October 2008 - Wine-Devel - list.winehq.org

Re: Opengl32: add a test case for WGL_ARB_create_context/opengl3 [try2]
by Christoph Frick Oct. 10, 2008

Oct. 10, 2008

On Fri, Oct 10, 2008 at 10:42:05AM +0200, Roderick Colenbrander wrote: > + /* Try to create an OpenGL 3.0 contet */ content? context? -- cu

1 0

Re: Opengl32: add a test case for WGL_ARB_create_context/opengl3 [try2]
by Paul Vriens Oct. 10, 2008

Oct. 10, 2008

Roderick Colenbrander wrote: > Hi, > > This is a resubmit of the same patch but not part of a series which contained a failing patch, so it should pass patchwatcher. > > Regards, > Roderick Colenbrander > > > > > ------------------------------------------------------------------------ > > Hi Roderick, Why did you add all the (unneeded) indentation? -- Cheers, Paul.

2 1

Re: Resend: mshtml.dll:Add implementation HTMLDocument_get_referrer
by Jacek Caban Oct. 10, 2008

Oct. 10, 2008

Ivan Sinitsin wrote: > Add implementation of HTMLDocument_get_referrer function, in unit `htmldoc.c` > (mshtml.dll). > > Changelog: > Add implementation of HTMLDocument_get_referrer function. > > > > + if(!This->nscontainer) > + return E_FAIL; > + > + nsres = nsIWebNavigation_GetDocument(This->nscontainer->navigation, &nsdoc); > + if(NS_FAILED(nsres) || (!nsdoc)) { > + ERR("GetDocument failed: %08x\n", nsres); > + return E_FAIL; > + } > + > + nsIDOMDocument_QueryInterface(nsdoc, &IID_nsIDOMHTMLDocument, (void**)&nshtmldoc); > + nsIDOMDocument_Release(nsdoc); > + Please use nsdoc stored in HTMLDocument. It currently doesn't make much difference, but with changes from my tree it will. Jacek

1 0

d3d: Bug in handling of depth-stencil surfaces, and potential patch
by Jim Cameron Oct. 10, 2008

Oct. 10, 2008

See bug 10758. If you create a depth-stencil surface with Direct3D 9 on Windows, select it into a D3D9 device with SetDepthStencilSurface() and then release it (so its reference count goes to zero), the surface will not be deleted. It will persist until its external reference count is zero /and/ it has been deselected with SetDepthStencilSurface() or Reset(). However, Wine's D3D implementation will delete the surface as soon as its refcount goes to zero, even if it is still selected into the device. This causes a crash with Myst IV. Attached is a patch that adds a private reference count to IDirect3DSurface9_Impl, plus a test case for it. The patch fixes the problem with Myst IV; could one of the resident D3D gurus have a look at it and see if I've got the approach right or if there's something else we should be doing? We'll probably also have to make the equivalent change in D3D8 as well, but I haven't considered that. Cheers, jim

2 1

Re: DrawTextEx parameter return value fix
by James Hawkins Oct. 10, 2008

Oct. 10, 2008

On Thu, Oct 9, 2008 at 9:13 PM, Dean <dean(a)dnaml.com> wrote: > What are you trying to fix? You need to add a test to the test suite to verify that this change is correct. -- James Hawkins

1 0

Re: configure.ac: add -Werror to default compiler options
by Dan Kegel Oct. 9, 2008

Oct. 9, 2008

Juan wrote: > Since Rob was noncommittal about whether -Werror was a good or a bad > idea, I'll take a stance: I think it's a bad one. ... > some warnings are just plain wrong. In some cases the "fix" > to quiet the warning is uglier than the warning. In those cases, we can disable the warning. > Having a code base that has hundreds of warnings is also a problem, > but we're fortunate enough not to be in that situation. Are you sure about that? When I configure with -Wall -Werror, I can't even configure properly; one gets the errors configure: libxrandr development files not found, XRandr won't be supported. configure: libxinerama development files not found, multi-monitor setups won't be supported. configure: WARNING: Old Mesa headers detected. Consider upgrading your Mesa libraries. OpenGL and Direct3D won't be supported. And if one blows past those, "make depend" fails. In fact, everything fails pretty spectacularly. We should probably have a look at -Wall errors in more detail and breadth before we assume that we're ok. - Dan

8 17

Re: [PATCH] Revert "comctl32: Fix a test on locales where default first day of week is not the max value."
by James Hawkins Oct. 9, 2008

Oct. 9, 2008

On Thu, Oct 9, 2008 at 4:27 PM, Marcus Meissner <marcus(a)jet.franken.de> wrote: > > Hi, > > This is breaking my "make test" run for 2 weeks now. > (LANG=de_DE.utf8) > > This reverts commit 2f8830d55b71d2fb8be912c7041d7adab7580391. > --- > dlls/comctl32/tests/monthcal.c | 3 +-- > 1 files changed, 1 insertions(+), 2 deletions(-) > > diff --git a/dlls/comctl32/tests/monthcal.c b/dlls/comctl32/tests/monthcal.c > index 6f2ae70..271dfd1 100644 > --- a/dlls/comctl32/tests/monthcal.c > +++ b/dlls/comctl32/tests/monthcal.c > @@ -673,8 +673,7 @@ static void test_monthcal_firstDay(HWND hwnd) > if (i == -1){ > expect(MAKELONG(fday, FALSE), res); > }else if (i >= 7){ > - /* out of range sets max first day of week, locale is ignored */ > - expect(MAKELONG(6, TRUE), res); > + expect(MAKELONG(fday, TRUE), res); > }else{ > expect(MAKELONG(i, TRUE), res); > } Reverting a change that makes the tests pass on all platforms is not the correct fix. Fixing the implementation to match the tests is the correct fix. Please don't apply this. I've sent a patch that fixes the problem for other locales. -- James Hawkins

1 0

Need help with the quartz filtermapper tests
by Juan Lang Oct. 9, 2008

Oct. 9, 2008

Hi Alexander, my recent patch to improve the GUIDs generated by rpcrt4[1] causes a failure in the quartz filtermapper tests. Specifically, I get the following error: quartz:filtermapper.c:142: Test failed: EnumMatchingFilters failed to return the test filter 2 The trouble is, I can't imagine how changing a generated GUID from one random sequence of bytes to another random sequence of bytes should cause this to fail. On Dan's suggestion, I inserted a Sleep(2000); in UuidCreate, and that causes the tests to pass once again. So, I suspect there's a race condition somewhere in quartz or the tests. I really know nothing at all about what quartz does or what the tests are trying to verify. Might you be able to shed some insight? Thanks, --Juan [1]: http://www.winehq.org/pipermail/wine-patches/2008-October/062833.html

1 0

Favor: review refactor of gdi32/freetype.c - AddFontToList
by Jeremy White Oct. 9, 2008

Oct. 9, 2008

So in my ongoing quest to get CreateScalableFontResource completed, I believe that I need to refactor gdi32/freetype.c - AddFontToList to expose an interface to allow me to retrieve a Face *. Attached is a series of 4 patches that accomplish what I believe to be a reasonable refactoring of that function. >From my analysis, the only material change in the logic flow is that the old code would skip a face that existed prior to building the new Face * structure. The new code goes ahead and builds up a proposed Face * before testing to see if the proposed new Face should be skipped. It passes the font tests and notepad still runs, but I recognize that this is a much more invasive change than I am really qualified to make. I would appreciate others with expertise in this area reviewing this series. Also, if there is advice on a standard bevy of tests I could run, I would appreciate it. There is a fair amount of special case logic in this function, and I know that our regression tests do not exercise it all. Thanks, Jeremy

1 0

Re: [PATCH 3/4] msi: Add initial implementation of MsiDatabaseMerge, with tests.
by Paul Vriens Oct. 9, 2008

Oct. 9, 2008

James Hawkins wrote: > --- > dlls/msi/database.c | 497 ++++++++++++++++++++++++++++++++++++++++++++++++++- > dlls/msi/msipriv.h | 2 + > dlls/msi/record.c | 71 ++++++++ > dlls/msi/tests/db.c | 449 +++++++++++++++++++++++++++++++++++++++++++++- > 4 files changed, 1014 insertions(+), 5 deletions(-) > > > ------------------------------------------------------------------------ > > Hi James, There seems to be 1 test failure introduced on W2K and below: http://test.winehq.org/data/tests/msi:db.html Could you have a look? -- Cheers, Paul.

1 0