On 11/13/2010 10:59 AM, Bernhard Übelacker wrote:
When I run both tests as single files from a "make crosstest" I was not able to reproduce neither the explorer crash or the explorer windows not opening.
I didn't see a crash but maybe this is because I'm testing on W2K. I can still reliably get the double-clicking out of order when just running the cross-compiled comctl32 tests ('make crosstest' on my own box).
There is also something 'fundamentally' wrong with the tests:
+ hdlg = (HWND)PropertySheetA(&psh); + if (hdlg == INVALID_HANDLE_VALUE) + { + win_skip("comctl32 4.70 needs dwSize adjustment\n"); + psh.dwSize = sizeof(psh) - sizeof(HBITMAP) - sizeof(HPALETTE) - sizeof(HBITMAP); + hdlg = (HWND)PropertySheetA(&psh); + } + ShowWindow(hdlg,SW_NORMAL); + + ok_sequence(sequences, PROPSHEET_SEQ_INDEX, property_sheet_seq, "property sheet with custom window proc", TRUE); + + DestroyWindow(hdlg);
The PropertySheet call will create a modal (not a modeless) property sheet and thus the return value is not a handle (see also MSDN). You can easily check by adding some traces (see http://testbot.winehq.org/JobDetails.pl?Key=7013). hdlg is always 0 and destroying the window will give you an error ERROR_INVALID_WINDOW_HANDLE.
Will investigate further but any help is welcome.