- Wine-Devel - list.winehq.org

[PATCH] kernelbase: Native machine is optional for IsWow64Process2.
by Dmitry Timoshkov Aug. 18, 2020

Aug. 18, 2020

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=49716 Signed-off-by: Dmitry Timoshkov <dmitry(a)baikal.ru> --- dlls/kernel32/tests/process.c | 15 +++++++++++++++ dlls/kernelbase/process.c | 11 ++++++++--- 2 files changed, 23 insertions(+), 3 deletions(-) diff --git a/dlls/kernel32/tests/process.c b/dlls/kernel32/tests/process.c index eaa5da2e95..3162ba3c53 100644 --- a/dlls/kernel32/tests/process.c +++ b/dlls/kernel32/tests/process.c @@ -2238,6 +2238,12 @@ static void test_IsWow64Process2(void) ok(machine == IMAGE_FILE_MACHINE_UNKNOWN, "got %#x\n", machine); ok(native_machine == expect_native, "got %#x\n", native_machine); + SetLastError(0xdeadbeef); + machine = 0xdead; + ret = pIsWow64Process2(pi.hProcess, &machine, NULL); + ok(ret, "IsWow64Process2 error %u\n", GetLastError()); + ok(machine == IMAGE_FILE_MACHINE_UNKNOWN, "got %#x\n", machine); + ret = TerminateProcess(pi.hProcess, 0); ok(ret, "TerminateProcess error\n"); @@ -2263,6 +2269,15 @@ static void test_IsWow64Process2(void) ok(machine == IMAGE_FILE_MACHINE_UNKNOWN, "got %#x\n", machine); ok(native_machine == expect_native, "got %#x\n", native_machine); } + + SetLastError(0xdeadbeef); + machine = 0xdead; + ret = pIsWow64Process2(GetCurrentProcess(), &machine, NULL); + ok(ret, "IsWow64Process2 error %u\n", GetLastError()); + if (is_wow64) + ok(machine == IMAGE_FILE_MACHINE_I386, "got %#x\n", machine); + else + ok(machine == IMAGE_FILE_MACHINE_UNKNOWN, "got %#x\n", machine); } static void test_SystemInfo(void) diff --git a/dlls/kernelbase/process.c b/dlls/kernelbase/process.c index 58fba52415..cb7488f0e4 100644 --- a/dlls/kernelbase/process.c +++ b/dlls/kernelbase/process.c @@ -898,8 +898,6 @@ BOOL WINAPI DECLSPEC_HOTPATCH IsWow64Process2( HANDLE process, USHORT *machine, if (wow64) { - GetNativeSystemInfo( &si ); - if (process != GetCurrentProcess()) { #if defined(__i386__) || defined(__x86_64__) @@ -915,15 +913,22 @@ BOOL WINAPI DECLSPEC_HOTPATCH IsWow64Process2( HANDLE process, USHORT *machine, nt = RtlImageNtHeader( NtCurrentTeb()->Peb->ImageBaseAddress ); *machine = nt->FileHeader.Machine; } + + if (!native_machine) return TRUE; + + GetNativeSystemInfo( &si ); } else { + *machine = IMAGE_FILE_MACHINE_UNKNOWN; + + if (!native_machine) return TRUE; + #ifdef _WIN64 GetSystemInfo( &si ); #else GetNativeSystemInfo( &si ); #endif - *machine = IMAGE_FILE_MACHINE_UNKNOWN; } switch (si.u.s.wProcessorArchitecture) -- 2.26.2

2 1

[PATCH] conhost: Build with -mwindows.
by Jacek Caban Aug. 18, 2020

Aug. 18, 2020

Signed-off-by: Jacek Caban <jacek(a)codeweavers.com> --- programs/conhost/Makefile.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

1 0

[PATCH] shell32: Free original_dos_name with correct function in TRASH_GetDetails() (Coverity).
by Gijs Vermeulen Aug. 18, 2020

Aug. 18, 2020

Signed-off-by: Gijs Vermeulen <gijsvrm(a)gmail.com> --- dlls/shell32/trash.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/dlls/shell32/trash.c b/dlls/shell32/trash.c index 93b4ee4588..f34b106bd6 100644 --- a/dlls/shell32/trash.c +++ b/dlls/shell32/trash.c @@ -662,7 +662,7 @@ static HRESULT TRASH_GetDetails(const TRASH_BUCKET *bucket, LPCSTR filename, WIN if (original_dos_name != NULL) { lstrcpynW(data->cFileName, original_dos_name, MAX_PATH); - SHFree(original_dos_name); + heap_free(original_dos_name); } else { -- 2.28.0

2 1

Translation problem with oledb32
by Julian Rüger Aug. 18, 2020

Aug. 18, 2020

Hi everyone! I stumbled upon a small problem while translating the not-quite-recent oledb32 changes: #: version.rc:92 winemine.rc:65 msgid "Advanced" msgstr "Fortgeschrittene" That translation makes sense for winemine, but the correct translation for oledb32 would be "Erweitert". That however would sound very weird in winemine, especially if I don't change the other difficulty strings as well. In fact, I can't come up with something that fits in both cases without sounding weird. Is it possible to split this into two separate msgids? Thanks and have a nice weekend!

2 2

[PATCH] explorerframe: return S_OK in ITaskbarList::RegisterTab
by Louis Lenders Aug. 18, 2020

Aug. 18, 2020

https://bugs.winehq.org/show_bug.cgi?id=49719 Signed-off-by: Louis Lenders <xerox.xerox2000x(a)gmail.com> --- dlls/explorerframe/taskbarlist.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/dlls/explorerframe/taskbarlist.c b/dlls/explorerframe/taskbarlist.c index 1a0eb7567d..6f24f35992 100644 --- a/dlls/explorerframe/taskbarlist.c +++ b/dlls/explorerframe/taskbarlist.c @@ -161,7 +161,7 @@ static HRESULT STDMETHODCALLTYPE taskbar_list_RegisterTab(ITaskbarList4 *iface, { FIXME("iface %p, hwndTab %p, hwndMDI %p stub!\n", iface, hwndTab, hwndMDI); - return E_NOTIMPL; + return S_OK; } static HRESULT STDMETHODCALLTYPE taskbar_list_UnregisterTab(ITaskbarList4 *iface, HWND hwndTab) -- 2.28.0

1 0

[PATCH RFC 0/5] Timer Resolution
by Arkadiusz Hiler Aug. 18, 2020

Aug. 18, 2020

TL;DR: Good learning experience, lead to a fix in winmm but IMO we are better off defaulting to 1ms resolution and avoid all the extra complications. The whole thing started because of miss-assigning blame for DOSBox regression to mismatched GetTickCount() and Sleep() resolutions. Initial attempts at fixing the issue just made Sleep() work in increments of 15.6ms. Then, I have learned about timer resolution and I have started implementing this series. As soon as I had working timeBeginPeriod() DOSBox regressed again. After some confusing debugging of "phantom" calls to GetTickCount() with 1-call stacktraces and building myself and unoptimized debug version of SDL and DOSBox I have found out that, in Wine, timeGetTime() is an alias for GetTickCount() and this has caused the problem. More details here: https://www.winehq.org/pipermail/wine-devel/2020-August/171965.html I have started questioning usefulness and feasibility of implementing resolution support: - introducing extra branching for each Sleep / Wait call - multiple places to maintain (some wait calls use server_wait, other call server directly) - a lot of Sleep(1) to assess across Wine's codebase - per-process implementation may be not sufficient for some multi-process software I did testing with popular software and a lot of things set resolution: * Chrome and a lot of Electron-based apps - 1ms with multimedia content, fluctuating otherwise * Firefox - 8ms when downloading, 5ms when playing music/videos * Visual Studio - 1ms all the time * Discord - 1ms all the time while being on a voice chat * Spotify - oscillating 1-5ms when playing music * EGS - 1ms * many many games Resolution is global - if one process requests high resolution (i.e. low update interval) the whole system gets it. Conjecture: If having 1ms resolution all the time / by default would cause any noticeable problems / performance issue with software, it would be seen on Windows and loudly complained about. I think it's worth abandoning the effort to implement proper resolution setting, default to 1ms and implement more convincing stubs for timeBeginPeriod() and NtSetTimerResolution() and friends. Any thoughts? Cc: Rémi Bernon <rbernon(a)codeweavers.com> Cc: Paul Gofman <pgofman(a)codeweavers.com> Arkadiusz Hiler (5): ntdll: Implmement NtQueryTimerResolution and NtSetTimerResolution ntdll: Make wait/delay calls respect the timer resolution server: Set the worst case time for updating user_shared_data to 15ms ntdll: Remove todo_wine from tick count vs sleep test winmm: Implement timeBeginPeriod() and timeEndPeriod() dlls/ntdll/tests/time.c | 187 ++++++++++++++++++++++++++++++++- dlls/ntdll/unix/server.c | 16 ++- dlls/ntdll/unix/sync.c | 84 +++++++++++++-- dlls/ntdll/unix/unix_private.h | 10 ++ dlls/winmm/tests/timer.c | 46 +++++++- dlls/winmm/time.c | 76 +++++++++++--- server/fd.c | 2 +- 7 files changed, 396 insertions(+), 25 deletions(-) -- 2.28.0

3 11

[PATCH] winapi: Update win16.api.
by Francois Gouget Aug. 18, 2020

Aug. 18, 2020

Wine no longer has winedos16.dll and wprocs.dll and display.drv, keyboard.drv and mouse.drv have a .drv extension, not a .dll one. Signed-off-by: Francois Gouget <fgouget(a)free.fr> --- tools/winapi/win16.api | 22 +++------------------- 1 file changed, 3 insertions(+), 19 deletions(-) diff --git a/tools/winapi/win16.api b/tools/winapi/win16.api index ea646c9ec8d..5e78e544dac 100644 --- a/tools/winapi/win16.api +++ b/tools/winapi/win16.api @@ -217,7 +217,7 @@ LPSTR WORD -%%display.dll +%%display.drv %long @@ -362,7 +362,7 @@ INT16 UINT16 WORD -%%keyboard.dll +%%keyboard.drv %long @@ -628,7 +628,7 @@ MMRESULT16 UINT16 WORD -%%mouse.dll +%%mouse.drv %ptr @@ -1408,12 +1408,6 @@ CONTEXT86 * void -%%winedos16.dll - -%void - -void - %%wineps16.drv %long @@ -1556,13 +1550,3 @@ HWTHOOK16 INT16 UINT16 WPARAM16 - -%%wprocs.dll - -%ptr - -CONTEXT86 * - -%void - -void -- 2.20.1

1 0

[PATCH] winapi: Remove obsolete entries from win32.api.
by Francois Gouget Aug. 18, 2020

Aug. 18, 2020

Wine no longer has d3dx8.dll, d3dxof.dll, glut32.dll, msacmmap.dll, winedos.dll, wineesd.drv, winenas.drv and winetty.drv dlls. Signed-off-by: Francois Gouget <fgouget(a)free.fr> --- tools/winapi/win32.api | 189 ----------------------------------------- 1 file changed, 189 deletions(-) diff --git a/tools/winapi/win32.api b/tools/winapi/win32.api index ed14ecc034e..d9b77dbce39 100644 --- a/tools/winapi/win32.api +++ b/tools/winapi/win32.api @@ -924,48 +924,6 @@ LPCWSTR void -%%d3dx8.dll - -%long - -DWORD -HFONT -HRESULT -UINT - -%float - -FLOAT - -%ptr - -D3DXCOLOR * -D3DXMATRIX * -D3DXPLANE * -D3DXQUATERNION * -D3DXVECTOR2 * -D3DXVECTOR3 * -D3DXVECTOR4 * -D3DVIEWPORT8 * -FLOAT * -LPCVOID -LPD3DXBUFFER * -LPD3DXFONT * -LPD3DXMATRIXSTACK * -LPDIRECT3DDEVICE8 - -%str - -LPCSTR - -%wstr - -LPCWSTR - -%void - -void - %%d3dxof.dll %long @@ -1761,31 +1719,6 @@ char * void -%%glut32.dll - -%double - -double - -%float - -float - -%long - -int -unsigned int - -%ptr - -void (*)(int) -void * -void ** - -%void - -void - %%gphoto2.ds %long @@ -2671,16 +2604,6 @@ PHACMSTREAM PWAVEFILTER PWAVEFORMATEX -%%msacmmap.dll - -%long - -DWORD -HDRVR -LONG -UINT -WORD - %%mscms.dll %long @@ -5692,61 +5615,6 @@ IUnknown * IWineD3D * IWineD3DClipper * -%%winedos.dll - -%long - -BOOL -BYTE -DWORD -UINT -WORD -int - -%ptr - -LPVOID -UINT16 * -void * - -%str - -LPCSTR - -%void - -void - -%%wineesd.drv - -%long - -DWORD -HDRVR -LPARAM -LRESULT -UINT -WORD - -%long --pointer - -DWORD_PTR - -%%winejack.drv - -%long - -DWORD -HDRVR -LPARAM -LRESULT -UINT -WORD - -%long --pointer - -DWORD_PTR - %%winejoystick.drv %long @@ -5773,21 +5641,6 @@ UINT DWORD_PTR -%%winenas.drv - -%long - -DWORD -HDRVR -LPARAM -LRESULT -UINT -WORD - -%long --pointer - -DWORD_PTR - %%wineoss.drv %long @@ -5857,48 +5710,6 @@ void LPCWSTR -%%winetty.drv - -%long - -BOOL -COLORREF -DWORD -HANDLE -HBITMAP -HDC -HFONT -HRGN -HWND -INT -LONG -UINT - -%ptr - -BITMAPINFO * -CREATESTRUCTA * -DEVMODEW * -DWORD * -INT * -LPCVOID -LPINT -LPPALETTEENTRY -LPPOINT -LPSIZE -LPTEXTMETRICW -LPVOID -POINT * -RECT * -TTYDRV_PDEVICE * -TTYDRV_PDEVICE ** -WINDOWPOS * -void * - -%wstr - -LPCWSTR - %%winex11.drv %long -- 2.20.1

1 0

[PATCH] fsutil: Initialize ret in hardlink() (Coverity).
by Gijs Vermeulen Aug. 18, 2020

Aug. 18, 2020

Signed-off-by: Gijs Vermeulen <gijsvrm(a)gmail.com> --- programs/fsutil/main.c | 2 +- programs/fsutil/tests/fsutil.c | 6 ++++++ 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/programs/fsutil/main.c b/programs/fsutil/main.c index bde07ac2df..0bc8817b88 100644 --- a/programs/fsutil/main.c +++ b/programs/fsutil/main.c @@ -97,7 +97,7 @@ static int create_hardlink(int argc, WCHAR *argv[]) static int hardlink(int argc, WCHAR *argv[]) { - int ret; + int ret = 0; if (argc > 2) { diff --git a/programs/fsutil/tests/fsutil.c b/programs/fsutil/tests/fsutil.c index 9867788274..1111b5b2fe 100644 --- a/programs/fsutil/tests/fsutil.c +++ b/programs/fsutil/tests/fsutil.c @@ -63,6 +63,12 @@ static void test_hardlink(void) ok(hfile != INVALID_HANDLE_VALUE, "failed to create a file\n"); CloseHandle(hfile); + rc = runcmd("fsutil"); + ok(rc == 0, "failed to run fsutil\n"); + + rc = runcmd("fsutil hardlink"); + ok(rc == 0, "failed to run fsutil hardlink\n"); + rc = runcmd("fsutil hardlink create link file"); ok(rc == 0, "failed to create a hardlink\n"); -- 2.28.0

2 1

[PATCH 1/2] riched20: Implement ITextRange::MoveStart() and ITextRange::MoveEnd() for tomCharacter
by Damjan Jovanovic Aug. 18, 2020

Aug. 18, 2020

Signed-off-by: Damjan Jovanovic <damjan.jov(a)gmail.com> --- dlls/riched20/richole.c | 77 +++++++++++++++++++++++++-- dlls/riched20/tests/richole.c | 98 ++++++++++++++++++++++++++++++++++- 2 files changed, 169 insertions(+), 6 deletions(-)

2 2