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

Re: user32: update DrawText tests to cover behavior differences on win98
by Dmitry Timoshkov Oct. 14, 2008

Oct. 14, 2008

"Aric Stewart" <aric(a)codeweavers.com> wrote: > diff --git a/dlls/user32/tests/text.c b/dlls/user32/tests/text.c > index 86655d6..3677d2f 100644 > --- a/dlls/user32/tests/text.c > +++ b/dlls/user32/tests/text.c > @@ -674,13 +674,20 @@ static void test_DrawState(void) > SetLastError(0xdeadbeef); > ret = DrawState(hdc, GetStockObject(DKGRAY_BRUSH), NULL, 0, strlen(text), > 0, 0, 10, 10, DST_TEXT); > - ok(!ret, "DrawState succeeded\n"); > + /* succeeds on windows 98 */ > + if (ret) > + win_skip("DrawState succeeds on Windows 98\n"); > + else > + ok(!ret, "DrawState succeeded\n"); > ok(GetLastError() == 0xdeadbeef, "not expected error %u\n", GetLastError()); > > SetLastError(0xdeadbeef); > ret = DrawState(hdc, GetStockObject(DKGRAY_BRUSH), NULL, 0, 0, > 0, 0, 10, 10, DST_TEXT); > - ok(!ret, "DrawState succeeded\n"); > + if (ret) > + win_skip("DrawState succeeds on Windows 98\n"); > + else > + ok(!ret, "DrawState succeeded\n"); > ok(GetLastError() == 0xdeadbeef, "not expected error %u\n", GetLastError()); > > ReleaseDC(hwnd, hdc); skip()/win_skip() are supposed to be used when the test(s) are going to be skipped due to some condition, in this case nothing is skipped, and broken() sounds like a better choice. -- Dmitry.

1 0

re: [user32: 2/2] Testcase for font size = 0x7fff
by Dan Kegel Oct. 13, 2008

Oct. 13, 2008

Koro wrote: > I wrote one this summer (http://www.korosoft.net/projects/tdemu/) Interesting. Would you be willing to contribute your Task Dialogs implementation to Wine under the terms of the LGPL? If not, well, thanks for the testcase and bug fix! - Dan

3 3

Re: ConvertStringSecurityDescriptorToSecurityDescriptor - more tests (pass)
by James Hawkins Oct. 13, 2008

Oct. 13, 2008

On Mon, Oct 13, 2008 at 4:23 PM, Paul Bryan Roberts <pbronline-wine(a)yahoo.co.uk> wrote: > A second attempt that, I hope, takes into account feedback from James > Hawkins. > > Original correction to ParseStringAclToAcl() to set missing error code > stands. > > Gratuitous setting of invalid parameter withdrawn. > > Replaced with check for NULL on StringSecurityDescriptor parameter in both A > and W forms of ConvertStringSecurityDescriptorToSecurityDescriptor. > > New checks backed up by additional test cases. > > Two instances of "highly superfluous" TRACE removed as suggested by JH, > > Issue - unable to add test cases that show > > if (GetVersion() & 0x80000000) > SetLastError(ERROR_CALL_NOT_IMPLEMENTED); > > should precede checks for NULL parameters. What to do ? Drop patch > altogether ? > + if (GetVersion() & 0x80000000) + SetLastError(ERROR_CALL_NOT_IMPLEMENTED); Where are you getting this from? -- James Hawkins

1 0

Re: loadperf: add stub dll
by Austin English Oct. 13, 2008

Oct. 13, 2008

On Mon, Oct 13, 2008 at 4:20 PM, Louis. Lenders <xerox_xerox2000(a)yahoo.co.uk> wrote: > as outlined in bug http://bugs.winehq.org/show_bug.cgi?id=12099, fixes a > crash in some apps that try install .net 1.1; also fixes a crash (Unhandled > page fault...) in dotnet20 installer (though that crash is not fatal there, > the dotnet 2.0 continues happily after it) > > > > > You've got changes to configure as well as configure.ac. Please remove the changes to configure. -- -Austin

1 0

Re: [try2] wordpad: do not try to close uninitialized handles
by James Hawkins Oct. 13, 2008

Oct. 13, 2008

On Mon, Oct 13, 2008 at 2:42 PM, Andrey Turkin <andrey.turkin(a)gmail.com> wrote: > [another approach with smaller changeset, as James suggested] > > Fix some functions where CloseHandle was called not matter if previous > open succeeded or not. > @@ -96,12 +96,13 @@ void registry_set_options(HWND hMainWnd) registry_set_pagemargins(hKey); } - RegCloseKey(hKey); + if (hKey) + RegCloseKey(hKey); Please get rid of this check for NULL hKey. We've spent a lot of time removing such checks. -- James Hawkins

1 0

RE: d3d9/tests: Fix test for old gfx-card
by Stefan Dösinger Oct. 13, 2008

Oct. 13, 2008

It is kinda strange, since the NULLREF device should be device independent, but I think the patch is OK. Catching device creation failures sounds ok even in this situation > -----Original Message----- > From: wine-patches-bounces(a)winehq.org [mailto:wine-patches- > bounces(a)winehq.org] On Behalf Of Detlef Riekenberg > Sent: Monday, October 13, 2008 1:55 AM > To: wine-patches(a)winehq.org > Subject: d3d9/tests: Fix test for old gfx-card > > > My "Radeon Mobility M7" is to old. > > > -- > > By by ... Detlef

2 1

Re: user32: Add the parameter bStatic in MDI_GetWindow to select theusing child list
by Dmitry Timoshkov Oct. 13, 2008

Oct. 13, 2008

"Ilya Shpigor" <shpigor(a)etersoft.ru> wrote: > The wine have problem with the Ctrl+Shift+Tab switching between the MDI > windows (switch only neghboring windows). I suppose, that reason of this is > using the WIN_ListChildren in the MDI_GetWindow function. WIN_ListChildren > return the child list with the variable sequence of elements. This sequence > depends on the current active window. > But in some cases we need this variable sequence. For example to determine > the focused window after destroy other window. > This patch add new parameter in MDI_GetWindow to check that child list are > needed. There is the clientInfo->child using instead the WIN_ListChildren. > Also some necessary changes in processing this list are added. A usual aproach to fixing Wine bugs is to create a test case to demonstrate the problem first, have it committed, and only then start making pass it. So far it's not clear (at least to me) what you are trying to fix. -- Dmitry.

1 0

Re: shlwapi: Implement SHSendMessageBroadcastA/W [try #2: Resend]
by Alexandre Julliard Oct. 13, 2008

Oct. 13, 2008

Detlef Riekenberg <wine.dev(a)web.de> writes: > Patch was not committed and got no additional comments. > > More Hints? It's supposed to handle non-responsive windows. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [1/4] mshtml: Implement IHTMLStyle get/put posLeft (try 3)
by Jacek Caban Oct. 13, 2008

Oct. 13, 2008

> {attrLeft, DISPID_IHTMLSTYLE_LEFT}, > + {attrLeft, DISPID_IHTMLSTYLE_POSLEFT}, You can't duplicate attr names in this table. We want to map 'left' to DISPID_IHTMLSTYLE_LEFT, not to random one. Jacek

1 0

Re: comctl32/tests: Fix tests to not depend on 96 dpi
by Dmitry Timoshkov Oct. 13, 2008

Oct. 13, 2008

"Detlef Riekenberg" <wine.dev(a)web.de> wrote: > #define DEFAULT_MIN_TAB_WIDTH 54 > -#define TAB_DEFAULT_WIDTH 96 > #define TAB_PADDING_X 6 > #define EXTRA_ICON_PADDING 3 > #define MAX_TABLEN 32 > @@ -507,11 +506,13 @@ static void test_tab(INT nMinTabWidth) > HDC hdc; > HFONT hOldFont; > INT i; > + DWORD dpi; > > hwTab = create_tabcontrol(TCS_FIXEDWIDTH, TCIF_TEXT|TCIF_IMAGE); > SendMessage(hwTab, TCM_SETMINTABWIDTH, 0, nMinTabWidth); > > hdc = GetDC(hwTab); > + dpi = GetDeviceCaps(hdc, LOGPIXELSY); If you are replacing 'width' shouldn't this be LOGPIXELSX? -- Dmitry.

1 0