June 2015 - Wine-Devel - list.winehq.org

Re: [3/3] advapi32/tests: Add preliminary event tracing tests.
by Nikolay Sivov June 9, 2015

June 9, 2015

On 08.06.2015 6:56, Zebediah Figura wrote: > + /* test ERROR_DISK_FULL */ > + properties->MaximumFileSize = 1000000000000; /* 1 EB */ > + SetLastError(0xdeadbeef); > + ret = StartTraceA(&handle, sessionname, properties); > + todo_wine > + { > + ok(ret != ERROR_SUCCESS, "Expected failure\n"); > + ok(GetLastError() == ERROR_DISK_FULL, "Expected ERROR_DISK_FULL, got %d\n", GetLastError()); > + } > + properties->MaximumFileSize = 0; Patches look fine to me, but I have to ask - is it possible that this test will actually try to write that much?

2 2

Re: gdiplus: Implement trivial case of GdipCreateRegionRgnData for empty and infinite regions
by Vincent Povirk June 9, 2015

June 9, 2015

+ if (region_header->magic != VERSION_MAGIC) + return InvalidParameter; We should probably accept 0xdbc01002 as well here.

2 1

Re: [1/4] msvcirt: Add implementation of streambuf::allocate (try 2, resend)
by Piotr Caban June 8, 2015

June 8, 2015

On 06/08/15 15:56, Iván Matellanes wrote: > + ok(sb.ebuf - sb.base == 512 , "wrong reserve area size, expected 512 got %d\n", sb.ebuf - sb.base); > + ok(sb2.ebuf - sb2.base == 512 , "wrong reserve area size, expected 512 got %d\n", sb2.ebuf - sb2.base); This will produce warning on 64-bit build. You can e.g. display both pointers in error message. Sorry for not spotting it earlier, Piotr

1 0

Wineconf proposal: discuss 64-bit?
by Josh DuBois June 8, 2015

June 8, 2015

I'm wondering whether it would be useful to discuss 64-bit support in Wine at wineconf this year. I'm interested in feedback on whether it's a useful topic. With games in particular, 64-bit seems to be becoming a more common requirement, especially with newer DirectX 11 games (which may start running in wine ... sometime). I guess the basics work on Intel Linux. Several issues remain, however: - OS X support: compiler toolchains are an issue, there are problems with thread-local storage. - ARM support: I think Andre H. believes that 64-bit will be hard to support on ARM. As far as I know there aren't yet a lot of real-world use-cases for that, but there are certainly a lot of ARM chips around nowdays and maybe Windows 10 will start appearing on ARM devices. - MSI: I think there are MSI limitations which prevent some 64-bit apps from installing (e.g., bug 29941) . I think this blocks 64-bit MS Office. - Mono / .NET ? - Other ... ? It might be that there's not a great deal to do at wineconf other than catalog the issues. If it's worth a short (not an hour - maybe if we have some 10-15 minute segments) discussion while everyone is in the same place, I'd be happy to try and catalog them.

2 2

Re: [5/7] qmgr: Implement IBackgroundCopyJob::Cancel.
by Jacek Caban June 8, 2015

June 8, 2015

Hi Hans, On 06/08/15 14:32, Hans Leidekker wrote: > --- > dlls/qmgr/file.c | 1 + > dlls/qmgr/job.c | 33 +++++++++++++++++++++++++++++++-- > 2 files changed, 32 insertions(+), 2 deletions(-) I don't know qmgr, but shouldn't it result in IBinding::Abort() call on underlying binding object? Jacek

2 2

Wineconf talk proposal: Wine and Valgrind
by Austin English June 8, 2015

June 8, 2015

Howdy all, If anyone is interested, I can give a presentation on Wine/Valgrind, i.e. current status, major outstanding issues on either side, issues found/fixed, etc. Please let me know if you'd be interested. -Austin

2 1

Re: [PATCH 1/2] d3dx9_36: Implemented D3DXSaveVolumeToFile including DDS files
by Nikolay Sivov June 8, 2015

June 8, 2015

On 08.06.2015 10:01, Sergey Isakov wrote: > It is hard to distill what part of codes comes from where. My patches depends on previous patches and so they must be included too. > OK, all copylefts must be included if I will know how to do this. > If Christian Costa will say what to do I will do this. Sure, let's hear his opinion. > > On 08.06.2015, at 9:43, Nikolay Sivov wrote: > >> On 07.06.2015 23:02, Sergey Isakov wrote: >>> I decided to improve builtin d3dx9_36 library with support of more file >>> formats, texture formats, volume textures. >> >> It's not your work though, it looks to me like 3 or so patches from Christian Costa squashed together. Why didn't you mention it? >

3 3

Re: xaudio2_7: Implement IXaudio2 interface as stub (try 4)
by Jacek Caban June 8, 2015

June 8, 2015

Hi Guillaume, On 06/08/15 10:06, Guillaume Charifi wrote: > Fixes https://bugs.winehq.org/show_bug.cgi?id=26808 > > --- > dlls/xaudio2_7/Makefile.in | 2 > dlls/xaudio2_7/xaudio_dll.c | 300 > ++++++++++++++++++++++++++++++++++++- > dlls/xaudio2_7/xaudio_classes.idl | 28 +++ > 3 files changed, 319 insertions(+), 11 deletions(-) > > > xaudio_2.7.dll.IXAudio2_stub.patch > > > diff --git a/dlls/xaudio2_7/Makefile.in b/dlls/xaudio2_7/Makefile.in > index 853d1d6..6bdf362 100644 > --- a/dlls/xaudio2_7/Makefile.in > +++ b/dlls/xaudio2_7/Makefile.in > @@ -3,3 +3,5 @@ IMPORTS = advapi32 kernel32 ole32 user32 uuid > > C_SRCS = \ > xaudio_dll.c > + > +IDL_SRCS = xaudio_classes.idl Adding registration IDL deserves a separated patch. > diff --git a/dlls/xaudio2_7/xaudio_dll.c b/dlls/xaudio2_7/xaudio_dll.c > index c8b7904..07c4780 100644 > --- a/dlls/xaudio2_7/xaudio_dll.c > +++ b/dlls/xaudio2_7/xaudio_dll.c > @@ -30,12 +30,15 @@ > #include "wine/debug.h" > #include <propsys.h> > #include "initguid.h" > + > #include "xaudio2.h" > > WINE_DEFAULT_DEBUG_CHANNEL(xaudio2); > > static HINSTANCE instance; > > +static HRESULT XAudio2_create(IUnknown *pUnkOuter, LPVOID *ppObj); > + > BOOL WINAPI DllMain(HINSTANCE hinstDLL, DWORD fdwReason, LPVOID lpvReserved) > { > TRACE("(%p, %d, %p)\n", hinstDLL, fdwReason, lpvReserved); > @@ -52,35 +55,310 @@ BOOL WINAPI DllMain(HINSTANCE hinstDLL, DWORD fdwReason, LPVOID lpvReserved) > return TRUE; > } > > -HRESULT WINAPI DllCanUnloadNow(void) > +typedef struct { > + IClassFactory IClassFactory_iface; > + HRESULT (*pfnCreateInstance)(IUnknown *pUnkOuter, LPVOID *ppObj); Please avoid using LP* types. This should be void**, but using IUnknown** would be even better. > +} IClassFactoryImpl; > + > +static inline IClassFactoryImpl *impl_from_IClassFactory(IClassFactory *iface) > { > - return S_FALSE; > + return CONTAINING_RECORD(iface, IClassFactoryImpl, IClassFactory_iface); > +} > + > +static HRESULT WINAPI XAudio2CF_QueryInterface(IClassFactory *iface, REFIID riid, void **ppobj) > +{ > + if(IsEqualGUID(riid, &IID_IUnknown) > + || IsEqualGUID(riid, &IID_IClassFactory)) > + { > + IClassFactory_AddRef(iface); > + *ppobj = iface; > + return S_OK; > + } > + > + *ppobj = NULL; > + WARN("(%p)->(%s, %p): interface not found\n", iface, debugstr_guid(riid), ppobj); > + return E_NOINTERFACE; > +} > + > +static ULONG WINAPI XAudio2CF_AddRef(IClassFactory *iface) > +{ > + return 2; > +} > + > +static ULONG WINAPI XAudio2CF_Release(IClassFactory *iface) > +{ > + return 1; > } > > +static HRESULT WINAPI XAudio2CF_CreateInstance(IClassFactory *iface, IUnknown *pOuter, > + REFIID riid, void **ppobj) > +{ > + IClassFactoryImpl *This = impl_from_IClassFactory(iface); > + HRESULT hr; > + LPUNKNOWN punk; Same here, use IUnknown* instead. > + > + TRACE("(%p)->(%p,%s,%p)\n", This, pOuter, debugstr_guid(riid), ppobj); > + > + *ppobj = NULL; > + hr = This->pfnCreateInstance(pOuter, (LPVOID *) &punk); > + if (FAILED(hr)) > + return hr; > + > + hr = IUnknown_QueryInterface(punk, riid, ppobj); > + IUnknown_Release(punk); > + return hr; > +} > + > +static HRESULT WINAPI XAudio2CF_LockServer(IClassFactory *iface, BOOL dolock) > +{ > + IClassFactoryImpl *This = impl_from_IClassFactory(iface); > + FIXME("(%p)->(%d): stub!\n", This, dolock); > + return S_OK; > +} > + > +static const IClassFactoryVtbl XAudio2CF_Vtbl = > +{ > + XAudio2CF_QueryInterface, > + XAudio2CF_AddRef, > + XAudio2CF_Release, > + XAudio2CF_CreateInstance, > + XAudio2CF_LockServer > +}; > + > +static IClassFactoryImpl xaudio2_cf = { { &XAudio2CF_Vtbl }, XAudio2_create }; > + > HRESULT WINAPI DllGetClassObject(REFCLSID rclsid, REFIID riid, LPVOID *ppv) > { > + IClassFactory *factory = NULL; > + > TRACE("(%s, %s, %p)\n", debugstr_guid(rclsid), debugstr_guid(riid), ppv); > > - if (ppv == NULL) { > - WARN("invalid parameter\n"); > - return E_INVALIDARG; > + if(IsEqualGUID(rclsid, &CLSID_XAudio2)) { > + factory = &xaudio2_cf.IClassFactory_iface; How about simply return IClassFactory_QueryInterface(...) here? > } > + if(!factory) return CLASS_E_CLASSNOTAVAILABLE; > > - *ppv = NULL; > + return IClassFactory_QueryInterface(factory, riid, ppv); > +} > > - WARN("(%s, %s, %p): no class found.\n", debugstr_guid(rclsid), > - debugstr_guid(riid), ppv); > - return CLASS_E_CLASSNOTAVAILABLE; > +HRESULT WINAPI DllCanUnloadNow(void) > +{ > + return S_FALSE; > } > > HRESULT WINAPI DllRegisterServer(void) > { > TRACE("\n"); > - return __wine_register_resources( instance ); > + return __wine_register_resources(instance); > } > > HRESULT WINAPI DllUnregisterServer(void) > { > TRACE("\n"); > - return __wine_unregister_resources( instance ); > + return __wine_unregister_resources(instance); > +} > + > +typedef struct { > + IXAudio2 IXAudio2Impl_iface; > + LONG ref; > +} IXAudio2Impl; > + > +static inline IXAudio2Impl *impl_from_IXAudio2(IXAudio2 *iface) > +{ > + return CONTAINING_RECORD(iface, IXAudio2Impl, IXAudio2Impl_iface); > } > + > +static const struct IXAudio2Vtbl XAudio2_Vtbl; This is not needed. Please move XAudio2_create below the vtbl. > +HRESULT XAudio2_create(IUnknown *pUnkOuter, LPVOID *ppObj) > +{ > + IXAudio2Impl* object; > + > + TRACE("(%p, %p)\n", pUnkOuter, ppObj); > + > + if(pUnkOuter) > + return CLASS_E_NOAGGREGATION; > + > + object = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, sizeof(IXAudio2Impl)); > + if(!object) > + return E_OUTOFMEMORY; > + > + object->IXAudio2Impl_iface.lpVtbl = &XAudio2_Vtbl; > + object->ref = 1; > + > + *ppObj = object; You shouldn't assign object type here, even if the pointer is the same ans interface. Use &object->IXAudio2Impl_iface instead. > + > + return S_OK; > +} > + > +/*** IUnknown methods ***/ > +static HRESULT WINAPI IXAudio2Impl_QueryInterface(IXAudio2 *iface, REFIID riid, void **ppvObject) > +{ > + IXAudio2Impl *This = impl_from_IXAudio2(iface); > + > + TRACE("(%p)->(%s, %p)\n", This, debugstr_guid(riid), ppvObject); > + > + if(IsEqualGUID(riid, &IID_IUnknown) || > + IsEqualGUID(riid, &IID_IXAudio2)) > + { > + IXAudio2_AddRef(iface); > + *ppvObject = iface; > + return S_OK; > + } > + > + ERR("(%p)->(%s,%p), not found\n", This,debugstr_guid(riid), ppvObject); This should be FIXME instead (or WARN, it depends on how likely apps are to call it with an interface that shouldn't be implemented). Also, QI is expected to set ppvObject to NULL on error. Thanks, Jacek

1 0

Re: dlls/winsta: add initial winsta.h (try 2)
by Alexandre Julliard June 8, 2015

June 8, 2015

Austin English <austinenglish(a)gmail.com> writes: > updated TS_ALL_PROCESSES_INFO to one from > https://msdn.microsoft.com/en-us/library/cc248681.aspx Looks like you forgot that part. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: kernel32: Fail properly for unsupported classes in GetFileInformationByHandleEx
by Marvin June 8, 2015

June 8, 2015

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=13968 Your paranoid android. === wvistau64 (32 bit file) === file.c:444: Test failed: attributes of file "testfile.xxx" are 0x0820 === wvistau64 (64 bit file) === file.c:444: Test failed: attributes of file "testfile.xxx" are 0x0820

2 1