October 2009 - Wine-Devel - list.winehq.org

Re: [PATCH 1/4] comctl32/tooltips: Fix an obviously wrong structure size check
by Paul Vriens Oct. 15, 2009

Oct. 15, 2009

On 10/15/2009 01:36 AM, Nikolay Sivov wrote: > Changelog: > - fix an obviously wrong structure size check > > > Hi Nikolay, +static void test_ttm_gettoolinfo(void) +{ + TTTOOLINFOA ti; + TTTOOLINFOW tiW; + HWND hwnd; + DWORD r; + + hwnd = CreateWindowExA(0, TOOLTIPS_CLASSA, NULL, 0, + 10, 10, 300, 100, + NULL, NULL, NULL, 0); + + ti.cbSize = TTTOOLINFOA_V2_SIZE; + ti.hwnd = NULL; + ti.hinst = GetModuleHandleA(NULL); + ti.uFlags = 0; + ti.uId = 0x1234ABCD; + ti.lpszText = NULL; + ti.lParam = 0xdeadbeef; + GetClientRect(hwnd, &ti.rect); + r = SendMessageW(hwnd, TTM_ADDTOOLA, 0, (LPARAM)&ti); The W-call won't work on Win9x/WinMe. diff --git a/dlls/comctl32/tooltips.c b/dlls/comctl32/tooltips.c index 54600e8..6b483df 100644 --- a/dlls/comctl32/tooltips.c +++ b/dlls/comctl32/tooltips.c @@ -1076,7 +1076,7 @@ TOOLTIPS_AddToolA (TOOLTIPS_INFO *infoPtr, const TTTOOLINFOA *lpToolInfo) } } - if (lpToolInfo->cbSize >= sizeof(TTTOOLINFOA)) + if (lpToolInfo->cbSize >= TTTOOLINFOW_V2_SIZE) Shouldn't that be "TTTOOLINFOA_V2_SIZE" ? @@ -1369,7 +1369,7 @@ TOOLTIPS_EnumToolsW (const TOOLTIPS_INFO *infoPtr, UINT uIndex, LPTTTOOLINFOW lp /* lpToolInfo->lpszText = toolPtr->lpszText; */ lpToolInfo->lpszText = NULL; /* FIXME */ - if (lpToolInfo->cbSize >= sizeof(TTTOOLINFOW)) + if (lpToolInfo->cbSize >= TTTOOLINFOA_V2_SIZE) lpToolInfo->lParam = toolPtr->lParam; And shouldn't this be "TTTOOLINFOW_V2_SIZE" ? -- Cheers, Paul.

2 3

Re: [PATCH 2/4] comctl32/tooltips: Test the way structure size field is validated
by Paul Vriens Oct. 15, 2009

Oct. 15, 2009

On 10/15/2009 01:36 AM, Nikolay Sivov wrote: > Changelog: > - test the way structure size field is validated > > > > Hi Nikolay, Some more W-calls that won't suceed on Win9x/WinMe in the added tests. -- Cheers, Paul.

1 0

Re: [Fwd: [Bug 20285] cmd.exe does not treat tabs as parameter delimiters]
by Juan Lang Oct. 14, 2009

Oct. 14, 2009

Hi Colin, > As requested by Austin, I'm forwarding this patch to > wine-patches at winehq.org. Please see the bug report for a test case. You just forwarded the wine-bugs email. This isn't the same as sending the patch. You must include the patch in your email. See http://wiki.winehq.org/SubmittingPatches for help with the process if you need more explicit instructions. Thanks, --Juan

1 0

Help me get a patch into Wine -> Mac cdrom issue
by Joerg-Cyril.Hoehle＠t-systems.com Oct. 14, 2009

Oct. 14, 2009

Chip, I had hoped that your patch would allow mcicda to work, cf. bug #20323, alas it is not so. Instead I get to see ERR("This version of Mac OS X does not support IOCDAudioControl" on Leopard. I googled slightly and found a message from 2004: http://lists.apple.com/archives/darwin-drivers/2004/Jul/msg00049.html >IOCDAudioControl is deprecated if not obsolete. Therefore I believe the user should not be bothered with a reference to something deprecated half a decade ago and unavailable. A subsequent patch to a simple FIXME("NIY\n") would be enough. Please leave IOCDAudioControl in a comment in the code at most. Reading further, it looks like there's nothing to read audio CDs anymore. Instead, developers are recommended to take advantage of the cddafs, which mounts all audio tracks as "/Volumes/Audio-CD/01 Titel 1.aiff" and to play those files with CoreAudio instead. I've checked that mixed CDs with both data files and audio tracks let 2 CD icons appear on the Finder's desktop, e.g. /Volumes/ABC/ and /Volumes/Audio-CD/ So I wonder whether mcicda on Mac OS should forward calls to waveaudio and play .aiff files instead. I don't know how ntdll and the mountmgr etc. could be changed to accomodate the separation in two volumes, i.e. how to map both /Volumes/ABC and /Volumes/Audio-CD/ to D:\ Regards, Jörg Höhle.

5 6

Re: secur32: check for variable before freeing it
by Henri Verbeet Oct. 14, 2009

Oct. 14, 2009

2009/10/13 Austin English <austinenglish(a)gmail.com>: > cred = schan_free_handle(i, SCHAN_HANDLE_CRED); > - pgnutls_certificate_free_credentials(cred->credentials); > + if(cred) pgnutls_certificate_free_credentials(cred->credentials); I don't think "cred" is ever supposed to be NULL. Could you add some traces to see where that handle comes from?

4 8

Locale related test request on Win7
by Nikolay Sivov Oct. 14, 2009

Oct. 14, 2009

Hi. Could anyone with Win7 access run this for me and send a log back? This test requires Russian localization support to be installed. Thanks.

1 0

Re: [1/3] (resend) mshtml: Add test suite for IHTMLLocation
by Alexandre Julliard Oct. 14, 2009

Oct. 14, 2009

"Andrew Eikum" <aeikum(a)codeweavers.com> writes: > + if(hres == S_OK){ > + if(test->hostname_ok) > + ok(str_eq_wa(str, test->hostname), "%s: expected retrieved hostname to be L\"%s\", was: %s\n", test->name, test->hostname, wine_dbgstr_w(str)); > + else > + todo_wine ok(str_eq_wa(str, test->hostname), "%s: expected retrieved hostname to be L\"%s\", was: %s\n", test->name, test->hostname, wine_dbgstr_w(str)); Please split the long lines like Detlef requested. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [1/10] Check for ms_hook_prologue attribute support, make a first function hookable
by Dmitry Timoshkov Oct. 14, 2009

Oct. 14, 2009

"Stefan Dösinger" <stefan(a)codeweavers.com> wrote: > -HRESULT WINAPI DirectInput8Create(HINSTANCE hinst, DWORD dwVersion, REFIID riid, LPVOID *ppDI, LPUNKNOWN punkOuter) { > +HRESULT WINAPI DECLSPEC_HOTPATCH DirectInput8Create(HINSTANCE hinst, DWORD dwVersion, REFIID riid, LPVOID *ppDI, > LPUNKNOWN punkOuter) { Probably DECLSPEC_HOTPATCH just should be a part of WINAPI. -- Dmitry.

2 1

Re: [4/10] OpenGL32: Give wglSwapBuffer a real function, make it hookable
by Roderick Colenbrander Oct. 14, 2009

Oct. 14, 2009

You should also patch make_opengl because openg32.spec is automatically generated. Roderick On Tue, Oct 13, 2009 at 10:04 PM, Stefan Dösinger <stefan(a)codeweavers.com> wrote: > >

2 1

Re: [PATCH 1/2] oleaut32/tests: Test return value of SysReAllocStringLen for the null terminator.
by Nikolay Sivov Oct. 13, 2009

Oct. 13, 2009

Alexander Kochetkov wrote: > This test works fine under windows 98, XP and under wine. > And allways check for the null terminator. > > see bug 20334. > Hi, Alexander. You patches should not break test in order they are supposed to be applied: your test currently misses todo_wine here: --- + ok (str[STRING_SIZE] == 0, + "Expected null terminator, got 0x%04X\n", str[STRING_SIZE]); --- After that your second patch should remove this todo_wine marker. Without such simple rule last patch (2/2) can't be reverted without breaking tests. Another small issue is a trailing space in the same line --- 'ok (str[STRING_SIZE] == 0, ' --- it's not a big problem cause I believe patches are auto processed for such things before checking to tree, it will lead to a merge conflict in your local repo, that's all.

1 0