November 2015 - Wine-devel - list.winehq.org

Request for help from Conservancy - be nice if we could respond
by Jeremy White 26 Nov '15

26 Nov '15

Hey Folks, The Software Freedom Conservancy is making a strong push to move away from dependence on large corporate donors to be sustained by more individual donations. They are asking individual projects to do a bit of advertising for them. The Conservancy has been nothing but good for us and to us; if we can help spread this word, I think that would be a good. So please share their plea, which is here: http://sfconservancy.org/supporter/ If someone found a good way to insert that into the web site as well, that might be a nice thing as well. Cheers, Jeremy -------- Forwarded Message -------- Subject: Conservancy fundraiser launching on Monday Date: Sun, 22 Nov 2015 00:46:16 -0500 From: Karen Sandler <karen(a)sfconservancy.org> Reply-To: karen(a)sfconservancy.org To: project-reps(a)sfconservancy.org Hi project reps! I wanted to give you all a heads up that Conservancy is launching a big fundraiser on Monday. As many of you know, it's been difficult to find financial stability for Conservancy... the 10% that you all contribute is very much appreciated but doesn't cover even one staffer plus basic overhead and the services we provide are greater than our peer fiscal sponsorship organizations. We need additional support to be able to survive, and while we've relied on corporate donations to do that (and will continue to seek that support) it's too capricious to serve as our primary funding source. We've concluded that we need a base of individual supporters we can rely on annually in order to close that gap. We launched our Supporter program last December but without an urgent message behind it we've been unable to get the number of donors that we need. We have structured the campaign with two levels - a lower level that will sustain the organization for a 'bare minimum' service plan to you, our member projects, and a separate, higher level to continue doing license enforcement and other more complex legal work in service to our projects. While our campaign focuses on GPL enforcement, we know many of you are not invested in copyleft at all. Not meeting this goal means we'll scale back much of the licensing, trademark and patent-related services we provide. The message we have for the campaign frames the problem as urgent, but the urgency relates to the next fiscal year (since that's how we plan our budget). Given the commitments we've lined up we have about a year to reassess if the fundraiser fails, and we'll have time to work with you to figure out how we transition to the next phase, whether it's reducing our services or moving to a volunteer model relying a lot more on all of you. We hope you will help us to meet our goal. Our effective number, given the small base we already have signed up as supporters, will be around 500 new sign ups. If each project can get just 15-20 people to sign up for just $10 per month, we'll meet the bottom goal easily. Conservancy still gets a lot done with very little, so our urgent needs really are attainable if you all help us get there. Please promote the campaign when it launches, blog and microblog about your positive experiences with Conservancy and encourage your contributors and users to sign up. Thanks so much for all of your hard work, and I hope we can continue with uninterrupted support for your excellent projects which inspire us every day! karen Karen M. Sandler Executive Director, Software Freedom Conservancy __________ Become a Supporter today! http://sfconservancy.org/supporter/

3 2

Re: Re: ieframe: Don't release the advise sink in IOleObject::SetClientSite.
by Marvin 26 Nov '15

26 Nov '15

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=18568 Your paranoid android. === build (build) === Patch failed to apply

1 0

Re: ieframe: Don't release the advise sink in IOleObject::SetClientSite.
by Jacek Caban 26 Nov '15

26 Nov '15

Hi Hans, On 11/26/15 10:53, Hans Leidekker wrote: > diff --git a/dlls/ieframe/tests/webbrowser.c b/dlls/ieframe/tests/webbrowser.c > index c8dfa5b..f62a943 100644 > --- a/dlls/ieframe/tests/webbrowser.c > +++ b/dlls/ieframe/tests/webbrowser.c > @@ -148,6 +148,7 @@ DEFINE_EXPECT(ControlSite_TranslateAccelerator); > DEFINE_EXPECT(OnFocus_TRUE); > DEFINE_EXPECT(OnFocus_FALSE); > DEFINE_EXPECT(GetExternal); > +DEFINE_EXPECT(AdviseSink_Release); > > static const WCHAR wszItem[] = {'i','t','e','m',0}; > > @@ -3955,6 +3956,7 @@ static ULONG WINAPI sink_AddRef(IAdviseSink *iface) > > static ULONG WINAPI sink_Release(IAdviseSink *iface) > { > + CHECK_EXPECT2(AdviseSink_Release); > return 1; > } IMO checking GetAdvise result after SetClientSite() call would be cleaner than checking Release calls. > @@ -4003,6 +4005,7 @@ static void test_SetAdvise(void) > IWebBrowser2 *browser; > IViewObject2 *view; > IAdviseSink *sink; > + IOleObject *oleobj; > DWORD aspects, flags; > > if (!(browser = create_webbrowser())) return; > @@ -4031,9 +4034,22 @@ static void test_SetAdvise(void) > ok(!flags, "got %08x\n", aspects); > ok(sink == &test_sink, "got %p\n", sink); > > + hr = IWebBrowser2_QueryInterface(browser, &IID_IOleObject, (void **)&oleobj); > + ok(hr == S_OK, "got %08x\n", hr); > + > + SET_EXPECT(GetContainer); > + SET_EXPECT(Site_GetWindow); > + SET_EXPECT(Invoke_AMBIENT_OFFLINEIFNOTCONNECTED); > + SET_EXPECT(Invoke_AMBIENT_SILENT); > + hr = IOleObject_SetClientSite(oleobj, &ClientSite); > + ok(hr == S_OK, "got %08x\n", hr); You can't do it that way. All SET_EXPECT()s need CHECK_CALLED() or alike. Otherwise you leave tests in an inconsistent state. > + > + SET_EXPECT(AdviseSink_Release); > hr = IViewObject2_SetAdvise(view, 0, 0, NULL); > ok(hr == S_OK, "got %08x\n", hr); > + CHECK_CALLED(AdviseSink_Release); It would be also interesting to see what happens in SetClientSite(NULL) call. Thanks, Jacek

1 0

Re: [PATCH] wininet/test: Add tests on host value for InternetConnectW (try 3)
by Marvin 26 Nov '15

26 Nov '15

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=18567 Your paranoid android. === w1064 (32 bit http) === http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1 http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1 http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1 http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1 http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1 http.c:1773: Test failed: Expected 2, got 4 http.c:1774: Test failed: Expected CRLF, got ' http.c:2408: Test failed: count was wrong http.c:2409: Test failed: http data wrong, got: Garbage: Header http.c:3072: Test failed: HttpAddRequestHeaders succeeded http.c:3073: Test failed: Expected error ERROR_HTTP_HEADER_NOT_FOUND, got 0 http.c:3078: Test failed: code = 200, expected 400 http.c:3078: Test failed: code = 200, expected 400 http.c:3078: Test failed: unexpected status code 200, expected 400 http.c:3078: Test failed: unexpected status code 200, expected 400 http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1

1 0

Re: [PATCH] wininet/test: Add tests on host value for InternetConnectW
by Marvin 26 Nov '15

26 Nov '15

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=18566 Your paranoid android. === w1064 (32 bit http) === http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1 http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1 http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1 http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1 http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1 http.c:1773: Test failed: Expected 2, got 4 http.c:1774: Test failed: Expected CRLF, got ' http.c:2408: Test failed: count was wrong http.c:2409: Test failed: http data wrong, got: Garbage: Header http.c:3072: Test failed: HttpAddRequestHeaders succeeded http.c:3073: Test failed: Expected error ERROR_HTTP_HEADER_NOT_FOUND, got 0 http.c:3078: Test failed: code = 200, expected 400 http.c:3078: Test failed: code = 200, expected 400 http.c:3078: Test failed: unexpected status code 200, expected 400 http.c:3078: Test failed: unexpected status code 200, expected 400 http.c:534: Test failed: HTTP_QUERY_RAW_HEADERS: expected length 0, but got 1

1 0

Re: ws2_32: Update WS_get_local_ips to not return a bogus IP address.
by Bruno Jesus 26 Nov '15

26 Nov '15

On Sat, Nov 21, 2015 at 2:12 PM, Jamie Taylor <Jamie.Taylor(a)pobox.com> wrote: > There were many shortcomings in the previous behavior, but most > importantly it resulted in the first IP address in the > returned hostent structure being the magic loopback address in > the default network configuration on Mac OS X. The new behavior > returns all IP addresses from non-loopback network adapters, and > ensures that the addresses from the adapter for the default route > are at the front of the list. Testing confirmed this behavior > on Mac OS X. > > This will most likely fix https://bugs.winehq.org/show_bug.cgi?id=37271 > and https://bugs.winehq.org/show_bug.cgi?id=22819 > but I do not have the programs under test in those bugs, so I can't > verify. > > I also do not have convenient access to a linux machine with an > equivalent network configuration (i.e., at least two interfaces > with default routes) to test. Hi, Jamie. Thanks for your effort, as you may have observed in the patches list your patch is marked as deferred because we are in code freeze for Wine 1.8 release. This code was touched a few times in attempts to fix this same issue you are addressing so maybe this time it will be the definitive fix =) I have a similar (not so robust) old patch that I used on FreeBSD (and also works in Linux), could you try it in OSX just out of curiosity, please? It has printf to help showing the results so a make test in the ws2_32/tests folder is enough to check if it works. About your patch I just would like to say that you should follow the surrounding code style, so the { goes in the next line. I tested it on Linux and it works for Heroes of Might & Magic 2 (https://bugs.winehq.org/show_bug.cgi?id=22819) Best wishes, Bruno

3 10

Re: [2/2] oleaut32: Implement a better stub for IPicture::SaveAsFile.
by Alexandre Julliard 25 Nov '15

25 Nov '15

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > Based on OLEPictureImpl_Save implementation. Surely you don't need to duplicate all of it. -- Alexandre Julliard julliard(a)winehq.org

3 7

Re: [PATCH 1/5] d3d10core/tests: Multisampled render targets are zeroed on creation.
by Henri Verbeet 25 Nov '15

25 Nov '15

On 24 November 2015 at 22:18, Stefan Dösinger <stefan(a)codeweavers.com> wrote: > +static void test_multisample_init(void) > +{ > + D3D10_TEXTURE2D_DESC desc; > + ID3D10Texture2D *single, *multi; > + ID3D10Device *device; > + ULONG refcount; > + DWORD color; > + HRESULT hr; > + unsigned int x, y; > + struct texture_readback rb; > + BOOL all_zero = TRUE; > + > + if (!(device = create_device())) > + { > + skip("Failed to create device, skipping tests.\n"); > + return; > + } > + > + desc.Width = 640; > + desc.Height = 480; > + desc.MipLevels = 1; > + desc.ArraySize = 1; > + desc.Format = DXGI_FORMAT_R8G8B8A8_UNORM; > + desc.SampleDesc.Count = 2; > + desc.SampleDesc.Quality = 0; > + desc.Usage = D3D10_USAGE_DEFAULT; > + desc.BindFlags = D3D10_BIND_RENDER_TARGET; > + desc.CPUAccessFlags = 0; > + desc.MiscFlags = 0; > + hr = ID3D10Device_CreateTexture2D(device, &desc, NULL, &multi); > + ok(SUCCEEDED(hr), "Failed to create texture, hr %#x.\n", hr); > + You should check if the sample count is supported with CheckMultisampleQualityLevels(). > + /* CopyResource from a multisampled to a sysmem texture works, but leaks memory on Windows > + * (Nvidia, not checked on AMD). Use a single sampled buffer to work around. */ Direct3D 10+ doesn't have sysmem textures. You can make resources accessible to the CPU, but you can't prevent the GPU from reading from them. How did you verify that the CopyResource() works? > + desc.SampleDesc.Count = 1; > + hr = ID3D10Device_CreateTexture2D(device, &desc, NULL, &single); > + ok(SUCCEEDED(hr), "Failed to create texture, hr %#x.\n", hr); Note that you're creating a texture here, contrary to what the comment above claims. > + ID3D10Device_CopyResource(device, (ID3D10Resource *)single, (ID3D10Resource *)multi); At least according to the documentation, CopyResource() doesn't work at all with multi-sampled resources, and this should use ResolveSubresource() instead. It's possible the documentation is wrong about that, but that would need a more convincing test. Right now it looks like the test only tests that the initial contents of "single" are zeroes.

2 1

Re: [PATCH 4/5] dxgi/tests: Test DXGI_SWAP_EFFECT_FLIP_DISCARD.
by Henri Verbeet 25 Nov '15

25 Nov '15

On 24 November 2015 at 22:18, Stefan Dösinger <stefan(a)codeweavers.com> wrote: > - {TRUE, 0, 4 /* undefined */, DXGI_ERROR_INVALID_CALL, DXGI_ERROR_INVALID_CALL, 0}, > - {TRUE, 1, 4 /* undefined */, DXGI_ERROR_INVALID_CALL, DXGI_ERROR_INVALID_CALL, 0}, > - {TRUE, 2, 4 /* undefined */, DXGI_ERROR_INVALID_CALL, DXGI_ERROR_INVALID_CALL, 0}, > + {TRUE, 0, DXGI_SWAP_EFFECT_FLIP_DISCARD, DXGI_ERROR_INVALID_CALL, DXGI_ERROR_INVALID_CALL, 0}, > + {TRUE, 1, DXGI_SWAP_EFFECT_FLIP_DISCARD, DXGI_ERROR_INVALID_CALL, DXGI_ERROR_INVALID_CALL, 0}, > + {TRUE, 2, DXGI_SWAP_EFFECT_FLIP_DISCARD, S_OK, DXGI_ERROR_INVALID_CALL, 0}, > + {TRUE, 0, 5 /* undefined */, DXGI_ERROR_INVALID_CALL, DXGI_ERROR_INVALID_CALL, 0}, > + {TRUE, 1, 5 /* undefined */, DXGI_ERROR_INVALID_CALL, DXGI_ERROR_INVALID_CALL, 0}, > + {TRUE, 2, 5 /* undefined */, DXGI_ERROR_INVALID_CALL, DXGI_ERROR_INVALID_CALL, 0}, I don't mind too much, but I'm not a real fan of testing undefined values like that, at least until applications start depending on an error value being returned. It's not that much of a stretch that "5", like "4", will become a valid value with a defined meaning in a future dxgi version.

1 0

Re: [PATCH 2/6 v2] msxml3: Update xmlNode reference count when nodes are merged in node_insert_before
by Nikolay Sivov 25 Nov '15

25 Nov '15

Let's wait until 1.8 with the rest.

1 0