Hi!

In "Split" view mode, when moving the mouse pointer from i.e. the "3D" view down to the "Right" view, then the "Area/Zoom/Pan/Reset" buttons change focus to the "Right" view (as expected), but when then rotating the mouse wheel, then (roughly half the time) the previous "3D" view gets zoomed instead of the "Right" view. When doing a left click in the "Right" view after moving from "3D" view, then mouse wheel zoom works as expected. However, I find this not to be intuitive. Interestingly, sometime the autofocus works as expected and sometimes it doesn't - appears to be pretty random.

Feedback is highly appreciated!

TIA and best regards

Bernhard


Software Setup:
Moi3D V4
Fedora 38
Kernel 6.4.6-200
Mesa 23.1.4
Wine 8.12
FVWM 2.7.0

Hardware Setup:
AMD Ryzen 7950X
ASUS ProArt Creator X670E
PowerColor Radeon RX7900XTX Red Devil

Hi Bernhard, it's most likely a bug in Wine. But I'll see if I can detect this situation and compensate for it in the next v5 beta.

- Michael

Hi, Could also depend on the desktop manager, not from Moi3D or from WineHQ, I'm using IceWM an X-manger and I sometimes have problems with the focus, try changing the desktop manager maybe try something different from the one you use.

Hi Michael,

thanks very much for your fast response and sorry for my delayed reply :-)

Appears it has something to do with multi-monitor configuration: when the MoI window spans multiple displays, then the mouse-wheel zoom focus works as expected - as long as the mouse pointer is on the primary display, only!

However, for efficient object editing, I maximize the windows that so each view is on a separate display, and so I need to move the mouse pointer to the other displays to edit objects and then mouse-wheel zoom focus doesn't work reliable any more.

As mentioned, the "Area/Zoom/Pan/Rotate/Reset" buttons always change focus as expected on any display. Would it be possible to use the same detection mechanism for the editing area? Or make it configurable in case it brakes Windows or OSX?

Feedback is highly appreciated!

TIA and best regards

Bernhard

BTW.: on more data point: mouse-wheel zoom focus works as expected with MoI V3 (MoI window spanning multiple displays and mouse pointer located on any screen).
Unfortunately, MoI V3 crashes when the model reaches a certain complexity ...

Hi Bernhard,

re:
> As mentioned, the "Area/Zoom/Pan/Rotate/Reset" buttons always change focus
> as expected on any display. Would it be possible to use the same detection
> mechanism for the editing area?

Not exactly because you're doing a click for the Area/Zoom/Pan/Rotate/Reset, and no click for the mouse wheel zoom.

But I should be able to do some extra checking just for the case of running under Wine, to manually re-route a
mouse wheel event to the viewport that the mouse is over top of.

- Michael

Hi Michael,

Thanks for your fast response!

>> As mentioned, the "Area/Zoom/Pan/Rotate/Reset" buttons always change focus
>> as expected on any display. Would it be possible to use the same detection
>> mechanism for the editing area?

> Not exactly because you're doing a click for the Area/Zoom/Pan/Rotate/Reset, and no click for the mouse wheel zoom.

Sorry for the miss-understanding - what I meant is: when I move the mouse pointer from i.e the Top view to the 3D view, then the "Area/Zoom/Pan/Rotate/Reset" buttons in the 3D view automatically show up without the need to do a click (and the buttons disappear in the Top view as expected). That works very reliable from any view to any view - I always assumed it's meant to work like that :-)

When the mouse pointer is moved outside any of the views, then the buttons do not show up in any of the views (kinda expected). When I enter a view from outside the MoI window, then the buttons show up in that specific view (as expected). For me, this does not require a mouse click. That why I thought this mechanism might be suitable to determine which view has the focus for the mouse-wheel zoom.

Feedback is highly appreciated!

TIA and best regards

Bernhard

Hi Bernhard,

re:
> That why I thought this mechanism might be suitable to determine which view has the focus
> for the mouse-wheel zoom.

I see. But that particular mechanism is part of the operating system - the way Windows is set up a mouse move message is sent to the window the mouse cursor is over (unless the mouse has been captured) while a mouse wheel message is sent to the window that has keyboard focus.

But it should not be difficult for me to cook up something similar.

- Michael

Hi Bernhard, so now that I dig into it a bit more I see that there is already something in place that is supposed to be routing mouse wheels to the viewport under the mouse.

So I guess the MoI function that retrieves the viewport under the mouse cursor point is not working right on your system.

Is your multi monitor configuration unusual in any way, like is each monitor in some separate coordinate space instead of being all one connected desktop or something like that?

- Michael

Hi Michael,

Thanks for your continued support!

My setup might be slightly unusual in two ways:

a) I'm using six 1920x1200 24" displays (3 columns, 2 rows) where the left two displays are directly connected to the video card (as usual) and the other four displays are connected via two "Matrox DualHead2Go Digital SE" to the same video card. Means, the OS detects two 1920x1200 and two 3840x1200 displays (forming a 5760x2400 desktop in total). The Matrox splitters are necessary because the Radeon RX7900XTX can only drive up to four displays. I temporarily disabled the outputs for the two DH2G so that there are only two 1920x1200 displays connected but there is no improvement.

b) The top three displays are operated in "inverted" mode because they are mounted upside down to have a thin bezel between the displays. I temporarily disabled the outputs for the two DH2G so that there are only two 1920x1200 displays connected and operated both displays in "normal" (not inverted) mode but there is no improvement.

BTW.: the same effect is observable for GNOME 44 with X11 and Wayland

I'll temporarily install V4 on a PC that has an nvidia card, just to make sure it's not an issue with the amdgpu driver.

Looking forward to try out the solution :-)

TIA and best regards

Bernhard

Just a quick update concerning the following statement:

> I'll temporarily install V4 on a PC that has an nvidia card, just to make sure it's not an issue with the amdgpu driver.

Unfortunately, the installation of MoI V4 fails: after executing the installer, a window with all black content shows up and the installer hangs.
I also tried installing MoI V3, but the installer hangs right after execution. This is with latest Fedora 38 packages and nvidia driver 535.54.03 (latest production driver).
MoI V3 used to run on that PC, many years back (f23). Probably some kind of regression in Fedora or the nvidia driver.

Hi all!

Just to keep you in the loop:

Michael fixed the mouse wheel zoom autofocus issue for Linux/Wine! Many, many thanks, I highly appreciate it!

Michael wrote that the fix will be included in the next beta release.

He also wrote that if other Linux users are running into the same specific problem (which is with windows extending across multiple monitors running under Wine in Linux) then they can e-mail him to get a patched v4.

best regards

Bernhard