Right CTRL automatically presses [Alt] in Emacs

[timo-a]

Hi,

when I open emacs and press e.g. C-x (that hold down ctrl and then press x as well) and I use the right ctrl (left works!) Emacs shows C-M-x (M usually maps to [Alt]) as being pressed and acts accordingly (no such command).

Also the capslock key does not work as intended, I think it is mapped to C-u.

Edit: I'm using v0.5.1 (6b211f) and Emacs 27.1

[MaxGyver83]

Hi @timo-a ,

thanks for your report. I'll try to reproduce it.

[MaxGyver83]

when I open emacs and press e.g. C-x (that hold down ctrl and then press x as well) and I use the right ctrl (left works!) Emacs shows C-M-x (M usually maps to [Alt]) as being pressed and acts accordingly (no such command).

I can confirm this.

Also the capslock key does not work as intended, I think it is mapped to C-u.

For me it works as expected (as a level 3 modifier). What happens when you enter for example CapsLock+f? When you hit CapsLock only, do you see C-u in Emacs's status bar?

[MaxGyver83]

when I open emacs and press e.g. C-x (that hold down ctrl and then press x as well) and I use the right ctrl (left works!) Emacs shows C-M-x (M usually maps to [Alt]) as being pressed and acts accordingly (no such command).

This bug does not occur in version 0.4. Maybe you can use it until this issue is fixed?
@fw623 : I suppose this bug was introduced with this pull request. Do you have an idea, how to fix it (without reverting your PR)?

[timo-a]

When I press Capslock alone, nothing is shown in the status bar.
When I press Capslock-a, Capslock-s, Capslock-d, Capslock-f, Capslog-g (where "asdfg" mean what is printed on the physical keys, not a virtual mapping) on either "neo" or "qwertz" the following happens:

Capslock +	statusbar
a	C-M-ß
s	/ is typed out
d	C-u 7-
f	C-u 0-
g	* is typed out

I can confirm that right ctrl works in v0.4, however the Capslock bug occurs there as well

[fw623]

@fw623 : I suppose this bug was introduced with this pull request. Do you have an idea, how to fix it (without reverting your PR)?

Honestly, I don't know without trying out stuff. But I currently don't have time for that.

Though one (possibly key) difference I noticed is that in master we call handleSystemKey before handling the neo modifiers. Whereas in v0.4 we handle mod3 and mod4 before looking at the system keys. Though quite a fair bit has changed about that handling in general, so that might not be the cause.

When I press Capslock-a, Capslock-s, Capslock-d, Capslock-f, Capslog-g (where "asdfg" mean what is printed on the physical keys, not a virtual mapping) on either "neo" or "qwertz" the following happens: [...]

As for this, the keypresses that are sent is actually what I'd expect. \ is AltGr + ß (= Ctrl + Alt + ß) on my keyboard (German QWERTZ) and {/} are on AltGr + 7/AltGr + 0. While / and * can be reached by just shifting.

I remember that I had some problems in VSCode in Windows with } and \ since they'd trigger a shortcut instead of typing the character. Maybe that's related to your situation.

Maybe it's worth looking into sendChar and trying around with how that's handled. A potential problem I see there is that the necessary modifier keys are pressed and then released once for the press and then again for the release. So maybe if we only send the modifiers on keyDown events (e.g. if (!isKeyUp && altgr) sendDown(VK_RMENU, 56, false); instead of if (altgr) sendDown(VK_RMENU, 56, false);). It could also be that some isExtended is incorrect in a sendDown/sendUp there.

[MaxGyver83]

Fixed in v0.5.2
Please re-open if it doesn't work as expected.