[PATCH resend] user32: Fix messages sent on a window without WS_CHILD, but with an invisible parent, when it is shown.

[Gabriel Ivăncescu]

Some applications depend on the fact that WM_WINDOWPOSCHANGING is sent
after WM_SHOWWINDOW when SetParent is called, even if the window ends up
not visible because its parent is not visible, and occurs when the window
itself does not have the WS_CHILD style set.

This also fixes a TODO message sequence, so that Wine matches Windows.

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=40262
Signed-off-by: Gabriel Ivăncescu <gabrielopcode(a)gmail.com>
---
 dlls/user32/tests/msg.c |  3 ++-
 dlls/user32/winpos.c    | 11 +++++++----
 2 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/dlls/user32/tests/msg.c b/dlls/user32/tests/msg.c
index bf20c5c..fb341cc 100644
--- a/dlls/user32/tests/msg.c
+++ b/dlls/user32/tests/msg.c
@@ -16397,6 +16397,7 @@ static const struct message WmSetParentSeq_2[] = {
     { HCBT_ACTIVATE, hook|optional },
     { EVENT_SYSTEM_FOREGROUND, winevent_hook|wparam|lparam, 0, 0 },
     { WM_WINDOWPOSCHANGING, sent|wparam|optional, SWP_NOSIZE|SWP_NOMOVE },
+    { WM_QUERYNEWPALETTE, sent|optional },
     { WM_NCACTIVATE, sent|wparam|optional, 1 },
     { WM_ACTIVATE, sent|wparam|optional, 1 },
     { HCBT_SETFOCUS, hook|optional },
@@ -16467,7 +16468,7 @@ static void test_SetParent(void)
 
     SetParent(popup, child);
     flush_events();
-    ok_sequence(WmSetParentSeq_2, "SetParent() visible WS_POPUP", TRUE);
+    ok_sequence(WmSetParentSeq_2, "SetParent() visible WS_POPUP", FALSE);
 
     ok(GetWindowLongA(popup, GWL_STYLE) & WS_VISIBLE, "WS_VISIBLE should be set\n");
     ok(!IsWindowVisible(popup), "IsWindowVisible() should return FALSE\n");
diff --git a/dlls/user32/winpos.c b/dlls/user32/winpos.c
index e3c3f5c..7b5c657 100644
--- a/dlls/user32/winpos.c
+++ b/dlls/user32/winpos.c
@@ -1122,8 +1122,8 @@ static BOOL show_window( HWND hwnd, INT cmd )
 
     swp = USER_Driver->pShowWindow( hwnd, cmd, &newPos, swp );
 
-    parent = GetAncestor( hwnd, GA_PARENT );
-    if (parent && !IsWindowVisible( parent ) && !(swp & SWP_STATECHANGED))
+    if ((style & WS_CHILD) && (parent = GetAncestor( hwnd, GA_PARENT )) &&
+        !IsWindowVisible( parent ) && !(swp & SWP_STATECHANGED))
     {
         /* if parent is not visible simply toggle WS_VISIBLE and return */
         if (showFlag) WIN_SetStyle( hwnd, WS_VISIBLE, 0 );
@@ -1958,8 +1958,11 @@ static BOOL fixup_flags( WINDOWPOS *winpos, const RECT *old_window_rect, int par
     if (winpos->cy < 0) winpos->cy = 0;
     else if (winpos->cy > 32767) winpos->cy = 32767;
 
-    parent = GetAncestor( winpos->hwnd, GA_PARENT );
-    if (!IsWindowVisible( parent )) winpos->flags |= SWP_NOREDRAW;
+    if (wndPtr->dwStyle & WS_CHILD)
+    {
+        parent = GetAncestor( winpos->hwnd, GA_PARENT );
+        if (!IsWindowVisible( parent )) winpos->flags |= SWP_NOREDRAW;
+    }
 
     if (wndPtr->dwStyle & WS_VISIBLE) winpos->flags &= ~SWP_SHOWWINDOW;
     else
-- 
2.21.0

[Marvin]

Hi,

While running your changed tests, I think I found new failures.
Being a bot and all I'm not very good at pattern recognition, so I might be
wrong, but could you please double-check?

Full results can be found at:
https://testbot.winehq.org/JobDetails.pl?Key=78790

Your paranoid android.


=== w10pro64 (64 bit report) ===

user32:
msg.c:13904: Test failed: CreateDialogParam_1: 15: the msg 0x001c was expected, but got msg 0x0047 instead
msg.c:13904: Test failed: CreateDialogParam_1: 16: the msg 0x0086 was expected, but got msg 0x001c instead
msg.c:13904: Test failed: CreateDialogParam_1: 17: the msg 0x0006 was expected, but got msg 0x0086 instead
msg.c:13904: Test failed: CreateDialogParam_1: 18: the msg 0x0007 was expected, but got msg 0x0006 instead
msg.c:13904: Test failed: CreateDialogParam_1: 20: the msg sequence is not complete: expected 0000 - actual 0007

=== debiant (32 bit report) ===

user32:
monitor: Timeout

=== debiant (32 bit Chinese:China report) ===

user32:
monitor: Timeout

=== debiant (32 bit WoW report) ===

user32:
monitor: Timeout

=== debiant (64 bit WoW report) ===

user32:
monitor: Timeout

[Gabriel Ivăncescu]

Hi,

Is there something wrong with this patch, or something I should add to 
it? I've resent it a few times now and always slept through the cracks. 
It fixes the bug, and fixes the existing message sequence to match 
Windows. (the WM_QUERYNEWPALETTE is optional and only because Wine sends 
it, Windows only does that in 256-color mode, which is completely 
unrelated to this patch)

FWIW, the patch just adds checks for WS_CHILD, because apparently that's 
what Windows does, and that's what fixes the bug, too.