Repository navigation
lanes: a read-only command that lists a repo's lanes and the next free position; --all for every repo; lane-start <repo> without a number lists and suggests (Brett Heap's rulings 2026-09-13) #28
Description
Lane: openrepotools-3 (openRepoTools-3)
Brett Heap's rulings, in session, 2026-09-13, interactive multi-choice, verbatim: command name "lanes (Recommended)"; default scope "This repo's lanes + next free (Recommended)"; lane-start <repo> with no number "Yes, list and suggest (Recommended)". Asked after his question: "are we going to have a openRepoLanes command that will show the lanes that are available for the current repo? or option --all that will show all repos and their lanes?"
The gap. The register (lanes/LANES.md in the workspace repository) is one appended file and every read of it today asks about ONE thing: lanes-edit.sh who --lane <lane>, who --landing owner/repo, live-holder <lane>; lane-start <repo> <n> refuses without the number. Nothing lists the lanes of a repo, so a person has to already know a lane's name to enter it — which is how openRepoTools-3 ran three days with no row.
lanes — the reader that pairs with the writer lanes-edit.sh, a bare name beside park, resume, status, lane-start, lane-end; installed by openRepoTools --install once #24 lands; READS AND CHANGES NOTHING.
lanesinside a repo checkout: that repo's lanes, one line each — lane name · state · window and workstation holding it · last session id (the row's LAST id, Amendment 6) · objects held (from the lane's log, aswho --lane) · handoff path — then the NEXT FREE position and the exactlane-start <repo> <n>line to take it.lanes --all: every repo under the projects root ($PROJECTS_ROOT, default~/projects), one block per repo.lanes <repo>: one named repo from anywhere.- States, three: LIVE here — a live session on this workstation holds it (
live-holder), the line names its window and profile; PARKED — the row exists, nothing live holds it, the line gives the resume command; UNKNOWN — the row's workstation (Rule 10 column) is another machine whose session records this one cannot read; the line says to ask@<lane>or read its handoff (the wordingwhoalready uses). - A lane's REPO is its object log's home line (
STARTED/RESUMED … → home owner/repo, Amendment 7), resolved throughrepos.tsv; the name's<repo>prefix is a label (Amendment 7). Where a lane's home and its prefix disagree, the line says so rather than filing it twice or once wrongly. - Reads the PUBLISHED register after a fetch (
origin/main, R19) like every other read in the tooling; the header says "as of ". A register that cannot be fetched is said, not guessed (falls back to the working tree and says so). - Exit 0 with a listing; exit 1 when the projects root has no register configured (
workspace.yaml), namingopenRepoTools wip init.
Two hooks into existing surfaces
lane-start <repo>with no<n>prints that repo's block and the next free position, then the exactlane-start <repo> <n>line — and still launches nothing without the number (ruling 3).- The name guard's offer (Amendment 12 clause (h), Name guard (lane protocol Amendment 12): lanes-edit.sh guard as a UserPromptSubmit hook, lane-start --name on every launch, --install places the hook — after #24 lands #25): when a person renames a session to a lane name that has no row, the offer shows that repo's block before asking
yes/no, so the person sees what already exists there.
Tests in tests/test_lane_helpers.sh (the sandbox already fakes tmux, the session records and the register): a repo with two lanes (one LIVE via a real sleep pid, one PARKED) and its next free position; --all grouping by HOME with a name whose prefix disagrees with its home; UNKNOWN for a row on the other workstation; an empty repo (no lanes, next free = 1); no register; lane-start <repo> with no number printing the list and exit 2 as an argument error still (or 0? — decide with the guard's tests; today it is exit 2).
Sequencing. Waits for #24 (the tooling's move here). Built by lane openRepoTools-3 together with #25's guard, since the guard's offer reuses the listing.
🤖 Generated with Claude Code
Activity
brettheap commentedon Sep 13, 2026
NARROWED. Amendment 11 clause (j) (brettheap/new-workstation#21, ratified by Brett Heap 2026-09-13T18:05:29Z, verbatim "Yes, build
lanesin A11") already defineslanes: every lane the register and the logs know, newest write first, ten columns (name, state, workstation, profile, window, last transcript uuid, directory, held objects, age of last write, restart line), read-only, local by default with--fetch,--repo <origin>/--dir <path>narrowing; installed by--install; built by lane openRepoProject-1 in #26 on #24's branch. This issue does NOT buildlanes. It holds the two deltas Brett ruled ~19:5xZ and the one conflict, for reconciliation with (j): (1) a NEXT FREE POSITION line ending the per-repo view, with the exactlane-start <repo> <n>to type; (2)lane-start <repo>with no number lists and suggests instead of refusing; (3) CONFLICT: barelanesinside a repo = "This repo's lanes + next free" (his 19:5xZ word) vs (j)'s every-lane default (his 18:05Z word) — put to him. Timings measured the same hour for the store question: register parse 3 ms; tooling reads 1.7–2.2 s (a git fetch of ~0.85 s each, then bash); one write 1.2 s; rows up to 96 KB.Lane: openrepotools-3 (openRepoTools-3)
brettheap commentedon Sep 14, 2026
Lane: openrepotools-3 (openRepoTools-3)
Superseded in part by Amendment 18 DRAFT clause (i) (brettheap/new-workstation#35, revision 3; tooling #38 act 8), on Brett Heap's answers of 2026-09-14: the DEFAULT listing becomes the lanes AVAILABLE TO BIND (free or provably dead binding) rather than all of this repository's lanes;
--boundadds the bound ones with their binding;--allkeeps its every-repository meaning. The 2026-09-13 rulings here (namelanes; narrow inside a checkout;lane-start <repo>with no number lists and suggests) stand; the new clause adds the question barelane-startasks on a terminal. The next-free line and thelane-start <repo> <n>lines stay aslanesprints them since #26.brettheap commentedon Oct 13, 2026
CLAIMED — lane openRepoTools-5, session 0d83dca2-c18c-4bb0-99ce-896e6cafea02@Eagle, 2026-10-13T18:23:26Z, for #28
Logged in
lanes/log/openRepoTools-5.mdat40df63b386c1e09298ae38c01e286dcd688bc88a.The three reads (lane-collision-protocol Rule 1):
brettheap commentedon Oct 13, 2026
Lane openRepoTools-5: closing #28 as fixed, checked at main f87fd5a. Brett Heap's word, recorded in lane 5's register log as a RULED line at 2026-10-11T16:5xZ, verbatim: "Close all five (Recommended)".
Fixed by #26 (63a74af, merged 2026-09-14). Both deltas this issue holds are on main:
lane-start <repo> <n>to type:lanes:702-780printsnext free position:andtake it: lane-start <repo> <n>(the position itself islanes-edit.sh next-free's answer sincelane: one word lists the lanes and binds the one you pick — available goes through the launcher in its own directory and profile, LIVE HERE attaches and never starts a second process, bound elsewhere names where; andrestartleaves the person's PATH in the same breath (Amendment 18 Addendum 1 and Addendum 2) #45).lane-start <repo>with no number lists and suggests instead of refusing:lane-start:858-886, asserted attests/test_lane_helpers.sh:6189-6192.lanesshows inside a repo) was settled by Amendment 11 Addendum 4 ruling 6, "Narrow inside a checkout" (Brett Heap, 2026-09-13):tests/test_lane_helpers.sh:6131-6138,README.md:145-148.laneword islane: one word lists the lanes and binds the one you pick — available goes through the launcher in its own directory and profile, LIVE HERE attaches and never starts a second process, bound elsewhere names where; andrestartleaves the person's PATH in the same breath (Amendment 18 Addendum 1 and Addendum 2) #45 (f98d734).One thing was never built: the original body's hook 2, the name guard's offer showing the repo's lane block before it asks. Nothing in the guard renders it at main, and this issue's own narrowing comment dropped it (it lists only the two deltas and the conflict). It can be refiled if wanted.