January 2009 - Wine-devel - list.winehq.org

Re: urlmon: Skip more tests on old IE and ignore BINDSTATUS_PROXYDETECTING notification.
by Paul Vriens 13 Jan '09

13 Jan '09

Jacek Caban wrote: > --- > dlls/urlmon/tests/url.c | 157 > +++++++++++++++++++++++++---------------------- > 1 files changed, 84 insertions(+), 73 deletions(-) > > > > ------------------------------------------------------------------------ > > Hi Jacek, As of this patch we skip running most of the tests on Wine (and Windows). I only see a few W2K boxes that run the full tests for url.c (+/- 3800 tests). The rest only run 137 tests -- Cheers, Paul.

2 3

Re: [2/5] wined3d: Fix texture creation error handling.
by Roderick Colenbrander 13 Jan '09

13 Jan '09

Hi Henri, Just a suggestion for a future patch. The current autogenmipmap code isn't correct. Even when this capability isn't supported the function might not return D3DERR_INVALIDCALL IF CheckDeviceFormat returns D3DOK_NOAUTOGEN for this usage flag. I stumbled on this when I started improving our caps code but didn't have time to fix it yet. Roderick -- Sensationsangebot verlängert: GMX FreeDSL - Telefonanschluss + DSL für nur 16,37 Euro/mtl.!* http://dsl.gmx.de/?ac=OM.AD.PD003K1308T4569a

1 0

Re: [3/5] WineD3D: Put vertex shader duplication infrastructure in place
by Henri Verbeet 13 Jan '09

13 Jan '09

2009/1/13 Stefan Dösinger <stefan(a)codeweavers.com>: > - This->baseShader.is_compiled = TRUE; > - This changes code in pixelshader.c, looks unrelated. > +void find_vs_compile_args(IWineD3DVertexShaderImpl *shader, IWineD3DStateBlockImpl *stateblock, struct vs_compile_args *args) { > + args->fog_src = stateblock->renderState[WINED3DRS_FOGTABLEMODE] == WINED3DFOG_NONE ? VS_FOG_COORD : VS_FOG_Z; > + args->swizzle_map = ((IWineD3DDeviceImpl *)shader->baseShader.device)->strided_streams.swizzle_map; > +} You need to mask swizzle_map with use_map like in the original code, or you'll end up creating new shaders if an attribute that was previously swizzled is undefined now. > + shader->gl_shaders = HeapReAlloc(GetProcessHeap(), 0, old_array, > + (shader->num_gl_shaders + 1) * sizeof(*shader->gl_shaders)); Please don't grow arrays by a single element like that. I noticed find_gl_pshader() does that as well, please fix. > - * > + * You're adding a trailing space here.

1 0

Re: [dxdiagn] Implemented retrieval of b3DAccelerationEnabled property
by Vitaliy Margolen 12 Jan '09

12 Jan '09

Markus wrote: > > ------------------------------------------------------------------------ > > + HKEY key_handle; > + LPCSTR searchSubKey = "SOFTWARE\\Microsoft\\DirectDraw"; > + LPCSTR dataSubKey = "EmulationOnly"; I don't see anywhere Wine using this key. How did you came up with this key? > + V_VT(&v) = VT_BOOL; V_BOOL(&v) = get_ddraw_acceleration(); > + IDxDiagContainerImpl_AddProp( pDisplayAdapterSubCont, b3DAccelerationEnabled, &v ); > + VariantClear(&v); Why don't you use add_prop_bool()? Vitaliy

3 2

Re: [PATCH 1/6] wininet: Make sure that we have data buffered before sending INTERNET_STATUS_REQUEST_COMPLETE notification.
by Hans Leidekker 12 Jan '09

12 Jan '09

On Monday 12 January 2009 18:19:54 Jacek Caban wrote: > +static void HTTP_ReceiveRequest(WININETHTTPREQW *req) > +{ > + INTERNET_ASYNC_RESULT iar; > + BYTE buffer[4048]; I think HTTP_ReceiveRequestData would be a better name for this function. And I guess this is also a good opportunity to round that buffer size up to the nearest power of two. -Hans

2 2

Re: comctl32: fix Itemtext in treeview.c (Bug #16466)
by Austin English 12 Jan '09

12 Jan '09

On Sun, Jan 11, 2009 at 10:03 AM, Nerv <nerv(a)dawncrow.de> wrote: > This fixes TREEVIEW_DoSetItemT if tvItem->pszText=0 > > > > Please use your full real name on patches. A testcase would help. -- -Austin

1 0

Comctl32 tests on version 6.0
by Nikolay Sivov 12 Jan '09

12 Jan '09

Running crossbuilt tests for comctl32 showed some failures when version 6.0 used: 27 - imagelist 25 - rebar 130 - toolbar (version 6.0 module used through external manifest xml in my test) Toolbar and rebar failures show different rectangle coordinates (all toolbar fails are about -2 shift of y coord). I've some questions: 0) What way is better to implement this 6.0 module - using WinSxS creating another separate version 6 and act like Win, or extend existing module (with version checking probably see next question)? 1) the main is about imagelist: why are we relying here on internal storage format? Tests show that on 6.0 this format's changed. If the reason is to get the same structure when writing to IStream (to be compatible with streams saved on native platform) here's another question => 2) Since wine module should be able to support "all" versions of native ones we need to run tests on both versions on Windows. Is there any way to do so easily maybe playing with activation contexts? The reason why I'm interested in that is the implementation of interfaces IImageList (followed by IImageList2) and related calls HIMAGELIST_QueryInterface from comctl32 (6.0) SHGetImageList from shell32 All this comes from http://bugs.winehq.org/show_bug.cgi?id=15885 Any thoughts are appreciated. Nikolay

2 1

Failing wine tests
by Rico Schüller 12 Jan '09

12 Jan '09

Hi, I've just done 4 runs on the same machine with the same wine version with the same test version (see http://test.winehq.org/data/8cf8e2c60a263e0b015b36dc36c9bf83e6767aa8/#group… - the rs-f10-...). Therefore I named the runs like: inst - test run from a wine version with an installed wine (make install) noinst - test run from a not installed wine version (run directly from the built dir) lo - (links oben) - which means the window from the winetest.exe is on the upper left side on the desktop (default position) ru - (rechts unten) - which means the window from the winetest.exe is on the lower right side on the desktop (just moved the window before I hit start) My result is, that some tests fail/ behave different on the some runs: The main difference is that two tests fail because the winetest.exe window is at the wrong position. These are: user32:msg and riched20:editor. These tests fail when the winetest.exe window as in the upper left corner. Any ideas? The next difference is that some fonts are not found if wine is run from the build directory. Running the installed version these test succeed: user32:combo , gdi32:font and comctl:rebar. Has anyone an idea where this comes from? (Probably a path problem?) There is also a test comctl32:treeview which has 6 todos more if it is run from the build directory. Also note: user32:msg and the user32:win test run a different number of tests each time. Is this intended? Cheers Rico

1 0

Re: [1/2] httpapi: add HttpInitialize and HttpTerminate stubs
by Austin English 12 Jan '09

12 Jan '09

On Mon, Jan 12, 2009 at 11:33 AM, Andrey Turkin <andrey.turkin(a)gmail.com> wrote: > > .Net 3.0 installer needs these > --- > dlls/httpapi/httpapi.spec | 2 + > dlls/httpapi/httpapi_main.c | 42 +++++++++++++++++++++++++++++++++++++ > include/http.h | 48 +++++++++++++++++++++++++++++++++++++++++++ > 3 files changed, 92 insertions(+), 0 deletions(-) > create mode 100644 include/http.h > > > > > > diff --git a/dlls/httpapi/httpapi.spec b/dlls/httpapi/httpapi.spec index aa8fea2..1033eb8 100644 --- a/dlls/httpapi/httpapi.spec +++ b/dlls/httpapi/httpapi.spec @@ -18,6 +18,7 @@ @ stub HttpFilterRawWriteAndAppRead @ stub HttpFlushResponseCache @ stub HttpGetCounters +@ stdcall HttpInitialize(long long ptr) @ stub HttpInitialize @ stub HttpInitializeServerContext @ stub HttpOpenAppPool You forgot to remove the stub entry. @@ -48,5 +49,6 @@ @ stub HttpShutdownAppPool @ stub HttpShutdownFilter @ stub HttpTerminate +@ stdcall HttpTerminate(long ptr) @ stub HttpWaitForDemandStart @ stub HttpWaitForDisconnect Same. -- -Austin

2 1

gecko regressions; debug build of gecko?
by Dan Kegel 12 Jan '09

12 Jan '09

I've been doing some random qa of popular apps lately, and think I've found some regressions caused by the gecko update: http://bugs.winehq.org/show_bug.cgi?id=16869 - Sketchup 7 crashes when bringing up 3d Warehouse http://bugs.winehq.org/show_bug.cgi?id=16892 - Yahoo Messenger 8.1 crashes after it puts up the instant message window Hmm, can we have a debug version of the gecko download so we can give more useful backtraces?

3 5