Position of chords is not respected (perhaps?)

[UdeRecife]

I discovered kanata recently and I've been in loved with it ever since.

But since love does not always translate into to knowledge, I'm been having difficulty in replicating the behavior of chords in two different layers. The chord keys are aliased with @ch_.

So I have

(deflayer default
  caps  _    _     _     _     _    _   f7    f8   f9 f10 f11 f12
  @ngrv _   _    _     _     _    _   _     _    _    _    _    _    _    @cmins @numhm _
  @fntb @mzq _   _     _     _    _   _     _    _    _    _    @fn] @fn\ _      @mzend _
  @ctlc @lma @las @rad @ndf     _    _   @ndj @rak @lal @lm; @ct' @ctr
  @osls @fnz _     _     _     _    _   _     _    _    @fn/ @osrs
  @oslc lmet @osla       @spcc      @osra rmet @osrc
)

(deflayer compact
  ∅  ∅    ∅    ∅    ∅    ∅    ∅    ∅    ∅    ∅     ∅    ∅    ∅
  ∅  ∅    f2   f3   f4   f5   ∅    @ch7 @ch8 @ch9  @ch0 @ch- ∅ bspc   @cmins @numhm _
  _  @chq @chw @che @chr @cht @chy @chu @chi @cho  @chp @ch[ _ _      _      @mzend _
  _  _    _    pgdn home end  left down up   right _    _    _
  _  _    ∅    ∅    spc    spc    @mwl @mwd @mwu @mwr  _    _
  _ _ _        _     _ _ _
)

These work ok.

But when replicate them in a colemak-dh-wide layer, like this:

(deflayer colemak_dh_wide
  caps  _    _     _    _ _ _ f7 f8 f9   f10  f11  f12
  @ngrv 1    2     3    4 5 6 =  7  8    9    0    -  _   @cmins @numhm _
  @fntb @mzq  w    f    p g [ j  l  u    y    ;    '  _   _      @mzend _
  @ctlc @lma @lar  @ras @ndt d ] h  @ndn @rae @lai @rmo @ctr
  _ @fnz _     c    v b / k  m  ,    .    _
  @oslc lmet @osla     @spccw     @osra rmet @osrc
)

(deflayer compactw
  ∅  ∅    ∅    ∅    ∅    ∅    ∅    ∅    ∅    ∅     ∅    ∅    ∅
  ∅  ∅    ∅    ∅    ∅    ∅    ∅    ∅    @ch7 @ch8 @ch9  @ch0 @ch- bspc   @cmins @numhm _
  _  @chq @chw @che @chr @cht @chy _    @chu @chi @cho  @chp @ch[ _      _      @mzend _
  _  _    _    pgdn home end  _    left down up   right _    _
  _  _    ∅    ∅    spc  spc  _    @mwl @mwd @mwu @mwr  _
  _ _ _        _     _ _ _
)

the chord keys behave as if they were still placed in the default layer.

The other changes are respected, as the left up down right keys moved one key to the right, to match the colemak-dh-wide layout.

Am I making missing some adjustment, or is this an issue?

[UdeRecife]

Well, I made it work by creating a new set of chords for the colemak-dh-wide layer.

So the issue is perhaps not really an issue. I feel I'm still failing to really grasp the design logic of chords. So this is probably on me.

[UdeRecife]

I made another observation.

When I use my qwerty config, all things work as expected.

But when I reverse the settings and use a colemak-dh-wide config, then things start to act weird.

For instance, I have this as src

(defsrc
  esc  f1 f2 f3 f4 f5 f6 f7 f8 f9 f10 f11 f12
  grv  1  2  3  4  5  6  =  7  8  9   0    -   bspc  ins home pgup
  tab  q  w  f  p  b  [  j  l  u  y   ;    '   \     del end  pgdn
  caps a  r  s  t  g  ]  m  n  e  i   o    ret
  lsft x  c  d  v  z  /  k  h  ,  .   rsft
  lctl lmet lalt  spc  ralt cmp rctl

and as default layer

(deflayer default
caps   _     _    _    _    _ _ _  _    _    _    _    _
@ngrv  1     2    3    4    5 6 =  7    8    9    0    -    bspc   @qwins @numhm _
@fnwtb @mzq  w    f    p    b [ j  l    u    y    ;    '    @fnw\  _      @mzend _
@ctlc  @lma  @lar @ras @ndt g ] m  @ndn @rae @lai @rmo @ctr
@fnwls @fnwx c    d    v    z / k  h    ,    .    @fnwrs
@oslc  lmet @osla     @spccw     @osra rmet @osrc

While this works ok with the qwerty, in this config, when I press the alias @lma (meaning left meta a), it works as expected. But the following alias, @lar (meaning left alt r) is parsed as a ralt for reasons that escape my understanding.

Am I missing something?

[jtroo]

What OS are you running? And if on Windows, which kanata variant are you using?

[UdeRecife]

What OS are you running? And if on Windows, which kanata variant are you using?

I'm running:

• Arch (Kernel: 6.9.6-zen1-1-zen).
• kanata 1.6.1

[jtroo]

On Linux defsrc always operates on usb scancodes which means defsrc should always look similar to a QWERTY layout unless deflocalkeys is used. Changing defsrc to colemak will give weird results as you found.

[UdeRecife]

Thanks for providing the explanation. Now I get it.

As I mentioned, I love kanata. It breathed a new and much more interesting life onto my keyboard. Kudos to you!