September 2014 - Wine-devel

Re: [PATCH 3/5] riched20: Implement ITextRange::GetDuplicate. (try 3)
by Huw Davies 17 Sep '14

17 Sep '14

On Tue, Sep 16, 2014 at 06:39:20PM +0800, Jactry Zeng wrote: > + hres = CreateITextRange(This->reOle, This->start, This->end, ppRange); > + if (FAILED(hres)) > + return E_FAIL; > + return hres; Actually since CreateITextRange can only fail in out-of-memory cases we can't easily test this behaviour and there's no guarantee that the docs are correct anyway. So just return CreateITextRange(..) here would be cleaner. Huw.

1 0

Re: [PATCH 4/8] riched20: Implement ITextRange::GetDuplicate. (try 2)
by Huw Davies 16 Sep '14

16 Sep '14

On Fri, Sep 12, 2014 at 08:33:42PM +0800, Jactry Zeng wrote: > diff --git a/dlls/riched20/richole.c b/dlls/riched20/richole.c > index a88aff2..b85ffda 100644 > --- a/dlls/riched20/richole.c > +++ b/dlls/riched20/richole.c > + hres = CreateITextRange(This->reOle, This->start, This->end, ppRange); > + if(hres) > + return E_FAIL; This isn't the way to test for failure. Use the FAILED() macro. Huw.

2 1

Re: [PATCH 1/2] riched20: Implement ITextRange::GetChar.
by Alexandre Julliard 15 Sep '14

15 Sep '14

Jactry Zeng <wine(a)jactry.com> writes: > +#define TEST_TXTRGE_GETCHAR(first, lim, expected_char) \ > + create_interfaces(&w, &reOle, &txtDoc, NULL); \ > + SendMessageA(w, WM_SETTEXT, 0, (LPARAM)test_text1); \ > + ITextDocument_Range(txtDoc, first, lim, &txtRge); \ > + pch = 0xdeadbeef; \ > + hres = ITextRange_GetChar(txtRge, &pch); \ > + ok(hres == S_OK, "ITextRange_GetChar\n"); \ > + ok(pch == expected_char, "got wrong char: %c\n", pch); \ > + ITextRange_Release(txtRge); \ > + release_interfaces(&w, &reOle, &txtDoc, NULL); Please avoid that sort of huge macro. If necessary, use helper functions instead. -- Alexandre Julliard julliard(a)winehq.org

2 1

Problem with CompareString/CompareStringEx - accented characters
by Virgo Pärna 15 Sep '14

15 Sep '14

I have been trying to figure out, what is wrong with wine's implementation of CompareString and CompareStringEx functions. From what I can figure out, those functions completely ignore locale passed to them. That alone is not a critical issue, but what causes problems to me, is that it seems always act, as if NORM_IGNORENONSPACE was passed (comparision shows, that A and Ä are same). I did try to rip relavent parts out of wine for test utility and it seems, that A and Ä have same value in collation_table which is used by compare_diacritic_weights. Could it be, that the make_unicode's READ_SORTKEYS_FILE looses relavent information, when generating collation.c file? My perl is even worse than my C, so figuring out those regexes is way beyond me. I did spam https://bugs.winehq.org/show_bug.cgi?id=9583 with my analysis. -- Virgo Pärna virgo.parna(a)mail.ee

1 0

Re: [PATCH 6/8] d3d8: Mark tests which no longer pass due to reverts as todo_wine.
by Marvin 15 Sep '14

15 Sep '14

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=8913 Your paranoid android. === w8 (32 bit device) === device.c:2398: Test failed: Expected foreground window 000B00CA, got 0002004C. device.c:3040: Test failed: Expected wndproc 0x412cf0, got 0x412cf0.

2 1

Re: [PATCH 5/8] Revert "wined3d: Restore the pixel format of the window whose pixel format was actually changed."
by Ken Thomases 15 Sep '14

15 Sep '14

It seems I accidentally sent this one twice. Please ignore the duplicate, but the series should be OK otherwise. -Ken On Sep 14, 2014, at 7:47 PM, Ken Thomases <ken(a)codeweavers.com> wrote: > > This reverts commit 4c4552c5a1910a9d5adf8eccff0ac62d89ffe376. > > For: > https://bugs.winehq.org/show_bug.cgi?id=35655 > https://bugs.winehq.org/show_bug.cgi?id=35718 > https://bugs.winehq.org/show_bug.cgi?id=35950 > https://bugs.winehq.org/show_bug.cgi?id=35975 > https://bugs.winehq.org/show_bug.cgi?id=36900 > --- > dlls/wined3d/context.c | 87 +++++++++++------------------------------- > dlls/wined3d/wined3d_private.h | 1 - > 2 files changed, 22 insertions(+), 66 deletions(-) > > <0005-Revert-wined3d-Restore-the-pixel-format-of-the-windo.patch>

1 0

Re: wininet: Avoid compiler warnings due to negation before binary operator. (RESEND)
by Alexandre Julliard 12 Sep '14

12 Sep '14

Gerald Pfeifer <gerald(a)pfeifer.com> writes: > As of last week, what will become GCC 5.0 does warn about this as > part of -Wall. > > There are a couple of cases in Wine, such as this, which I'd like > to address. Please find a way that doesn't make the code unnecessarily ugly. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 1/8] riched20: Stub for ITextRange interface and implement ITextDocument::Range. (try 5)
by Huw Davies 12 Sep '14

12 Sep '14

On Fri, Sep 12, 2014 at 08:33:29PM +0800, Jactry Zeng wrote: > Thanks Huw and Sebastian. > > Superseded patch 106499. > > ChangeLog: > try 5: > - Alloc txtRge in CreateITextRange() > try 4: > - Simplify with macros in include/wine/list.h > - Using LONG instead of int in This->start and This->end > - Limit cp1 to (0, TextLength) > --- > dlls/riched20/richole.c | 730 > +++++++++++++++++++++++++++++++++++++++++- > dlls/riched20/tests/richole.c | 35 +- > 2 files changed, 762 insertions(+), 3 deletions(-) This first one looks ok to me. You'll need to re-send the rest of the series, but it would be better to wait until this gets in. Huw.

1 0

Re: [PATCH 1/8] riched20: Stub for ITextRange interface and implement ITextDocument::Range. (resend try 4)
by Huw Davies 12 Sep '14

12 Sep '14

On Thu, Sep 11, 2014 at 05:35:38PM +0800, Jactry Zeng wrote: > +static HRESULT CreateITextRange(IRichEditOleImpl *reOle, LONG start, LONG end, > + ITextRangeImpl *txtRge, ITextRange** ppRange) > +{ > + txtRge->ITextRange_iface.lpVtbl = &trvt; > + txtRge->ref = 1; > + txtRge->reOle = reOle; > + txtRge->start = start; > + txtRge->end = end; > + list_add_head(&reOle->rangelist, &txtRge->entry); > + *ppRange = &txtRge->ITextRange_iface; > + return S_OK; > +} This looks very suspicious to me. You're storing a reference to 'reOle' so you should AddRef it somewhere. It may be better to pass the interface ptr, AddRef and store that. Also, is the purpose of rangelist just to be able to NULL out reOle when the parent object is destroyed? If so, this can go away if you correctly handle the ref counting. Huw.

3 13

Re: reg: Add reg.exe test suite. (try 3)
by Marvin 11 Sep '14

11 Sep '14

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=8896 Your paranoid android. === build (build) === Patch failed to apply

1 0