March 2015 - Wine-Devel - list.winehq.org

Re: [PATCH] null.sys: Added stub dll.
by Alexandre Julliard March 12, 2015

March 12, 2015

Qian Hong <qhong(a)codeweavers.com> writes: > First part for https://bugs.winehq.org/show_bug.cgi?id=38107 I don't see why you need a driver for this. -- Alexandre Julliard julliard(a)winehq.org

5 4

My Ideas for GSoC 2015
by Zhenbo Li March 12, 2015

March 12, 2015

Hello, everyone I'm Zhenbo Li, a college student who participated in GSoC 2014 for mshtml.dll. Here are some ideas for GSoC 2015. 1. IDropTargetHelper ================== This interface is to display images [1]. I'm sure that it is necessary for VPT[2], and I guess implementing this interface will contribute to programs like 7zip. Although IDropTargetHelper is a useful interface, I haven't found a proper demo(most demos depend on MFC). So I'm writing a simple demo for this interface[3], with no MFC or .NET code. 2. IXMLHttpRequest ================== XMLHttpRequest is used heavily in AJAX programming.[4] Wine has implemented most of it in msxml component, but it's a stub in mshtml. Most js code doesn't treat wine IE as IE6 or earlier, so they will not choost ActiveX[5] to work around it(like Firebug Lite[6]) 3. sessionStorage =============== IHTMLStorage[7] is used widely, including Firebug Lite[6]. A well-implemented IHTMLStorage interface should support both sessionStorage and localStorage. I tried to write a hack, but found that it may be necessary to adjust wine-gecko code to support that(Thank Jacek). Therefore, implementing sessionStorage can be the first step of implementing localStorage, and it can be an example for WebIDL binding for further development. These are my humble ideas for GSoC 2015, and I'm glad to choose one of them as my proposal, if they meet the criteria for GSoC. Also, it is my honor if my ideas make a positive effect on other GSoC students. Thank you for reading this mail. I'm looking forward to your feedback. [1]: https://msdn.microsoft.com/en-us/library/windows/desktop/bb762028%28v=vs.85… [2]: http://virtualplaytable.com/ [3]: https://gitcafe.com/endle/IDrop_Helper_Example [4]: https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest [5]: http://en.wikipedia.org/wiki/XMLHttpRequest#Support_in_Internet_Explorer_ve… [6]: https://getfirebug.com/firebuglite [7]: https://msdn.microsoft.com/en-us/library/cc288313%28v=vs.85%29.aspx

1 1

Re: [PATCH 5/5] ddraw/tests: r200 ignores the alpha channel when matching color keys.
by Marvin March 12, 2015

March 12, 2015

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=12108 Your paranoid android. === w8 (32 bit ddraw1) === ddraw1.c:6292: Test failed: Got unexpected color 0x000000ff. ddraw1.c:6296: Test failed: Got unexpected color 0x00000080.

2 1

Re: [PATCH] vbscript: Implemented Asc(try 4)
by Piotr Caban March 12, 2015

March 12, 2015

On 03/04/15 19:59, Shuai Meng wrote: > Change log: > Free 'bstr' only when it comes from to_string(). > --- > dlls/vbscript/global.c | 26 ++++++++++++++++++++++++-- > dlls/vbscript/tests/api.vbs | 42 > ++++++++++++++++++++++++++++++++++++++++++ > 2 files changed, 66 insertions(+), 2 deletions(-) This function needs to convert the WCHAR to correct character set. In order to easily write a test for it ChrW needs to be implemented first. The test may e.g. look like this: Asc(ChrW(261)) Your implementation will return 261 in this case, on windows it produces 97. Cheers, Piotr

2 2

Re: setupapi: Implement SetupOpenLog(), SetupLogErrorA(), SetupLogErrorW(), SetupCloseLog() (resend)
by Nikolay Sivov March 12, 2015

March 12, 2015

On 11.03.2015 22:46, Pierre Schweitzer wrote: > +BOOL WINAPI SetupLogErrorW(LPCWSTR MessageString, LogSeverity Severity) > +{ > + BOOL ret; > + static const WCHAR null[] = {'(','n','u','l','l',')',0}; > + DWORD written; > + DWORD len; > + > + EnterCriticalSection(&setupapi_cs); > + > + if (setupact == INVALID_HANDLE_VALUE || setuperr == INVALID_HANDLE_VALUE) > + { > + LeaveCriticalSection(&setupapi_cs); > + SetLastError(ERROR_FILE_INVALID); > + return FALSE; > + } > + > + if (MessageString == NULL) > + MessageString = null; > + > + len = lstrlenW(MessageString) * sizeof(WCHAR); > + > + ret = WriteFile(setupact, MessageString, len, &written, NULL); > + if (!ret) > + { > + LeaveCriticalSection(&setupapi_cs); > + return ret; > + } > + > + if (Severity >= LogSevMaximum) > + { > + LeaveCriticalSection(&setupapi_cs); > + return FALSE; > + } > + > + if (Severity > LogSevInformation) > + ret = WriteFile(setuperr, MessageString, len, &written, NULL); > + > + LeaveCriticalSection(&setupapi_cs); > + > + return ret; > +} It'd be probably shorter if you had return with unlock call and one place and just jumped there. Also it's questionable if you need to return when first WriteFile() failed without trying to write to error log. I don't remember if I said it last time, but I don't see a reason to add a new file just for couple of functions. > + GetWindowsDirectoryW(win, MAX_PATH); > + lstrcpyW(path, win); > + lstrcatW(path, setupactlog); This never changes, so you can init it once. > + setupact = CreateFileW(path, GENERIC_WRITE, FILE_SHARE_WRITE | FILE_SHARE_READ, > + NULL, OPEN_ALWAYS, FILE_ATTRIBUTE_NORMAL, NULL); > + if (setupact == INVALID_HANDLE_VALUE) > + { > + LeaveCriticalSection(&setupapi_cs); > + return FALSE; > + } > + > + SetFilePointer(setupact, 0, NULL, FILE_END); Will FILE_APPEND_DATA work in this case? > + if (setupact != INVALID_HANDLE_VALUE && setuperr != INVALID_HANDLE_VALUE) > + { > + LeaveCriticalSection(&setupapi_cs); > + SetLastError(ERROR_ALREADY_INITIALIZED); > + return TRUE; > + } Setting error here makes sense as tests show. > + SetLastError(ERROR_ALREADY_EXISTS); > + return TRUE; But this one looks redundant. > + ret = SetupLogErrorW((LPCWSTR)L"Test without opening\r\n", LogSevInformation); Don't use L"" please. > + SetLastError(0xdeadbeef); > + ret = SetupLogErrorW(NULL, LogSevInformation); > + error = GetLastError(); > + if (!ret) ok(error == ERROR_INVALID_PARAMETER, "got wrong error: %d\n", error); /* Win Vista + */ > + else ok(ret, "SetupLogError failed\n"); This makes no sense, is it supposed to fail or not? Same for new test file - I think you can add those test in some existing file.

2 3

Re: [PATCH 3/5] d3d9/tests: Relax the fog range test a bit.
by Marvin March 12, 2015

March 12, 2015

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=12106 Your paranoid android. === w8 (32 bit visual) === visual.c:926: Test failed: Expected clear value 0xefefefef, got 0x00000000, fmt=D3DFMT_P8.

1 0

Re: [PATCH 2/5] d3d9/tests: Extend the colorfill test.
by Marvin March 12, 2015

March 12, 2015

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=12105 Your paranoid android. === w8 (32 bit visual) === visual.c:926: Test failed: Expected clear value 0xefefefef, got 0x00000000, fmt=D3DFMT_P8.

1 0

Re: [PATCH 9/9] gdiplus: Add pixel format test for transparent non-animated gif (try3)
by Dmitry Timoshkov March 12, 2015

March 12, 2015

Piotr Caban <piotr(a)codeweavers.com> wrote: > --- > dlls/gdiplus/tests/image.c | 30 ++++++++++++++++++++++++++++++ > 1 file changed, 30 insertions(+) Please add the tests similar to the animated case which checks pixel data, also add the tests for the palette to see that it actually has a trasparent color. Otherwise it's hard to see what is going on, and my skills of reading binary gif data becomes rusty with time. -- Dmitry.

3 3

Re: [PATCH 7/7] gdiplus/tests: Add gif frames composition tests
by Dmitry Timoshkov March 12, 2015

March 12, 2015

Piotr Caban <piotr(a)codeweavers.com> wrote: > --- > dlls/gdiplus/tests/image.c | 99 > +++++++++++++++++++++++++++++++++++++++++++++- > 1 file changed, 97 insertions(+), 2 deletions(-) It would be interesting to test this on the WIC side first. -- Dmitry.

2 1

Re: [PATCH 6/7] gdiplus: Fix animated gif frames composition
by Dmitry Timoshkov March 12, 2015

March 12, 2015

Piotr Caban <piotr(a)codeweavers.com> wrote: > + hr = IWICBitmapFrameDecode_QueryInterface(frame, &IID_IWICBitmapSource, (void**)&bmp_source); > + if(FAILED(hr)) > + return hr; > + hr = IWICBitmapSource_GetSize(bmp_source, width, height); > + IWICBitmapSource_Release(bmp_source); What's wrong with calling appropriate IWICBitmapFrameDecode method directly? -- Dmitry.

2 2

Re: [PATCH 5/7] gdiplus: Convert animated gif to PixelFormat32bppARGB on load
by Dmitry Timoshkov March 12, 2015

March 12, 2015

Piotr Caban <piotr(a)codeweavers.com> wrote: > + status = decode_frame_wic(decoder, frame_count>1 ? TRUE : FALSE, > + 0, gif_metadata_reader, image); The format probably doesn't depend on a frame count but on a presence of transparent color. -- Dmitry.

2 1

Re: [PATCH 4/7] gdiplus: Don't store stream reference in GpImage, it's done by IWICBitmapDecoder
by Dmitry Timoshkov March 12, 2015

March 12, 2015

Piotr Caban <piotr(a)codeweavers.com> wrote: > @@ -3808,12 +3816,6 @@ GpStatus WINGDIPAPI GdipImageSelectActiveFrame(GpImage *image, GDIPCONST GUID *d > if (image->current_frame == frame) > return Ok; > > - if (!image->stream) > - { > - TRACE("image doesn't have an associated stream\n"); > - return Ok; > - } It doesn't make much sence to add this piece of code in previous patch. -- Dmitry.

2 1

Re: [PATCH 6/7] gdiplus: Fix animated gif frames composition (try2)
by Vincent Povirk March 12, 2015

March 12, 2015

I'm pretty sure you can just cast IWICBitmapFrameDecode to IWICBitmapSource, and you don't need to use QueryInterface.

2 1

Re: [PATCH 1/7] gdiplus: Don't create new instance of IWICBitmapDecoder to encode every frame of animation (try2)
by Vincent Povirk March 12, 2015

March 12, 2015

@@ -4288,7 +4341,8 @@ static const struct image_codec codecs[NUM_CODECS] = { /* SigMask */ emf_sig_mask, }, NULL, - decode_image_olepicture_metafile + decode_image_olepicture_metafile, + select_frame_wic }, { { /* WMF */ @@ -4307,7 +4361,8 @@ static const struct image_codec codecs[NUM_CODECS] = { /* SigMask */ wmf_sig_mask, }, NULL, - decode_image_olepicture_metafile + decode_image_olepicture_metafile, + select_frame_wic }, You have a later patch changing these to NULL, I'm guessing you meant to do that in this one?

2 1

Re: ws2_32/tests: Add more select() tests related to OOB handling
by Marvin March 12, 2015

March 12, 2015

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=12094 Your paranoid android. === w2008s64 (64 bit sock) === No test summary line found

1 0

CMD - implement more robust parser
by Péter Urbanovics March 11, 2015

March 11, 2015

Hi, I am Peter. I read about this (Google Summer of Code) project "CMD - implement more robust parser" and I am interested. What should I do to prepare for this project? How can I contact with the mentor Jason Edmeades?

2 1

CMD - implement more robust parser
by Abhishek Srivastava March 11, 2015

March 11, 2015

I am Abhishek and I am interested in project titled "CMD - implement more robust parser". I am new to The wine project and trying to fix few available bugs on BugList. Which kind of bugs should I approach first? And how can I contact mentor for project Mr. Jason Edmeades?

2 1

Re: [PATCH 5/7] gdiplus: Convert animated gif to PixelFormat32bppARGB on load
by Marvin March 11, 2015

March 11, 2015

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=12083 Your paranoid android. === wxppro (32 bit image) === image.c:2463: Test failed: Expected 24, got 12 === w2003std (32 bit image) === image.c:2463: Test failed: Expected 24, got 12 === wvistau64 (32 bit image) === image.c:2463: Test failed: Expected 24, got 12 === w2008s64 (32 bit image) === image.c:2463: Test failed: Expected 24, got 12 === wvistau64 (64 bit image) === image.c:2463: Test failed: Expected 24, got 12 === w2008s64 (64 bit image) === image.c:2463: Test failed: Expected 24, got 12

1 0

Re: [PATCH] user32: Merge SCROLL_DrawInterior and SCROLL_DrawInterior_9x.
by Alexandre Julliard March 11, 2015

March 11, 2015

Mark Harmstone <hellas(a)burntcomma.com> writes: > The scrollbar-drawing code in user32 contains two nearly-identical functions, > SCROLL_DrawInterior and SCROLL_DrawInterior_9x, the latter of which is only > called when drawing a moving thumb. This patch cleans up the code by merging > them into one function. It's not really cleaner with the extra boolean and the reentrant calls to the same function. You probably want to restructure it some more. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: msvcp90: Add istream::operator>>complex implementation and tests.( try 3)
by YongHao Hu March 11, 2015

March 11, 2015

Hi, I have a better idea to write the test case just after sending this patch. Now I'm planing to submit a new patch with current version of implementation and an improved version of test. Please don't commit this version. However, I think the implementation itself is fine now, could you comment on the implementation part? On 2015年03月11日 14:47, YongHao Hu wrote: > Try3: Thanks QianHong > - Superseded Patch 109782, 109684 > - Avoid unnecessary else block. > > Try2: Thanks Piotr > - Delete useless check. > - Make the code more readable. > > --- > dlls/msvcp71/msvcp71.spec | 12 ++-- > dlls/msvcp80/msvcp80.spec | 12 ++-- > dlls/msvcp90/ios.c | 177 > ++++++++++++++++++++++++++++++++++++++++++++++ > dlls/msvcp90/msvcp90.spec | 12 ++-- > dlls/msvcp90/tests/ios.c | 151 +++++++++++++++++++++++++++++++++++++++ > 5 files changed, 346 insertions(+), 18 deletions(-) > >

2 1

Re: msvcp90: Add istream::operator>>complex implementation and tests.
by Piotr Caban March 11, 2015

March 11, 2015

Hi, On 03/07/15 12:13, YongHao Hu wrote: > +/* ??$?5MDU?$char_traits(a)D@std@@@std@@YAAAV?$basic_istream(a)DU?$char_traits(a)D@std@@@0(a)AAV10@AAV?$complex(a)M@0@@Z */ > +/* ??$?5MDU?$char_traits(a)D@std@@@std@@YAAEAV?$basic_istream(a)DU?$char_traits(a)D@std@@@0(a)AEAV10@AEAV?$complex(a)M@0@@Z */ > +basic_istream_char* __cdecl basic_istream_char_read_complex_float(basic_istream_char *this, complex_float *v) > +{ > + float r; > + basic_ios_char *base = basic_istream_char_get_basic_ios(this); > + if(ios_base_good(&base->base)) { There's no need to check the stream state here, it will be done by stream reading functions. > + ws_basic_istream_char(this); > + if(basic_istream_char_peek(this) == '(') { > + basic_istream_char_get(this); > + basic_istream_char_read_float(this, &r); > + if(!ios_base_fail(&base->base)) { I think that the implementation will be more readable if you just return on error here. The if(!ios_base_fail(...)) line can be changed to: if(ios_base_fail(&base->base)) return this; You don't need to set failbit in this case because it's already set by basic_istream_char_read_float. The same applies to some other ios_base_fail() checks. Thanks, Piotr

2 1

Re: include: don't redefine _EVENT_TYPE or _TIMER_TYPE (resend)
by Alexandre Julliard March 10, 2015

March 10, 2015

Austin English <austinenglish(a)gmail.com> writes: > Previous try was dropped without comment. I guess I don't see where you are going with that. ntdef.h is supposed to contain a ton of definitions that are already in other headers, are you planning to add a #ifdef around every single one of them? That doesn't seem very workable. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 2/5] d3d9/tests: Add a test for UpdateTexture.
by Marvin March 10, 2015

March 10, 2015

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=12038 Your paranoid android. === w8 (32 bit visual) === visual.c:18055: Test failed: Got unexpected color 0x00000000, case 0, 6. visual.c:18055: Test failed: Got unexpected color 0x00303000, case 0, 7. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 0, 10. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 0, 12. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 1, 1. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 1, 2. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 1, 3. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 1, 4. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 1, 5. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 1, 6. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 1, 7. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 1, 8. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 1, 10. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 1, 12. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 2, 1. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 2, 2. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 2, 3. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 2, 4. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 2, 5. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 2, 6. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 2, 7. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 2, 8. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 2, 10. visual.c:18055: Test failed: Got unexpected color 0xdeadbeec, case 2, 12.

2 1

Re: [PATCH 3/5] d3d8/tests: Add a test for UpdateTexture.
by Marvin March 10, 2015

March 10, 2015

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=12039 Your paranoid android. === w8 (32 bit visual) === visual.c:5983: Test failed: Got unexpected color 0x00000000, case 0, 6. visual.c:5983: Test failed: Got unexpected color 0x00303000, case 0, 7. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 0, 10. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 0, 12. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 1, 1. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 1, 2. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 1, 3. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 1, 4. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 1, 5. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 1, 6. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 1, 7. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 1, 8. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 1, 10. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 1, 12. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 2, 1. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 2, 2. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 2, 3. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 2, 4. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 2, 5. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 2, 6. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 2, 7. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 2, 8. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 2, 10. visual.c:5983: Test failed: Got unexpected color 0xdeadbeec, case 2, 12.

1 0

Re: [PATCH 2/3] gdiplus: Convert gif frame to DIB on load
by Vincent Povirk March 10, 2015

March 10, 2015

I don't like that we essentially have another instance here of the bitmap create/destroy logic. I think I'd rather we create another bitmap than use the existing one with new image data. It's not clear to me that decode_image_wic is even worth calling. Most of the logic in there is for reading image bits, and couldn't we in theory use blit_gif_frame or select_image_gif from the next patch for that? Or, if we wanted to add another hook to it, we could ask decode_image_wic to not attempt to find a matching format, and instead fall back on the conversion to 32bppARGB.

2 1