July 2017 - Wine-devel - list.winehq.org

Re: [PATCH 2/5] regedit: Use 'return' statements in delete_registry_key() to allow for memory clean-up
by Alexandre Julliard 11 Jul '17

11 Jul '17

Hugh McMaster <hugh.mcmaster(a)outlook.com> writes: > @@ -1593,11 +1593,11 @@ void delete_registry_key(WCHAR *reg_key_name) > > if (!(key_class = parse_key_name(reg_key_name, &key_name))) { > output_message(STRING_INCORRECT_REG_CLASS, reg_key_name); > - exit(1); > + return; > } > if (!*key_name) { > output_message(STRING_DELETE_REG_CLASS_FAILED, reg_key_name); > - exit(1); > + return; If the process is going to exit anyway, freeing memory is a waste of time. -- Alexandre Julliard julliard(a)winehq.org

2 3

Re: [PATCH 3/9] msvcr120: Remove precompiler check for asinh.
by Piotr Caban 11 Jul '17

11 Jul '17

On 07/10/17 06:55, Alex Henrie wrote: > asinh is part of C89, so it is guaranteed to be present. Are you sure about it? As far as I know it's part of C99. The manual for asinh looks strange (it says that it conforms to C89 and gives test macro that checks for never standards). Thanks, Piotr

4 7

Re: [PATCH 6/7] d3dx9: Implement d3dx_shader_get_raw_constant_info() function.
by Matteo Bruni 10 Jul '17

10 Jul '17

2017-07-06 13:15 GMT+02:00 Paul Gofman <gofmanp(a)gmail.com>: > Signed-off-by: Paul Gofman <gofmanp(a)gmail.com> > --- > dlls/d3dx9_36/d3dx9_private.h | 3 +++ > dlls/d3dx9_36/shader.c | 18 ++++++++++++++++++ > 2 files changed, 21 insertions(+) > > diff --git a/dlls/d3dx9_36/d3dx9_private.h b/dlls/d3dx9_36/d3dx9_private.h > index db80490..bfb3909 100644 > --- a/dlls/d3dx9_36/d3dx9_private.h > +++ b/dlls/d3dx9_36/d3dx9_private.h > @@ -355,4 +355,7 @@ HRESULT d3dx_param_eval_set_shader_constants(ID3DXEffectStateManager *manager, s > struct d3dx_param_eval *peval, BOOL update_all) DECLSPEC_HIDDEN; > BOOL is_param_eval_input_dirty(struct d3dx_param_eval *peval, ULONG64 update_version) DECLSPEC_HIDDEN; > > +const D3DXSHADER_CONSTANTINFO *d3dx_shader_get_raw_constant_info(ID3DXConstantTable *iface, > + D3DXHANDLE constant) DECLSPEC_HIDDEN; > + > #endif /* __WINE_D3DX9_PRIVATE_H */ > diff --git a/dlls/d3dx9_36/shader.c b/dlls/d3dx9_36/shader.c > index 4dbecb1..11527c8 100644 > --- a/dlls/d3dx9_36/shader.c > +++ b/dlls/d3dx9_36/shader.c > @@ -682,6 +682,7 @@ HRESULT WINAPI D3DXPreprocessShaderFromResourceW(HMODULE module, const WCHAR *re > } > > struct ctab_constant { > + const D3DXSHADER_CONSTANTINFO *raw_constant_info; > D3DXCONSTANT_DESC desc; > struct ctab_constant *constants; > }; > @@ -946,6 +947,22 @@ static HRESULT WINAPI ID3DXConstantTableImpl_GetDesc(ID3DXConstantTable *iface, > return D3D_OK; > } > > +const D3DXSHADER_CONSTANTINFO *d3dx_shader_get_raw_constant_info(ID3DXConstantTable *iface, D3DXHANDLE constant) > +{ > + struct ID3DXConstantTableImpl *const_table = impl_from_ID3DXConstantTable(iface); > + struct ctab_constant *c = get_valid_constant(const_table, constant); > + > + TRACE("const_table %p, constant %p.\n", const_table, constant); > + > + if (!c) > + { > + WARN("Invalid argument specified.\n"); > + return NULL; > + } > + > + return c->raw_constant_info; > +} > + > static HRESULT WINAPI ID3DXConstantTableImpl_GetConstantDesc(ID3DXConstantTable *iface, D3DXHANDLE constant, > D3DXCONSTANT_DESC *desc, UINT *count) > { > @@ -2052,6 +2069,7 @@ HRESULT WINAPI D3DXGetShaderConstantTableEx(const DWORD *byte_code, DWORD flags, > { > object->constants[i].desc.RegisterCount = constant_info[i].RegisterCount; > } > + object->constants[i].raw_constant_info = &constant_info[i]; > } > > *constant_table = &object->ID3DXConstantTable_iface; I don't like this all that much. If you need access to the original D3DXSHADER_CONSTANTINFO only for the Reserved field, as in the next patch, I'd say just store that one in struct ctab_constant. Then, for the retrieval part, I don't think it makes much sense to hide our internals from ourselves. You could move the definition of struct ctab_constant to d3dx9_private.h and introduce an internal function to return it from the constant table. You could then entirely replace the use of ID3DXConstantTable_GetConstantDesc(), if you want. Otherwise you can keep that part as-is for the time being and make use of the new function only to get to the Reserved data. Actually, if you go with the latter, it probably makes sense to merge this patch into the next.

1 0

Re: [PATCH 4/7] d3dx9: Get rid of init_set_constants() function.
by Matteo Bruni 10 Jul '17

10 Jul '17

2017-07-06 13:15 GMT+02:00 Paul Gofman <gofmanp(a)gmail.com>: > Signed-off-by: Paul Gofman <gofmanp(a)gmail.com> > --- > dlls/d3dx9_36/preshader.c | 58 ++++++++++++++--------------------------------- > 1 file changed, 17 insertions(+), 41 deletions(-) > > diff --git a/dlls/d3dx9_36/preshader.c b/dlls/d3dx9_36/preshader.c > index cabddcd..1aaf59d 100644 > --- a/dlls/d3dx9_36/preshader.c > +++ b/dlls/d3dx9_36/preshader.c > @@ -305,7 +305,8 @@ static unsigned int get_reg_components(unsigned int table) > > #define PRES_BITMASK_BLOCK_SIZE (sizeof(unsigned int) * 8) > > -static HRESULT init_set_constants(struct d3dx_const_tab *const_tab, ID3DXConstantTable *ctab); > +static HRESULT init_set_constants_param(struct d3dx_const_tab *const_tab, ID3DXConstantTable *ctab, > + D3DXHANDLE hc, struct d3dx_parameter *param); With some function reordering this forward declaration shouldn't be necessary at all.

1 0

Re: [PATCH 3/7] d3dx9: Simplify init and cleanup in get_constants_desc().
by Matteo Bruni 10 Jul '17

10 Jul '17

2017-07-06 13:15 GMT+02:00 Paul Gofman <gofmanp(a)gmail.com>: > Signed-off-by: Paul Gofman <gofmanp(a)gmail.com> > --- > dlls/d3dx9_36/preshader.c | 17 +++-------------- > 1 file changed, 3 insertions(+), 14 deletions(-) > > diff --git a/dlls/d3dx9_36/preshader.c b/dlls/d3dx9_36/preshader.c > index 2fb14ab..cabddcd 100644 > --- a/dlls/d3dx9_36/preshader.c > +++ b/dlls/d3dx9_36/preshader.c > @@ -618,10 +618,6 @@ static HRESULT get_constants_desc(unsigned int *byte_code, struct d3dx_const_tab > D3DXHANDLE hc; > unsigned int i; > > - out->inputs = cdesc = NULL; > - out->inputs_param = NULL; > - out->input_count = 0; > - inputs_param = NULL; > hr = D3DXGetShaderConstantTable(byte_code, &ctab); > if (FAILED(hr) || !ctab) > { > @@ -635,8 +631,8 @@ static HRESULT get_constants_desc(unsigned int *byte_code, struct d3dx_const_tab > goto err_out; > } > > - cdesc = HeapAlloc(GetProcessHeap(), 0, sizeof(*cdesc) * desc.Constants); > - inputs_param = HeapAlloc(GetProcessHeap(), 0, sizeof(*inputs_param) * desc.Constants); > + out->inputs = cdesc = HeapAlloc(GetProcessHeap(), 0, sizeof(*cdesc) * desc.Constants); > + out->inputs_param = inputs_param = HeapAlloc(GetProcessHeap(), 0, sizeof(*inputs_param) * desc.Constants); > if (!cdesc || !inputs_param) > { > hr = E_OUTOFMEMORY; > @@ -678,16 +674,9 @@ static HRESULT get_constants_desc(unsigned int *byte_code, struct d3dx_const_tab > } > } > out->input_count = desc.Constants; > - out->inputs = cdesc; > - out->inputs_param = inputs_param; > hr = init_set_constants(out, ctab); > - ID3DXConstantTable_Release(ctab); > - return hr; > err_out: > - HeapFree(GetProcessHeap(), 0, cdesc); > - HeapFree(GetProcessHeap(), 0, inputs_param); > - if (ctab) > - ID3DXConstantTable_Release(ctab); > + ID3DXConstantTable_Release(ctab); > return hr; > } It would be a nice cleanup to rename the err_out: label to cleanup:, done: or similar.

1 0

Re: [PATCH 1/2] widl: Only generate Proxy Stubs when functions have the call_as attribute
by Nikolay Sivov 10 Jul '17

10 Jul '17

On 10.07.2017 7:43, Alistair Leslie-Hughes wrote: > Fixes: https://bugs.winehq.org/show_bug.cgi?id=43302 > > Signed-off-by: Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> > --- > tools/widl/header.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/tools/widl/header.c b/tools/widl/header.c > index 7b97158..e48b488 100644 > --- a/tools/widl/header.c > +++ b/tools/widl/header.c > @@ -1172,7 +1172,7 @@ static void write_method_proto(FILE *header, const type_t *iface) > { > const var_t *func = stmt->u.var; > > - if (!is_local(func->attrs)) { > + if (is_callas(func->attrs)) { > const char *callconv = get_attrp(func->type->attrs, ATTR_CALLCONV); > if (!callconv) callconv = "STDMETHODCALLTYPE"; > /* proxy prototype */ > This is wrong, 'call_as' is used to map different method to be used for remote calls. Midl certainly does not limit prototype generation like that. You mentioned 'odl' attribute on bug report, it's irrelevant too.

2 2

Re: [v2 PATCH] widl: Only generate Proxy Stubs when functions have the call_as attribute
by Huw Davies 10 Jul '17

10 Jul '17

On Mon, Jul 10, 2017 at 10:11:42AM +0000, Alistair Leslie-Hughes wrote: > Fixes: https://bugs.winehq.org/show_bug.cgi?id=43302 > > v2: Merge two matches together. > > Signed-off-by: Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> > --- > include/unknwn.idl | 10 ++++++++++ > tools/widl/header.c | 2 +- > 2 files changed, 11 insertions(+), 1 deletion(-) > > diff --git a/include/unknwn.idl b/include/unknwn.idl > index 78a2d9db18..7856a7f504 100644 > --- a/include/unknwn.idl > +++ b/include/unknwn.idl > @@ -46,6 +46,16 @@ interface IUnknown > ULONG Release(); > } > > +cpp_quote("HRESULT STDMETHODCALLTYPE IUnknown_QueryInterface_Proxy(IUnknown* This, REFIID riid, void **ppvObject);") > +cpp_quote("void __RPC_STUB IUnknown_QueryInterface_Stub(IRpcStubBuffer* This, IRpcChannelBuffer* pRpcChannelBuffer,") > +cpp_quote(" PRPC_MESSAGE pRpcMessage, DWORD* pdwStubPhase);") > +cpp_quote("ULONG STDMETHODCALLTYPE IUnknown_AddRef_Proxy(IUnknown* This);") > +cpp_quote("void __RPC_STUB IUnknown_AddRef_Stub(IRpcStubBuffer* This, IRpcChannelBuffer* pRpcChannelBuffer,") > +cpp_quote(" PRPC_MESSAGE pRpcMessage, DWORD* pdwStubPhase);") > +cpp_quote("ULONG STDMETHODCALLTYPE IUnknown_Release_Proxy(IUnknown* This);") > +cpp_quote("void __RPC_STUB IUnknown_Release_Stub(IRpcStubBuffer* This, IRpcChannelBuffer* pRpcChannelBuffer,") > +cpp_quote(" PRPC_MESSAGE pRpcMessage, DWORD* pdwStubPhase);") > + > cpp_quote("#ifdef SAVE_WIDL_C_INLINE_WRAPPERS") > cpp_quote("#define WIDL_C_INLINE_WRAPPERS") > cpp_quote("#endif") > diff --git a/tools/widl/header.c b/tools/widl/header.c > index 7b971582db..e48b488ae3 100644 > --- a/tools/widl/header.c > +++ b/tools/widl/header.c > @@ -1172,7 +1172,7 @@ static void write_method_proto(FILE *header, const type_t *iface) > { > const var_t *func = stmt->u.var; > > - if (!is_local(func->attrs)) { > + if (is_callas(func->attrs)) { > const char *callconv = get_attrp(func->type->attrs, ATTR_CALLCONV); > if (!callconv) callconv = "STDMETHODCALLTYPE"; > /* proxy prototype */ On further testing with midl v7, /Os does emit Proxy/Stub prototypes for all functions. Since widl essentially runs in this mode, I'd prefer to keep this behaviour if possible. In the bug you mention redefinition errors in qmgrprxy. How are you getting these? Huw.

1 1

Wine appimage request
by Dpto Catastro Topografia MuniSantaCruz 10 Jul '17

10 Jul '17

Hello, I have used this software and it has been very useful to me, I am a surveyor engineer and free software user. A year ago in the world linux have agreed to generate universal packages self-contained with everything they need to install on any linux distribution. Of which there are today 3; Simon Peter's appimage, Gnome flatpak and Canonical snaps. The advantage of these packages is that they use everything in the same package to work, libraries and so on. Hence they can be installed in any linux distribution. It would be great to see the WINE in appimage or in flatpak or snap, and avoid using the terminal for new users. I somehow defend myself with the terminal but sometimes it is still somewhat annoying and tedious. I leave you the links of these new packaging technologies, so that you can analyze them and may in the not too distant future offer some of these simple packages. In fact I prefer appimage, I downloaded several of these applications with this type of package and is simple to use and distribute. Flatpak and snap are still using the terminal to install them, the snaps are fairly simple (it will be installed in the gnome and kde software center in 2018), flatpak the gnome desktop and kde install it from the software center to install packages Flatpak. Appimage has its own charm and does not use the terminal, just make executable and run. Thank you Links Http://appimage.org/ Http://flatpak.org/ Https://snapcraft.io/ atte Luis V --

1 0

Re: [PATCH 5/5] d3drm: Assign animation key ids
by Henri Verbeet 10 Jul '17

10 Jul '17

On 10 July 2017 at 12:40, Nikolay Sivov <nsivov(a)codeweavers.com> wrote: > +static BOOL d3drm_animation_get_next_key_id(struct d3drm_animation_keys *keys, DWORD *id) > +{ > + DWORD i; > + > + if (keys->next_unused <= keys->last_id) > + { > + *id = keys->next_unused++; > + return TRUE; > + } > + > + if (keys->next_free <= keys->last_id) > + { > + *id = keys->next_free; > + keys->next_free = ~0u; > + return TRUE; > + } > + > + if (keys->count == keys->last_id - keys->first_id + 1) > + return FALSE; > + > + /* Worst case, look for the spot from the beginning. */ > + for (i = 0, *id = keys->first_id; i < keys->count; i++, ++*id) > + { > + if (*id != keys->keys[i].id) > + break; > + } Does that work? The keys aren't necessarily ordered by ID, right? Regardless, it may be easier to make use of the fact that we never shrink the keys array.

1 0

Re: server: Don't allow APCs on objects associated with completion port.
by Marvin 10 Jul '17

10 Jul '17

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=32203 Your paranoid android. === w7u (32 bit sock) === sock.c:6121: Test failed: Expected event sequence [FD_WRITE(0)], got [] sock.c:6142: Test failed: Expected event sequence [FD_READ(0) FD_CLOSE(0)], got [FD_WRITE(0) FD_CLOSE(10053)] sock.c:6145: Test failed: Failed to empty buffer: -1 - 10054 sock.c:6146: Test failed: Expected event sequence [FD_READ(0)], got [] sock.c:6149: Test failed: Failed to empty buffer: -1 - 10054

1 0