July 2014 - Wine-Devel - list.winehq.org

Re: [PATCH 1/2] riched20: Implement ITextSelection::GetText. (try 3)

by Jactry Zeng

2014-07-04 19:00 GMT+08:00 Jactry Zeng <jactry92(a)gmail.com>: > > 1. end-of-story is different from end-of-paragraph > From document of ITextDocument::GetStoryCount: Rich edit controls have only one story.[1] > So ‘story’ may is an unit like 'paragraph' in Rich edit control. We may need add a MERF_ENDSTORY to implement it.[2] > > 2. What does the end-of-story looks like? > When we run the wordpad with native riched20, we can use ‘ctrl + shitf + →’ to select the ‘space’ at the end of the Rich edit control. But we cannot set the cursor behind it.(a picture here[3]) > And wine’s riched20 didn’t implement this ‘space’. This 'space' may is the end-of-story. > > A crash will happen when we copy the ‘space’ from a native Rich edit control to a builtin one.(backtrace.txt[4]) > > BTW, there is a know bug of EM_SETSEL: cannot select the last character of Rich edit control.[5] > > 3. > From the last paragraph of this blog: http://blogs.msdn.com/b/murrays/archive/2008/11/22/paragraphs-and-paragraph… > It is sure there is a final EOP in a Rich edit control. > > There seems to be every indication that we should implement the final EOP of Rich edit control rather than just append a ‘\r’ in ITextSelection::GetText. > So I just let a FIXME comment there. I am not sure if I am right. I missed some links here: [1] http://msdn.microsoft.com/en-us/library/windows/desktop/bb774027(v=vs.85).a… [2] http://source.winehq.org/git/wine.git/blob/HEAD:/dlls/riched20/editstr.h#l1… [3] http://img.vim-cn.com/6f/dbc090928d97b2cdf53838278087fa95660b9b.png [4] https://gist.github.com/Jactry/cae7d739162fca4f072f [5] http://source.winehq.org/git/wine.git/blob/HEAD:/dlls/riched20/tests/editor… -- Regards, Jactry Zeng

July 15, 2014

Re: [2/2] ddraw: add dsurface dimension tests, try 11

by Stefan Dösinger

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am 2014-07-15 10:20, schrieb Patrick Rudolph: > Stefan pointed out that on old cards this test might fail. If I remember correctly the problem was that some cards have a pitch limit of something slightly smaller than 2^17. So a width = 0xffff surface fails on some cards unless the format is a 8 bpp format. Testing a height = 0xffff surface should be fine though. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQIcBAEBAgAGBQJTxTB3AAoJEN0/YqbEcdMwM44P/2qYCGmCojro8ce1YCriT6vk AEu6LzICoU65//UNQuTsravgHyGMWZldTmcNvoGzFPHN+qsk+17iPxCqzFVYQ7Gx IPtfFB49gpxJzo/n3n4EKSqusUrNIru3g/uR7FAHVgcRk2I5VUnU4eNDbat57HCl Z8/8yHAtqII1q/xMyWRFbCU6q10j0byKop5lT9tSQdSRwR13G+syku3OZ+1KIhld q3H6C7ovdGUoKykZRSni1H0TEQMmC2Eo2WKdL5wJHClHi53PIctb9clKj5MRA7vj Mw5zfrISyL8MOcpVqo1C3borWWEsLCyLpsBB2qgoEky6B7iDpMI4wR0YSBze9s5G cf4oQBIWa+o3HNEZDIEvdw3LVD0Mbhh9dRM5Z2CXOhi+r8c8DT6Xs1SFoQ/We1U0 kLt3QfVJBR8dP8N2JKPj6+Y6Wd0GGOmqeljFKX6USq8yhJfqyP0fZzvv1myFBIco w8umGm93n6TCXz/gTz3H+zqtbICb/UEsoqGkd+jFXqOdI3fSy5eejo7LUhmNifNl zRA+cNIAtlO4T6Ghirs9yNoP1fQP0rgDP6Qem+4fWXq1KR9WqD/98Sfa4Qn7TVB6 SC7pVxRYbMpfS75FPFSnCU48x+9mPqQjMZXCcUznAZLaS04/XGTaXT09u4HHnO+7 e6qQ1p1QRn+KShQ82nFq =YMbh -----END PGP SIGNATURE-----

July 15, 2014

Re: [PATCH 1/2] riched20: Implement ITextSelection::GetText. (try 3)

by Jactry Zeng

2014-07-15 21:29 GMT+08:00 Huw Davies <huw(a)codeweavers.com>: > > Can you at least point to a test that shows this bug? Do you mean tests about bug of S/GetSelection? It seem is a known bug: http://source.winehq.org/git/wine.git/blob/HEAD:/dlls/riched20/tests/editor… dlls/riched20/tests/editor.c:4352 struct exsetsel_s { LONG min; LONG max; LRESULT expected_retval; int expected_getsel_start; int expected_getsel_end; int _getsel_todo_wine; }; const struct exsetsel_s exsetsel_tests[] = { /* sanity tests */ {5, 10, 10, 5, 10, 0}, {15, 17, 17, 15, 17, 0}, /* test cpMax > strlen() */ {0, 100, 18, 0, 18, 1}, /* test cpMin == cpMax */ {5, 5, 5, 5, 5, 0}, /* test cpMin < 0 && cpMax >= 0 (bug 4462) */ {-1, 0, 5, 5, 5, 0}, {-1, 17, 5, 5, 5, 0}, {-1, 18, 5, 5, 5, 0}, /* test cpMin < 0 && cpMax < 0 */ {-1, -1, 17, 17, 17, 0}, {-4, -5, 17, 17, 17, 0}, /* test cMin >=0 && cpMax < 0 (bug 6814) */ {0, -1, 18, 0, 18, 1}, {17, -5, 18, 17, 18, 1}, {18, -3, 17, 17, 17, 0}, /* test if cpMin > cpMax */ {15, 19, 18, 15, 18, 1}, {19, 15, 18, 15, 18, 1} }; -- Regards, Jactry Zeng

July 15, 2014

Re: [PATCH 1/2] riched20: Implement ITextSelection::GetText. (try 3)

by Huw Davies

On 15 Jul 2014, at 14:13, Jactry Zeng wrote: > > 2014-07-15 21:00 GMT+08:00 Huw Davies <huw(a)codeweavers.com>: > > > > My point is, we should fix the bug in S/GetSelection first. There's no point in working around this bug. > > I still have no idea about this bug. :( > Can I just let a FIXME there now? So I can continue work of ITextSelection and ITextRange. No, as whatever you write will need fixing once the underlying bug is fixed. Can you at least point to a test that shows this bug? Huw.

July 15, 2014

Re: [PATCH 1/2] riched20: Implement ITextSelection::GetText. (try 3)

by Jactry Zeng

2014-07-15 21:00 GMT+08:00 Huw Davies <huw(a)codeweavers.com>: > > My point is, we should fix the bug in S/GetSelection first. There's no point in working around this bug. I still have no idea about this bug. :( Can I just let a FIXME there now? So I can continue work of ITextSelection and ITextRange. -- Regards, Jactry Zeng

July 15, 2014

Re: [8/8] msi: Avoid Wow6432Node usage in registry table handling.

by Marvin

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=8000 Your paranoid android. === w2000pro (32 bit install) === The test timed out === wxppro (32 bit install) === The test timed out

July 15, 2014

Re: [PATCH 1/2] riched20: Implement ITextSelection::GetText. (try 3)

by Huw Davies

On 15 Jul 2014, at 13:51, Jactry Zeng wrote: > Thanks for your reply. > I am not sure if my understanding is right or not... > > + ME_GetSelection(This->reOle->editor, &start, &end); > + nChars = ME_GetCursorOfs(end) - ME_GetCursorOfs(start); > + bEOP = (end->pRun->type == diTextEnd); > > Here I would like to use ME_GetSelection() to get the end of the selection range. And than judge if it is > the end of the text buffer and pass bEOP into ME_GetTextW. > If bEOP is true, ME_GetTextW will append a '\r' in the end of the return string. > But because of bug of ME_GetSelection/ME_SetSelection, I can't got a ME_Cursor of the end of the text buffer. My point is, we should fix the bug in S/GetSelection first. There's no point in working around this bug. Huw.

July 15, 2014

Re: [PATCH 1/2] riched20: Implement ITextSelection::GetText. (try 3)

by Jactry Zeng

Hi, Thanks for your reply. I am not sure if my understanding is right or not... + ME_GetSelection(This->reOle->editor, &start, &end); + nChars = ME_GetCursorOfs(end) - ME_GetCursorOfs(start); + bEOP = (end->pRun->type == diTextEnd); Here I would like to use ME_GetSelection() to get the end of the selection range. And than judge if it is the end of the text buffer and pass bEOP into ME_GetTextW. If bEOP is true, ME_GetTextW will append a '\r' in the end of the return string. But because of bug of ME_GetSelection/ME_SetSelection, I can't got a ME_Cursor of the end of the text buffer. 2014-07-15 17:23 GMT+08:00 Huw Davies <huw(a)codeweavers.com>: > On 15 Jul 2014, at 02:18, Jactry Zeng wrote: > > But because of bug of EM_SETSEL or EM_GETSEL, this feature > > still cannot be implemented completely. > > For example: > > When the text in Rich Edit control is: "TestSomeText" > > 01234567890123 > > We cannot distinguish between SendMessageA(w, EM_SETSEL, 0, 12) and > SendMessageA(w, EM_SETSEL, 0, 13). > > Attachment is the patch, I just added wine_todo for failed tests cause > by this bug. > > But I still have no idea about fixing this bug. Could you give me some > suggestion? > > Or just mark them as todo now is OK? > > Do you fully understand this bug yet? In other words under what > conditions does the current implementation fail? > It should be trivial to fix up ME_SetSelection(). > > Huw. > > -- Regards, Jactry Zeng

July 15, 2014

Re: [7/8] msi/tests: Add more registry tests.

by Marvin

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=7999 Your paranoid android. === wxppro (32 bit install) === The test timed out === w7u (32 bit install) === The test timed out === w8 (32 bit install) === The test timed out

July 15, 2014

Re: [PATCH 5/5] d3d8/tests: Get rid of struct nvertex.

by Henri Verbeet

These are all existing failures, unfortunately they don't always happen consistently.

July 15, 2014