February 2011 - Wine-Devel - list.winehq.org

Re: [PATCH 2/4] urlmon: Read remaining protocol data if report_data was called after binding finished.
by Marvin Feb. 15, 2011

Feb. 15, 2011

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=9272 Your paranoid android. === WVISTAADM (32 bit url) === Failure running script in VM: The specified guest user must be logged in interactively to perform this operation

1 0

Re: [PATCH] Fixed several links in the documentation...
by Alexandre Julliard Feb. 15, 2011

Feb. 15, 2011

max(a)mtew.isa-geek.net writes: > @@ -204,7 +204,7 @@ int CDECL MSVCRT__chdir(const char * newdir) > /********************************************************************* > * _wchdir (MSVCRT.@) > * > - * Unicode version of _chdir. > + * Unicode version of MSVCRT__chdir. You don't want to use the internal name here, it's not useful for users. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: gdi32: Fixed GetDIBits for top-down bitmaps (try 2).
by Alexandre Julliard Feb. 15, 2011

Feb. 15, 2011

Jack Edmonds <pocketcookies2(a)gmail.com> writes: > gdi32/dib.c: Reversed source bitmap pointer instead of destination bitmap pointer. > Removed check for destination bitmap height when determining whether to copy the source bitmap upside down. > gdi32/tests/bitmap.c: Added a test for copying between top-down bitmaps. The rest of the code won't cope so well with a negative source width. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [3/5] msxml3/xsl: Basic put_input() method for IXSLProcessor
by Jacek Caban Feb. 15, 2011

Feb. 15, 2011

Hi Nikolay, On 2/14/11 9:04 PM, Nikolay Sivov wrote: > Basic put_input() method for IXSLProcessor > + IXMLDOMNode *old_input = This->input; + HRESULT hr; - FIXME("(%p): stub\n", This); - return E_NOTIMPL; + TRACE("(%p)->(type=%d)\n", This, V_VT(&input)); + + /* try IXMLDOMNode directly first */ + if (V_VT(&input) == VT_UNKNOWN) + hr = IUnknown_QueryInterface(V_UNKNOWN(&input),&IID_IXMLDOMNode, (void**)&This->input); + else if (V_VT(&input) == VT_DISPATCH) + hr = IDispatch_QueryInterface(V_DISPATCH(&input),&IID_IXMLDOMNode, (void**)&This->input); + + if (hr != S_OK) + { + IXMLDOMDocument *doc; + + hr = DOMDocument_create(&CLSID_DOMDocument, NULL, (void**)&doc); + if (hr == S_OK) + { + hr = IXMLDOMDocument_QueryInterface(doc,&IID_IXMLDOMNode, (void**)&This->input); + IXMLDOMDocument_Release(doc); + } + } + + if (hr == S_OK&& old_input) IXMLDOMNode_Release(old_input); Failure of QueryInterface call will set This->input to NULL (you probably want it not changed and even if not, that's a leak). It's probably better idea to have local pointer to the new input instead of old one and change it once you know the function will succeed. Jacek

2 2

Re: [4/5] msxml3/xsl: Store output stream for processor
by Jacek Caban Feb. 15, 2011

Feb. 15, 2011

Hi Nikolay, On 2/14/11 9:04 PM, Nikolay Sivov wrote: > + case VT_UNKNOWN: > + hr = IUnknown_QueryInterface(V_UNKNOWN(&output),&IID_IStream, (void**)&This->output); > + break; You leak This->output here. Jacek

1 0

Re: [1/2] winex11.drv: Add PS_USERSTYLE support to pens (take 2)
by Daniel Feb. 15, 2011

Feb. 15, 2011

grr, please disregard previous patches. I've discovered a problem that can lead to use of un-initialized values. Daniel On 02/15/2011 01:09 AM, Daniel wrote: > Clean up per Vitaliy's direction. Also found a FIXME comment that needed > to be removed. > > Daniel > > >

1 1

Re: gdi32: correct substitutions (take 2)
by Dmitry Timoshkov Feb. 15, 2011

Feb. 15, 2011

Rafał Mużyło <galtgendo(a)gmail.com> wrote: > Corrected version of the previous patch. > This time I'm not adding a new helper function, just putting it in one > place. > Corrections: > - as oppossed to get_font_subst, reverse check can have more than one > result, so changed to adjust all of them > - just in case a->a substitution would result, explicitly prevent that > (perhaps gdi32 handles it gracefully, but it's still redundant) Did you test that Windows handles font substitutions this way? -- Dmitry.

1 0

Re: Re: [PATCH] Implement software DXT1 compression
by alberto＠mattea.info Feb. 14, 2011

Feb. 14, 2011

Hi Stefan, thanks for your answer. I've checked the nvidia texture tools but, sadly, I think it's impossible to use them in wine: while they have c bindings, those bindings lack the essential function nvttSetOutputOptionsOutputHandler(NvttOutputOptions * outputOptions, nvttOutputHandler outputHandler, nvttImageHandler imageHandler) (commented out in the headers), so the output can only be written to a file. What about going the mesa way, where the drivers aren't linked with any s3tc library, but load it at runtime if it is present? Any idea is welcome.

2 1

Re: wine is passing make test again :-)
by Greg Geldorp Feb. 14, 2011

Feb. 14, 2011

> From: Andre Hentschel > and now without the 9x stuff the result page looks better again, someone > just need to fix his user32 test :) Frankly, I'm a bit surprised and disappointed to see that people are not taking responsibility for fixing failures they introduced in tests. I don't expect to have time to clean up after them for the next month or so. Greg.

3 2

kernel32/process: Add stderr redirection for native linux programs started from windows program
by Igor Egorov Feb. 14, 2011

Feb. 14, 2011

A week ago I submitted a patch for stderr redirection (http://www.winehq.org/pipermail/wine-patches/2011-February/098578.html) is there still something wrong with it? Will it be accepted?

2 1

Re: Austin English : shell32/tests: Don't check return values inside of if(0) ( LLVM/Clang).
by Charles Davis Feb. 14, 2011

Feb. 14, 2011

Austin, Static analyzer checks for unreachable assignments (like the ones that were removed in this patch) were turned off in Clang trunk recently (as of r125415). Looks like Clang was wrong ;). Chip

2 1

Re: winex11.drv: Add PS_USERSTYLE support to pens
by Vitaliy Margolen Feb. 14, 2011

Feb. 14, 2011

On 02/14/2011 03:55 AM, Daniel wrote: > Hello all. This only covers the screen. Support still needs to be added > to printers. > + case PS_USERSTYLE: { > + for(i = 0; i < physDev->pen.dash_len ; i++) { Please follow file style - curly braces on separate line. No braces for single line blocks. > + /* Should be limited to 16 or less */ > + char x11_user_style[16]; > + physDev->pen.dash_len = user_style_size > 16 ? 16 : user_style_size; Should use MAX_DASHLEN instead of magic numbers. Could use "min()" for simplicity. > + x11_user_style[i] = (char)(user_style[i] > 255 ? 255 : user_style[i]); Why not copy directly into physDev->pen.dashes? No need for typecast - it's done automatically. Vitaliy.

1 0

Re: kernel32: Implement and test GetVolumePathNamesForVolumeName.
by Alexandre Julliard Feb. 14, 2011

Feb. 14, 2011

Hans Leidekker <hans(a)codeweavers.com> writes: > BOOL WINAPI GetVolumePathNamesForVolumeNameW(LPCWSTR volumename, LPWSTR volumepathname, DWORD buflen, PDWORD returnlen) > { > - FIXME("(%s, %p, %d, %p), stub!\n", debugstr_w(volumename), volumepathname, buflen, returnlen); > - SetLastError(ERROR_CALL_NOT_IMPLEMENTED); > - return FALSE; > + static const WCHAR prefixW[] = > + {'\\','\\','?','\\','V','o','l','u','m','e','{','0','0','0','0','0','0','0','0','-','0','0','0','0','-', > + '0','0','0','0','-','0','0','0','0','-','0','0','0','0','0','0','0','0','0','0'}; > + static const WCHAR fmtW[] = {'%','c',':','\\',0}; > + > + TRACE("%s, %p, %u, %p\n", debugstr_w(volumename), volumepathname, buflen, returnlen); > + > + if (!volumename || strlenW( volumename ) != 49 || > + volumename[47] != '}' || memcmp( volumename, prefixW, sizeof(prefixW) )) > + { > + SetLastError( ERROR_INVALID_NAME ); > + return FALSE; > + } > + if (returnlen) > + { > + *returnlen = 5; > + } > + if (buflen < 5 || !volumepathname) > + { > + SetLastError( ERROR_MORE_DATA ); > + return FALSE; > + } > + sprintfW( volumepathname, fmtW, strtolW( volumename + 45, NULL, 16 ) ); > + volumepathname[4] = 0; > + return TRUE; Please don't make assumption about volume guids. This should be implemented by retrieving the correct data from mountmgr. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [4/4] mshtml: Add IDispatchEx support to IXMLHTTPRequest
by Marvin Feb. 14, 2011

Feb. 14, 2011

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=9250 Your paranoid android. === W2KPROSP4 (32 bit domdoc) === domdoc.c:3449: Test failed: ret 80004002 === WXPPROSP3 (32 bit domdoc) === domdoc.c:3449: Test failed: ret 80004002 === W2K3R2SESP2 (32 bit domdoc) === domdoc.c:3449: Test failed: ret 80004002 === WVISTAADM (32 bit domdoc) === domdoc.c:3449: Test failed: ret 80004002 === W2K8SE (32 bit domdoc) === domdoc.c:3449: Test failed: ret 80004002 === W7PRO (32 bit domdoc) === domdoc.c:3449: Test failed: ret 80004002 === W7PROX64 (32 bit domdoc) === domdoc.c:3449: Test failed: ret 80004002 === W7PROX64 (64 bit domdoc) === domdoc.c:3449: Test failed: ret 80004002

1 0

wine is passing make test again :-)
by Austin English Feb. 13, 2011

Feb. 13, 2011

Looks like Jacek and Gentoo have the magic combination. It's passed 4 days in a row: http://test.winehq.org/data/ae3e768a2e1f9c21b355798c8fbf473db3d83434/wine_j… http://test.winehq.org/data/bf640445c748b0588616da201e8ff8f3b1c778e9/wine_j… http://test.winehq.org/data/780f9b3b4fe8ff17784860a2a65b4ecab268436a/wine_j… http://test.winehq.org/data/745a50c9dbb03d37db58ff53136e93f00943bc0d/wine_j… now let's see if we can spread the love to someone else :-). -- -Austin

2 2

D3DXCreateTorus patch... thoughts?
by Misha Koshelev Feb. 13, 2011

Feb. 13, 2011

This is for http://bugs.winehq.org/show_bug.cgi?id=23930 Based on my prior mesh patches. Test bot reports success: https://testbot.winehq.org/JobDetails.pl?Key=9240 Thank you Misha

1 0

Re: gdi32/dib.c: Write true bitmap height into info and copy scanlines in the correct order.
by Nicolas Le Cam Feb. 13, 2011

Feb. 13, 2011

2011/2/8 Jack Edmonds <jack(a)jack-desktop.> > Fixes bug 25698 and addresses one of the test cases in bug 24278 (possibly > resolves bug 24278 but more testing is required to be sure). > > dib.c: Fills BITMAPINFOHEADER with correct width and height when bits == > NULL and copying from a top-to-bottom bitmap. > For top-to-bottom bitmaps, reverses the source pointer rather than the > destination pointer. > Ignoring the case when both source and destination are top-to-bottom > bitmaps and treating this as if the destination is only a top-to-bottom > bitmap. > --- > dlls/gdi32/dib.c | 113 > ++++++++++++++++++++++++--------------------- > dlls/gdi32/tests/bitmap.c | 112 > ++++++++++++++++++++++++++++++++++++++++++++ > 2 files changed, 173 insertions(+), 52 deletions(-) > > diff --git a/dlls/gdi32/dib.c b/dlls/gdi32/dib.c > index 5067e49..bf7a525 100644 > --- a/dlls/gdi32/dib.c > +++ b/dlls/gdi32/dib.c > @@ -636,54 +636,6 @@ INT WINAPI GetDIBits( > > switch (bpp) > { > - case 0: /* query bitmap info only */ > - if (core_header) > - { > - BITMAPCOREHEADER* coreheader = (BITMAPCOREHEADER*) info; > - coreheader->bcWidth = bmp->bitmap.bmWidth; > - coreheader->bcHeight = bmp->bitmap.bmHeight; > - coreheader->bcPlanes = 1; > - coreheader->bcBitCount = bmp->bitmap.bmBitsPixel; > - } > - else > - { > - info->bmiHeader.biWidth = bmp->bitmap.bmWidth; > - info->bmiHeader.biHeight = bmp->bitmap.bmHeight; > - info->bmiHeader.biPlanes = 1; > - info->bmiHeader.biSizeImage = > - DIB_GetDIBImageBytes( bmp->bitmap.bmWidth, > - bmp->bitmap.bmHeight, > - bmp->bitmap.bmBitsPixel ); > - if (bmp->dib) > - { > - info->bmiHeader.biBitCount = bmp->dib->dsBm.bmBitsPixel; > - switch (bmp->dib->dsBm.bmBitsPixel) > - { > - case 16: > - case 32: > - info->bmiHeader.biCompression = BI_BITFIELDS; > - break; > - default: > - info->bmiHeader.biCompression = BI_RGB; > - break; > - } > - } > - else > - { > - info->bmiHeader.biCompression = (bmp->bitmap.bmBitsPixel > > 8) ? BI_BITFIELDS : BI_RGB; > - info->bmiHeader.biBitCount = bmp->bitmap.bmBitsPixel; > - } > - info->bmiHeader.biXPelsPerMeter = 0; > - info->bmiHeader.biYPelsPerMeter = 0; > - info->bmiHeader.biClrUsed = 0; > - info->bmiHeader.biClrImportant = 0; > - > - /* Windows 2000 doesn't touch the additional struct members if > - it's a BITMAPV4HEADER or a BITMAPV5HEADER */ > - } > - lines = abs(bmp->bitmap.bmHeight); > - goto done; > - > case 1: > case 4: > case 8: > @@ -877,10 +829,10 @@ INT WINAPI GetDIBits( > LPBYTE dbits = bits, sbits = (LPBYTE) bmp->dib->dsBm.bmBits + > (startscan * srcwidthb); > unsigned int x, y, width, widthb; > > - if ((height < 0) ^ (bmp->dib->dsBmih.biHeight < 0)) > + if (bmp->dib->dsBmih.biHeight < 0) > { > - dbits = (LPBYTE)bits + (dstwidthb * (lines-1)); > - dstwidthb = -dstwidthb; > + sbits += (srcwidthb * (abs(bmp->dib->dsBmih.biHeight) - > startscan - 1)); > + srcwidthb = -srcwidthb; > } > > switch( bpp ) { > @@ -1086,7 +1038,64 @@ INT WINAPI GetDIBits( > } > } > } > - else lines = abs(height); > + else /* query bitmap info only */ > + { > + if (core_header) > + { > + BITMAPCOREHEADER* coreheader = (BITMAPCOREHEADER*) info; > + if (bmp->dib) > + { > + coreheader->bcWidth = bmp->dib->dsBmih.biWidth; > + coreheader->bcHeight = bmp->dib->dsBmih.biHeight; > + } > + else > + { > + coreheader->bcWidth = bmp->bitmap.bmWidth; > + coreheader->bcHeight = bmp->bitmap.bmHeight; > + } > + coreheader->bcPlanes = 1; > + coreheader->bcBitCount = bmp->bitmap.bmBitsPixel; > + } > + else > + { > + if (bmp->dib) > + { > + info->bmiHeader.biHeight = bmp->dib->dsBmih.biHeight; > + info->bmiHeader.biWidth = bmp->dib->dsBmih.biWidth; > + info->bmiHeader.biBitCount = bmp->dib->dsBm.bmBitsPixel; > + switch (bmp->dib->dsBm.bmBitsPixel) > + { > + case 16: > + case 32: > + info->bmiHeader.biCompression = BI_BITFIELDS; > + break; > + default: > + info->bmiHeader.biCompression = BI_RGB; > + break; > + } > + } > + else > + { > + info->bmiHeader.biWidth = bmp->bitmap.bmWidth; > + info->bmiHeader.biHeight = bmp->bitmap.bmHeight; > + info->bmiHeader.biCompression = (bmp->bitmap.bmBitsPixel > > 8) ? BI_BITFIELDS : BI_RGB; > + info->bmiHeader.biBitCount = bmp->bitmap.bmBitsPixel; > + } > + info->bmiHeader.biPlanes = 1; > + info->bmiHeader.biSizeImage = > + DIB_GetDIBImageBytes( bmp->bitmap.bmWidth, > + bmp->bitmap.bmHeight, > + bmp->bitmap.bmBitsPixel ); > + info->bmiHeader.biXPelsPerMeter = 0; > + info->bmiHeader.biYPelsPerMeter = 0; > + info->bmiHeader.biClrUsed = 0; > + info->bmiHeader.biClrImportant = 0; > + > + /* Windows 2000 doesn't touch the additional struct members if > + it's a BITMAPV4HEADER or a BITMAPV5HEADER */ > + } > + lines = abs(bmp->bitmap.bmHeight); > + } > > /* The knowledge base article Q81498 ("DIBs and Their Uses") states > that > if bits == NULL and bpp != 0, only biSizeImage and the color table > are > diff --git a/dlls/gdi32/tests/bitmap.c b/dlls/gdi32/tests/bitmap.c > index d0d7a9c..a76d437 100644 > --- a/dlls/gdi32/tests/bitmap.c > +++ b/dlls/gdi32/tests/bitmap.c > @@ -3050,6 +3050,116 @@ static void test_32bit_bitmap_blt(void) > DeleteDC(hdcScreen); > } > > +static void test_GetDIBits_info_query(void) > +{ > + BITMAPINFO bi; > + HBITMAP bmp; > + HDC hdc; > + int *picture; > + > + hdc = GetDC(NULL); > + > + bi.bmiHeader.biSize=sizeof(bi.bmiHeader); > + bi.bmiHeader.biWidth=2; > + bi.bmiHeader.biHeight=2; > + bi.bmiHeader.biPlanes=1; > + bi.bmiHeader.biBitCount=32; > + bi.bmiHeader.biCompression=BI_RGB; > + bmp = CreateDIBSection(hdc, &bi, DIB_RGB_COLORS, (void**)&picture, > NULL, 0); > + /*Clear the structure to make sure that we really are testing that the > structure is being written to.*/ > + memset(&bi, 0, sizeof(bi)); > + /*Test with a BITMAPINFOHEADER.*/ > + bi.bmiHeader.biSize = sizeof(BITMAPINFOHEADER); > + GetDIBits(hdc, bmp, 0, 0, NULL, &bi, DIB_RGB_COLORS); > + ok(bi.bmiHeader.biHeight == 2, "Height should be 2."); > + ok(bi.bmiHeader.biWidth == 2, "Width should be 2."); > + /*Clear the structure to make sure that we really are testing that the > structure is being written to.*/ > + memset(&bi, 0, sizeof(bi)); > + /*Now test with a BITMAPCOREHEADER.*/ > + bi.bmiHeader.biSize=sizeof(BITMAPCOREHEADER); > + GetDIBits(hdc, bmp, 0, 0, NULL, &bi, DIB_RGB_COLORS); > + ok(((BITMAPCOREHEADER*)&bi)->bcHeight == 2, "Height should be 2."); > + ok(((BITMAPCOREHEADER*)&bi)->bcWidth == 2, "Width should be 2."); > + /*Clean up.*/ > + DeleteObject(bmp); > + > + /*Test top to bottom bitmap.*/ > + bi.bmiHeader.biSize=sizeof(bi.bmiHeader); > + bi.bmiHeader.biWidth=2; > + bi.bmiHeader.biHeight=-2; > + bi.bmiHeader.biPlanes=1; > + bi.bmiHeader.biBitCount=32; > + bi.bmiHeader.biCompression=BI_RGB; > + bmp = CreateDIBSection(hdc, &bi, DIB_RGB_COLORS, (void**)&picture, > NULL, 0); > + /*Clear the structure to make sure that we really are testing that the > structure is being written to.*/ > + memset(&bi, 0, sizeof(bi)); > + /*Test with a BITMAPINFOHEADER.*/ > + bi.bmiHeader.biSize=sizeof(BITMAPINFOHEADER); > + GetDIBits(hdc, bmp, 0, 0, NULL, &bi, DIB_RGB_COLORS); > + ok(bi.bmiHeader.biHeight == -2, "Height should be -2."); > + ok(bi.bmiHeader.biWidth == 2, "Width should be 2."); > + /*Clear the structure to make sure that we really are testing that the > structure is being written to.*/ > + memset(&bi, 0, sizeof(bi)); > + /*Now test with a BITMAPCOREHEADER.*/ > + bi.bmiHeader.biSize=sizeof(BITMAPCOREHEADER); > + GetDIBits(hdc, bmp, 0, 0, NULL, &bi, DIB_RGB_COLORS); > + ok(((BITMAPCOREHEADER*)&bi)->bcHeight == (WORD)-2, "Height should be > -2."); > + ok(((BITMAPCOREHEADER*)&bi)->bcWidth == (WORD)2, "Width should be > 2."); > + /*Clean up.*/ > + DeleteObject(bmp); > +} > + > +static void test_GetDIBits_single_pixel_destination(void) > +{ > + BITMAPINFO bi; > + HBITMAP bmptb, bmpbt; > + HDC hdc; > + int pixelOut; > + int *picture; > + bi.bmiHeader.biSize=sizeof(bi.bmiHeader); > + bi.bmiHeader.biWidth=2; > + bi.bmiHeader.biHeight=2; > + bi.bmiHeader.biPlanes=1; > + bi.bmiHeader.biBitCount=32; > + bi.bmiHeader.biCompression=BI_RGB; > + > + /*Get the device context for the screen.*/ > + hdc=GetDC(NULL); > + > + /*Create the bottom to top image (image's bottom scan line is at the > top in memory).*/ > + bmpbt=CreateDIBSection(hdc, &bi, DIB_RGB_COLORS, (void**)&picture, > NULL, 0); > + /*Now that we have a pointer to the pixels, we write to them.*/ > + picture[0]=1; > + picture[1]=2; > + picture[2]=3; > + picture[3]=4; > + /*Create the top to bottom image (images' bottom scan line is at the > bottom in memory).*/ > + bi.bmiHeader.biHeight=-2; /*We specify that we want a top to bottom > image by specifying a negative height.*/ > + bmptb=CreateDIBSection(hdc, &bi, DIB_RGB_COLORS, (void**)&picture, > NULL, 0); > + /*Write to this top to bottom bitmap.*/ > + picture[0]=1; > + picture[1]=2; > + picture[2]=3; > + picture[3]=4; > + > + bi.bmiHeader.biWidth=1; > + > + bi.bmiHeader.biHeight=2; > + GetDIBits(hdc, bmpbt, 0, 1, &pixelOut, &bi, DIB_RGB_COLORS); > + ok(pixelOut==1, "Bottom-up -> bottom-up should have first scanline."); > + GetDIBits(hdc, bmptb, 0, 1, &pixelOut, &bi, DIB_RGB_COLORS); > + ok(pixelOut==3, "Top-down -> bottom-up should have last scanline."); > + > + bi.bmiHeader.biHeight=-2; > + GetDIBits(hdc, bmpbt, 0, 1, &pixelOut, &bi, DIB_RGB_COLORS); > + ok(pixelOut==1, "Bottom-up -> top-down should have first scanline."); > + GetDIBits(hdc, bmptb, 0, 1, &pixelOut, &bi, DIB_RGB_COLORS); > + ok(pixelOut==3, "Top-down -> top-down should have last scanline."); > + > + DeleteObject(bmpbt); > + DeleteObject(bmptb); > +} > + > START_TEST(bitmap) > { > HMODULE hdll; > @@ -3080,4 +3190,6 @@ START_TEST(bitmap) > test_bitmapinfoheadersize(); > test_get16dibits(); > test_clipping(); > + test_GetDIBits_single_pixel_destination(); > + test_GetDIBits_info_query(); > } > -- > 1.7.1 > > > > Hi Jack, You forgot trailing \n in the ok() calls. -- Nicolas Le Cam

2 1

Wine and Summer of Code 2011
by Dan Kegel Feb. 13, 2011

Feb. 13, 2011

http://www.google-melange.com/document/show/gsoc_program/google/gsoc2011/fa… says projects can apply starting on Feb 28. In preparation, I moved the 2010 projects from http://wiki.winehq.org/SummerOfCode to http://wiki.winehq.org/SummerOfCode/PreviousProjects I also added a DSS Provider project idea, removed the gstreamer idea since that's implemented, and tidied up the dxdiag idea a bit. Who's interested in mentoring this year? And can y'all think of any fresh ideas for the suggested projects list, or ideas that should be retired? - Dan

6 5

vblank timestamps and page flipping in DRI2
by Saulius Krasuckas Feb. 13, 2011

Feb. 13, 2011

There are some news on the $Subject: [1]. (have a look at the chapter "Fine tuned") I thought it could help Wine fixing such bugs as 6086, 13085, 17260, 22362 [2]. I'll paste the text in just for archiving purposes: --- quote --- Fine tuned The kernel hackers have extended the DRM infrastructure to include "precision vblank timestamps", which are explained in detail in the commit comment and are relevant to OpenML (Open Media Library) conformity. This feature enables the kernel to provide timestamps with an accuracy of a few microseconds, allowing applications to detect when the rendering of a new image has started ("double-buffer swap"). This can, for instance, be relevant for applications where it is important to precisely co-ordinate the audio output, or the control of other devices, with the rendering of the image. The technology was first implemented for experimental psychology and neuroscience applications, but it can also be applied to provide (more) jerk-free video playback, animations or desktop effects. Such timestamps can also be utilised by the new page flipping feature available via DRI2. This feature improves the way the kernel synchronises the release of an updated image with the moment the graphics hardware starts rendering the new image (Vsync). Synchronising the release avoids such distortions as flickering, tearing or incomplete rendering, as it prevents situations where one half of a displayed image is part of the old image while the other half is part of the new image . for instance with individual movie frames. In kernel version 38, the graphics driver support required for both features will be available via the AMD drivers (1, 2) and the Intel drivers (1, 2). The NVIDIA drivers currently only support page flipping (1, 2). --- quote --- S. [1] http://www.h-online.com/open/features/Kernel-Log-Coming-in-2-6-38-Part-1-Gr… [2] http://bugs.winehq.org/show_bug.cgi?id=6086 http://bugs.winehq.org/show_bug.cgi?id=13085 http://bugs.winehq.org/show_bug.cgi?id=17260 http://bugs.winehq.org/show_bug.cgi?id=22362

2 1

Tiny imap message appender written in perl
by Krzysztof Nikiel Feb. 13, 2011

Feb. 13, 2011

It's a kind of side effect of my Wine patching. Sending those patches traditional way was an epic fail. Mail client wrapped lines, git-imap-send provided by Debian don't support SSL. I have modified my tiny imap downloader to append new messages to specified box. It uses Net::SSLeay module. Maybe it will be useful to someone.

2 2

Tiny imap message appender written in perl different attachment
by Krzysztof Nikiel Feb. 13, 2011

Feb. 13, 2011

This attachment may be better: #!/usr/bin/perl # SSL IMAP message appender # Copyright (C) 2011 Krzysztof Nikiel use Socket; use Net::SSLeay qw(die_now die_if_ssl_error); Net::SSLeay::load_error_strings(); Net::SSLeay::SSLeay_add_ssl_algorithms(); Net::SSLeay::randomize(); $g_port=993; $g_tag=1; $g_ret=""; if ($#ARGV < 3) { print "usage: $0 server user box pass file1 ...\n", "\tAppends files to the specified IMAP box.\n\n"; exit; } $g_serv=shift @ARGV; $g_user=shift @ARGV; $g_box=shift @ARGV; $g_pass=shift @ARGV; sub imapclient { my $dest_serv, $port, $dest_ip, $dest_serv_params; my $ctx, $res; my $res; my @boxlist; my $boxok = 0; $port = $g_port; $port = getservbyname($port, 'tcp') unless $port =~ /^\d+$/; $dest_ip = gethostbyname($g_serv); $dest_serv_params = sockaddr_in($port, $dest_ip); socket(S, &AF_INET, &SOCK_STREAM, 0) or die "socket: $!"; connect(S, $dest_serv_params) or die "connect: $!"; select(S); $| = 1; select(STDOUT); # The network connection is now open, lets fire up SSL $ctx = Net::SSLeay::CTX_new() or die_now("Failed to create SSL_CTX $!"); $g_ssl = Net::SSLeay::new($ctx) or die_now("Failed to create SSL $!"); Net::SSLeay::set_fd($g_ssl, fileno(S)); # Must use fileno $res = Net::SSLeay::connect($g_ssl); print "Cipher '" . Net::SSLeay::get_cipher($g_ssl) . "'\n"; # IMAP client starts here print "\nconnecting to $g_serv\n"; imapread(); showresp(); unless ($g_ret =~ /^\* OK/) {return}; print "logging in as $g_user\n"; $res = imapcmd("LOGIN $g_user $g_pass"); imapread(); showresp(); print "reading mailbox list\n"; $res = imapcmd("LIST \"\" \*"); imapread(); unless (testresp($g_ret)) { print "LIST failed\n"; goto logout; } #print "list ret: '$g_ret'\n"; @boxlist=(); foreach (split /\n/,$g_ret) { if (/^* LIST [^"]*"[^"]*"[^"]*"([^"]*)"/) { push @boxlist,$1; } } foreach (@boxlist) { print "$_\n"; if ("$_" eq "$g_box") {$boxok = 1} } if (!$boxok) {print "Warning: can't find '$g_box' in mailbox names\n"} foreach (@ARGV) { imapappend("$g_box", $_); } logout: print "logging out\n"; $res = imapcmd("LOGOUT"); imapread(); showresp(); if ($g_ret =~ /^\* BYE/) {print "LOGOUT OK\n"}; do { imapread(); showresp(); } while ($g_ret ne ""); #end of IMAP code shutdown S, 1; # Half close --> No more output, sends EOF to server $res = Net::SSLeay::read($g_ssl); # Perl returns undef on failure print $res; Net::SSLeay::free($g_ssl); # Tear down connection Net::SSLeay::CTX_free($ctx); close S; } sub imapappend { my $res, $resp; my $box = @_[0]; my $file = @_[1]; my @msg = (); print "reading from '$file'..."; if (!open(MSG,"<",$file)) { print "Warning: can't open $file\n"; return; } $size = 0; while(<MSG>) { chomp; push @msg,$_; $size += length; $size += 2; } close MSG; print $#msg + 1, " lines; $size bytes\n"; print "appending to \"$box\"\n"; $res = imapcmd("APPEND \"$box\" {$size}"); imapread(); showresp(); unless ($g_ret =~ /^+/) {return}; foreach (@msg) { $res = sslwrite("$_"); } $res = sslwrite(""); imapread(); showresp(); } sub showresp { print $g_ret; } sub imapread { $g_ret = Net::SSLeay::read($g_ssl); $g_ret =~ s/\r//g; } sub imapcmd { my ($msg) = @_; my $res; $g_tag += 1; return sslwrite("TAG$g_tag $msg"); } sub sslwrite { my ($msg) = @_; my $res; $res = Net::SSLeay::ssl_write_all($g_ssl, "$msg\r\n"); return $res; } sub testresp { if (@_[0] =~ /TAG(\d+) OK/) { if ($1 == $g_tag) {return 1} } return 0; } imapclient();

1 0

Re: winex11.drv: map_window should honor iconic state
by Alexandre Julliard Feb. 13, 2011

Feb. 13, 2011

Alexander Lakhin <exclusion(a)gmail.com> writes: > + data->iconic = (new_style & WS_MINIMIZE) != 0; > + > if (!data->embedded) > { > update_net_wm_states( display, data ); > sync_window_style( display, data ); > wine_tsx11_lock(); > XMapWindow( display, data->whole_window ); > + if (data->iconic) > + XIconifyWindow( display, data->whole_window, DefaultScreen(display) ); This is handled by setting the initial window state in the wm hints. If there's somewhere that the state isn't set correctly, that's what should be fixed. -- Alexandre Julliard julliard(a)winehq.org

2 3

Re: gdi32: Add an EnumFontFamilies test to see how it handles various font faces.
by Marvin Feb. 13, 2011

Feb. 13, 2011

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=9230 Your paranoid android. === WNT4WSSP6 (32 bit font) === font.c:3495: Test failed: expected Arial Bold got Arial font.c:3545: Test failed: font Arial Bold should not be enumerated font.c:3495: Test failed: expected Arial Bold Italic got Arial font.c:3553: Test failed: font Arial Bold Italic should not be enumerated === W2KPROSP4 (32 bit font) === font.c:3495: Test failed: expected Arial Bold got Arial font.c:3545: Test failed: font Arial Bold should not be enumerated font.c:3495: Test failed: expected Arial Bold Italic got Arial font.c:3553: Test failed: font Arial Bold Italic should not be enumerated === WXPPROSP3 (32 bit font) === font.c:3495: Test failed: expected Arial Bold got Arial font.c:3545: Test failed: font Arial Bold should not be enumerated font.c:3495: Test failed: expected Arial Bold Italic got Arial font.c:3553: Test failed: font Arial Bold Italic should not be enumerated === W2K3R2SESP2 (32 bit font) === font.c:3495: Test failed: expected Arial Bold got Arial font.c:3545: Test failed: font Arial Bold should not be enumerated font.c:3495: Test failed: expected Arial Bold Italic got Arial font.c:3553: Test failed: font Arial Bold Italic should not be enumerated === WVISTAADM (32 bit font) === font.c:3495: Test failed: expected Arial Bold got Arial font.c:3545: Test failed: font Arial Bold should not be enumerated font.c:3495: Test failed: expected Arial Bold Italic got Arial font.c:3553: Test failed: font Arial Bold Italic should not be enumerated === W2K8SE (32 bit font) === font.c:3495: Test failed: expected Arial Bold got Arial font.c:3545: Test failed: font Arial Bold should not be enumerated font.c:3495: Test failed: expected Arial Bold Italic got Arial font.c:3553: Test failed: font Arial Bold Italic should not be enumerated === W7PRO (32 bit font) === font.c:3495: Test failed: expected Arial Bold got Arial font.c:3545: Test failed: font Arial Bold should not be enumerated font.c:3495: Test failed: expected Arial Bold Italic got Arial font.c:3553: Test failed: font Arial Bold Italic should not be enumerated

1 0

Re: wined3d: Remove an old hack around Mesa choking on protected memory from flush_to_framebuffer_drawpixels.
by Stefan Dösinger Feb. 13, 2011

Feb. 13, 2011

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Yes, go ahead and kill this hack Am 13.02.2011 um 00:46 schrieb Gerald Pfeifer: > This somehow got stuck from the original discussion we had a year ago. > > This is in response to Stefan's feedback > (http://www.winehq.org/pipermail/wine-devel/2010-May/083480.html) > for a previous version of this patch of mine > (http://www.winehq.org/pipermail/wine-patches/2010-May/088229.html) > > Either my original patch or this new one (or a variation thereof) > should be applied. Stefan, I hope it got you right? > > Gerald > > --- > dlls/wined3d/surface.c | 17 ----------------- > 1 files changed, 0 insertions(+), 17 deletions(-) > > diff --git a/dlls/wined3d/surface.c b/dlls/wined3d/surface.c > index 4377f10..b7f4b1c 100644 > --- a/dlls/wined3d/surface.c > +++ b/dlls/wined3d/surface.c > @@ -1960,23 +1960,6 @@ static void flush_to_framebuffer_drawpixels(IWineD3DSurfaceImpl *This, > glRasterPos3i(rect.left, rect.top, 1); > checkGLcall("glRasterPos3i"); > > - /* Some drivers(radeon dri, others?) don't like exceptions during > - * glDrawPixels. If the surface is a DIB section, it might be in GDIMode > - * after ReleaseDC. Reading it will cause an exception, which x11drv will > - * catch to put the dib section in InSync mode, which leads to a crash > - * and a blocked x server on my radeon card. > - * > - * The following lines read the dib section so it is put in InSync mode > - * before glDrawPixels is called and the crash is prevented. There won't > - * be any interfering gdi accesses, because UnlockRect is called from > - * ReleaseDC, and the app won't use the dc any more afterwards. > - */ > - if ((This->flags & SFLAG_DIBSECTION) && !(This->flags & SFLAG_PBO)) > - { > - volatile BYTE read; > - read = This->resource.allocatedMemory[0]; > - } > - > /* If not fullscreen, we need to skip a number of bytes to find the next row of data */ > glPixelStorei(GL_UNPACK_ROW_LENGTH, This->currentDesc.Width); > > -- > 1.7.2.2 -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.17 (Darwin) iQIcBAEBAgAGBQJNV2TrAAoJEN0/YqbEcdMwb5EQAIhXIczJL8JYsElaRDXH0kmc k7knkDu35OE0Rp4wBlqt51BdO9V/bcUzJsNFU5OFmtFgDQsHabNZV/EYvp2mU62D PDqvbsSWPM1L+LeWw/atjZb4Q5nYg9MH01a2JsDct1YtpkN9CNxakxalWwF6HpSX r2sMFn38N5/awWqcOS5GQGn348Ak1igzdQGwcHSM/LK/LeuuNZE5r3OOgAgzxheq /GISFCQ8T0G+UJDcOpGptW7ZjHVeHY1CNFA+HbXcfPJYiUDsZZhnsjy1MehEy0gE hvrdikWkwnTiAKRjqjriXuz0M7ktZXpzfnjipuZC5MSPqzhncck4GE9JuAeRx8HF eDX/O+gLsJ+TBNfpxBKmLi/ZfFMa/x12yLk76UULDmzQ1kpSHs5UIkSptYUs3Bwg IfQAoMydXH6HMcXWD+la3N3O2MdVVvIVaKGz1GSqPPLA4EVz1UobDmXQBcJ0YTzf RO5dlu6HmbiuR8dD89CdH3eguOldLK0u4BcK/3wJjzgjquspLO24cnlK+/AfIUnw w+oXcPFULbyiok99kUmQ01DkvLasMj4GVFHRxjN69iPgL8Ofed7Yfp/eX0qRu5MK k6bZnx/XqgeuQaSeNzKXgAtQgRM0oPKsA4dgVcda8gHdZRVUxdFDjlF+0gB+8N5d KD0hY4wjYd/67FwtDhEP =B0mG -----END PGP SIGNATURE-----

1 0

Re: Fixed GetDIBits for top-down bitmaps.
by Marvin Feb. 13, 2011

Feb. 13, 2011

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=9225 Your paranoid android. === WNT4WSSP6 (32 bit bitmap) === bitmap.c:3090: Test failed: Bottom-up -> bottom-up should have first scanline. bitmap.c:3096: Test failed: Bottom-up -> top-down should have first scanline. === W2KPROSP4 (32 bit bitmap) === bitmap.c:3090: Test failed: Bottom-up -> bottom-up should have first scanline. bitmap.c:3096: Test failed: Bottom-up -> top-down should have first scanline. === WXPPROSP3 (32 bit bitmap) === bitmap.c:3090: Test failed: Bottom-up -> bottom-up should have first scanline. bitmap.c:3096: Test failed: Bottom-up -> top-down should have first scanline. === W2K3R2SESP2 (32 bit bitmap) === bitmap.c:3090: Test failed: Bottom-up -> bottom-up should have first scanline. bitmap.c:3096: Test failed: Bottom-up -> top-down should have first scanline. === WVISTAADM (32 bit bitmap) === bitmap.c:3090: Test failed: Bottom-up -> bottom-up should have first scanline. bitmap.c:3096: Test failed: Bottom-up -> top-down should have first scanline. === W2K8SE (32 bit bitmap) === bitmap.c:3090: Test failed: Bottom-up -> bottom-up should have first scanline. bitmap.c:3096: Test failed: Bottom-up -> top-down should have first scanline. === W7PRO (32 bit bitmap) === bitmap.c:3090: Test failed: Bottom-up -> bottom-up should have first scanline. bitmap.c:3096: Test failed: Bottom-up -> top-down should have first scanline. === W7PROX64 (32 bit bitmap) === bitmap.c:3090: Test failed: Bottom-up -> bottom-up should have first scanline. bitmap.c:3096: Test failed: Bottom-up -> top-down should have first scanline. === W7PROX64 (64 bit bitmap) === bitmap.c:3090: Test failed: Bottom-up -> bottom-up should have first scanline. bitmap.c:3096: Test failed: Bottom-up -> top-down should have first scanline.

1 0