DXGI_FRAME_PRESENTATION_MODE; You have some stray tabs there. We generally prefer adding a comma after the last enum element as well, it makes adding elements a little easier. That's pretty minor of course. Applies to DXGI_OVERLAY_SUPPORT_FLAG as well. > +interface IDXGIDevice3 : IDXGIDevice2 > +{ > + void Trim(); > +}; Superfluous semicolon. Applies to all the other interfaces as well. > \ No newline at end of file Alexandre fixes these when committing patches, but please end source files with a newline.

1 0

Re: [PATCH v2 2/5] dxgi1_2.idl: Add missing structs and interfaces
by Henri Verbeet 17 Sep '17

17 Sep '17

On 17 September 2017 at 07:39, Ihsan Akmal <ihsanapps(a)gmail.com> wrote: > +interface IDXGIOutputDuplication : IDXGIObject > +{ ... > +}; The semicolon is superfluous. Applies to the IDXGIDisplayControl and IDXGIOutput1 interfaces as well.

1 0

Re: [PATCH] includes: CHOOSECOLOR.lCustData is an LPARAM.
by Nikolay Sivov 17 Sep '17

17 Sep '17

On 17.09.2017 11:48, Stefan Dösinger wrote: > This matches the PSDK and MinGW headers. > > Signed-off-by: Stefan Dösinger <stefan(a)codeweavers.com> > --- > include/commdlg.h | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/include/commdlg.h b/include/commdlg.h > index df56c143c0..eeefce2bfe 100644 > --- a/include/commdlg.h > +++ b/include/commdlg.h > @@ -183,7 +183,7 @@ typedef struct { > DWORD rgbResult; > LPDWORD lpCustColors; > DWORD Flags; > - DWORD lCustData; > + LPARAM lCustData; > LPCCHOOKPROC lpfnHook; > LPCSTR lpTemplateName; > } CHOOSECOLORA; > @@ -196,7 +196,7 @@ typedef struct { > DWORD rgbResult; > LPDWORD lpCustColors; > DWORD Flags; > - DWORD lCustData; > + LPARAM lCustData; > LPCCHOOKPROC lpfnHook; > LPCWSTR lpTemplateName; > } CHOOSECOLORW; > Another difference, not near as important, is that SDK and mingw are using COLORREFs for color fields.

1 0

Re: [PATCH 1/5] d3d11.idl: Add missing D3D11_DEVICE_CONTEXT_TYPE enums and D3D11_FEATURE_DATA_* declaration
by Alex Henrie 17 Sep '17

17 Sep '17

Hello Ihsan, and thank you for your interest in contributing to Wine! Please note that when submitting patches, you must use your real name in the From and Signed-off-by lines. You can find this and other helpful instructions at <https://wiki.winehq.org/Submitting_Patches>. -Alex

1 0

Re: [PATCH v2 1/3] mshtml: Add a null parameter check in navigate_new_window().
by Jacek Caban 15 Sep '17

15 Sep '17

On 15.09.2017 07:34, Zebediah Figura wrote: > v2: check in caller instead of do_query_service(). > Signed-off-by: Zebediah Figura <zfigura(a)codeweavers.com> > --- > dlls/mshtml/navigate.c | 6 ++++++ > 1 file changed, 6 insertions(+) > > diff --git a/dlls/mshtml/navigate.c b/dlls/mshtml/navigate.c > index 423d650a8c..06425c3f2b 100644 > --- a/dlls/mshtml/navigate.c > +++ b/dlls/mshtml/navigate.c > @@ -2246,6 +2246,12 @@ HRESULT navigate_new_window(HTMLOuterWindow *window, IUri *uri, const WCHAR *nam > nsChannelBSC *bsc; > HRESULT hres; > > + if (!window->doc_obj->client) > + { > + FIXME("no attached client site; cannot query INewWindowManager\n"); > + return E_NOTIMPL; > + } This doesn't look right. Look at HlinkTarget_Navigate. We explicitly choose this code path because client is NULL. This seems to be a regression from commit 63205bf01d6 and the right solution seems to be simply ignoring the fact that we don't have INewWindowManager and continue. Thanks, Jacek

2 1

Re: [PATCH v4 2/2] ieframe: Add stub implementation of InternetExplorerManager.
by Jacek Caban 15 Sep '17

15 Sep '17

On 15.09.2017 07:43, Zebediah Figura wrote: > +HRESULT WINAPI InternetExplorerManager_Create(IClassFactory *iface, IUnknown *pOuter, REFIID riid, void **ppv) > +{ > + InternetExplorerManager *ret; > + HRESULT hr; > + > + TRACE("(%p %s %p)\n", pOuter, debugstr_guid(riid), ppv); > + > + if (!(ret = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, sizeof(*ret)))) > + return E_OUTOFMEMORY; > + > + ret->IInternetExplorerManager_iface.lpVtbl = &InternetExplorerManager_vtbl; > + ret->ref = 1; > + > + hr = IInternetExplorerManager_QueryInterface(&ret->IInternetExplorerManager_iface, riid, ppv); > + IInternetExplorerManager_Release(&ret->IInternetExplorerManager_iface); > + > + return hres; That doesn't compile. Jacek

2 1

Re: [PATCH] dsquery: Register CommonQuery coclass.
by Alexandre Julliard 15 Sep '17

15 Sep '17

Zebediah Figura <zfigura(a)codeweavers.com> writes: > + > +[ > + threading(apartment), > + uuid(83bc5ec0-6f2a-11d0-a1c4-00aa00c16e65) > +] > +coclass CommonQuery { interface ICommonQuery; } This should most likely go in the public header. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH] d3d9: Ask for texture GetDC() capability GET_DC based on texture format instead of initial swapchain buffer format
by Henri Verbeet 15 Sep '17

15 Sep '17

On 15 September 2017 at 07:14, Nikolay Sivov <nsivov(a)codeweavers.com> wrote: > Initially application sets buffer format to UNKNOWN that is resolved later > in wined3d to something compatible with GetDC() functionality. > A small test would be great.

1 0

Re: [PATCH 2/2] d2d1: Implement GetBounds() for path geometries
by Henri Verbeet 15 Sep '17

15 Sep '17

On 14 September 2017 at 14:10, Nikolay Sivov <nsivov(a)codeweavers.com> wrote: > + /* GetBounds() with bezier segments. */ > + hr = ID2D1Factory_CreatePathGeometry(factory, &geometry); > + ok(SUCCEEDED(hr), "Failed to create path geometry, hr %#x.\n", hr); > + hr = ID2D1PathGeometry_Open(geometry, &sink); > + ok(SUCCEEDED(hr), "Failed to open geometry sink, hr %#x.\n", hr); > + fill_geometry_sink_bezier(sink); > + hr = ID2D1GeometrySink_Close(sink); > + ok(SUCCEEDED(hr), "Failed to close geometry sink, hr %#x.\n", hr); > + ID2D1GeometrySink_Release(sink); > + > + set_rect(&rect, 0.0f, 0.0f, 0.0f, 0.0f); > + hr = ID2D1PathGeometry_GetBounds(geometry, NULL, &rect); > + ok(SUCCEEDED(hr), "Failed to get geometry bounds, hr %#x.\n", hr); > + match = compare_rect(&rect, 5.0f, 20.0f, 75.0f, 752.0f, 0); > + ok(match, "Got unexpected rectangle {%.8e, %.8e, %.8e, %.8e}.\n", > + rect.left, rect.top, rect.right, rect.bottom); Note that the way fill_geometry_sink_bezier() is constructed, the bounds tend to coincide with the control points, probably even under a carefully chosen rotation. The figure from test_bezier_intersect() under various rotations may be more interesting.

1 0