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

Re: [PATCH 2/2] hnetcfg/tests: Add INetFwAuthorizedApplication tests

by Nikolay Sivov

On 29.11.2016 9:58, Alistair Leslie-Hughes wrote: > Since put_ProcessImageFileName appears to validate the image name that is > supplied, the normal put/get/compare was excluded because it returned > E_INVALIDARG even for a fully qualified filename. An investigation for another > day. > > Signed-off-by: Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> > --- > dlls/hnetcfg/tests/Makefile.in | 2 +- > dlls/hnetcfg/tests/policy.c | 32 +++++++++++++++++++++++++++++++- > 2 files changed, 32 insertions(+), 2 deletions(-) > > diff --git a/dlls/hnetcfg/tests/Makefile.in b/dlls/hnetcfg/tests/Makefile.in > index aa4062a..5e1ce91 100644 > --- a/dlls/hnetcfg/tests/Makefile.in > +++ b/dlls/hnetcfg/tests/Makefile.in > @@ -1,5 +1,5 @@ > TESTDLL = hnetcfg.dll > -IMPORTS = ole32 uuid > +IMPORTS = ole32 uuid oleaut32 > > C_SRCS = \ > policy.c > diff --git a/dlls/hnetcfg/tests/policy.c b/dlls/hnetcfg/tests/policy.c > index a3ed4b2..a05b02e 100644 > --- a/dlls/hnetcfg/tests/policy.c > +++ b/dlls/hnetcfg/tests/policy.c > @@ -65,6 +65,36 @@ static void test_interfaces(void) > INetFwMgr_Release(manager); > } > > +static void test_NetFwAuthorizedApplication(void) > +{ > + INetFwAuthorizedApplication *app; > + HRESULT hr; > + > + hr = CoCreateInstance(&CLSID_NetFwAuthorizedApplication, NULL, CLSCTX_INPROC_SERVER|CLSCTX_INPROC_HANDLER, > + &IID_INetFwAuthorizedApplication, (void**)&app); > + ok(hr == S_OK, "got: %08x\n", hr); > + if(hr == S_OK) > + { It's not really necessary to protect it like that. If creation consistently fails test or all tests should be skipped. > + static WCHAR empty[] = {0}; > + BSTR bstr; > + > + hr = INetFwAuthorizedApplication_get_ProcessImageFileName(app, NULL); > + ok(hr == E_POINTER, "got: %08x\n", hr); > + > + hr = INetFwAuthorizedApplication_get_ProcessImageFileName(app, &bstr); > + ok(hr == S_OK, "got: %08x\n", hr); > + ok(!bstr, "got: %s\n", wine_dbgstr_w(bstr)); > + > + hr = INetFwAuthorizedApplication_put_ProcessImageFileName(app, NULL); > + ok(hr == E_INVALIDARG, "got: %08x\n", hr); > + > + hr = INetFwAuthorizedApplication_put_ProcessImageFileName(app, empty); > + ok(hr == E_INVALIDARG, "got: %08x\n", hr); Please use actual BSTR here. > + > + INetFwAuthorizedApplication_Release(app); > + } > +} > + > START_TEST(policy) > { > INetFwMgr *manager; > @@ -84,7 +114,7 @@ START_TEST(policy) > INetFwMgr_Release(manager); > > test_interfaces(); > - > + test_NetFwAuthorizedApplication(); > > CoUninitialize(); > } >

Nov. 29, 2016

Re: [v2 PATCH 1/2] hnetcfg: Improve INetFwAuthorizedApplication::get_ProcessImageFileName stub.

by Nikolay Sivov

On 29.11.2016 9:58, Alistair Leslie-Hughes wrote: > From: Michael Müller <michael(a)fds-team.de> > > v2: > Corrected error return values. > > Signed-off-by: Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> > --- > dlls/hnetcfg/apps.c | 23 +++++++++++++++++++++-- > 1 file changed, 21 insertions(+), 2 deletions(-) > > diff --git a/dlls/hnetcfg/apps.c b/dlls/hnetcfg/apps.c > index 2b8d78c..a89e6a7 100644 > --- a/dlls/hnetcfg/apps.c > +++ b/dlls/hnetcfg/apps.c > @@ -38,6 +38,7 @@ typedef struct fw_app > { > INetFwAuthorizedApplication INetFwAuthorizedApplication_iface; > LONG refs; > + BSTR filename; > } fw_app; > > static inline fw_app *impl_from_INetFwAuthorizedApplication( INetFwAuthorizedApplication *iface ) > @@ -60,6 +61,7 @@ static ULONG WINAPI fw_app_Release( > if (!refs) > { > TRACE("destroying %p\n", fw_app); > + if (fw_app->filename) SysFreeString( fw_app->filename ); There's no need for this check. > HeapFree( GetProcessHeap(), 0, fw_app ); > } > return refs; > @@ -252,7 +254,18 @@ static HRESULT WINAPI fw_app_get_ProcessImageFileName( > fw_app *This = impl_from_INetFwAuthorizedApplication( iface ); > > FIXME("%p, %p\n", This, imageFileName); > - return E_NOTIMPL; > + > + if (!imageFileName) > + return E_POINTER; > + > + if (!This->filename) > + { > + *imageFileName = NULL; > + return S_OK; > + } > + > + *imageFileName = SysAllocString( This->filename ); > + return *imageFileName ? S_OK : E_OUTOFMEMORY; > } You can reduce this to --- *imageFileName = SysAllocString( This->filename ); return *imageFileName || !This->filename ? S_OK : E_OUTOFMEMORY; --- > > static HRESULT WINAPI fw_app_put_ProcessImageFileName( > @@ -262,7 +275,12 @@ static HRESULT WINAPI fw_app_put_ProcessImageFileName( > fw_app *This = impl_from_INetFwAuthorizedApplication( iface ); > > FIXME("%p, %s\n", This, debugstr_w(imageFileName)); > - return S_OK; > + > + if (!imageFileName || !imageFileName[0]) > + return E_INVALIDARG; > + > + This->filename = SysAllocString( imageFileName ); > + return This->filename ? S_OK : E_OUTOFMEMORY; > } What happens if you call put_ProcessImageFileName() repeatedly?

Nov. 29, 2016

Re: hnetcfg: Improve INetFwAuthorizedApplication::get_ProcessImageFileName stub.

by Alistair Leslie-Hughes

On 24/11/16 07:22, Bruno Jesus wrote: > From: Michael Müller <michael(a)fds-team.de> > > Manually tested using vbscript. > > @@ -250,7 +252,18 @@ static HRESULT WINAPI fw_app_get_ProcessImageFileName( > fw_app *This = impl_from_INetFwAuthorizedApplication( iface ); > > FIXME("%p, %p\n", This, imageFileName); > - return E_NOTIMPL; > + > + if (!imageFileName) > + return E_INVALIDARG; The return value here is incorrect. I've sent an updated patch+tests to wine-patches Best Regards Alistair Leslie-Hughes

Nov. 29, 2016

Re: [PATCH 2/2] hnetcfg/tests: Add INetFwAuthorizedApplication tests

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=26959 Your paranoid android. === wxppro (32 bit policy) === policy.c:85: Test failed: got: 80070008 policy.c:89: Test failed: got: 80070003 policy.c:92: Test failed: got: 80070003 === w2003std (32 bit policy) === policy.c:85: Test failed: got: 80070008 policy.c:89: Test failed: got: 80070003 policy.c:92: Test failed: got: 80070003 === wvistau64 (32 bit policy) === policy.c:92: Test failed: got: 80070003 === w2008s64 (32 bit policy) === policy.c:92: Test failed: got: 80070003 === w7u (32 bit policy) === policy.c:92: Test failed: got: 80070003 === w7pro64 (32 bit policy) === policy.c:92: Test failed: got: 80070003 === w8 (32 bit policy) === policy.c:92: Test failed: got: 80070003 === w864 (32 bit policy) === policy.c:92: Test failed: got: 80070003 === w1064 (32 bit policy) === policy.c:92: Test failed: got: 80070003

Nov. 29, 2016

Re: [PATCH] wincodecs: Trigger conversion to target format in WriteSource()

by Dmitry Timoshkov

Vincent Povirk <madewokherd(a)gmail.com> wrote: > Signed-off-by: Vincent Povirk <vincent(a)codeweavers.com> Does this even fix a real application? What exactly this patch is supposed to fix? -- Dmitry.

Nov. 28, 2016

Re: [PATCH 4/5] d3d11/tests: Add test for shader input register limits.

by Józef Kucia

On Fri, Nov 25, 2016 at 4:05 PM, Henri Verbeet <hverbeet(a)gmail.com> wrote: > On 25 November 2016 at 12:17, Józef Kucia <jkucia(a)codeweavers.com> wrote: >> Signed-off-by: Józef Kucia <jkucia(a)codeweavers.com> >> --- >> >> The "ps_many_resources_code" shader bytecode is very big because texture >> resources cannot be indexed and the loop is unrolled. >> >> --- >> dlls/d3d11/tests/d3d11.c | 221 +++++++++++++++++++++++++++++++++++++++++++++++ >> 1 file changed, 221 insertions(+) >> > > This doesn't work here. In particular: > > fixme:d3d_shader:print_glsl_info_log Info log received from GLSL shader #5: > fixme:d3d_shader:print_glsl_info_log error: Too many fragment > shader texture samplers That's bad. Generally, SM4 hardware is expected to support 32 samplers. In order to prevent this OpenGL error, the number of samplers used in generated GLSL code could be limited to the number of samplers supported by OpenGL implementation. However, the test will fail when wined3d starts to correctly handle more than 16 samplers in pixel shaders. > ... > fixme:d3d:context_bind_shader_resources Shader 0x18bb10 needs 21 > samplers, but only 16 are supported. This one is expected. The two instances of "todo_wine" in the test are related to this issue.

Nov. 28, 2016

Re: [v3] cmd/tests: chaining rules of pipe and else

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=26946 Your paranoid android. === wvistau64 (32 bit) === batch.c:313: Test failed: unexpected char 0x27 position 0 in line 534 (got ''--a--c---'', wanted ''--a------'@or_broken@'--a--------'@or_broken@'%~ai'') === wvistau64 (64 bit) === batch.c:313: Test failed: unexpected char 0x27 position 0 in line 534 (got ''--a--c---'', wanted ''--a------'@or_broken@'--a--------'@or_broken@'%~ai'')

Nov. 28, 2016

[PATCH] apphelp: Stubs for GetPermLayers and SetPermLayers.

by Lauri Kenttä

Yes, I know this is wine-devel and not wine-patches. I'm just archiving this here as a future reference, in case someone wants to implement these functions later. I did some small tests, see the example below. I also tested getting the values after setting them manually through the "Compability" tab of file properties. Wine has all kinds of compability options, so it would be possible (in theory) to support these functions quite well. My test program: #include <windows.h> #include <stdio.h> BOOL WINAPI GetPermLayers(const WCHAR* wszPath, WCHAR* wszBuffer, DWORD* pdwBytes, DWORD dwFlags); BOOL WINAPI SetPermLayers(const WCHAR* wszPath, const WCHAR* wszBuffer, BOOL bGlobal); /* WCHAR helpers */ void wc(WCHAR* dest, const char* src) { while (*src) *dest++ = *src++; *dest++ = *src++; } void cw(char* dest, const WCHAR* src) { while (*src) *dest++ = *src++; *dest++ = *src++; } const char* c(const WCHAR* src) { static char buf[4][1024]; static int i; char* dest = buf[i++ % 4]; cw(dest, src); return dest; } int main() { WCHAR buf[64]; WCHAR name[64]; DWORD size; BOOL ret; wc(name, "C:\\a.exe"); wc(buf, "640x480 lol fail"); ret = SetPermLayers(name, buf, 0); printf("ret %x\n", ret); /* ret 1 */ wc(buf, "DEADBEEF"); size = 0; ret = GetPermLayers(name, buf, &size, 3); printf("ret %x, size %d, buf '%s'\n", ret, size, c(buf)); /* ret 0, size 34, buf 'DEADBEEF' */ wc(buf, "DEADBEEF"); size = 2; ret = GetPermLayers(name, buf, &size, 3); printf("ret %x, size %d, buf '%s'\n", ret, size, c(buf)); /* ret 0, size 34, buf 'DEADBEEF' */ wc(buf, "DEADBEEF"); size = sizeof(buf); ret = GetPermLayers(name, buf, &size, 3); printf("ret %x, size %d, buf '%s'\n", ret, size, c(buf)); /* ret 1, size 34, buf '640x480 lol fail'; works just like 640X480, though */ wc(buf, "640X480"); ret = SetPermLayers(name, buf, 0); printf("ret %x\n", ret); /* ret 1 */ wc(buf, "DEADBEEF"); size = sizeof(buf); ret = GetPermLayers(name, buf, &size, 3); printf("ret %x, size %d, buf '%s'\n", ret, size, c(buf)); /* ret 1, size 16, buf '640X480' */ wc(buf, ""); ret = SetPermLayers(name, buf, 0); printf("ret %x\n", ret); /* ret 1 */ wc(buf, "DEADBEEF"); size = sizeof(buf); ret = GetPermLayers(name, buf, &size, 3); printf("ret %x, size %d, buf '%s'\n", ret, size, c(buf)); /* ret 0, size 0, buf 'DEADBEEF' */ return 0; } Signed-off-by: Lauri Kenttä <lauri.kentta(a)gmail.com> --- dlls/apphelp/apphelp.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/dlls/apphelp/apphelp.c b/dlls/apphelp/apphelp.c index 0ff74aa..7a09052 100644 --- a/dlls/apphelp/apphelp.c +++ b/dlls/apphelp/apphelp.c @@ -117,3 +117,23 @@ void WINAPI SdbGetAppPatchDir(HSDB hsdb, WCHAR *path, DWORD size) FIXME("stub: %p %p %d\n", hsdb, path, size); if (size && path) *path = 0; } + +BOOL WINAPI GetPermLayers(const WCHAR *path, WCHAR *buffer, DWORD *bytes, DWORD flags) +{ + /* Get the compability mode (Windows version, 640x480 etc.) of path.exe + * + * path = absolute(?) path to the executable + * buffer = where to put the result + * bytes = buffer size; if it's too small, set to required size and return FALSE. + * flags = local(1) | global(2). + */ + FIXME("stub: %s %p %p %x\n", debugstr_w(path), buffer, bytes, flags); + return FALSE; +} + +BOOL WINAPI SetPermLayers(const WCHAR *path, const WCHAR *buffer, BOOL global) +{ + /* Set the compability mode (Windows version, 640x480 etc.) for path.exe */ + FIXME("stub: %s %s %x\n", debugstr_w(path), debugstr_w(buffer), global); + return FALSE; +} -- 2.10.2

Nov. 28, 2016

Re: [PATCH 3/3] wintrust/tests: Remove a test that crashes on Windows 8 amd 10.

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=26944 Your paranoid android. === w1064 (64 bit softpub) === softpub.c:254: Test failed: Expected S_OK, got 00000001 softpub.c:257: Test failed: Expected S_OK, got 00000001

Nov. 28, 2016

Re: [PATCH 2/3] wintrust/tests: Initialize the pSigState member to prevent Windows 8 and 10 from crashing.

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=26943 Your paranoid android. === w8 (32 bit softpub) === softpub: unhandled exception c0000005 at 7511A63B === w864 (32 bit softpub) === softpub: unhandled exception c0000005 at 7772A63B === w1064 (32 bit softpub) === softpub: unhandled exception c0000005 at 768B0B90 === w864 (64 bit softpub) === softpub: unhandled exception c0000005 at 00007FFB7AFF9B8F === w1064 (64 bit softpub) === softpub.c:254: Test failed: Expected S_OK, got 00000001 softpub.c:257: Test failed: Expected S_OK, got 00000001 softpub: unhandled exception c0000005 at 00007FFC77A7F593

Nov. 28, 2016