Auto-sync claude-code 2026-06-23 20:05

2026-06-23 20:05:55 -05:00
parent 8a35ebb095
commit 44f146d4f7
2 changed files with 30 additions and 0 deletions
@@ -2,3 +2,4 @@
 - [Community user avdp](project_community_user_avdp.md) — Docker/NAS user on Fedora with virtual serial port, may PR package manager detection
 - [Vestaboard API](project_vestaboard_api.md) — issue #48, implement Vestaboard Local API for HA integration
 - [Always bump VERSION](feedback_version_bump.md) — VERSION file must match release tag before creating a release
 - [Open PRs June 2026](project_open_prs_june2026.md) — PRs #50/#51/#52 reviewed, awaiting contributor responses
@@ -0,0 +1,29 @@
 ---
 name: open-prs-june2026
 description: Status of open PRs
 metadata: 
  node_type: memory
  type: project
  originSessionId: 6b819400-585a-4c35-b6da-9dd655a36f60
 ---
 Three open PRs reviewed on 2026-06-23. Plan: merge #52 first, then #50, then #51 with beta cycle.
 ## PR #52 — Fix calibration adjustments to move immediately (InertiaImpact)
 - **Status:** Approved with question
 - **Waiting on:** Confirmation that firmware `g` command is "goto absolute step position" (not character index)
 - **Plan:** Merge once confirmed
 ## PR #50 — Add SplitFlap Gateway MQTT transport (avandeputte)
 - **Status:** Comment posted, changes requested
 - **Waiting on:** Fix for silent publish on disconnect — `write()` should check `_connected` before publishing
 - **Plan:** Merge to main after fix (feature is opt-in, no beta needed)
 ## PR #51 — Universal Firmware provisioning (InertiaImpact)
 - **Status:** Comment posted, needs architectural clarification
 - **Waiting on:** Explanation of how serial read contention is avoided (background reader vs calibration transactions)
 - **Plan:** Tag `0.3.0-beta.1` off this branch for soak testing, merge to main after ~1-2 weeks of beta feedback
 **Why:** These PRs are from community contributors. Serial contention in #51 is the main risk — it could cause intermittent calibration timeouts that only surface on real hardware.
 **How to apply:** When contributors respond, check their answers against the concerns above. For #51, look for confirmation that calibration holds serial_lock across the full write+read cycle.