November 2015 - Wine-devel - list.winehq.org

Re: [PATCH v4] ucrtbase: Hook up _seh_filter_exe and _seh_filter_dll
by Piotr Caban 09 Nov '15

09 Nov '15

On 11/09/15 12:36, Martin Storsjo wrote: > +@ cdecl _seh_filter_dll(long ptr) __CppXcptFilter > +@ cdecl _seh_filter_exe(long ptr) __CppXcptFilter _seh_filter_dll should behave like __CppXcptFilter and _seh_filter_exe like _XcptFilter. You also need to copy -arch=i386,x86_64,arm parameter. Otherwise you break compilation for other architectures because __CppXcptFilter is not defined there. Thanks, Piotr

2 1

Re: [PATCH 4/4] ntdll: Added NtNotifyChangeMultipleKeys implementation.
by Dmitry Timoshkov 09 Nov '15

09 Nov '15

Hi Jacek, Jacek Caban <jacek(a)codeweavers.com> wrote: > diff --git a/dlls/ntdll/ntdll.spec b/dlls/ntdll/ntdll.spec > index 464b233..d9bab90 100644 > --- a/dlls/ntdll/ntdll.spec > +++ b/dlls/ntdll/ntdll.spec > @@ -205,7 +205,7 @@ > # @ stub NtModifyBootEntry > @ stdcall NtNotifyChangeDirectoryFile(long long ptr ptr ptr ptr long long long) > @ stdcall NtNotifyChangeKey(long long ptr ptr ptr long long ptr long long) > -# @ stub NtNotifyChangeMultipleKeys > +@ stdcall NtNotifyChangeMultipleKeys(long long ptr long ptr ptr ptr long long ptr long long) You forgot to update the Zw* .spec entry. -- Dmitry.

1 0

Re: [PATCH v5] ucrtbase: Hook up _seh_filter_exe and _seh_filter_dll
by Piotr Caban 09 Nov '15

09 Nov '15

On 11/09/15 12:46, Martin Storsjo wrote: > +@ cdecl -arch=i386,x86_64,arm _seh_filter_dll(long ptr) __CppXcptFilter > +@ cdecl -arch=i386,x86_64,arm _seh_filter_exe(long ptr) __CppXcptFilter This is still wrong.

2 3

Re: [PATCH v3] ucrtbase: Implement _seh_filter_exe and _seh_filter_dll
by Piotr Caban 09 Nov '15

09 Nov '15

On 11/05/15 13:52, Martin Storsjo wrote: > +int CDECL _seh_filter_dll(MSVCRT_ulong ex, PEXCEPTION_POINTERS ptr) > +{ > + TRACE("(%08x,%p)\n", ex, ptr); > + /* I assume ptr->ExceptionRecord->ExceptionCode is the same as ex */ > + return msvcrt_exception_filter(ptr); > +} This function should behave like __CppXcptFilter. Thanks, Piotr

2 2

Re: [PATCH] oledb32/tests: Add more GetConversionSize tests
by Nikolay Sivov 09 Nov '15

09 Nov '15

On 09.11.2015 12:49, Alistair Leslie-Hughes wrote: > It appears that if the type being converted is of a fixed > size then it returns a constant value of 110. > I don't think it makes sense to have such magic size in tests, until there's an understanding what DataConvert returns as destination in such cases.

1 0

Re: [4/5] gdi32: Embolden width direction only if ppem <= 50. (resend)
by Huw Davies 09 Nov '15

09 Nov '15

On Fri, Nov 06, 2015 at 09:25:11PM +0900, Akihiro Sagawa wrote: > > In other words, avoid increasing the thickness of height direction if > the size is small. This prevents distorting fonts as in Bug 34884. > This feature requires newer FreeType library (2.4.10 and/or later). What does Windows do here? Does it really have a cutoff at 50 ppem? I suspect we should always just stretch in the x-direction irrespective of the ppem (at least in the case where we have FT_Outline_EmboldenXY). Huw.

1 0

Re: [3/5] gdi32: Implement synthesizing bold glyphs for bitmap fonts. (resend)
by Huw Davies 09 Nov '15

09 Nov '15

On Fri, Nov 06, 2015 at 09:25:09PM +0900, Akihiro Sagawa wrote: > @@ -6970,7 +6971,16 @@ static DWORD get_glyph_outline(GdiFont *incoming_font, UINT glyph, UINT format, > INT w = min( pitch, (ft_face->glyph->bitmap.width + 7) >> 3 ); > INT h = min( height, ft_face->glyph->bitmap.rows ); > while(h--) { > - memcpy(dst, src, w); > + if (!font->fake_bold) > + memcpy(dst, src, w); > + else { > + INT x; > + dst[0] = 0; > + for (x = 0; x < w; x++) { > + dst[x ] = (dst[x] & 0x80) | (src[x] >> 1) | src[x]; > + dst[x+1] = (src[x] & 0x01) << 7; This could write past the end of dst (and similarly for ones below). Huw.

1 0

Re: [PATCH 3/3] quartz: Fix error code when chekcing circular connection.
by Qian Hong 09 Nov '15

09 Nov '15

Please drop this patch, I find the actual define for 80040217 this time, will resend soon.

1 0

Re: [PATCH] gdi32: Initialize system font link registry.
by Huw Davies 09 Nov '15

09 Nov '15

On Wed, Nov 04, 2015 at 09:40:38PM +0800, Qian Hong wrote: > +static void update_font_system_link_info(UINT current_ansi_codepage) > + if (RegCreateKeyA(HKEY_LOCAL_MACHINE, font_link_reg_key, &hkey) == ERROR_SUCCESS) > + { > + HKEY hsubkey; > + if (RegCreateKeyA(hkey, system_link_subkey, &hsubkey) == ERROR_SUCCESS) > + { You could create this key in one call. Please use the unicode registry functions for both key creation and value setting (yes, I know there are many ansi versions already there, but let's avoid adding more!). You'll find the system link unicode string already defined in 'system_link'. Huw.

2 1

Re: taskschd/tests: Fix memory leaks (Valgrind).
by Marvin 08 Nov '15

08 Nov '15

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=18033 Your paranoid android. === w1064 (32 bit scheduler) === scheduler.c:159: Test failed: expected ERROR_INVALID_NAME, got 0x80070002 scheduler.c:204: Test failed: expected ERROR_PATH_NOT_FOUND, got 0x80070002 scheduler.c:354: Test failed: expected ERROR_FILE_NOT_FOUND, got 0 scheduler.c:767: Test failed: expected ERROR_PATH_NOT_FOUND, got 0x80070002 scheduler.c:773: Test failed: expected ERROR_PATH_NOT_FOUND, got 0x80070002 scheduler.c:941: Test failed: expected ERROR_FILE_NOT_FOUND, got 0

1 0