Fix ortho layouts and add planck_mit

[opulpo]

This PR is targeted at orthogonal layouts.

Il removes extra trailing commas in ortho layouts, and a missing closing parenthese in planck_grid, that would make qmk compile fail.

It also updates planck_grid which I believe incorrectly had an extra top row and a 2u space key.

Il finally adds the actual 2u space key Planck layout, planck_mit.

Tested to work with KPrepublic’s CSTC40 rev3, which has a planck_mit layout.

[severindupouy]

todo:

Hello, thanks for your contribution!

Some points to complete the PR :

• update the docs/supported_keyboards.md, add the new LAYOUT_planck_mit in the list.
• if you know enough about git, can you split in 2 commits, 1 for the fix, 1 for introducing planck_mit ?
• please reword the commit's messages following https://www.conventionalcommits.org (so 1 commit fix: blabla and 1 commit feat: support planck_mit layout)

If you have question or need help, tell me.

[opulpo]

options that seemed more relevant to Selenium were removed: LEFT_HAND_SPACE and RESTORE_SPACE, the latter causing the left thumb key mapped to LSFT_T(KC_BSPC) trigger spaces on tap, even without holding another key beforehand.

Actually I overlooked that one, and realized afterwards that RESTORE_SPACE was actually an Arsenik option only. I’m not sure how it applies to Arsenik however, since Arsenik makes sense with a space bar in the middle :

#define RESTORE_SPACE
/* Having Space accessible to only one thumb may create some problems, especially
 * when trying to type Shift + Space or Lafayette / AltGr + Space (depending if
 * your space key is on your left or right hand). When active, backspace gets
 * temporarily replaced by space when the original space key is held
 *
 * (Requires `VIM_NAVIGATION`)
 */

In practice anyway, leaving the option enabled would make my LTHUMB_TUCK key mapped to LSFT_T(KC_BSPC) actually fire Space rather than Backspace on tap (without holding any other key).

Not sure what to do then, the code that triggers with RESTORE_SPACE enabled is still there in keymap.c. Is this option a leftover of something initially intended for Selenium?

bool process_record_user(uint16_t keycode, keyrecord_t *record) {
#ifdef RESTORE_SPACE
    static bool thumb_mod_same_hand_as_space_held = false;
    if ((keycode & 0xff) == KC_SPC && record->tap.count == 0) thumb_mod_same_hand_as_space_held = record->event.pressed;
#endif

    // Let QMK do its thing on key releases.
    if (!record->event.pressed) return true;

#ifdef RESTORE_SPACE
    if ((keycode & 0xff) == KC_BSPC && !thumb_mod_same_hand_as_space_held && record->tap.count > 0) {
        tap_code(KC_SPC);
        return false;
    }
#endif

[opulpo]

One last comment :) this conversation seems to raise the issue of having all supported layouts in a shared folder. Shouldn’t Arsenik and Selenium each have their own supported layouts folder, i.e. only layouts with a middle space bar for the former, and only layouts with mirrored thumb clusters for the latter?

[severindupouy]

Hi @opulpo.

I think you don't have a clear view of the aekeynox project approach. No problem with that, aekeynox is complex and very opinionated.

So please, follow my lead :

• the compile fixes and planck_mit are good — let's land only those. Please scope this PR down to them (two focused commits, conventional messages, closing Generating an Arsenik layout seems to output a Selenium layout #21), and drop the changes to arsenik/internals.h and arsenik/options.h. The planck_mit should have the correction I pointed out (k52 placement).

The other points you're raising looks to be questions, they belongs in separate issues if you want to open discussions. But as I read them, they rest on misreadings of the current code:

• shared/layouts.h is keyed by board geometry, not by keymap. Arsenik already runs on split/mirrored boards (output/crkbd, lily58, sofle), so it isn't "middle-spacebar only," and a per-keymap layout folder would duplicate every board.
• RESTORE_SPACE is a default-on Arsenik option, not a Selenium leftover. By design it makes the thumb Backspace keys type Space on tap — real Backspace is the dedicated corner key, and the thumbs revert to Backspace while a space-layer key is held. The LSFT_T(KC_BSPC)-tap-emits-Space you flagged is the option working as intended, not a malfunction.

For questions, the place to have discussions is in issues. Then we could agree whether it's actually a problem, and eventually open a related PR.

Happy to review focused contributions — let's get the fixes + planck_mit merged.

[opulpo]

Hi @severindupouy , I understand that my additional changes were out of place and somewhat careless.

• the compile fixes and planck_mit are good — let's land only those. Please scope this PR down to them (two focused commits, conventional messages, closing Generating an Arsenik layout seems to output a Selenium layout #21), and drop the changes to arsenik/internals.h and arsenik/options.h. The planck_mit should have the correction I pointed out (k52 placement).

I updated the PR as requested. Please note there’s an additional merge commit since I mistakenly opened the PR with my fork’s main instead of my dedicated fix branch.

I’m still unsure that it closes #21, since the keymap obtained does not match Arsenik on a middle-spacebar-layout as described and illustrated on its dedicated Github page and webpage, but that may be by design and a non-issue.