Repository navigation
The cross-profile transcript-holder scan tests only kill -0, so a recycled pid refuses a lane the estate's one liveness contract would let through #56
Description
Found by Copilot round 1 on #47 (lane-start:1725), on the live_holders_of_id this estate gained with Amendment 17 adoption act 6 (the transcript follows the lane). The read is fail-closed, so the cost is a refusal and never a corruption — which is why it is filed rather than held back.
The two contracts
lanes-edit.sh has ONE liveness contract and states it in as many words (record_is_live, lanes-edit.sh:2469): a record is live when its status is not ended, its pid answers kill -0, and /proc/<pid>/stat field 22 still equals the procStart the record wrote down — the third so that a RECYCLED PID cannot hold a name.
lane-start's new scan needs a different question answered — "is any live process, in ANY profile, holding this transcript id" (Amendment 18(h)) — and live-holder answers per LANE, so the scan was written beside it:
case "$lho_blob" in *"\"sessionId\":\"$lho_id\""*) : ;; *) continue ;; esac
lho_pid="$(printf '%s' "$lho_blob" | grep -o '"pid":[0-9][0-9]*' | head -n1 | sed 's/.*://')"
[ -n "$lho_pid" ] || continue
kill -0 "$lho_pid" 2>/dev/null || continueOne of the three tests, not three. An ENDED record whose pid has been reused, or a pid recycled onto anything at all, reads as a live holder — and the lane is then refused with "a LIVE process there holds it" and the retire act, about a process that is not that session.
The act
Not a fourth copy of the contract. lanes-edit.sh is where it lives, so the clean shape is a read of its own — lanes-edit.sh holders-of <transcript-id>, printing <pid>\t<where>\t<profile> per live holder across every profile's sessions/, built on record_is_live — with lane-start's live_holders_of_id becoming a call to it and its $HOME/.claude-profiles walk (and $CLAUDE_PROFILES_HOME override) going with it, since lanes-edit.sh already walks exactly those paths for session_files.
Cases in tests/test_lane_helpers.sh: an ENDED record whose pid is alive is not a holder; a record whose procStart does not match the pid's is not a holder; and the live one still is.
Lane: openRepoTools-3.
Activity
brettheap commentedon Sep 14, 2026
CLAIMED — lane openRepoTools-3, session 4135b2c9-34a6-47e3-a156-5adebb632fae@Eagle, 2026-09-14T21:31:24Z, for #56
Logged in
lanes/log/openRepoTools-3.mdat8e14335da3c2a43f20a755ac0a9779a826da545d.The three reads (lane-collision-protocol Rule 1):
A pane is never respawned over a record that is only part written, a …
Amendment 17 tooling: a handoff is the swap under one name — `lane-ha…
brettheap commentedon Oct 11, 2026
TAKEOVERopenRepoTools-3 — lane openRepoTools-5, session 0d83dca2-c18c-4bb0-99ce-896e6cafea02@Eagle, 2026-10-11T16:52:26Z, for #56
Logged in
lanes/log/openRepoTools-5.mdatdeb3918f1ec1aa20753bb8f274e38fe455cc5329(takeover of the stale claim held by lane openRepoTools-3).The three reads (lane-collision-protocol Rule 1):
lane-start's transcript follow counts a holder only when transcript-h…