[Bug 55417] New: user32:msg - test_showwindow() sometimes fails due to an unexpected WM_SYNCPAINT message on Windows 10

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55417

Bug ID: 55417
           Summary: user32:msg - test_showwindow() sometimes fails due to
                    an unexpected WM_SYNCPAINT message on Windows 10
           Product: Wine
           Version: unspecified
          Hardware: x86-64
                OS: Windows
            Status: NEW
          Severity: normal
          Priority: P2
         Component: user32
          Assignee: [e-mail]
          Reporter: [e-mail]

user32:msg - test_showwindow() sometimes fails due to an unexpected
WM_SYNCPAINT message on Windows 10:

msg.c:5016: Test failed: ShowWindow(SW_SHOWNA) on visible top level window: 0:
the msg 0x0018 was expected, but got msg 0x0088 instead
msg.c:5016: Test failed: ShowWindow(SW_SHOWNA) on visible top level window: 1:
the msg 0x0046 was expected, but got msg 0x0018 instead
msg.c:5016: Test failed: ShowWindow(SW_SHOWNA) on visible top level window: 6:
the msg sequence is not complete: expected 0000 - actual 0046

See https://test.winehq.org/data/patterns.html#user32:msg

Where: 0x18 == WM_SHOWWINDOW
       0x46 == WM_WINDOWPOSCHANGING
       0x88 == WM_SYNCPAINT

Maybe the WM_SYNCPAINT should just be added as optional.

This only happened 4 times in the past 8 months:
* 2022-12-09 win21H1_newtb-w10pro64-zh-CN-64
* 2023-03-27 win21H1_newtb-w10pro64-ko-64
* 2023-06-09 win21H1_newtb-w10pro64-ja-64
* 2023-08-04 win21H1_newtb-w10pro64-ja-64

[Bug 55417] user32:msg - test_showwindow() sometimes fails due to an unexpected WM_SYNCPAINT message on Windows 10

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55417

François Gouget  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]
           Keywords|                            |source, testcase

[Bug 55417] user32:msg - test_showwindow() sometimes fails due to an unexpected WM_SYNCPAINT message on Windows 10

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55417

--- Comment #1 from Esme Povirk  ---
Do you have a sense of whether WM_SYNCPAINT usually happens at consistent
times? From what I understand, it must have been caused by something happening
in another thread. Since this particular test doesn't involve multiple threads,
I assume it's caused by another process (perhaps a window being hidden).

[Bug 55417] user32:msg - test_showwindow() sometimes fails due to an unexpected WM_SYNCPAINT message on Windows 10

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55417

--- Comment #2 from Esme Povirk  ---
I'm noticing these are all CJK configurations. I wonder if this has something
to do with the input method.

[Bug 55417] user32:msg - test_showwindow() sometimes fails due to an unexpected WM_SYNCPAINT message on Windows 10

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55417

--- Comment #3 from Esme Povirk  ---
My best guess: A tooltip or other temporary window was displayed with
WS_EX_TOPMOST. The first ShowWindow call causes it to hide itself (for example,
if the mouse is no longer hovering over the owner of a tooltip), reveal some
part of the newly-shown window, and send WM_SYNCPAINT. The second call
processes sent messages before showing the window.

I'm not sure how to test that theory though.

[Bug 55417] user32:msg - test_showwindow() sometimes fails due to an unexpected WM_SYNCPAINT message on Windows 10

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55417

--- Comment #4 from François Gouget  ---
Unfortunately I don't have any insight into the timing of the WM_SYNCPAINT
messages. I normally come in when the WineTest report lands on test.winehq.org
and by then there is no record of what was happening on the screen at the time
user32:msg was running (screenshots might help here but that's probably dead).