Skip to content

Correct Ally press-test paddle sides and restore inputplumber - #81

Merged
DanielJBurbridge merged 1 commit into
feat/ally-parity-integration-20261111from
fix/ally-press-test-20261117
Nov 17, 2026
Merged

DanielJBurbridge merged 1 commit into
feat/ally-parity-integration-20261111from
fix/ally-press-test-20261117

Conversation

@DanielJBurbridge

Copy link
Copy Markdown
Owner

What changes

The Ally button-capture script labelled M1 as the left back paddle and M2 as the right. M2 is the left paddle and M1 is the right. Prompts in docs/handheld/research/tools/ally-press-test.sh and the same side labels in the handheld docs are corrected.

The script now stops inputplumber.service before the capture and starts it again on EXIT, INT and TERM, only if that unit was active, and the trap runs the restart once. Volume keys are read from the AT Translated Set 2 keyboard, found by name in /proc/bus/input/devices, not from a fixed event node.

Daniel's 17 Nov RC71L readings are recorded in docs/handheld/01-controller-button-map.md as observed with inputplumber stopped, pending a re-run with this script: M2 (left) sent F15, M1 (right) sent nothing, A was not reported while a paddle was held, and a 1 s Command Centre hold sent nothing.

Claims L230–L239. L230 is the re-run.

Model: grok-4.7

Verification

bash test/shell.d/ally-press-script-test.sh passed: bash -n on the script, the paddle prompts name M2 on the left and M1 on the right, and the trap stops and restarts inputplumber.service.

Not run on the Ally. The 17 Nov readings are from the previous script and stay pending a re-run.

Screenshots

No visible change: research script and docs only. The desktop shell is untouched.

Open in Web Open in Cursor 

M2 is the left back paddle and M1 is the right. The capture stops
inputplumber.service and starts it again afterwards if it was active.
Volume keys come from the AT Translated Set 2 keyboard, found by name.
The 17 Nov readings are recorded as pending a re-run. L230.

Model: grok-4.7
Co-authored-by: Daniel Burbridge <DanielJBurbridge@users.noreply.github.com>

@DanielJBurbridge DanielJBurbridge left a comment

Copy link
Copy Markdown
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

dev-red review: PASS at f5a65c9

  • The trap is installed before inputplumber is stopped, and it only restarts the unit if it was active, once (IP_RESTORED). INT exits with 130 and TERM with 143. If the stop fails, the script exits before the capture.
  • VOL is found by name, and an empty device is skipped. M2/M1 sides match the 17 Nov results in every doc this touches.
  • bash -n and ally-press-script-test pass locally. Integ a1fc4bb is an ancestor of this head. There's a Model trailer (grok-4.7). Nothing is removed from the follow-ups.
  • Actions test is green.

Low follow-up (L230 range): add HUP to the traps. This tool is run over ssh rog-ally. If the SSH session drops mid-capture, the shell gets SIGHUP, which isn't trapped, so inputplumber stays stopped and the pad loses its greeter pointer until reboot or a manual systemctl start. trap 'restore_capture; exit 129' HUP fixes it.

Cleared to merge at this sha.

@DanielJBurbridge
DanielJBurbridge merged commit e9fca47 into feat/ally-parity-integration-20261111 Nov 17, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants