Does devoverrides "kill" autorepeat?

[rpnfan]

I use devoverrides in the following way. Then a key (+ and = )does not autorepeat any longer. Neither on the base, nor on the shift layer. Is that a bug or normal and is there a workaround?

;; Swap/ assign some characters, when shifted/ unshifted (from US International layout to the ENDe position)
(defoverrides
  (=) (lsft =)  ;; swap = and + (first part)
  (lsft =) (=)  ;; swap = and + (second part)
)

my defcfg looks like that in case that is relevant

(defcfg
  process-unmapped-keys yes
  override-release-on-activation yes  ;; needed to remap shifted output 
  concurrent-tap-hold yes
)

[jtroo]

https://jtroo.github.io/config.html#override-release-on-activation

Overrides won't repeat if you have this config enabled.

[rpnfan]

Thanks, that is true. When I do not use that it will autorepeat, which I want to have.

But then I get e problem with my one-shot shift. It will work for swapping + and =, but the second part of my definition will not work. When I type

tap Shift
tap ,

the ouput is the wanted semicolon ; but in addition is followed by a comma , which is wrong. Is that a bug/ problem or is that by design and I do not understand why = will work fine, but not ;?

(defoverrides
  (=) (lsft =)  ;; swap = and + (first part)
  (lsft =) (=)  ;; swap = and + (second part)
  (lsft ,) (;)  ;; output ; on shifted ,-key 
  (lsft .) (lsft ;) ;; output : on shifted .-key 
)

[jtroo]

It is a design limitation that the release on activation is used as a workaround for, with the disadvantage being loss of key repeat.

It is conceivable that override behaviour can be fixed somehow without the loss of key repeat, but that is not implemented.

[rpnfan]

Thanks for the explanation jtroo, so I have to live with that limitation. At least for now. When you find a solution to implement that would be great :)

[jtroo]

Can you test out if the latest main (80189bd) works as desired, when excluding override-release-on-activation yes?