June 2016 - Wine-Devel - list.winehq.org

Re: [PATCH 3/9] d3d11: Implement d3d11_immediate_context_OMGetBlendState().

by Henri Verbeet

On 24 June 2016 at 13:21, Józef Kucia <joseph.kucia(a)gmail.com> wrote: > On Fri, Jun 24, 2016 at 1:07 PM, Henri Verbeet <hverbeet(a)gmail.com> wrote: >> Yeah, but that's just because in most cases there's no reason to >> expose them outside that file. E.g. impl_from_ID3D11Device() is in >> d3d11_private.h, and in ddraw several of them are. > > My impression is that we want to keep them private unless there is a > good reason to expose them. I guess my impression is mostly based on > this comment https://www.winehq.org/pipermail/wine-devel/2015-August/108939.html. > I can change to impl_from_ID3D11BlendState(). It doesn't make much > different to me. Yeah, it's a bit of a trade-off. An alternative could be to introduce a helper function that just returns a d3d_blend_state structure.

June 24, 2016

Re: [PATCH 7/9] d3d11: Implement ID3D11UnorderedAccessView interface.

by Józef Kucia

On Fri, Jun 24, 2016 at 12:56 PM, Henri Verbeet <hverbeet(a)gmail.com> wrote: > On 24 June 2016 at 11:46, Józef Kucia <jkucia(a)codeweavers.com> wrote: >> + switch (desc->ViewDimension) >> + { >> + case D3D11_UAV_DIMENSION_TEXTURE1DARRAY: >> + if (desc->u.Texture1DArray.ArraySize == -1 && desc->u.Texture1DArray.FirstArraySlice < layer_count) > I noticed too late for the other normalisation functions, but > "ArraySize" is unsigned. Should I use 0xffffffff in tests as well?

June 24, 2016

Re: [PATCH 3/9] d3d11: Implement d3d11_immediate_context_OMGetBlendState().

by Józef Kucia

On Fri, Jun 24, 2016 at 1:07 PM, Henri Verbeet <hverbeet(a)gmail.com> wrote: > Yeah, but that's just because in most cases there's no reason to > expose them outside that file. E.g. impl_from_ID3D11Device() is in > d3d11_private.h, and in ddraw several of them are. My impression is that we want to keep them private unless there is a good reason to expose them. I guess my impression is mostly based on this comment https://www.winehq.org/pipermail/wine-devel/2015-August/108939.html. I can change to impl_from_ID3D11BlendState(). It doesn't make much different to me.

June 24, 2016

Re: [PATCH 3/9] d3d11: Implement d3d11_immediate_context_OMGetBlendState().

by Henri Verbeet

On 24 June 2016 at 13:04, Józef Kucia <joseph.kucia(a)gmail.com> wrote: > On Fri, Jun 24, 2016 at 12:56 PM, Henri Verbeet <hverbeet(a)gmail.com> wrote: >> On 24 June 2016 at 11:45, Józef Kucia <jkucia(a)codeweavers.com> wrote: >>> @@ -4044,16 +4053,17 @@ static void STDMETHODCALLTYPE d3d10_device_OMGetBlendState(ID3D10Device1 *iface, >>> ID3D10BlendState **blend_state, FLOAT blend_factor[4], UINT *sample_mask) >>> { >>> struct d3d_device *device = impl_from_ID3D10Device(iface); >>> + ID3D11BlendState *d3d11_blend_state; >>> + struct d3d_blend_state *object; >>> >>> TRACE("iface %p, blend_state %p, blend_factor %p, sample_mask %p.\n", >>> iface, blend_state, blend_factor, sample_mask); >>> >>> - if ((*blend_state = device->blend_state ? (ID3D10BlendState *)&device->blend_state->ID3D10BlendState1_iface : NULL)) >>> - ID3D10BlendState_AddRef(*blend_state); >>> - wined3d_mutex_lock(); >>> - memcpy(blend_factor, device->blend_factor, 4 * sizeof(*blend_factor)); >>> - *sample_mask = wined3d_device_get_render_state(device->wined3d_device, WINED3D_RS_MULTISAMPLEMASK); >>> - wined3d_mutex_unlock(); >>> + d3d11_immediate_context_OMGetBlendState(&device->immediate_context.ID3D11DeviceContext_iface, >>> + &d3d11_blend_state, blend_factor, sample_mask); >>> + >>> + object = unsafe_impl_from_ID3D11BlendState(d3d11_blend_state); >> Why not just impl_from_ID3D11BlendState() here? > > Mostly because we keep impl_from_* functions private to a file in > which an interface is implemented. Yeah, but that's just because in most cases there's no reason to expose them outside that file. E.g. impl_from_ID3D11Device() is in d3d11_private.h, and in ddraw several of them are.

June 24, 2016

Re: [PATCH 3/9] d3d11: Implement d3d11_immediate_context_OMGetBlendState().

by Józef Kucia

On Fri, Jun 24, 2016 at 12:56 PM, Henri Verbeet <hverbeet(a)gmail.com> wrote: > On 24 June 2016 at 11:45, Józef Kucia <jkucia(a)codeweavers.com> wrote: >> @@ -4044,16 +4053,17 @@ static void STDMETHODCALLTYPE d3d10_device_OMGetBlendState(ID3D10Device1 *iface, >> ID3D10BlendState **blend_state, FLOAT blend_factor[4], UINT *sample_mask) >> { >> struct d3d_device *device = impl_from_ID3D10Device(iface); >> + ID3D11BlendState *d3d11_blend_state; >> + struct d3d_blend_state *object; >> >> TRACE("iface %p, blend_state %p, blend_factor %p, sample_mask %p.\n", >> iface, blend_state, blend_factor, sample_mask); >> >> - if ((*blend_state = device->blend_state ? (ID3D10BlendState *)&device->blend_state->ID3D10BlendState1_iface : NULL)) >> - ID3D10BlendState_AddRef(*blend_state); >> - wined3d_mutex_lock(); >> - memcpy(blend_factor, device->blend_factor, 4 * sizeof(*blend_factor)); >> - *sample_mask = wined3d_device_get_render_state(device->wined3d_device, WINED3D_RS_MULTISAMPLEMASK); >> - wined3d_mutex_unlock(); >> + d3d11_immediate_context_OMGetBlendState(&device->immediate_context.ID3D11DeviceContext_iface, >> + &d3d11_blend_state, blend_factor, sample_mask); >> + >> + object = unsafe_impl_from_ID3D11BlendState(d3d11_blend_state); > Why not just impl_from_ID3D11BlendState() here? Mostly because we keep impl_from_* functions private to a file in which an interface is implemented.

June 24, 2016

Re: [PATCH 7/9] d3d11: Implement ID3D11UnorderedAccessView interface.

by Henri Verbeet

On 24 June 2016 at 11:46, Józef Kucia <jkucia(a)codeweavers.com> wrote: > + switch (desc->ViewDimension) > + { > + case D3D11_UAV_DIMENSION_TEXTURE1DARRAY: > + if (desc->u.Texture1DArray.ArraySize == -1 && desc->u.Texture1DArray.FirstArraySlice < layer_count) I noticed too late for the other normalisation functions, but "ArraySize" is unsigned.

June 24, 2016

Re: [PATCH 3/9] d3d11: Implement d3d11_immediate_context_OMGetBlendState().

by Henri Verbeet

On 24 June 2016 at 11:45, Józef Kucia <jkucia(a)codeweavers.com> wrote: > @@ -4044,16 +4053,17 @@ static void STDMETHODCALLTYPE d3d10_device_OMGetBlendState(ID3D10Device1 *iface, > ID3D10BlendState **blend_state, FLOAT blend_factor[4], UINT *sample_mask) > { > struct d3d_device *device = impl_from_ID3D10Device(iface); > + ID3D11BlendState *d3d11_blend_state; > + struct d3d_blend_state *object; > > TRACE("iface %p, blend_state %p, blend_factor %p, sample_mask %p.\n", > iface, blend_state, blend_factor, sample_mask); > > - if ((*blend_state = device->blend_state ? (ID3D10BlendState *)&device->blend_state->ID3D10BlendState1_iface : NULL)) > - ID3D10BlendState_AddRef(*blend_state); > - wined3d_mutex_lock(); > - memcpy(blend_factor, device->blend_factor, 4 * sizeof(*blend_factor)); > - *sample_mask = wined3d_device_get_render_state(device->wined3d_device, WINED3D_RS_MULTISAMPLEMASK); > - wined3d_mutex_unlock(); > + d3d11_immediate_context_OMGetBlendState(&device->immediate_context.ID3D11DeviceContext_iface, > + &d3d11_blend_state, blend_factor, sample_mask); > + > + object = unsafe_impl_from_ID3D11BlendState(d3d11_blend_state); Why not just impl_from_ID3D11BlendState() here?

June 24, 2016

Re: Andrey Gusev : msvcp120/tests: Fix a typo in ok() message.

by Ken Thomases

On Jun 23, 2016, at 11:15 AM, Alexandre Julliard <julliard(a)winehq.org> wrote: > > diff --git a/dlls/msvcp120/tests/msvcp120.c b/dlls/msvcp120/tests/msvcp120.c > index 7c94906..354b2d3 100644 > --- a/dlls/msvcp120/tests/msvcp120.c > +++ b/dlls/msvcp120/tests/msvcp120.c > @@ -592,7 +592,7 @@ static void test_xtime_get(void) > diff = p__Xtime_diff_to_millis2(&after, &before); > > ok(diff >= tests[i], > - "xtime_get() not functioning correctly, test: %d, expect: ge %d, got: %d\n", > + "xtime_get() not functioning correctly, test: %d, expect: %d, got: %d\n", > i, tests[i], diff); > } Too late on this one, but I think the "ge" was supposed to mean "greater than or equal to", given the actual test being performed. Probably best to replace it with ">=". -Ken

June 23, 2016

Re: [PATCH 1/3] comdlg32: Optimize FILEDLG95_FILENAME_FillFromSelection.

by Henri Verbeet

On 23 June 2016 at 13:29, Lauri Kenttä <lauri.kentta(a)gmail.com> wrote: > This trades some memory for speed, but since currently selecting > 1000 files takes over a minute but the required memory for those > 1000 filenames is only 262 kB (worst case), this seems a reasonable > trade-off. > That's probably fine, but is there any reason this can't use the standard approach of just doubling the buffer size when you run out of space? If this is performance critical code, the multiple copies of the string related to GetName() probably don't help either.

June 23, 2016

Re: kernel32/tests: Fix a typo in skip() message.

by Marvin

Hi, While running your changed tests on Windows, 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=23833 Your paranoid android. === w1064 (32 bit locale) === locale.c:2370: Test failed: LCMapStringEx should fail with bad locale name locale.c:2371: Test failed: unexpected error code -559038737 locale.c:2462: Test failed: Expected lcid == 0, got got 00001000, error -559038737 locale.c:4192: Test failed: IsValidLocaleName should have failed locale.c:4194: Test failed: IsValidLocaleName should have failed locale.c:4577: Test failed: For id 7, expected ret == 4, got 1, error 87 locale.c:4579: Test failed: For id 7, Expected IVC, got '' === w1064 (64 bit locale) === locale.c:2370: Test failed: LCMapStringEx should fail with bad locale name locale.c:2371: Test failed: unexpected error code -559038737 locale.c:2462: Test failed: Expected lcid == 0, got got 00001000, error -559038737 locale.c:4192: Test failed: IsValidLocaleName should have failed locale.c:4194: Test failed: IsValidLocaleName should have failed locale.c:4577: Test failed: For id 7, expected ret == 4, got 1, error 87 locale.c:4579: Test failed: For id 7, Expected IVC, got ''

June 23, 2016