September 2010 - Wine-Devel

Re: user32: Implement support for getting class data GCLP_HBRBACKGROUND on other process window
by Alexandre Julliard Sept. 20, 2010

Sept. 20, 2010

Aleksey Bragin <aleksey(a)reactos.org> writes: > diff --git a/server/class.c b/server/class.c > index 0e402ba..90ab436 100644 > --- a/server/class.c > +++ b/server/class.c > @@ -47,6 +47,7 @@ struct window_class > int local; /* local class? */ > atom_t atom; /* class atom */ > mod_handle_t instance; /* module instance */ > + user_handle_t background; /* background brush handle */ > unsigned int style; /* class style */ > int win_extra; /* number of window extra bytes */ > client_ptr_t client_ptr; /* pointer to class in client address space */ Brushes are not user handles, and it doesn't make much sense to share them since the handles are process-local currently. -- Alexandre Julliard julliard(a)winehq.org

1 0

[PATCH] shell32: use flexible arrays to avoid fortify failures
by Joerg-Cyril.Hoehle＠t-systems.com Sept. 20, 2010

Sept. 20, 2010

Hi, Alexandre Julliard wrote: >> it needs to be "foo[]". ANYSIZE_ARRAY is defined as 1 >True, you don't want to change the public one, but you can define a >shell32-specific version of it. At least then it looks like a normal >array declaration. Thank you for pointing to ANYSIZE_ARRAY. Why not cope with the way MS people recommend and use foo[ANYSIZE_ARRAY], together with the FIELD_OFFSET() macro? I found this article useful: http://blogs.msdn.com/b/oldnewthing/archive/2004/08/26/220873.aspx and made a note to use that in future patches. It explains why ANYSIZE_ARRAY MUST be 1. Should somebody submit patches to replace foo[1] with foo[ANYSIZE_ARRAY] where needed in Wine's include/*.h ? E.g. wingdi.h: BITMAPINFO comes to mind, it uses RGBQUAD bmiColors[1]; http://source.winehq.org/source/include/wingdi.h#L1912 ANYSIZE_ARRAY would be a perfect match here. Perhaps one can teach fortify about ANYSIZE_ARRAY? Regards, Jörg Höhle

1 0

Re: [PATCH 1/6] d3dcompiler: Add D3DGetBlobPart() stub.
by Henri Verbeet Sept. 20, 2010

Sept. 20, 2010

2010/9/19 Rico Schüller <kgbricola(a)web.de>: > + FIXME("data %p, data_size %lu, part %#x, flags %u, blob %p stub!\n", data, data_size, part, flags, blob); I think %#x is more appropriate than %u for flags.

1 0

Re: [msi/tests] Prevent a popup on Windows NT
by testbot＠testbot.winehq.org Sept. 20, 2010

Sept. 20, 2010

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 http://testbot.winehq.org/JobDetails.pl?Key=5346 Your paranoid android. === W2KPROSP4 (32 bit install) === Timeout === WXPPROSP3 (32 bit install) === Timeout

1 0

Re: ipconfig: Add the Polish translation.
by Andrew Nguyen Sept. 19, 2010

Sept. 19, 2010

2010/9/19 Łukasz Wojniłowicz <lukasz.wojnilowicz(a)gmail.com>: > +/* > + * IP configuration utility > + * English language support > + * Pedantically, that should probably be "Polish language support".

2 2

Re: kernel32: Remove unused variable
by Vitaliy Margolen Sept. 19, 2010

Sept. 19, 2010

On 09/19/2010 01:47 AM, Luca Bennati wrote: > > Found with scan-build provided by clang/llvm > > diff --git a/dlls/kernel32/virtual.c b/dlls/kernel32/virtual.c > index dcd5b85..a10bb6c 100644 > --- a/dlls/kernel32/virtual.c > +++ b/dlls/kernel32/virtual.c > @@ -648,17 +648,13 @@ BOOL WINAPI IsBadReadPtr( LPCVOID ptr, UINT size ) > __TRY > { > volatile const char *p = ptr; > - char dummy; > UINT count = size; > > while (count > page_size) > { > - dummy = *p; > p += page_size; > count -= page_size; > } > - dummy = p[0]; > - dummy = p[count - 1]; You can't do that. That's the whole point of this function - to test if particular address is readable. Vitaliy.

1 0

Wine without X? Failing in dlls/winex11.drv
by Gerald Pfeifer Sept. 19, 2010

Sept. 19, 2010

Trying to build Wine on a new tester which initially had too few packages installed I ran into the following. In file included from bitblt.c:33: x11drv.h:30:22: error: X11/Xlib.h: No such file or directory x11drv.h:31:27: error: X11/Xresource.h: No such file or directory x11drv.h:32:23: error: X11/Xutil.h: No such file or directory x11drv.h:33:23: error: X11/Xatom.h: No such file or directory x11drv.h:44:21: error: X11/Xmd.h: No such file or directory x11drv.h:45:24: error: X11/Xproto.h: No such file or directory Indeed include/config.h has HAVE_X11_XLIB_H undefined since, well, all these headers indeed are missing. On the other hand, configure did not halt. In fact, configure did not even warn about this situation unlike it does in other cases. How to fix this? Should dlls/winex11.drv just not be built in such a case? Should configure error out? (We do have --without-x, too, though.) Anything else? Currently we do pass configure and then run into a compile failure later on. Gerald

3 3

Re: ws2_32: Fail when select is called with an invalid fd. (try 2)
by Alexandre Julliard Sept. 19, 2010

Sept. 19, 2010

"Vincent Povirk" <vincent(a)codeweavers.com> writes: > Fixed the pollfds leak from try 1. You are still leaking the file descriptors. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [urlmon/tests] Fix test failures on Win9x/WinMe by using more A-functions
by testbot＠testbot.winehq.org Sept. 19, 2010

Sept. 19, 2010

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 http://testbot.winehq.org/JobDetails.pl?Key=5328 Your paranoid android. === W2K8SE (32 bit url) === url.c:1337: Test failed: unexpected call OnProgress_SENDINGREQUEST

1 0

Re: user32: warnings (around uninitialized variables).
by Alexandre Julliard Sept. 19, 2010

Sept. 19, 2010

Gerald Pfeifer <gerald(a)pfeifer.com> writes: > @@ -388,6 +388,8 @@ static int DIB_GetBitmapInfo( const BITMAPINFOHEADER *header, LONG *width, > *compr = header->biCompression; > return 1; > } > + > + *width = *height = 0; > ERR("(%d): unknown/wrong size for header\n", header->biSize ); > return -1; > } 0 is not valid for a bitmap, these should never be used in the error case. -- Alexandre Julliard julliard(a)winehq.org

2 1