Skip to content

wire: align comments, names and the telemetry request size with the firmware (LAB-360) - #299

Open
mbreiser wants to merge 1 commit into
mainfrom
fix/wire-review-arena-interface
Open

mbreiser wants to merge 1 commit into
mainfrom
fix/wire-review-arena-interface

Conversation

@mbreiser

Copy link
Copy Markdown
Contributor

Summary

Small corrections to the G6 wire module, found by reviewing Will Dickson's Python library (willdickson/arena-interface), whose discrepancy table compares our JS with firmware 12bba7e. Tracked in LAB-360; firmware-side items are on LAB-343.

  • GET_TELEMETRY_BLOCK defaults to 178 record bytes (the firmware's budget), not 180. The firmware clamps, so nothing changes on the wire.
  • GET_PATTERN_INFO byte 11 is decoded as dutyCycle (the firmware name); stretch stays as an alias. The Studio and Console read dutyCycle and fall back to stretch, so a stale cached wire module still works.
  • Comment fixes: 01 C1 / 01 C2 on the IP and controller-info encoders, the GET_CRASHREPORT gate is 0xCB flags bit 3 (not the telemetry bit), GET_AO_VOLTAGE is a DAC register readback.
  • CLAUDE.md: crash-report gate text corrected.

Tier

next candidate

Release notes

none

Test plan

  • pixi run test (73/73 files)
  • pixi run prettier --check on the changed JS files
  • Not opened in a browser: the HTML change is one display string in each of arena_studio.html and arena_console.html

In-lab testing (on Next, between sessions)

  • Studio Console, connected controller: select a pattern on the SD card. The info line reads "duty cycle N" with a number, not "undefined".
  • Telemetry-capable firmware: run a short protocol with Run logging on. The run log still has controller rows (cc / cf / cs).
  • Old Console (arena_console.html): pattern info shows a "duty cycle" row.

The wider bench plan for the open follow-ups (stray input at connect, g6-encoding.js row order, new encoders, Will's hardware suite) is in LAB-360.

🤖 Generated with Claude Code

…irmware (LAB-360)

From the review of Will Dickson's arena-interface library, whose discrepancy
table compares our JS with firmware 12bba7e.

- GET_TELEMETRY_BLOCK defaults to 178 record bytes (the firmware's budget),
  not 180. The firmware clamps, so behavior on the wire is unchanged.
- GET_PATTERN_INFO byte 11 is decoded as dutyCycle (the firmware name);
  stretch stays as an alias. The Studio and Console read dutyCycle and fall
  back to stretch for a stale cached wire module.
- Comments: 01 C1 / 01 C2 on the IP and controller-info encoders, the
  GET_CRASHREPORT gate is 0xCB flags bit 3 (not the telemetry bit),
  GET_AO_VOLTAGE is a DAC register readback.
- CLAUDE.md: crash-report gate text corrected.

In-lab testing (after this is on Next, between sessions):
1. Studio Console, connected controller: select a pattern on the SD card and
   check the info line reads "duty cycle N" (a number, not "undefined").
2. Telemetry-capable firmware: run any short protocol with Run logging on and
   check the run log still has controller rows (cc / cf / cs).
3. Old Console (arena_console.html): pattern info shows a "duty cycle" row.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>

This branch has not been deployed

No deployments
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.

1 participant