Alt_R key is sending ISO_Level3_Shift

[atoddcci]

I am using Ubuntu 20.04 inside WSL2 on Windows 10.
On Windows 10, I'm running Xpra 4.0.4.
In Ubuntu, I'm running v4.0.4-r27557.

I'm new to Xpra but I'm really impressed by how good it looks relative to some of the other options out there for WSL, and I think that the Xpra approach is probably more secure as well. Thanks for this project.

I have a standard US keyboard, which has Alt keys on the left and right. The left one works fine, however, the right one is not sending Alt_R, it's sending ISO_Level3_Shift. The problem happens both with and without keyboard sync.

When I run keyboard-test on the Windows side, it reports Alt_R. However, when I run it on the WSL side, it reports ISO_Level3_Shift.

I'm attaching some initial debugging files and can provide more later; I'm hoping that since this is such a common setup I'm missing something obvious. Thank you.

I see the same in xev:

KeyPress event, serial 33, synthetic NO, window 0x800001,
root 0x388, subw 0x0, time 16631071, (151,0), root:(288,166),
state 0x0, keycode 92 (keysym 0xfe03, ISO_Level3_Shift),
same_screen YES,
XLookupString gives 0 bytes:
XmbLookupString gives 0 bytes:
XFilterEvent returns: False 

Keymap looks normal:

xkb_keymap {
xkb_keycodes  { include "evdev+aliases(qwerty)" };
xkb_types     { include "complete"      };
xkb_compat    { include "complete"      };
xkb_symbols   { include "pc+us+inet(evdev)"     };
xkb_geometry  { include "pc(pc104)"     };
}; 

rules:      evdev
model:      evdev
layout:     us

[atoddcci]

Output from the bug report UI

[atoddcci]

xmodmap -pm output

[atoddcci]

xmodmap -pke output

[Antoine Martin]

I'm new to Xpra but I'm really impressed by how good it looks relative to some of the other options out there for WSL

Thanks. The problem is that the default version shipped with Ubuntu, which is what most people will end up trying, is fundamentally broken and barely works at all.

I can confirm your findings, I'm also seeing some spurious NumLock key events firing.
This comes from an ancient workaround for non-us layouts: r1760 (8 years ago..).
I am very reluctant to touch this code, as fixing AltGr proved to be very hard.
But it may well be that the workarounds are no longer needed, as I am not seeing the spurious events we used to detect.
But it is too late for for the 4.1 release.

[migration script]

this ticket has been moved to: ​https://github.com/Xpra-org/xpra/issues/2946