Hello, thanks for the patch. I have some comments inlined:
=== debiant (32 bit report) ===
user32: combo.c:698: Test failed: 00000002: got 00000080 combo.c:704: Test failed: 00000002: got 00000080 combo.c:698: Test failed: 00000003: got 00000080 combo.c:704: Test failed: 00000003: got 00000080
Unfortunately I expect you'll need to debug and fix this. Fortunately, the other errors are probably spurious...
win.c:3042: Test succeeded inside todo block: 001000D4: expected next 00000000, got 00000000 win.c:3042: Test succeeded inside todo block: 001000D4: expected prev 000D00D6, got 000D00D6 win.c:3042: Test succeeded inside todo block: 001000D4: expected owner 00000000, got 00000000 win.c:3048: Test succeeded inside todo block: 001000D4: expected next 000E00DE, got 000E00DE win.c:3048: Test succeeded inside todo block: 001000D4: expected prev 00000000, got 00000000 win.c:3048: Test succeeded inside todo block: 001000D4: expected owner 00000000, got 00000000 win.c:3054: Test succeeded inside todo block: 001000D4: expected next 000E00DE, got 000E00DE win.c:3054: Test succeeded inside todo block: 001000D4: expected prev 000D00D6, got 000D00D6 win.c:3054: Test succeeded inside todo block: 001000D4: expected owner 00000000, got 00000000 win.c:3060: Test succeeded inside todo block: 001000D4: expected next 00000000, got 00000000 win.c:3060: Test succeeded inside todo block: 001000D4: expected prev 000D00D6, got 000D00D6 win.c:3060: Test succeeded inside todo block: 001000D4: expected owner 00000000, got 00000000 win.c:3066: Test succeeded inside todo block: 001000D4: expected next 00000000, got 00000000 win.c:3066: Test succeeded inside todo block: 001000D4: expected prev 000D00D6, got 000D00D6 win.c:3066: Test succeeded inside todo block: 001000D4: expected owner 00000000, got 00000000 win.c:3072: Test succeeded inside todo block: 001000D4: expected next 00000000, got 00000000 win.c:3072: Test succeeded inside todo block: 001000D4: expected prev 000D00D6, got 000D00D6 win.c:3072: Test succeeded inside todo block: 001000D4: expected owner 00000000, got 00000000 win.c:3078: Test succeeded inside todo block: 001000D4: expected next 00000000, got 00000000 win.c:3078: Test succeeded inside todo block: 001000D4: expected prev 000D00D6, got 000D00D6 win.c:3078: Test succeeded inside todo block: 001000D4: expected owner 00000000, got 00000000 win.c:3084: Test succeeded inside todo block: 001000D4: expected next 000E00DE, got 000E00DE win.c:3084: Test succeeded inside todo block: 001000D4: expected prev 00000000, got 00000000 win.c:3084: Test succeeded inside todo block: 001000D4: expected owner 00000000, got 00000000 win.c:3090: Test succeeded inside todo block: 001000D4: expected next 000E00DE, got 000E00DE win.c:3090: Test succeeded inside todo block: 001000D4: expected prev 000D00D6, got 000D00D6 win.c:3090: Test succeeded inside todo block: 001000D4: expected owner 00000000, got 00000000 win.c:3096: Test succeeded inside todo block: 001000D4: expected next 00000000, got 00000000 win.c:3096: Test succeeded inside todo block: 001000D4: expected prev 000D00D6, got 000D00D6 win.c:3096: Test succeeded inside todo block: 001000D4: expected owner 00000000, got 00000000
If you're introducing the fix at the same time as the tests, you don't want to use todo_wine.
That said, it's often better to split patches up, sending the tests first and the fix second, so that you can show exactly what gets fixed.
diff --git a/dlls/user32/winpos.c b/dlls/user32/winpos.c index b92a20df18..9c8576e07b 100644 --- a/dlls/user32/winpos.c +++ b/dlls/user32/winpos.c @@ -1983,6 +1983,12 @@ static BOOL fixup_flags( WINDOWPOS *winpos, const RECT *old_window_rect, int par } }
- if ((wndPtr->dwStyle & (WS_CHILD | WS_POPUP)) == WS_CHILD
&& winpos->hwndInsertAfter == HWND_TOPMOST )- {
winpos->flags |= SWP_NOZORDER;- }
- /* Check hwndInsertAfter */ if (winpos->flags & SWP_NOZORDER) goto done;
This solution seems suspicious, and is possibly even demonstrably (in)correct:
* If SetWindowPos() seems to do nothing, does it even succeed in this case? (I.e. what is the return value, and last error if applicable?)
* The fixed up flags are passed to WM_WINDOWPOSCHANGED. Is SWP_NOZORDER really set?
Finally, should we add similar treatment for HWND_NOTOPMOST?
-- 2.20.1