[svanlabs]

What generated this?

Generated-by: claude-code/claude-sonnet-5-5

Motivation

Closes #968.

report_unsettled names a saved open hand "lost" whenever the resync snapshot has moved past it, without asking the store. Three false warnings in 35 hours on build 7a1494c (2026-10-10 11:00 to 2026-10-11 22:12 UTC), each for a hand whose row is in the hands table:

• SvanBotV10 at 11:00:04, the hot swap: hand 08305a02 was stored at 10:59:57 by the process that was swapped out.
• SurSvan at 13:34:43 and Svanism at 14:03:59, at reconnects hours after the swap: hands 1e8862bf and 7abdd813 were stored at 11:00:23 and 11:00:39 by the live path, which left their saved starts in shared.resumable (resume_open_hand returns early when the tracker already knows the hand).

The warning exists so that a row lost at a hot swap is not silent (0315); a false one at every swap hides a true one, and the next release is a swap.

Solution

One check in the one reporter, in crates/apps/bot/src/client/recover.rs: after the entry is dropped, report_unsettled asks shared.store.hand and returns without a warning on Ok(Some(_)). A store that cannot answer still warns, because an unreadable store says nothing about the row. It covers all three causes.

Test: a_saved_hand_the_store_already_holds_is_not_reported_lost stores the saved hand's row, then resyncs onto another hand; it failed on origin/main with "the row is stored: nothing is lost" before the fix. The existing a_saved_hand_the_replay_cannot_settle_is_reported_lost_once still passes: with no stored row the warning stays.

Log-only change off the decision path, so no paired simulation and no golden change. Not released: the running build is 7a1494c.

Checklist

• scripts/check.sh full passes on this exact commit (e842e1b), run in a worktree on the shared warm target/dev:

test.py: 897 passed, 0 failed, 4 ignored in 35 binaries; build 0.2s, run 4.4s (6 jobs x 2 threads, profile gate, peak test RSS 195 MB)
== ok (full, 154 s)
check.sh full: exit 0

• This is one change.
• Not a behaviour change on the decision path: no paired simulation; the golden snapshot is untouched.
• Not a performance change.
• No document names the changed behaviour.
• The AI-provenance rule was followed: Generated-by: is in the commit footer and in the section above.

🤖 Generated with Claude Code