Mouse wheel stalls custom session until stop scrolling

[halsafar]

Using NoMachine Enterprise Client 5.3.26_1

I am using NoMachine for remote development, all the NoMachine end points are Linux.

If I open a Gnome Desktop session, my mouse wheel scrolls appropriately and everything is responsive in any app I launch within the gnome-desktop session.

If I open a custom session, say launch: xterm.  My mouse wheel scrolls okay in xterm.  I then launch say an IntelliJ editor, a web browser, doesn’t matter what every app is affected.  If I now try to scroll using my mouse scroll wheel the whole custom session locks up, makes a ton of beeps and then when the scrolling stops everything goes back to normal.  I am guessing the mouse scroll event is flooding NoMachine.  Anything I can do?

TLDR;  Custom Session -> xterm -> launch anything -> scroll mouse wheel -> app stalls until scrolling stops.  This behavior is not seen if I use a Gnome-Desktop session instead.

[halsafar]

Update.  If I launch a custom session that directly launches say an IntelliJ editor, Google-Chrome, (tried lots of apps) the problem still exists.  Mouse Scrolling appears to flood the event stack.  It only seems if I use a virtual desktop session like gnome or kde does this problem go away.

[graywolf]

What if you ran xterm and then you launched xev from there? Are such scrolling events logged by xev?

[halsafar]

So I launched Custom Session -> xterm.

Filled my terminal with junk so I could test scrolling.

Mouse wheel scrolls xterm perfectly, I can unlock my mouse wheel so it is floating and just spin it extremely fast.  Everything is fine.

Launch xev.  Spin the wheel. Here is an example of what scrolling each direction looks like to xev.  I can see it is recording every tiny movement of the mouse wheel.

 

ButtonRelease event, serial 28, synthetic NO, window 0x600001,

root 0xf1, subw 0x0, time 734318546, (170,8), root:(180,41),

state 0x800, button 4, same_screen YES

 

ButtonPress event, serial 28, synthetic NO, window 0x600001,

root 0xf1, subw 0x0, time 734318639, (170,8), root:(180,41),

state 0x0, button 5, same_screen YES

 

I might have to put together a video of this happening.

[graywolf]

Difference in “time” field of those events is 93 ms. It doesn’t look so short. Events are produced and consumed in the same way for any application. So it looks something different than an “event flooding” (the same flooding would happen for xterm as well as any other app).

Maybe a video could help us to understand what is happening.

Could you provide some detail about the Linux OS of client and server host? I’d like to try to reproduce the same issue.

[Britgirl]

Knowing the NoMachine server version installed on the remote host would also be useful. Is it the latest, 6.1.6?

[Author]

Posts