qmk/qmk_firmware

[Bug] Debounce on ErgoDash regressed somewhere between 0.10.35 and 0.12.18

Open

#12,285 opened on 2021年3月18日

GitHub で見る
 (12 comments) (0 reactions) (0 assignees)C (43,867 forks)batch import
bughelp wanted

Repository metrics

Stars
 (20,368 stars)
PR merge metrics
 (平均マージ 20d 9h) (30d で 27 merged PRs)

説明

Describe the Bug

I rebased my keymap from 0.10.35 to 0.12.18, then reflashed my board. Now I get a lot of key repeats, especially on the pinkie finger keys, which makes the board almost unusable. If I rebase back onto 0.10.35 and rebuild/flash everything works fine. You can see my keymap here: https://github.com/qmk/qmk_firmware/compare/master...tcrs:ergosnap-isolike - I don't think there is anything "interesting" going on in the keymap.

System Information

  • Keyboard: ErgoSnap (same matrix as ErgoDash)
    • Revision (if applicable: rev1
  • Operating system: Arch linux
  • AVR GCC version: 10.2.0
  • ARM GCC version: 10.2.0
  • QMK Firmware version: 0.12.18
  • Any keyboard related software installed?
    • AutoHotKey
    • Karabiner
    • Other:

Additional Context

I found this issue: https://github.com/qmk/qmk_firmware/issues/6269 Which sounds similar. However the conclusion seemed to be just to fiddle with DEBOUNCE/DEBOUNCE_TYPE until stuff starts working again. I would rather understand why this has changed and how to avoid the same problem happening again in the future. DEBOUNCE is set to 5 in ergodash/rev1/config.h in both 0.10.35 and 0.12.18, and it doesn't look like the default algorithm has changed in common_features.mk (sym_defer_g or sym_g as it used to be named). It doesn't look like there have been any material changes in quantum/debounce in between 0.10.35 and 0.12.18.

Could anyone give me some tips on how to debug what is happening here? It seems like a clear regression which probably doesn't affect only ErgoDash...

コントリビューターガイド