[Bug 55047] New: d3d test_cursor_clipping() fails

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55047

Bug ID: 55047
           Summary: d3d test_cursor_clipping() fails
           Product: Wine
           Version: 7.21
          Hardware: x86-64
                OS: Linux
            Status: NEW
          Keywords: regression, testcase
          Severity: normal
          Priority: P2
         Component: win32u
          Assignee: [e-mail]
          Reporter: [e-mail]
                CC: [e-mail]
   Regression SHA1: af902c188ebaf7d6dda3b628409d8c390ab410d5
      Distribution: ---

I can observe this on two different machines, with at least ddraw and d3d8
tests. It happens regardless of whether the virtual desktop is used. The
failure looks like this:

ddraw:ddraw1:

ddraw1.c:14352: Test failed: SetDis.layMode failed, hr 0x80004001

d3d8:device:

device.c:10815: Test failed: Adapter 0: Expect clip rect (0,0)-(640,480), got
(1,1)-(639,479).

It's not quite consistent, though. One machine can reproduce the ddraw1 failure
pretty consistently, enough for me to be able to bisect, but the other seems to
fail only intermittently. As a result I can't be 100% sure of the bisect
result, but it seems to make sense.

[Bug 55047] d3d test_cursor_clipping() fails

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55047

--- Comment #1 from Zeb Figura  ---
(In reply to Zeb Figura from comment #0)
> ddraw1.c:14352: Test failed: SetDis.layMode failed, hr 0x80004001

Sorry, copy-pasted the wrong line here. The line I meant to copy was:

ddraw1.c:14359: Test failed: Expect clip rect (0,0)-(320,200), got
(1,1)-(319,199).

[Bug 55047] d3d test_cursor_clipping() fails

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55047

--- Comment #2 from Rémi Bernon  ---
> device.c:10815: Test failed: Adapter 0: Expect clip rect (0,0)-(640,480), got (1,1)-(639,479).

This failure makes sense if the clip fullscreen window option is enabled, which
is not the default. Can you confirm that?

It is also a Wine-specific extension, so I'm not sure we have to consider it a
valid test scenario.
> ddraw1.c:14352: Test failed: SetDis.layMode failed, hr 0x80004001

There I don't see how the change can trigger such failure, but I will have a
look.

[Bug 55047] d3d test_cursor_clipping() fails

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55047

--- Comment #3 from Rémi Bernon  ---
> Sorry, copy-pasted the wrong line here. The line I meant to copy was:

Ah, yeah ofc. Then it makes sense.

[Bug 55047] d3d test_cursor_clipping() fails

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55047

--- Comment #4 from Zeb Figura  ---
(In reply to Rémi Bernon from comment #2)
> > device.c:10815: Test failed: Adapter 0: Expect clip rect (0,0)-(640,480), got (1,1)-(639,479).
>
> This failure makes sense if the clip fullscreen window option is enabled,
> which is not the default. Can you confirm that?

These should be stock prefixes (except with virtual desktop maybe applied). I
double checked and "Automatically capture the mouse in full-screen windows"
isn't enabled.

[Bug 55047] d3d test_cursor_clipping() fails

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55047

--- Comment #5 from Rémi Bernon  ---
The virtual desktop mode should be fixed after
b3c1bd33e0c913def8f214baf7875f2ef8a46e6e, but there's still some failures when
neither virtual desktop mode nor fullscreen window clipping options are used.

In this case clip_fullscreen_window still tries to clip the cursor to the
entire virtual screen. I'm not completely sure to understand why it was doing
that, because I don't think the cursor can go beyond the virtual screen anyway,
and I'm tempted to simply not clip fullscreen windows if the option is
disabled.

[Bug 55047] d3d test_cursor_clipping() fails

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55047

Rémi Bernon  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED
      Fixed by SHA1|                            |8d2de5dbe1f317ba8f9f7e2f88f
                   |                            |b3a0629e5e174

--- Comment #6 from Rémi Bernon  ---
Fully fixed with 8d2de5dbe1f317ba8f9f7e2f88fb3a0629e5e174.

[Bug 55047] d3d test_cursor_clipping() fails

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55047

--- Comment #7 from Zeb Figura  ---
(In reply to Rémi Bernon from comment #6)
> Fully fixed with 8d2de5dbe1f317ba8f9f7e2f88fb3a0629e5e174.

I can confirm this is fixed locally, thanks!

[Bug 55047] d3d test_cursor_clipping() fails

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=55047

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #8 from Alexandre Julliard  ---
Closing bugs fixed in 8.11.