Session Handoff [hw-cli]: liquidraver/ZephCore#100 merged; #101 (RX8130CE + YSN8900 weekday) open, awaiting owner review #47

Next steps, in priority order

1. Monitor rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101 and answer the owner's feedback.
   • rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101 is the RX8130CE and YSN8900 weekday fix, clean branch rx8130ce-week-one-hot at 2621782, fork iteration PR rtc: write the RX8130CE weekday one-hot #45 at 44ed7e3.
   • List everything posted on rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101 and rtc: write the RX8130CE weekday one-hot #45 after 2026-10-05T21:00Z: issue comments, review comments and reviews.
   • Fix on the dirty branch first, run local review, then re-squash and force-push the clean branch with a lease, on Pieter's go-ahead.
   • Done means the owner has merged it or asked for nothing more.
2. When rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101 merges, close fork rtc: write the RX8130CE weekday one-hot #45 unmerged and close RTC save writes the weekday in binary; the RX8130CE's WEEK register is one-hot #39 (it was fixed on both boards). Delete work/rx8130ce-week-one-hot and rx8130ce-week-one-hot on Pieter's go, as was done for cli: bound sensor list's start index and its pages #44.
3. Retire the finished rak4631: use a fitted RAK12002 RTC, and keep its backup switchover enabled liquidraver/ZephCore#98 contribution, if Pieter agrees (asked at close; see the comments below). Close fork rak4631: use a fitted RAK12002 RTC, and keep its backup switchover enabled (iteration branch) #35 unmerged and delete work/rak4631-rak12002-rtc and rak4631-rak12002-rtc. Keep the rak4631: use a fitted RAK12002 RTC, and keep its backup switchover enabled liquidraver/ZephCore#98 work worktree's Zephyr git worktree. The r6-base and rx8130ce-week worktrees reach it through their zephyr symlinks. Move it, or re-point those symlinks, before removing the worktree that holds it.
4. Hold companion: mark a CLI reply cut to the app frame (iteration branch) #27-companion: never treat an over-the-air CLI command as local (iteration branch) #32 until Pieter says go (his answer of 2026-10-04 on Session Handoff [hw-cli]: liquidraver/ZephCore #98 and #99 upstream; monitor them, reconnect with MeshCore and Zephyr sessions #41).
5. hw CLI on hardware (Add hw hardware-report CLI command (iteration branch) #1, Add the hw hardware-report CLI command (rebased, hw CLI only) #16). Not re-run since it was separated out.

External blockers

• Pieter's go-ahead is needed for any further force-push of a clean upstream branch, any new upstream PR, any branch deletion, and any write to meshcore-dev/MeshCore or zephyrproject-rtos/zephyr.
• liquidraver's review of rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101.
• CodeRabbit on the fork allows one review an hour and must be requested. Copilot reviews upstream PRs on its own.

Internal dependencies

• Step 2 waits on rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101 merging, since the iteration PR and its branches are the working copy until then.
• Step 3's worktree removal waits on re-homing the shared Zephyr tree. Otherwise the two worktrees that symlink to it stop building.
• RV3028 rules are shared with MeshCore (#3545, #3421) and Zephyr (#121252). A change to any of them is agreed across the three sessions first.

State (re-derive: git worktree list, gh pr list)

• Fork dev = upstream dev = 108b96d (a CLI docs rename after cli: reject a negative sensor list start index liquidraver/ZephCore#100; nothing in it overlaps rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101). rak4631: use a fitted RAK12002 RTC, and keep its backup switchover enabled liquidraver/ZephCore#98 merged as cdd6cc6, cli: reject a negative sensor list start index liquidraver/ZephCore#100 as d1cc647.
• Open upstream PRs from this track: rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101 only.
• Fork PRs: rtc: write the RX8130CE weekday one-hot #45 44ed7e3, rak4631: use a fitted RAK12002 RTC, and keep its backup switchover enabled (iteration branch) #35 7c75633 (its upstream PR has merged), and Add hw hardware-report CLI command (iteration branch) #1, Add the hw hardware-report CLI command (rebased, hw CLI only) #16, Elecrow ThinkNode M7: board, CH390 ethernet driver, wired TCP companion (iteration branch) #17 and companion: mark a CLI reply cut to the app frame (iteration branch) #27-companion: never treat an over-the-air CLI command as local (iteration branch) #32 unchanged. cli: bound sensor list's start index and its pages #44 is closed and its branches deleted.
• Worktrees: rx8130ce-week (dirty) and rx8130ce-week-clean are rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101's. r6-base is detached at d1cc647 and holds the restore build below. The rak-rtc-* and up-rak-rtc worktrees are rak4631: use a fitted RAK12002 RTC, and keep its backup switchover enabled liquidraver/ZephCore#98's.
• Hardware: this session's RAK4631 (the one without the extra sensor modules) runs a repeater debug build of d1cc647: handle lpxz, 927.875, tx 1, advert.interval 0, RTC time correct. It was lent to the Zephyr session for its RV3028 BASE/FIX demo and returned. Its 35h is restored to 40, and nothing transmitted during the loan.
• Sibling sessions: MeshCore (meshcore-dev-meshcore-*) cut Reject a negative sensor list start index 🤖🤖 meshcore-dev/MeshCore#3433 to the same one-liner as cli: reject a negative sensor list start index liquidraver/ZephCore#100 (head bbacf47a). Zephyr (zephyrproject-rtos-zephyr-*) ran its demo on our board.

The parked decision queue

0 issues carry decision on this fork. The question put to Pieter at close is commented below.

What the last round did

• rak4631: use a fitted RAK12002 RTC, and keep its backup switchover enabled liquidraver/ZephCore#98 merged upstream (cdd6cc6).
• cli: reject a negative sensor list start index liquidraver/ZephCore#100 cut to one line, then merged. liquidraver asked whether cli: reject a negative sensor list start index liquidraver/ZephCore#100 was over-engineered, and it was. A single sensor setting makes the negative start the only reachable fault. cli: reject a negative sensor list start index liquidraver/ZephCore#100 was cut to start < 0 || with Pieter's approval, and it merged as d1cc647. Fork cli: bound sensor list's start index and its pages #44 and sensor list <start> wraps a large start negative and prints NULL names #43 are closed, and their branches and worktrees deleted.
• Fork rtc: write the RX8130CE weekday one-hot #45 went upstream as rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101, after dev was merged into it and the ternary moved into rtc_time_block(). Pieter then had the meshtracker_x1 YSN8900 folded in. The basis is the Epson RX8900 manual, ETM45E-11 §8.2.2 Table 15 p. 16 (WEEK 03h one-hot). The YSN8900's own sheet has no register map, and its programming guide is unavailable. Evidence is on RTC save writes the weekday in binary; the RX8130CE's WEEK register is one-hot #39.
• Copilot on rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101: its one finding was DT_PROP on an absent boolean, and it was disproven in the thread.
• MeshCore told: the MeshCore session heard about both changes, and #3433 now matches cli: reject a negative sensor list start index liquidraver/ZephCore#100. Zephyr was not affected.

What not to repeat

• A reply's build claim outran the builds. The Copilot reply on rtc: write the RX8130CE and YSN8900 weekday one-hot liquidraver/ZephCore#101 said all four boards build "at this head", when two of them had been built one commit earlier. It was caught and rebuilt within minutes. When a reply cites builds, rebuild every board named at the exact head first.
• Recollection versus code on cli: reject a negative sensor list start index liquidraver/ZephCore#100. The intended argument was a buffer overflow in the parsing. Reading _atoi() showed it only wraps through unsigned arithmetic, so the owner's simpler fix was right. Read the code before defending a design to an upstream owner.

New learnings

• None that are durable beyond this chain.