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

Re: [PATCH 4/4] mshtml: Add IHTMLXMLHttpRequest::statusText property implementation.
by Jacek Caban 22 Jul '15

22 Jul '15

Hi Zhenbo, On 22/07/15 17:34, Zhenbo Li wrote: > + len = MultiByteToWideChar(CP_ACP, 0, str, -1, NULL, 0); > + *p = SysAllocStringLen(NULL, len); > + if(!*p) { > + nsACString_Finish(nscstr); > + return E_OUTOFMEMORY; > + } > + MultiByteToWideChar(CP_ACP, 0, str, -1, *p, len); I didn't catch it earlier, but this should use CP_UTF8. Cheers, Jacek

1 0

1.7.48 release delayed
by Alexandre Julliard 22 Jul '15

22 Jul '15

Folks, I'm out to my brother's wedding for a few days, so there won't be a release this Friday. I'm planning to release 1.7.48 early next week, probably on Tuesday. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [2/3] setupapi: SetupDiGetClassDevsEx returns INVALID_HANDLE_VALUE on error.
by Nikolay Sivov 22 Jul '15

22 Jul '15

On 22.07.2015 12:57, Dmitry Timoshkov wrote: > --- > dlls/setupapi/devinst.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/dlls/setupapi/devinst.c b/dlls/setupapi/devinst.c > index b084a5e..5a18937 100644 > --- a/dlls/setupapi/devinst.c > +++ b/dlls/setupapi/devinst.c > @@ -2340,7 +2340,7 @@ HDEVINFO WINAPI SetupDiGetClassDevsExW(const GUID *class, PCWSTR enumstr, HWND p > if (!(flags & DIGCF_ALLCLASSES) && !class) > { > SetLastError(ERROR_INVALID_PARAMETER); > - return NULL; > + return INVALID_HANDLE_VALUE; > } > if (flags & unsupportedFlags) > WARN("unsupported flags %08x\n", flags & unsupportedFlags); > How about a test for this and next one?

2 1

Re: winhttp: Implement WINHTTP_OPTION_UNLOAD_NOTIFY_EVENT.
by Dmitry Timoshkov 22 Jul '15

22 Jul '15

Huw Davies <huw(a)codeweavers.com> wrote: > + case WINHTTP_OPTION_UNLOAD_NOTIFY_EVENT: > + TRACE("WINHTTP_OPTION_UNLOAD_NOTIFY_EVENT: %p\n", *(HANDLE *)buffer); > + session->unload_event = *(HANDLE *)buffer; > + return TRUE; I admit, that adding another case that doesn't check input buffer size doesn't change much, but it has to stop at some point, rather sooner than later. -- Dmitry.

1 0

Skyrim hang on exit: Any news?
by Theodore Dubois 22 Jul '15

22 Jul '15

When you quit to desktop from Skyrim, it hangs. This is bug 34041. It's still a bug, but it has not been touched by anyone from Codewaevers since 2013 (as far as I can tell.) Clearly, it's low priority. Any news? ~Theodore

2 1

Re: user32/tests: Test for caption width of 13 on japanese W7
by Marvin 21 Jul '15

21 Jul '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=15482 Your paranoid android. === w8 (32 bit sysparams) === sysparams.c:2489: Test failed: Waiting for the WM_DISPLAYCHANGE message timed out sysparams.c:2502: Test failed: Set bpp 32, but WM_DISPLAYCHANGE reported bpp -1

1 0

Re: ntdll: Don't touch SecurityCookie when the pointer is outside of the image
by Michael Müller 21 Jul '15

21 Jul '15

Am 21.07.2015 um 20:33 schrieb André Hentschel: > + if (loadcfg && loadcfg_size >= sizeof(*loadcfg) && > + (ULONG_PTR)ptr > loadcfg->SecurityCookie &&loadcfg->SecurityCookie < (ULONG_PTR)ptr + total_size) > set_security_cookie((ULONG_PTR *)loadcfg->SecurityCookie); Your check doesn't take the size of the security cookie into account. If the cookie would only be partially inside of the image area, the check would still succeed although not all memory is writable. The check for the lower limit should be (<= instead of >): (ULONG_PTR)ptr <= loadcfg->SecurityCookie Is it possible to have an unmapped space between mapped sections? According to the NT section header it should be possible, but I am not sure if Wine fills this area.

2 1

Re: dwrite: Remove unused initializer (Clang)
by Nikolay Sivov 21 Jul '15

21 Jul '15

Looks good, thanks.

1 0

Re: user32/tests: Test for caption width of 13 on japanese W7
by Huw Davies 21 Jul '15

21 Jul '15

On 21 Jul 2015, at 19:32, André Hentschel wrote: > --- a/dlls/user32/tests/sysparams.c > +++ b/dlls/user32/tests/sysparams.c > @@ -1554,8 +1554,9 @@ static void test_SPI_SETNONCLIENTMETRICS( void ) /* 44 */ > > ok( Ncmcur.iCaptionWidth == 8 || > Ncmcur.iCaptionWidth == 12 || /* Vista, W7b */ > + Ncmcur.iCaptionWidth == 13 || /* Win7 japanese */ > Ncmcur.iCaptionWidth == Ncmstart.iCaptionWidth, /* with windows XP theme, the value never changes */ > - "CaptionWidth: %d expected 8, 12 or %d\n", Ncmcur.iCaptionWidth, Ncmstart.iCaptionWidth); > + "CaptionWidth: %d expected 8, 12, 13 or %d\n", Ncmcur.iCaptionWidth, Ncmstart.iCaptionWidth); > ok( Ncmcur.iScrollWidth == 8, > "ScrollWidth: %d expected 8\n", Ncmcur.iScrollWidth); > ok( Ncmcur.iScrollHeight == 8, I deliberately haven't fixed it like this because it has already become untenable. I suspect it should be compared to the height of the default gui font, or something like that. Huw.

1 0

[PATCH v2 00/15] AppUserModelID support
by Jonas Kümmerlin 21 Jul '15

21 Jul '15

Changes since the last version: - Feedback by Vincent Porvik, Nikolay Sivov and GOUJON Alexandre has been incorporated. - Where appropriate, bigger patches have been split into smaller chunks. - All required features are now implemented: You can try it out by installing LibreOffice using the MSI package, set the windows version to Win7 in winecfg and observe the new fancy desktop file matching performed by your desktop shell. (assuming your shell does this; Gnome3 and Unity do, XFCE and LXDE don't) - While Gnome3 works great, KDE hasn't been exactly cooperative so far. It now gets the grouping right, but ignores the StartupWMClass key in the desktop files, so pinning and fancy high-res icons won't work. This is most likely not fixable on our side, at least not without further research. I am still not happy with the PROPVARIANT (de)serialization code (patches 4 and 5, hope the mailing list won't eat it this time). The ole32 <-> propsys interfacing has been implemented as Nikolay Sivov suggested, but I'm not really happy with it. The new (de)serialization code also caused a weird regression in MSI.dll which I haven't been able to track down yet. I'm currently leaning towards moving the new code to propsys.dll and leaving ole32.dll untouched, even though that would result in two different implementations of basically the same thing. Does anyone have strong opinions on this matter? Rationale: AppUserModelIDs can be used to match open windows to shortcuts in order to make the Win7 Taskbar happen. The idea is similar to WM_CLASS in X11/freedesktop.org environments: Add the AppUserModelID to every window and every shortcut and let the shell figure out the matching. The shell can then do fancy things like pinning the app or showing a high quality icon. While most windows apps rely on the fallback algorithm of doing the shortcut<->window mapping via the full path to the executable, some applications, like LibreOffice, are making use of the AppUserModelID feature to appear as a set of different applications (Writer, Calc, ...) while sharing the same exe file (soffice.bin). Modern freedesktop.org shells work very similar to the Win7 Taskbar in this regard (using the X11 WM_CLASS attribute and the StartupWMClass= deskop entry key), and the goal is to have wine apps appear like any native app: With a high-res icon and pinnable. Jonas Kümmerlin (15): shell32: Implement process-wide explicit AppUserModelID propsys: Implement several PropVariantInit*() functions propsys: Extend PropVariantCompareEx() ole32: Rework property (de)serialization propsys: Implement Stg(De)SerializePropVariant propsys: Implement named properties on the in-memory property store propsys: Implement serialization of the in-memory property store propsys: Implement IPersistStream on the in-memory property store propsys: Partially implement property stringification shell32: Properly implement IPropertyStore for shell links shell32: Implement window property store (SHGetPropertyStoreForWindow) winex11.drv,user32,shell32 Use AppUserModelID to populate WM_CLASS propsys: Partially implement/stub IPropertyDescription msi: Write shortcut properties winemenubuilder: Save AppUserModelID as StartupWMClass dlls/msi/Makefile.in | 2 +- dlls/msi/action.c | 120 ++ dlls/msi/tests/Makefile.in | 2 +- dlls/msi/tests/action.c | 75 +- dlls/msi/tests/automation.c | 1 - dlls/msi/tests/suminfo.c | 4 +- dlls/ole32/stg_prop.c | 1717 +++++++++++++++++++++++----- dlls/ole32/tests/propvariant.c | 86 +- dlls/propsys/Makefile.in | 4 +- dlls/propsys/propdesc.c | 537 +++++++++ dlls/propsys/propdesc_builtin_data.c | 85 ++ dlls/propsys/propstore.c | 1062 ++++++++++++++++- dlls/propsys/propsys.spec | 36 +- dlls/propsys/propsys_classes.idl | 7 +- dlls/propsys/propsys_main.c | 6 - dlls/propsys/propsys_private.h | 23 + dlls/propsys/propvar.c | 889 +++++++++++++- dlls/propsys/tests/Makefile.in | 2 +- dlls/propsys/tests/propstore.c | 403 ++++++- dlls/propsys/tests/propsys.c | 709 +++++++++++- dlls/shell32/Makefile.in | 1 + dlls/shell32/shell32.spec | 2 + dlls/shell32/shell32_main.c | 109 +- dlls/shell32/shelllink.c | 218 +++- dlls/shell32/tests/Makefile.in | 1 + dlls/shell32/tests/appusermodel.c | 272 +++++ dlls/shell32/tests/shell32_test.h | 9 + dlls/shell32/tests/shelllink.c | 123 ++ dlls/shell32/winpropstore.c | 303 +++++ dlls/user32/message.c | 2 +- dlls/user32/user_private.h | 5 +- dlls/winex11.drv/window.c | 160 ++- include/propidl.idl | 6 + include/propsys.idl | 1 + include/propvarutil.h | 105 +- include/shellapi.h | 1 + include/shobjidl.idl | 6 + include/strsafe.h | 17 + include/winbase.h | 2 + include/wine/server_protocol.h | 59 +- programs/winemenubuilder/winemenubuilder.c | 123 +- server/protocol.def | 38 + server/request.h | 35 + server/trace.c | 47 + server/window.c | 148 +++ 45 files changed, 7098 insertions(+), 465 deletions(-) create mode 100644 dlls/propsys/propdesc.c create mode 100644 dlls/propsys/propdesc_builtin_data.c create mode 100644 dlls/shell32/tests/appusermodel.c create mode 100644 dlls/shell32/winpropstore.c -- 2.4.3

4 22