June 2022 - Wine-Devel - list.winehq.org

Re: [PATCH v2 0/3] MR216: uiautomationcore: More UiaProviderFromIAccessible work.

by Connor McAdams (＠cmcadams)

On Fri Jun 10 07:31:21 2022 +0000, Huw Davies wrote: > Err, obviously comparing `!name[0]` directly with `!name[1]` would also work. Yeah, since wcscmp doesn't handle NULL inputs, I was basically trying to do what lstrcmpW does in a single if statement, i.e: ``` if (!str1 && !str2) return 0; if (!str1) return -1; if (!str2) return 1; ``` but splitting it out into multiple if statement would definitely make things clearer. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/216#note_1831

June 10, 2022

Re: [PATCH v3 0/6] MR217: windows.media.speech: ISpeechSynthesisStream stub implementation.

by Bernhard Kölbl

On Fri Jun 10 12:42:13 2022 +0000, Rémi Bernon wrote: > Shouldn't we return an empty vector view instead? Either create the > vector on the fly, or probably better in the stream constructor and > return a view to it here? Yup, that's probably a better idea than this. I think returning NULL is actually not good anyway. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/217#note_1830

June 10, 2022

Re: [PATCH v3 0/6] MR217: windows.media.speech: ISpeechSynthesisStream stub implementation.

by Rémi Bernon

Rémi Bernon (@rbernon) commented about dlls/windows.media.speech/synthesizer.c: > +HRESULT WINAPI synthesis_stream_GetRuntimeClassName( ISpeechSynthesisStream *iface, HSTRING *class_name ) > +{ > + FIXME("iface %p, class_name %p stub.\n", iface, class_name); > + return E_NOTIMPL; > +} > + > +HRESULT WINAPI synthesis_stream_GetTrustLevel( ISpeechSynthesisStream *iface, TrustLevel *trust_level ) > +{ > + FIXME("iface %p, trust_level %p stub.\n", iface, trust_level); > + return E_NOTIMPL; > +} > + > +HRESULT WINAPI synthesis_stream_get_Markers( ISpeechSynthesisStream *iface, IVectorView_IMediaMarker **value ) > +{ > + FIXME("iface %p, value %p stub!\n", iface, value); > + *value = NULL; Shouldn't we return an empty vector view instead? Either create the vector on the fly, or more probably better in the stream constructor and return a view to it here? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/217#note_1829

June 10, 2022

Re: [PATCH v3 0/6] MR217: windows.media.speech: ISpeechSynthesisStream stub implementation.

by Bernhard Kölbl

On Fri Jun 10 12:37:58 2022 +0000, Rémi Bernon wrote: > This should probably be `IID_IAsyncOperation_SpeechSynthesisStream` instead? Good, catch! -- https://gitlab.winehq.org/wine/wine/-/merge_requests/217#note_1828

June 10, 2022

Re: [PATCH v3 0/6] MR217: windows.media.speech: ISpeechSynthesisStream stub implementation.

by Bernhard Kölbl

On Fri Jun 10 10:25:14 2022 +0000, Rémi Bernon wrote: > I'd suggest to simply remove that racy `GetResult` check, but otherwise > what about changing the check to `ok(hr == E_ILLEGAL_METHOD_CALL || hr > == S_OK, ...)`, and then and putting the rest in an `if (hr == > E_ILLEGAL_METHOD_CALL)` block? > Gotos are more often used to skip tests after a todo_wine, so that > implementing the call only require to remove the todo_wine and the goto, > but for normal control flow I think ifs are better. Yeah, an if block sounds cleaner. I'd like to keep the check though. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/217#note_1827

June 10, 2022

Re: [PATCH v3 0/6] MR217: windows.media.speech: ISpeechSynthesisStream stub implementation.

by Rémi Bernon

Rémi Bernon (@rbernon) commented about dlls/windows.media.speech/synthesizer.c: > return S_OK; > } > > +static HRESULT CALLBACK ssml_to_stream_operation( IInspectable *invoker, IInspectable **result ) > +{ > + return S_OK; > +} > + > static HRESULT WINAPI synthesizer_SynthesizeSsmlToStreamAsync( ISpeechSynthesizer *iface, HSTRING ssml, > IAsyncOperation_SpeechSynthesisStream **operation ) > { > FIXME("iface %p, text %p, operation %p stub.\n", iface, ssml, operation); > - return E_NOTIMPL; > + async_operation_inspectable_create(&IID_ISpeechSynthesisStream, NULL, ssml_to_stream_operation, (IAsyncOperation_IInspectable **)operation); Here too, `IID_IAsyncOperation_SpeechSynthesisStream` I guess. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/217#note_1826

June 10, 2022

Re: [PATCH v3 0/6] MR217: windows.media.speech: ISpeechSynthesisStream stub implementation.

by Rémi Bernon

Rémi Bernon (@rbernon) commented about dlls/windows.media.speech/synthesizer.c: > return E_NOTIMPL; > } > > +static HRESULT CALLBACK text_to_stream_operation( IInspectable *invoker, IInspectable **result ) > +{ > + return S_OK; > +} > + > static HRESULT WINAPI synthesizer_SynthesizeTextToStreamAsync( ISpeechSynthesizer *iface, HSTRING text, > IAsyncOperation_SpeechSynthesisStream **operation ) > { > FIXME("iface %p, text %p, operation %p stub.\n", iface, text, operation); > - return E_NOTIMPL; > + async_operation_inspectable_create(&IID_ISpeechSynthesisStream, NULL, text_to_stream_operation, (IAsyncOperation_IInspectable **)operation); This should probably be `IID_IAsyncOperation_SpeechSynthesisStream` instead? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/217#note_1825

June 10, 2022

Re: [PATCH 6/6] dinput/tests: Merge keyboard.c tests into device8.c.

by Marvin

Hi, While running your changed tests, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at: https://testbot.winehq.org/JobDetails.pl?Key=116665 Your paranoid android. === w8 (32 bit report) === dinput: driver_bus.c:290: Test failed: id 1: force_feedback.c:4020 got spurious packet driver_bus.c:1098: Test failed: id 1: force_feedback.c:4020 expect[0]: got 0xb000f, expected 0 driver_bus.c:1099: Test failed: id 1: force_feedback.c:4020 expect[0]: got id 6 driver_bus.c:1100: Test failed: id 1: force_feedback.c:4020 expect[0]: got len 2 === w8 (32 bit report) === dinput: joystick8.c:1836: Test failed: 0x700: state[0]: WaitForSingleObject succeeded

June 10, 2022

Re: [PATCH 0/4] MR218: advapi32: Implement RegRenameKey()

by Alexandre Julliard (＠julliard)

Alexandre Julliard (@julliard) commented about server/registry.c: > + name.len = (get_req_data_size() / sizeof(WCHAR)) * sizeof(WCHAR); > + > + if (!name.len) > + { > + set_error( STATUS_INVALID_PARAMETER ); > + } > + else if (!key->parent || (subkey = find_subkey( key->parent, &name, &index ))) > + { > + set_error( STATUS_CANNOT_DELETE ); > + } > + else if ((ptr = memdup( name.str, name.len ))) > + { > + free( key->name ); > + key->name = ptr; > + key->namelen = name.len; > + } The key needs to be properly sorted into the parent's subkey array. You'd also need to check if the name contains a backslash, handle renaming Wow6432Node, set the dirty flag, send notifications, etc. Basically you have to replicate most of what create_key() does. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/218#note_1823

June 10, 2022

Re: [PATCH 0/7] MR212: d3d11: Create dxgi resource objects for textures and buffers.

by Nikolay Sivov (＠nsivov)

On Fri Jun 10 11:11:22 2022 +0000, Jan Sikorski wrote: > I agree with Henri's point though, perhaps it would be better to reorder > initialization so that we create dxgi_resource first? You can't reorder right now, because dxgi side needs wined3d_resource pointer which you get from earlier wined3d_texture_create(). -- https://gitlab.winehq.org/wine/wine/-/merge_requests/212#note_1822

June 10, 2022