May 2010 - Wine-Devel - list.winehq.org

Re: yesterday's benchmark results: xubuntu makes 3dmark happier
by Dan Kegel May 21, 2010

May 21, 2010

On Fri, May 21, 2010 at 7:10 AM, Erich Hoover <ehoover(a)mines.edu> wrote: > Do you run your normal-Ubuntu tests with Compiz enabled? Apparently with > the nVidia drivers in particular that can make a big difference: > > http://www.phoronix.com/scan.php?page=article&item=compiz_speed_test&num=5 No, I think I had that solidly disabled for all runs. It wouldn't hurt for my script to check all these things, though.

1 0

yesterday's benchmark results: xubuntu makes 3dmark happier
by Dan Kegel May 21, 2010

May 21, 2010

Yesterday's results were 8 hours late because of an internet outage here. http://kegel.com/wine/yagmarkdata/00index.html http://kegel.com/wine/yagmarkdata/E8400-GT_220-Ubuntu_10.04_LTS-e8400-wine-… Looking at the graphs, it seems that all benchmarks are happier under xubuntu with nonessential services shut down as described yesterday. 3dmark (all versions) is affected lots more than Heaven.

1 0

Re: [2/2] winemenubuilder: add thumbnail mode for .lnk files
by Hans Leidekker May 21, 2010

May 21, 2010

On Fri, 2010-05-21 at 07:52 +0200, Damjan Jovanovic wrote: > --- winemenubuilder00.c 2010-05-21 07:44:48.195864397 +0200 > +++ winemenubuilder.c 2010-05-21 07:47:25.591878489 +0200 This will not apply from the root. This patch set needs your previous patches to make sense, right? > Changelog: > * winemenubuilder: add thumbnail mode for .lnk files winemenubuilder is a winelib app, it may not run in the prefix that the .lnk file was created in. So while these patches get rid of duplicate launchers on the desktop, we lose support for multiple prefixes.

2 1

Trying to fix a bug, stuck at cross-compiling
by Alexandru Băluț May 21, 2010

May 21, 2010

Hi, I'm trying to fix bug 12804, and made some changes to winecfg, and I need to test how it works on Windows. I followed the instructions from http://wiki.winehq.org/CompilingDLLsUsingMingw - Cross Compiling the Whole Tree, but I get an error: $ ../wine-crossc/configure --host=i586-mingw32msvc --with-wine-tools=../build-native --without-freetype --without-x [...] checking for the directory containing the Wine tools... configure: error: could not find Wine tools in ../build-native This seems to happen because the ../build-native/tools/winebuild directory does not exist. When I ran "make tools", in ../build-native, no error occurred. Any idea why this happens? Is there another way to build winecfg as an exe? Thanks, Alex

4 5

Re: wined3d: Avoid using the long type.
by Henri Verbeet May 21, 2010

May 21, 2010

This will conflict with my patches from earlier today.

1 0

Re: [4/6] d3dx9: Implement D3DXCreateTexture plus tests
by Henri Verbeet May 21, 2010

May 21, 2010

On 20 May 2010 23:05, Owen Rudge <orudge(a)codeweavers.com> wrote: > + hr = D3DXCreateTexture(device, D3DX_DEFAULT, 63, 0, 0, 0, D3DPOOL_DEFAULT, &texture); > + ok(hr == D3D_OK, "D3DXCreateTexture returned %#x, expected %#x\n", hr, D3D_OK); > + > + if (texture) > + { > + hr = IDirect3DTexture9_GetLevelDesc(texture, 0, &desc); > + > + ok(desc.Width == 63, "Returned width %d, expected %d\n", desc.Width, 63); > + ok(desc.Height == 63, "Returned height %d, expected %d\n", desc.Height, 63); Doesn't that depend on NPOT texture support in the device?

1 0

Re: [2/6] d3dx9: Implement D3DXCheckTextureRequirements
by Henri Verbeet May 21, 2010

May 21, 2010

On 20 May 2010 23:05, Owen Rudge <orudge(a)codeweavers.com> wrote: > +BOOL is_pow2(UINT num); > +UINT make_pow2(UINT num); I'm not so convinced these are going to be very useful for anything outside texture.c. > +BOOL is_pow2(UINT num) > +{ > + if (!num) return FALSE; > + > + return !(num & (num - 1)); > +} Do you really care that 0 is not a real power of two for these texture functions? It's not a valid width / height anyway. > + if ((caps.TextureCaps & D3DPTEXTURECAPS_POW2) && (!is_pow2(*width))) > + make_pow2(*width); ... > + if ((caps.TextureCaps & D3DPTEXTURECAPS_POW2) && (!is_pow2(*height))) > + make_pow2(*height); You have to assign to result to something for it to be of any use. > + max_mipmaps = 9; Where does that 9 come from? (I can easily guess, but...) > + hr = IDirect3DDevice9_GetDirect3D(device, &d3d); > + > + if (SUCCEEDED(hr)) > + IDirect3DDevice9_GetCreationParameters(device, &params); > + > + if (SUCCEEDED(hr)) > + IDirect3DDevice9_GetDisplayMode(device, 0, &mode); > + > + if (SUCCEEDED(hr)) It's probably nicer to check for failure and jump to some cleanup code. Also, you should check all the calls for errors, not just GetDirect3D(). The width / height handling looks a bit complicated to me. You should probably first calculate the width and height using local variables, and only assign those to *width/*height afterwards. Something along these lines: { UINT w = (width && *width) ? *width : 1; UINT h = (height && *height) ? *height : 1; if (w == D3DX_DEFAULT && h == D3DX_DEFAULT) w = h = 256; else if (w == D3DX_DEFAULT) w = h; else if (h == D3DX_DEFAULT) h = w; ... if (width) *width = w; if (height) *height = h; }

1 0

Re: [5/6] d3dx9: Add basic implementation of D3DXFilterTexture, plus tests
by testbot＠testbot.winehq.org May 20, 2010

May 20, 2010

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 http://testbot.winehq.org/JobDetails.pl?Key=2207 Your paranoid android. === WINEBUILD (build) === Make failed

1 0

Re: [2/2] shell32: Export ILFree by name
by testbot＠testbot.winehq.org May 20, 2010

May 20, 2010

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 http://testbot.winehq.org/JobDetails.pl?Key=2202 Your paranoid android. === WNT4WSSP6 (32 bit shlfileop) === shlfileop.c:2202: Test failed: SHGetNewLinkInfoA succeeded

1 0

Re: comctl32/listview: Don't forward WM_COMMAND to parent if we don't have an Edit box (resent)
by testbot＠testbot.winehq.org May 20, 2010

May 20, 2010

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 http://testbot.winehq.org/JobDetails.pl?Key=2201 Your paranoid android. === W7PRO (32 bit listview) === listview.c:4393: Test failed: NM_HOVER allow test: the msg sequence is not complete: expected 0087 - actual 0000 listview.c:4398: Test failed: NM_HOVER block test: the msg sequence is not complete: expected 0087 - actual 0000

1 0