[Bug 36873] New: Chronology (.NET 4.0 game) based on 'OpenTK': mouse input non-functional, mouse cursor invisible

https://bugs.winehq.org/show_bug.cgi?id=36873

Bug ID: 36873
           Summary: Chronology (.NET 4.0 game) based on 'OpenTK': mouse
                    input non-functional, mouse cursor invisible
           Product: Wine
           Version: 1.7.21
          Hardware: x86
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: -unknown
          Assignee: [e-mail]
          Reporter: [e-mail]

Hello folks,

the GOG.com version installs .NET Framework 4.0 client profile on its own in
clean 32-bit WINEPREFIX (without Mono).

If the mouse is moved (either in fullscreen or virtual desktop mode), the
console is spammed with:

--- snip ---
fixme:win:GetMouseMovePointsEx (16 0x33ea10 0x33e5cc 64 1) stub
fixme:win:GetMouseMovePointsEx (16 0x33ea10 0x33e5cc 64 1) stub
fixme:win:GetMouseMovePointsEx (16 0x33ea10 0x33e5cc 64 1) stub
--- snip ---

Source:
http://source.winehq.org/git/wine.git/blob/bdc9b147b948b3476815c52f7fc53f4c2...

--- snip ---
1427 int WINAPI GetMouseMovePointsEx(UINT size, LPMOUSEMOVEPOINT ptin,
LPMOUSEMOVEPOINT ptout, int count, DWORD res) {
1428
1429     if((size != sizeof(MOUSEMOVEPOINT)) || (count < 0) || (count > 64)) {
1430         SetLastError(ERROR_INVALID_PARAMETER);
1431     return -1;
1432     }
1433
1434     if(!ptin || (!ptout && count)) {
1435         SetLastError(ERROR_NOACCESS);
1436         return -1;
1437     }
1438
1439     FIXME("(%d %p %p %d %d) stub\n", size, ptin, ptout, count, res);
1440
1441     SetLastError(ERROR_POINT_NOT_FOUND);
1442     return -1;
1443 }
--- snip ---

MSDN:
http://msdn.microsoft.com/en-us/library/windows/desktop/ms646259%28v=vs.85%2...

That noisy stub (although annoying) doesn't seem to be the problem here.

See here for discussion and how it's implemented in 'OpenTK':

https://github.com/opentk/opentk/issues/76

https://github.com/opentk/opentk/blob/develop/Source/OpenTK/Platform/Windows...

--- snip ---
        void HandleMouseMove(IntPtr handle, WindowMessage message, IntPtr
wParam, IntPtr lParam)
        {
            unsafe
            {
                Point point = new Point(
                    (short)((uint)lParam.ToInt32() & 0x0000FFFF),
                    (short)(((uint)lParam.ToInt32() & 0xFFFF0000) >> 16));

// GetMouseMovePointsEx works with screen coordinates
                Point screenPoint = point;
                Functions.ClientToScreen(handle, ref screenPoint);
                int timestamp = Functions.GetMessageTime();

// & 0xFFFF to handle multiple monitors
http://support.microsoft.com/kb/269743
                MouseMovePoint movePoint = new MouseMovePoint()
                {
                    X = screenPoint.X & 0xFFFF,
                    Y = screenPoint.Y & 0xFFFF,
                    Time = timestamp,
                };

// Max points GetMouseMovePointsEx can return is 64.
                const int numPoints = 64;
                MouseMovePoint* movePoints = stackalloc
MouseMovePoint[numPoints];

// GetMouseMovePointsEx fills in movePoints so that the most
                // recent events are at low indices in the array.
                int points = Functions.GetMouseMovePointsEx(
                    (uint)MouseMovePoint.SizeInBytes,
                    &movePoint, movePoints, numPoints,
                    Constants.GMMP_USE_DISPLAY_POINTS);

int lastError = Marshal.GetLastWin32Error();

// No points returned or search point not found
                if (points == 0 || (points == -1 && lastError ==
Constants.ERROR_POINT_NOT_FOUND))
                {
                    // Just use the mouse move position
                    OnMouseMove(point.X, point.Y);
                }
                else if (points == -1)
                {
                    throw new System.ComponentModel.Win32Exception(lastError);
                }
                else
                {
...
--- snip ---

That information might be still useful for 'GetMouseMovePointsEx'
implementation though.

$ wine --version
wine-1.7.21-75-g0a4c786

Regards

[Bug 36873] Chronology (.NET 4.0 game) based on 'OpenTK': mouse input non-functional, mouse cursor invisible

https://bugs.winehq.org/show_bug.cgi?id=36873

Anastasius Focht  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |dotnet

[Bug 36873] Chronology (.NET 4.0 game) based on 'OpenTK': mouse input non-functional, mouse cursor invisible

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #1 from Austin English  ---
This is your friendly reminder that there has been no bug activity for over a
year. Is this still an issue in current (1.7.51 or newer) wine?

[Bug 36873] Chronology (.NET 4.0 game) based on 'OpenTK': mouse input non-functional, mouse cursor invisible

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #2 from Anastasius Focht  ---
Hello folks,

obviously still present.

https://source.winehq.org/git/wine.git/blob/HEAD:/dlls/user32/input.c#l1458

$ wine --version
wine-1.7.51

Regards

[Bug 36873] Chronology (.NET 4.0 game) based on 'OpenTK': mouse input non-functional, mouse cursor invisible

https://bugs.winehq.org/show_bug.cgi?id=36873

Anastasius Focht  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|-unknown                    |user32

--- Comment #3 from Anastasius Focht  ---
Hello folks,

revisiting, still present.

$ sha1sum setup_chronology_2.0.0.3.exe 
f254cd745c4b7879df7d1503ea556fecbf94795f  setup_chronology_2.0.0.3.exe

$ du -sh setup_chronology_2.0.0.3.exe 
344M    setup_chronology_2.0.0.3.exe

$ wine --version
wine-4.4

Regards

[Bug 36873] Chronology (.NET 4.0 game) based on 'OpenTK': mouse input non-functional, mouse cursor invisible

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

Tom Hobson  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 36873] Chronology (.NET 4.0 game) based on 'OpenTK': mouse input non-functional, mouse cursor invisible

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #4 from Tom Hobson  ---
This is now affecting Mount & Blade II: Bannerlord.

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

Anastasius Focht  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Chronology (.NET 4.0 game)  |Multiple games have
                   |based on 'OpenTK': mouse    |non-functional mouse input,
                   |input non-functional, mouse |invisible mouse cursor
                   |cursor invisible            |(Chronology, Mount & Blade
                   |                            |II: Bannerlord)

--- Comment #5 from Anastasius Focht  ---
Hello Tom,

thanks for the information. I've updated the summary line to be more generic
and mention your game. I've just retested Chronology 2.0.0.3 from GOG.com and
it's still present.

$ wine --version
wine-5.5-126-g5c0e699dba

Regards

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #6 from Tom Hobson  ---
Hi Anastasius,

Unsure if this will aid you in getting the game running (but it should combat
the error), but here is a patch that helped get M&B:II going:
https://github.com/ValveSoftware/Proton/issues/3706#issuecomment-607021113

Cheers

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

Ryan S. Northrup  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #7 from Ryan S. Northrup  ---
Created attachment 66797
  --> https://bugs.winehq.org/attachment.cgi?id=66797
GetMouseMovePointsEx

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

Ryan S. Northrup  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]
                   |                            |, [e-mail]

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #8 from Ryan S. Northrup  ---
Patch attached (as an actual patch).  Also CC'ing Alistair and Zebediah for
inclusion in wine-staging.

The patch partially implements GetMouseMovePointsEx by storing the previous
mouse cursor position and returning the current (and, if count > 1, previous)
cursor positions.  This is a deviation from the documented behavior (where
Windows actively tracks the last 64 cursor positions), but it seems to be Good
Enough™ in practice (at least for M&B2).

Alistair/Zebediah: given the deviation from Windows' behavior, can we get this
into wine-staging?

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

Alistair Leslie-Hughes  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
    Staged patchset|                            |https://github.com/wine-sta
                   |                            |ging/wine-staging/tree/mast
                   |                            |er/patches/user32-GetMouseM
                   |                            |ovePointsEx
             Status|NEW                         |STAGED

--- Comment #9 from Alistair Leslie-Hughes  ---
(In reply to Ryan S. Northrup from comment #8)
> Patch attached (as an actual patch).  Also CC'ing Alistair and Zebediah for
> inclusion in wine-staging.

Thanks Ryan.  I've based the patch in staging on your patch, can you please
test it to see that I didn't break anything?

Thanks

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #10 from Ryan S. Northrup  ---
Hi Alistair,

Thanks for the quick turnaround!

At a quick glance I don't think the patch as it exists in the wine-staging repo
is going to work; one of our findings in the Proton GitHub thread on this issue
(https://github.com/ValveSoftware/Proton/issues/3706) was that the two points
returned have to be different from one another for the program to actually
recognize it as a mouse movement.  That's why I opted to store the coordinates
from the previous call in those static variables (so that any mouse movement
changes would register an actual change in mouse position).

If there's a policy against static variables in functions, then that's
understandable, but if that's the case then we'd need to generate that second
position some other way.  We previously tried just incrementing the x and y
coordinates by 1, which mostly worked except for camera rotation in the game's
world map (it would only turn counterclockwise).

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #11 from Alistair Leslie-Hughes  ---
There is no policy as to not use statics. I just don't like them unless
required  and in this case it might be.

I would be interested to know, if the GetCursorPos has the same values as ptin?

If you add a fixme line to output the four values (ptin.x/y and pos.x/y) and
then attached the log here?

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #12 from Ryan S. Northrup  ---
Happy to get a log for you, but I can (and previously did) indeed confirm that
the cursor position the game's passing into the function is consistently
identical to the one from GetCursorPos.

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

Ryan S. Northrup  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #66797|0                           |1
        is obsolete|                            |

--- Comment #13 from Ryan S. Northrup  ---
Created attachment 66799
  --> https://bugs.winehq.org/attachment.cgi?id=66799
GetMouseMovePointsEx v2

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #14 from Ryan S. Northrup  ---
Created attachment 66800
  --> https://bugs.winehq.org/attachment.cgi?id=66800
Input/output coordinate log

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #15 from Ryan S. Northrup  ---
Both the requested log and the revised patch which generated it (reintroducing
the static vars for the previous cursor position) are now attached.  The
"ptin:" line is the position the application's passing in, the "ptout[0]" line
is the mouse position returned from GetCursorPos, and the "ptout[1]" line is
the previous position (i.e. the values of last_x and last_y at the time of
function call).

So, from a sample:

168045.639:0035:0073:trace:win:GetMouseMovePointsEx (24 0x4e43f298 0x4e43f2b0
64 1) semi-stub
168045.639:0035:0073:trace:win:GetMouseMovePointsEx     ptin: 466 775
168045.639:0035:0073:trace:win:GetMouseMovePointsEx     ptout[0]: 466 775
168045.639:0035:0073:trace:win:GetMouseMovePointsEx     ptout[1]: 479 787

We can see that "ptin" and "ptout[0]" are identical, even though the former
came from the function args and the latter came from GetCursorPos.

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #16 from Alistair Leslie-Hughes  ---
Thanks for confirming.

I've updated the patch with the one you attached.  Thanks.

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #17 from Ryan S. Northrup (RyNo)  ---
Perfect!  Thank you so much for your help on this (and in general; y'all rock).

Anastasius: can you confirm whether or not the issue persists on your end with
the current wine-staging?  I don't have a copy of Chronology on hand.

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

Arek Hiler  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #18 from Arek Hiler  ---
https://www.winehq.org/pipermail/wine-devel/2020-September/173639.html

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #19 from Arek Hiler  ---
I've tried to fully implement GMMPEx():

https://www.winehq.org/pipermail/wine-devel/2020-October/174989.html
https://source.winehq.org/patches/data/193956

I've tested it with M&B2 and the mouse movement feels right. I've got one round
of review from Julliard and after that I had hard time getting his attention
again. It may be worth giving it a spin in wine-staging.

Ryan, Alistair and Zebediah what do you think?

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

--- Comment #20 from Arek Hiler  ---
Nevermind. This just got in upstream as
f7edc328184c4755c81dec5372b3ae5ac7b8c5df

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

Julian Rüger  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

Gijs Vermeulen  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|STAGED                      |RESOLVED
      Fixed by SHA1|                            |f7edc328184c4755c81dec5372b
                   |                            |3ae5ac7b8c5df
         Resolution|---                         |FIXED

--- Comment #21 from Gijs Vermeulen  ---
(In reply to Arek Hiler from comment #20)
> Nevermind. This just got in upstream as
> f7edc328184c4755c81dec5372b3ae5ac7b8c5df

In that case, marking FIXED.

If anyone still has issues with either of these games, please open a new bug.

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

Alexandre Julliard  changed:

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

--- Comment #22 from Alexandre Julliard  ---
Closing bugs fixed in 5.20.

[Bug 36873] Multiple games have non-functional mouse input, invisible mouse cursor (Chronology, Mount & Blade II: Bannerlord)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=36873

erkankarabulut  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #23 from erkankarabulut  ---
Hello,

There is a similar bug on the game Drakensang, which is opened a year ago. Link
for the bug issue:https://bugs.winehq.org/show_bug.cgi?id=47386.

In short, cursor is visible but it is not possible to graphically move it,
although it workes on the background. Because I can see that buttons are
highlighted when I blindly move the mouse and I am also able to click on them
and everything works normally. Any quick solution to this issue?

Thanks in advance.