July 2008 - Wine-Devel - list.winehq.org

Re: mswsock:Implement 3 functions by calling through SIO_GET_EXTENSION_FUNCTION_POINTER (try 3)
by Alexandre Julliard July 8, 2008

July 8, 2008

Jon Griffiths <jon_p_griffiths(a)yahoo.com> writes: > +BOOL WINAPI AcceptEx(SOCKET listener, SOCKET acceptor, PVOID dest, DWORD dest_len, > + DWORD local_addr_len, DWORD rem_addr_len, LPDWORD received, > + LPOVERLAPPED overlapped) > { > - FIXME("(listen=%ld, accept=%ld, %p, %d, %d, %d, %p, %p), not implemented\n", > - sListenSocket,sAcceptSocket,lpOutputBuffer,dwReceiveDataLength, > - dwLocalAddressLength,dwRemoteAddressLength,lpdwBytesReceived,lpOverlapped > - ); > + EnterCriticalSection(&MSWSOCK_CritSect); > + if (!acceptex_fn) > + { > + /* Get pointers to the ws2_32 implementations. > + * NOTE: This assumes that ws2_32 contains only one implementation > + * of these functions, i.e. that you cannot get different functions > + * back by passing another socket in. If that ever changes, we'll need > + * to think about associating the functions with the socket and > + * exposing that information to this dll somehow. > + */ > + const GUID acceptex_guid = WSAID_ACCEPTEX; > + const GUID acceptexsockaddrs_guid = WSAID_GETACCEPTEXSOCKADDRS; > + DWORD len; > + WSAIoctl(acceptor, SIO_GET_EXTENSION_FUNCTION_POINTER, > + (void*)&acceptex_guid, sizeof(acceptex_guid), > + &acceptex_fn, sizeof(acceptex_fn), &len, NULL, NULL); > + WSAIoctl(acceptor, SIO_GET_EXTENSION_FUNCTION_POINTER, > + (void*)&acceptexsockaddrs_guid, sizeof(acceptexsockaddrs_guid), > + &acceptexsockaddrs_fn, sizeof(acceptexsockaddrs_fn), &len, NULL, NULL); > + } > + LeaveCriticalSection(&MSWSOCK_CritSect); You don't need a crit section since you assume that the pointers are always the same, but you do need to fetch them into local variables and check for error before setting the global ones. > + static const GUID guid = WSAID_TRANSMITFILE; > + LPFN_TRANSMITFILE fn; > + DWORD len; > > - return FALSE; > + /* Use the implementation provided by ws2_32 */ > + if (WSAIoctl(s, SIO_GET_EXTENSION_FUNCTION_POINTER, > + (void*)&guid, sizeof(guid), &fn, sizeof(fn), &len, NULL, NULL)) > + return FALSE; > + > + return fn(s, file, total_len, chunk_len, overlapped, buffers, flags); You should store this one globally too, no need to do the ioctl on every call. -- Alexandre Julliard julliard(a)winehq.org

1 0

[1/3] [d3d8] Add testcase for IDirect3DDevice8::CreateImageSurface
by Tobias Jakobi July 8, 2008

July 8, 2008

Intended, see patch 2/3 - the testcase is activated there (since it pass then).

1 0

Re: shell32[1/2]: RunFileDlg: allow custom icon
by Alexandre Julliard July 8, 2008

July 8, 2008

Mikołaj Zalewski <mikolaj(a)zalewski.pl> writes: > diff --git a/dlls/shell32/dialogs.c b/dlls/shell32/dialogs.c > index 0e858b8..a30cf33 100644 > --- a/dlls/shell32/dialogs.c > +++ b/dlls/shell32/dialogs.c > @@ -139,9 +139,11 @@ static INT_PTR CALLBACK RunDlgProc (HWND hwnd, UINT message, WPARAM wParam, LPAR > if (prfdp->uFlags & RFF_CALCDIRECTORY) > FIXME("RFF_CALCDIRECTORY not supported\n"); > > + if (prfdp->hIcon == NULL) > + prfdp->hIcon = (WPARAM)LoadIconW(NULL, (LPCWSTR)IDI_WINLOGO); > SetClassLongPtrW (hwnd, GCLP_HICON, (LPARAM)prfdp->hIcon) ; > - SendMessageW (GetDlgItem (hwnd, IDC_RUNDLG_ICON), STM_SETICON, > - (WPARAM)LoadIconW (NULL, (LPCWSTR)IDI_WINLOGO), 0); > + SendMessageW(GetDlgItem(hwnd, IDC_RUNDLG_ICON), STM_SETICON, (WPARAM)prfdp->hIcon, 0); > + The SetClassLongPtrW call is very wrong, you should probably fix that too while you're at it. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: kernel32: GetPrivateProfileString should set buffer to "" if default value is not set (resend)
by Alexandre Julliard July 8, 2008

July 8, 2008

"Dan Kegel" <dank(a)kegel.com> writes: > diff --git a/dlls/kernel32/profile.c b/dlls/kernel32/profile.c > index 9b6b1a1..e52cd66 100644 > --- a/dlls/kernel32/profile.c > +++ b/dlls/kernel32/profile.c > @@ -1114,9 +1114,10 @@ static int PROFILE_GetPrivateProfileString( LPCWSTR section, LPCWSTR entry, > } else if (buffer && def_val) { > lstrcpynW( buffer, def_val, len ); > ret = strlenW( buffer ); > - } > - else > + } else { > + buffer[0] = 0; /* buffer must be a null terminated empty string */ > ret = 0; What if len is 0? What if buffer is NULL? -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: wined3d: Remove const qualifier from output pointers (Resend)
by Alexandre Julliard July 8, 2008

July 8, 2008

Andrew Talbot <andrew.talbot(a)talbotville.com> writes: > Anything wrong with this patch? It does accord with GL/gl.h. You should fix the other functions too, there are many more similar cases. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [1/3] [d3d8] Add testcase for IDirect3DDevice8::CreateImageSurface
by H. Verbeet July 8, 2008

July 8, 2008

2008/7/7 Tobias Jakobi <liquid.acid(a)gmx.net>: > The test doesn't get run this way. If it did it would fail of course, so you should either submit the test after your second patch, or put the ok that tests the pool type inside a todo_wine block and remove it again in your second patch.

1 0

rpcrt4: Compile error
by Austin Lund July 8, 2008

July 8, 2008

Does anyone else get the following compile error at commit 2734fb44e0f4065179044b9eb93f87a7630f292d: make[2]: Entering directory `/home/lund/src/wine/wine-git/dlls/rpcrt4' make[2]: *** No rule to make target `../../include/wine/rpcss_shared.h', needed by `ndr_contexthandle.o'. Stop. make[2]: Leaving directory `/home/lund/src/wine/wine-git/dlls/rpcrt4' make[1]: *** [rpcrt4] Error 2 make[1]: Leaving directory `/home/lund/src/wine/wine-git/dlls' make: *** [dlls] Error 2

2 1

Re: wintrust/test: Verify array length before calling memcmp in test (RESEND)
by James Hawkins July 7, 2008

July 7, 2008

On Mon, Jul 7, 2008 at 5:05 PM, Roy Shea <royshea(a)gmail.com> wrote: > This revision uses a more concise check of hash equality than prior > patch archived at: > > http://www.winehq.org/pipermail/wine-patches/2008-June/056139.html > > Call to memcmp in test_calchash assumes length of hash and expectedhash > are the same. This need not be true in general, and is not true for the > current stub implementation of CryptCATAdminCalcHashFromFileHandle. > This patch adds an explicit test of array length. In doing so it > prevents the following Valgrind warning: > @@ -335,7 +335,9 @@ static void test_calchash(void) { ok(GetLastError() == ERROR_SUCCESS, "Expected ERROR_SUCCESS, got %d\n", GetLastError()); - ok(!memcmp(hash, expectedhash, sizeof(expectedhash)), "Hashes didn't match\n"); + ok(hashsize == sizeof(expectedhash) && + !memcmp(hash, expectedhash, sizeof(expectedhash)), + "Hashes didn't match\n"); What's with the weird spacing? It doesn't match the spacing of the ok call right above it. -- James Hawkins

2 1

CUDA wrapper
by Seth Shelnutt July 7, 2008

July 7, 2008

I just have a makefile. No makefile.in. Perhaps I used winemaker wrong? I simply did "winemaker --dll cuda". On Mon, Jul 7, 2008 at 5:00 PM, John Klehm <xixsimplicityxix(a)gmail.com> wrote: > > > Check the IMPORT variable in Makefile.in and see if everything you > need is listed. > > --John >

1 0

Re: mlang/tests: don't assume output of ConvertStringFromUnicode is null terminated
by James Hawkins July 7, 2008

July 7, 2008

2008/7/6 Dan Kegel <dank(a)kegel.com>: > The mlang test assumed that the output of ConvertStringFromUnicode > was null terminated, but it seems not to be. > > Fixes the valgrind warning: > Conditional jump or move depends on uninitialised value(s) > at strlen (mc_replace_strmem.c:242) > by lstrlenA (string.c:364) > by ConvertINetMultiByteToUnicode (mlang.c:526) > by ConvertINetString (mlang.c:633) > by fnIMultiLanguage2_ConvertString (mlang.c:2197) > by check_convertible (mlang.c:287) > by test_EnumCodePages (mlang.c:407) > Uninitialised value was created by a stack allocation > at check_convertible (mlang.c:270) > > I believe James was trying to fix this with > http://www.winehq.org/pipermail/wine-patches/2008-June/056454.html > but missed (he thought the problem was on the destination > len, but really it was on the source len?). > No, this is hiding a bug. The test code conforms with the API. The problem is that ConvertINetMultiByteToUnicode uses the value of an out-only parameter (NULL pDstStr, non-NULL pcDstSize). Check out the code block in mlang.c:632. -- James Hawkins

2 2