Session Handoff [rv3028-eeprom]: All upstream PRs answered; watch for maintainer review #18

Next steps, in priority order

1. Sweep every upstream PR and issue authored by ptr727 for new comments, not just the three below. Done when each new comment has a reply, or is waiting on Pieter. To check:

   gh pr list --repo meshcore-dev/MeshCore --author ptr727 --state all --limit 50 --json number,state,updatedAt,headRefOid,title
   gh issue list --repo meshcore-dev/MeshCore --author ptr727 --state all --limit 50 --json number,state,updatedAt
   gh pr view <n> --repo meshcore-dev/MeshCore --json comments,reviews   # conversation + review bodies
   gh api repos/meshcore-dev/MeshCore/pulls/<n>/comments                  # inline threads

   Anything with updatedAt after 2026-10-04T18:00Z is new. Check each commenter's author_association. Only a maintainer's verdict decides a PR, but every human question gets an answer. Clear the reply text with Pieter before posting it.
2. Watch meshcore-dev/MeshCore#3434 in particular. A community member (association NONE) asked whether the change fixes anything real. On 2026-10-04 Pieter approved a reply (comment): it changes no behaviour in-tree, it is a dead-knob fix, and we're happy to close it. If a maintainer says close, close upstream Pass TELEM_RAK12500_ADDRESS to the RAK12500 I2C probe 🤖🤖 meshcore-dev/MeshCore#3434 and fork PR Pass TELEM_RAK12500_ADDRESS to the RAK12500 I2C probe (iteration branch) #9 unmerged, and delete work/rak12500-i2c-address and its clean branch.
3. Respond to review on Adopt an RTC only if its time registers read like that chip 🤖🤖 meshcore-dev/MeshCore#3544, Store RV3028 backup switchover and trickle charger config in EEPROM 🤖🤖 meshcore-dev/MeshCore#3545 and Write and read the RX8130CE WEEK register as one-hot 🤖🤖 meshcore-dev/MeshCore#3550 per the upstream-contribution-workflow skill, exactly as round 2 laid out (link Session Handoff [rv3028-eeprom]: Three MeshCore PRs upstream; watch and respond to review #17, step 2):
   • Fix on the work branch, run and record local-strict-review, then re-squash onto the fix branch.
   • Push with --force-with-lease, then update the permalinks.
   • Reply and resolve with a live thread id.
   • Ask Pieter to re-request Copilot in the web UI.
   • Check the finding against ZephCore (see below).
4. Cross-firmware coordination: the ZephCore counterparts are upstream now, as liquidraver/ZephCore#98 (RV3028 EEPROM, the counterpart of Store RV3028 backup switchover and trickle charger config in EEPROM 🤖🤖 meshcore-dev/MeshCore#3545) and #99 (probe identity, the counterpart of Adopt an RTC only if its time registers read like that chip 🤖🤖 meshcore-dev/MeshCore#3544).
   • The ZephCore maintainer requires both firmwares to change any shared rule together. Before changing "a failed read proves nothing" or the two-agreeing-reads rule on Adopt an RTC only if its time registers read like that chip 🤖🤖 meshcore-dev/MeshCore#3544, tell the ZephCore session first.
   • Pieter's standing instruction for this contribution still holds: send any finding that may apply to ZephCore over to the ZephCore session.
5. If upstream dev moves: follow link Session Handoff [rv3028-eeprom]: Three MeshCore PRs upstream; watch and respond to review #17's step 3. Merge into the work branches, never rebase, then re-squash. Adopt an RTC only if its time registers read like that chip 🤖🤖 meshcore-dev/MeshCore#3544 and Store RV3028 backup switchover and trickle charger config in EEPROM 🤖🤖 meshcore-dev/MeshCore#3545 conflict in AutoDiscoverRTCClock::begin(), so whichever merges second needs that conflict fixed on its work branch first.
6. When a PR merges or is declined: follow link Session Handoff [rv3028-eeprom]: Three MeshCore PRs upstream; watch and respond to review #17's step 4. Close the fork iteration PR, delete its branches, close the upstream issue by hand (Fixes does not auto-close on dev), and comment on the fork issue.
7. Cleanup once Pieter is done with hardware: link Session Handoff [rv3028-eeprom]: Three MeshCore PRs upstream; watch and respond to review #17's step 5. That covers the local/rv3028-probe-* and local/rtc-probe-test worktrees, this track's detached hub worktree, and optionally reflashing the second board.

External blockers

• MeshCore maintainers: review and merge of Adopt an RTC only if its time registers read like that chip 🤖🤖 meshcore-dev/MeshCore#3544, Store RV3028 backup switchover and trickle charger config in EEPROM 🤖🤖 meshcore-dev/MeshCore#3545, Write and read the RX8130CE WEEK register as one-hot 🤖🤖 meshcore-dev/MeshCore#3550, and the older Read and write the RV3028 clock atomically, and reject switchover-garbled transfers 🤖🤖 meshcore-dev/MeshCore#3421, Do not feed an invalid GPS time into the RTC 🤖🤖 meshcore-dev/MeshCore#3422, Make GPS status reporting say what was actually checked 🤖🤖 meshcore-dev/MeshCore#3423, Treat the GPS enable pin as a pin number, not a truth value 🤖🤖 meshcore-dev/MeshCore#3425 (draft), Reject a negative sensor list start index 🤖🤖 meshcore-dev/MeshCore#3433 and Pass TELEM_RAK12500_ADDRESS to the RAK12500 I2C probe 🤖🤖 meshcore-dev/MeshCore#3434. Nothing is owed from our side on any of them as of 2026-10-04.
• Copilot re-review upstream needs Pieter to re-request it in the web UI after any push. --add-reviewer @copilot is refused for this account, and a force-push does not trigger it.

Internal dependencies

• Adopt an RTC only if its time registers read like that chip 🤖🤖 meshcore-dev/MeshCore#3544 and Store RV3028 backup switchover and trickle charger config in EEPROM 🤖🤖 meshcore-dev/MeshCore#3545 touch the same begin() block, so whichever merges second takes the conflict fix.
• Fork PRs Store RV3028 backup switchover and trickle charger config in EEPROM (iteration branch) #11, Adopt an RTC only if its time registers read like that chip (iteration branch) #14 and Write and read the RX8130CE WEEK register as one-hot (iteration branch) #16 (and Bound the sensor list start index and its paging reserve (iteration branch) #8 and Pass TELEM_RAK12500_ADDRESS to the RAK12500 I2C probe (iteration branch) #9 for the older pair) stay open until their upstream PR resolves, so drift is absorbed there first.

State

• Older upstream PRs, all open, no reply owed:
  • Read and write the RV3028 clock atomically, and reject switchover-garbled transfers 🤖🤖 meshcore-dev/MeshCore#3421, Do not feed an invalid GPS time into the RTC 🤖🤖 meshcore-dev/MeshCore#3422, Reject a negative sensor list start index 🤖🤖 meshcore-dev/MeshCore#3433: approval recommended.
  • Make GPS status reporting say what was actually checked 🤖🤖 meshcore-dev/MeshCore#3423: Copilot's last "needs a closer look" on the gps reply-format change opened no thread, and the body's "Note on the format change" covers it.
  • Treat the GPS enable pin as a pin number, not a truth value 🤖🤖 meshcore-dev/MeshCore#3425: draft, tracked by issue GPS enable pin is stored unsigned and assigned from a truth test, so PIN_GPS_EN=-1 becomes a bogus pin and pin 0 is rejected meshcore-dev/MeshCore#3426.
  • Pass TELEM_RAK12500_ADDRESS to the RAK12500 I2C probe 🤖🤖 meshcore-dev/MeshCore#3434: answered, see step 2.
• Upstream dev is still 3e3150c8, and the fork's dev mirror matches it.
• Worktrees: re-derive with git worktree list. The primary checkout is clean on main.
• Hardware: unchanged from link Session Handoff [rv3028-eeprom]: Three MeshCore PRs upstream; watch and respond to review #17. The second RAK4631 runs the Store RV3028 backup switchover and trickle charger config in EEPROM 🤖🤖 meshcore-dev/MeshCore#3545 diagnostic build, and the ZephCore session owns the first board.

The parked decision queue

0 parked: the fork has no open issue labelled decision. One question went to Pieter at close: whether to add the upstream ZephCore numbers (meshcore-dev#98, meshcore-dev#99) beside the fork links in meshcore-dev#3544's and meshcore-dev#3545's bodies. The fork links stay either way, because the raw logs and tested heads live there. The answer is commented on this link.

What the last round did

• Swept all nine upstream PRs and four issues. Only Pass TELEM_RAK12500_ADDRESS to the RAK12500 I2C probe 🤖🤖 meshcore-dev/MeshCore#3434 owed a reply, a human question unanswered since 2026-09-17, and its Pieter-approved answer is posted.
• Took the ZephCore session's report of Added support for LilyGo TBeam meshcore-dev/MeshCore#98 and Feature Request: Graph Traversal & Signal Quality Mapping for MeshCore Networks meshcore-dev/MeshCore#99. Neither item it flagged needs mirroring:
  • MeshCore's rtcCheck() already rules out an all-0xFF read on every read, and Adopt an RTC only if its time registers read like that chip 🤖🤖 meshcore-dev/MeshCore#3544's body records the asymmetry.
  • The failed-read rule is unchanged, and the ZephCore session was told so.

What not to repeat

• Don't sweep only this track's three PRs. The Pass TELEM_RAK12500_ADDRESS to the RAK12500 I2C probe 🤖🤖 meshcore-dev/MeshCore#3434 question sat for 17 days because each round watched only the handoff's own list. Step 1's author-wide listing is the fix.
• Everything in link Session Handoff [rv3028-eeprom]: Three MeshCore PRs upstream; watch and respond to review #17's "What not to repeat" still holds: Copilot triggering, no fork #N in upstream text, no "the maintainer" meaning Pieter, and no output-suppressing tails on writes.

New learnings

• Sweep scope and the Pass TELEM_RAK12500_ADDRESS to the RAK12500 I2C probe 🤖🤖 meshcore-dev/MeshCore#3434 status are in project memory meshcore-upstream-pr-state-2026-09-16 (update 2026-10-04).