bug#65068: 29.1; xkb-interception interaction causes problems with key combinations

["Alexander Prähauser" <alexander.praehauser@gmx.at>]

Sorry for the late reply.

> Which virtual modifier key have you assigned to the apostrophe 
> key?

I've made the apostrophe key (AC11 in xkb) a level 3 switch 
through the following function:

> // The AC11 key (while pressed) chooses the third shift level.
> partial modifier_keys
> xkb_symbols "ac11_switch" {
>   key <AC11> {
>     type[Group1]="ONE_LEVEL",
>     symbols[Group1] = [ ISO_Level3_Shift ]
>   };
> };

If tapped it sends an F14 signal (which I use in Emacs to insert 
math symbols in cdlatex).

> Nit: Emacs is not developed on GitHub, but on Savannah:

I see. What I meant was that I built it using the emacs-git entry 
in the AUR.

Po Lu [2023-08-09 Wed 08:02] wrote:

> Alexander Prähauser <alexander.praehauser@gmx.at> writes:
>
>> I'm using the interception-`dual-function-keys`-plugin
>>
>> https://gitlab.com/interception/linux/plugins/dual-function-keys
>>
>> for my keyboard layout. In version 28.2 it worked perfectly, 
>> but
>> now, if I try to use combinations using keys that have
>> a dual function, the combination breaks off sending the wrong
>> signals. For instance, my CapsLock key has a dual function,
>> sending a F15 (XF86Launch5) if tapped and acting as the
>> apostrophe-key (AC11 in xkb) if held, which is configured in 
>> xkb
>> to act as an xkb level 3 modifier.
>
> Which virtual modifier key have you assigned to the apostrophe 
> key?
>
>> However, now when I try to use
>> CapsLock for a key combination using a level 3 symbol, for
>> instance if I hold Control, then press CapsLock, it seems to 
>> send
>> simply the apostrophe-key, so that I receive the
>> message "C-' is undefined" and the key sequence is broken 
>> off. If
>> I continue to hold both Control and CapsLock, it seems
>> to access level 3 normally though, so that I can still use
>> combinations consisting of C-[level 3 character]. If I only 
>> want
>> to access a level 3 symbol by holding CapsLock, it works 
>> normally,
>> it seems
>> only combinations are affected. While typing this message I
>> noticed an even weirder thing: my F-key is configured to
>> have "a" on level and "Right" on level 5, and Rightalt is
>> configured to act as a level 5 modifier if held, and if I
>> press the F key normally or while holding Rightalt this works, 
>> but
>> if I first hold Control, then Rightalt, then tap the
>> F-key, it seems to interpret the result as C-f, even though on 
>> no
>> xkb-level the F-key is configured to actually send f.
>>
>> The problem persists if I instead build the latest 
>> Emacs-version
>> (30.1) from Github
>
> Nit: Emacs is not developed on GitHub, but on Savannah:
>
>   https://git.sv.gnu.org/cgit/emacs.git