[brettheap]

Lane: openreposhape-2 (openRepoShape-2)
Claim: opensoft/openRepoTools#198 (comment)

Re-homed from opensoft/openRepoTools#316

• Old head: 77e94b9 on opensoft/openRepoTools feat/198-a23-slice-2b-start-handoff, the archived original. That branch is not pushed again.
• New head: 4031100 on opensoft/openRepoTools-code feat/198-a23-slice-2b-start-handoff. The old head's rebuild is 31990c4; this round's commits follow it, and the last two are merges of slice 2a' (a44b9d2, then 4031100 with A61).
• Why: Brett Heap's answer to Q28 (Lane writers commit first and sync second, so a pull that refuses or conflicts strands the commit and later writes stack on it: fast-forward before the edit, and refuse before committing (fail closed) openRepoTools#198, comment 6722244541; Amendment 22 act 2: the SQL register under lanes-edit.sh — mode=primary writers in one transaction each, the estate switch and lanes/STORE, the write journal and the renderer (one mirror commit per write), exit 69, fail-closed readers, lanes-index import/rollback, SERIALIZABLE claim order, and the agreement tests openRepoTools#165). Act 2's six drafts move to the code leg in their stack order, by cherry-pick and not by merge (Triad migration (feature 004-migrate-to-triad): execution tracking under Brett Heap's word of 2026-10-07 openRepoTools#186, comment 6548291952 S18).
• Method: each old commit's tree was cut to the code leg's 26 owned top-level paths (the cutover plan's leg: code entries), and the commit was made again on its rebuilt parent. Author, dates and message are kept, with a (cherry picked from commit …) line added. Old main's f18b818, which the old stack stood on, is the code leg's 0cefa1f.
• Range-diff of the old slice against the rebuild: all 7 pairs are !, and each differs only by the cherry-pick line. No hunk of code differs, and this slice changed no spec-leg path, so nothing was dropped.

Receipt: the 7 commits, old opensoft/openRepoTools sha to code-leg sha

Each subject ends with the slice's reference to the act 2 issue, left out here.

#	old (opensoft/openRepoTools)	new (opensoft/openRepoTools-code)	subject
1	f1b19e9	068f190	lane-start writes the Rule 3 stamp with one write-handoff --stamp through the lane's checkout, fast-forwards W before and after its writes, and is refused after its first write only by a restart's own fences
2	84633df	ce73c92	lane-start reads each restart-intent field to the end of the answer, so a launch no longer dies with 141 when an early-exiting awk closes the pipe under printf
3	a1f658e	b848bd8	lane-handoff writes the PAUSED top block with one write-handoff --prepend through the file handoff-path names, so its handoff is committed and pushed under the checkout's lock and never by its own pull and push in W
4	37b8806	3ce57f6	The supervised-restart cases run /ctx through the lane's own checkout, and a stamp that could not be published, or that the checkout refused, launches with the bound file at its pre-image digest
5	c8ac6ca	dbfbec6	lane-start launches with no first prompt where the lane's own copy of its handoff is not there, naming that file and the command that makes the checkout, and never reads W's copy in its place
6	127ac75	dc50eb5	lane-start and lane-handoff say, under LANES_NO_GIT=1, that the handoff is written and nothing is committed, where their exit-0 words claimed a commit and a push
7	77e94b9	31990c4	lane-start says, where the lane's checkout refused a plain launch's stamp, that the first prompt is the checkout's copy and origin's differs, naming both copies and the command that brings the checkout up to date

Stacked on

Slice 2a' on the code leg, feat/198-a23-slice-2a-handoff-verbs at 56a9b87: 2a's round head 1320847 (old 7bc572c, A60) with slice 1b at 34799b7, which brings slice 1a's merge of the code leg's main (577110e), merged as 2b1a232, and A61 on top of it (2a's decision 32: the fix for the two cases CI failed at 2b1a232). This branch merged 1320847 at e12b369, 2a' at 2b1a232 as a44b9d2, and 2a' at 56a9b87 as 4031100, all clean. This PR's own diff is git diff 56a9b87 HEAD: lane-start, lane-handoff and their tests.

Why one PR for both scripts (the plan's §3, slice 2b): /ctx's round trip compares paths with -ef and by digest. lane-handoff binds the restart intent, and lane-start checks it. If either script moved alone, every /ctx under the opt-in would refuse as "different handoffs".

This round: lane openRepoTools-3's review of opensoft/openRepoTools#316 at 37b8806, and its review of opensoft/openRepoTools#311

Lane openRepoTools-3's review of opensoft/openRepoTools#311 (comment 6376170218) touched this slice in three places (982, question 6, 990), and its review of opensoft/openRepoTools#316 said LAND AFTER FIXES.

Each fix has a case that was red before it, or, where the case pins what already held, a mutant the case kills; each has its own commit. The docstring change (question 3) needed neither. Commits before 31990c4 were made on the old root and are named by their old shas.

item	what it found	answered in	decision
727, question 2	5d fell back to W's copy, against S2's "never through W", and kept 723 open	old c8ac6ca (dbfbec6)	24
721, 728, question 4, rv2b SHOULD 1	the exit-0 words claimed a commit under LANES_NO_GIT=1	old 127ac75 (dc50eb5)	25
725	a refused plain stamp delivered the checkout's older block and did not say so	old 77e94b9 (31990c4)	26
726	the printed re-run of a refused stamp carries none of the restart's fences	e97ad3f	27
729	the --restart respawn dropped LANES_WORKSPACE_ROOT, LANES_REPO and LANES_FILE	67e6da5	28
730, 731	the duplicate-row refusal sent a merge to W under the opt-in; retired_word's pipe was missing from "Not changed"	2bd3ea2	29
question 3	the t018 docstrings still said "byte for byte"	fb4a7b0	30
opensoft/openRepoTools#303's fit (c); 982, question 6 and 990 of the review of opensoft/openRepoTools#311	the register's PAUSED line said handoff refreshed whatever happened; the 6 arm said "kept on HEAD"	c1f2928, 439699f	31
rv2b's NITs 1, 2, 3, 5	5d read a path a peer had moved; composite asserts; two copies of each helper; WORKBENCHES_ not scrubbed	2e4d7dd	32
733 (in part), the battery's survivors	26 of m2b's 59 mutants survived the bench	e5c3326	33
merge	2a's round head, which 723, 722 and 724 rest on	e12b369
MUST 723	a sibling lane's cell was stamped, written and delivered	58309ba	34
724	after 2a's 603, a changed pre-image would read as lost ownership	615cfdd	35
722, and 721's last case	an 8(d) handoff at both call sites	5bf9bbb	36
733 (M1, M2, M4)	5c's two restart dies and a why with a tab had no case	c12f078	37
merge	2a' (2a with slices 1b and 1a and the code leg's main)	a44b9d2
2a's A61	CI's git 2.56.0 read the clone that 2a's other-owner case makes unreadable, so two cases of tests/test_a23_write_handoff.py failed on #18 at 2b1a232 (2a's decision 32), and this slice carries the same file	merge 4031100

Decisions

1. Routing stays opt-in. 2b makes the calls, and the verbs route themselves.
   • Without LANES_ROUTE_WRITES=1, write-handoff writes W under W's lock, with S3 and S4.
   • So lane-start's and lane-handoff's unlocked W commits are gone in both modes. Those are lane openRepoTools-3's finding 282 on Amendment 23 slice 1b: every writer fetches and fast-forwards under its lock or refuses with nothing written, pushes with no pull and reads origin back, a refused push replays the one edit, and commit with nothing staged is the publish act (#198) openRepoTools#307, its first two writers.
2. The refreshes run only under the opt-in with the default resolution, and never in a dry run (which prints a PLAN line for each).
   • Without the opt-in, W is the write checkout, and every write's S3 fast-forwards it under its lock. A refresh outside that lock could race a writer there.
   • An explicit LANES_REPO or LANES_FILE is the caller's register (the seam), so W is not what is read.
   • Every exit of refresh-workspace is a note.
3. rows_for_lane reads W after the first refresh (the plan's §2.7).
   • It reads W's working tree where the refresh exited 0. Otherwise it reads origin/main's register as W's tracking ref has it, and falls back to the working tree, with a note, where W has no such ref. No temporary file is used, since exec skips the EXIT trap.
   • This replaces slice 1a's lane_register_follow in lane-start. lane-end keeps its own follow (slice 4's).
4. The lane's checkout is made at "5. the row" by lane-checkout <lane> --create, under the opt-in with the default resolution, and never in a dry run. A refusal is a note, and the launch goes on (1a's SHOULD 2). After the row, the writable handoff path that handoff-path names is printed beside the row's spelling.
5. 5c is one LANES_LANE=<lane> write-handoff <lane> --stamp <line> --subject <subject>, with no LANES_NO_FETCH. Under an active restart it adds --restart <op> <gen> <attempt> starting <transcript> --expected-digest <digest>.
   • Exits 0, 3 and 6 mean the file was written, and ri_expected_digest becomes the digest on stdout.
   • Exit-code notes: 3 says "look before writing it again". 6 said "kept on HEAD" (superseded by 31), and names the publish act. Any other exit says "written and was NOT published". None promises a carry (S3).
6. hs_git_net and hs_carry_note are deleted, as the plan says. That settles lane openRepoTools-3's request (282) that hs_carry_note's middle branch name the publish act: there is no carry note left to reword, and no commit of lane-start's outside a lock.
7. The reads follow the write. (Its fallback to W's copy is superseded by 24.) The restart binding and 5d read the file that handoff-path names, which is the one write-handoff writes.
   • Under an active restart this is strict.
   • For any other launch, a path that is not a file falls back to hs_resolve (W's copy), with a note. A checkout that could not be made then does not cost the first prompt.
8. "Never refuses the launch" is made true: after the first write, only the restart fences die.
   • With no stdout line, a 7 dies ("lost ownership of its handoff before the resume stamp").
   • With no stdout line and the bound file no longer at the bound digest, the launch dies with the old words ("the restart handoff changed before its owned resume stamp").
   • Any other exit with no stdout line is a note that prints the stamp line and the exact re-run. 5d's own -ef and digest fences still stand behind it.
   • So a stamp refused under a restart launches with the unstamped, bound file, and the intent goes pending → starting at the pre-image digest (Lane writers commit first and sync second, so a pull that refuses or conflicts strands the commit and later writes stack on it: fast-forward before the edit, and refuse before committing (fail closed) openRepoTools#198 A2).
9. The stale-checkout recoveries name no pull and no bare push.
   • Where the register read is W, they name lanes-edit.sh refresh-workspace. Where it is a caller's checkout, they name git -C <checkout> fetch origin main && git -C <checkout> merge --ff-only origin/main.
   • "Push the register from here" becomes "publish the register from here, then re-run: LANES_LANE=<lane> lanes-edit.sh commit "<what it is>"", the publish act, which also carries a pending hand edit.
10. request-handoff gains LANES_LANE in its env. The printed step line is unchanged.
11. lane-handoff's file is handoff-path <lane> --create, run with its stderr shown and without LANES_NO_FETCH, so the cell is read from the freshest origin/main.
    • 8 is "no handoff named": the old note, with the block printed.
    • Any other exit is a note naming it, with the block printed.
    • It does not check first that the file exists: write-handoff's S3 fast-forwards a stale checkout, and its 8 means there is nothing to write into.
12. One LANES_LANE=<lane> write-handoff <lane> --prepend <block> --subject "handoff(<lane>@<ws>): PAUSED, <why>", with [--restart <op> <gen> 0 preparing none] under a restart, no LANES_NO_FETCH and no --expected-digest, replaces publish-handoff and the commit block.
    • The block file is top_block's output alone. write-handoff adds --- and a blank line, the same bytes as before.
    • handoff_written=1 exactly where stdout carries the path<TAB>digest line: exits 0, 3, 6, and S3's 2, which writes and does not commit. handoff_file becomes that path, so the pending intent and its read-back bind the file that was written.
    • Each exit gets a note in place of the commit block's two step lines. The helper's own words and git's are printed, no longer sent to /dev/null.
13. append-line gains LANES_LANE. The skill test's append-line ("PAUSED — lane …") regex still finds the one call.
14. A --restart respawn line carries LANES_ROUTE_WRITES when it is set, so the supervisor and the lane-start it runs resolve the checkout the intent bound. Without it, an opted-in /ctx binds the lane's checkout, and its respawn reads W and refuses with "different handoffs". The new round-trip case shows that refusal.
15. hs_resolve's case (c) is no longer a stamp target. The stamp goes only where write-handoff resolves the row's cell, in the checkout writes go through (S5: never W). The shell suite's CN-2 becomes "not stamped, the line printed".
16. handoff-path after the row also runs in a dry run, with LANES_NO_FETCH=1 there. An active restart's --no-launch is a dry run, and 5d's strict branch needs the bound file in it. Elsewhere it fetches, so a new lane's first start still finds the row it pushed from its checkout before W's ref moved.
17. The restart binding's handoff-path runs with LANES_NO_FETCH=1, like every other restart read there. It keeps lane-start's two refusals word for word: "the restart handoff the row names is missing" and "the restart intent and register name different handoffs".
18. Under a restart, the preparing bind names the file handoff-path gave, at handoff_digest of it now (none where it is absent), before the write.
    • A file that is not there is then write-handoff's 8: the block is printed, and the gate refuses with "the HANDOFF WAS NOT REFRESHED", as before (the shell suite's repoHF-12).
19. The PAUSED subject spells every control character of the why as a space (subject_text, LC_ALL=C). write-handoff refuses a subject that carries one (64, nothing written). The door already refuses a newline and a carriage return, so this only keeps a why with a tab refreshing the handoff, as the old git commit -m did.
20. lane-start reads the restart intent's answer to its end (intent_field), for eri_state, eri_op and the twelve ri_* fields.
    • CI run 39806108007 at f1b19e9 had test_restart_missing_checksum_is_not_launch_authority exit 141 (SIGPIPE): printf … | awk '… {print $2; exit}' under pipefail lost a race against a writer that was not done.
    • intent_field keeps the FIRST value, as the early exit did.
    • This is lane-start: a reader that stops early, under set -euo pipefail, can end lane-start with no message, or take the minted-transcript branch for a session it has recorded (the shape of #280) openRepoTools#281's class at the restart reads lane openRepoTools-3 named. The rest of lane-start: a reader that stops early, under set -euo pipefail, can end lane-start with no message, or take the minted-transcript branch for a session it has recorded (the shape of #280) openRepoTools#281 is listed under "Not changed".
21. The atomic-stamp case is rewritten, not deleted. test_restart_atomic_stamp_failure_preserves_original_for_retry asserted exit 2. Now a stamp that could not be published atomically is a note (decision 8): the launch exits 0, the bound file is unchanged, and the intent is starting at the pre-image digest. The retry stamps it once.
22. The t018 single-repository goldens move exactly where lane-handoff's commit act printed (t018-handoff/single-repository-decoys.golden:34-36,108-110 and t018/single-repository.golden:142-144).
    • handoff commit: <subject> and handoff pushed to origin/main become write-handoff's own lanes-edit: committed: <subject> and lanes-edit: pushed origin/main (attempt 1), before handoff refreshed: …. After it comes lane-handoff's line naming the exit it read.
    • The subject carries the why, as decision 12 sends it. The old printed subject left it out.
    • Nothing else in either golden moves. Both were re-captured from this branch's tools by each module's own --capture, normalised as before.
23. Hygiene covers both scripts. _NO_PULL_FILES gains lane-start and lane-handoff. The new test_no_lane_script_promises_the_carry[lane-start|lane-handoff] reads both scripts. At the base, lane-handoff had a carry promise ("captured by the next lane's commit") and two pull --rebase lines.
24. No first prompt where the lane's own copy is not there (727, question 2; supersedes 7's fallback). 5d reads only the file handoff-path names, for every launch. Where it is missing, the launch goes on with no first prompt, and one note names the missing file and, under the opt-in, LANES_ROUTE_WRITES=1 lanes-edit.sh lane-checkout <lane> --create. W's copy is never read in its place. hs_resolve and hs_workspace_root lost their last caller and are deleted, which retires 7's fallback and 15's case (c).
25. The exit-0 words say what was done (721, 728, question 4, rv2b SHOULD 1). Under LANES_NO_GIT=1, lane-start says "Rule 3's stamp is written in and NOT committed: LANES_NO_GIT=1 commits and pushes nothing", and lane-handoff says "the block is written in and nothing is committed or pushed". 36 words every other 0.
26. A refused plain stamp names a checkout copy that is behind origin (725). Where 5c wrote nothing on a plain launch, 5d compares the checkout's file with its own origin/<branch>, which S3 has just fetched (blob ids). Where they differ, one note names both copies and git -C <checkout> merge --ff-only origin/<branch>. The launch goes on with the checkout's copy, the file this lane's writes go to.
27. The printed re-run waits for ready under a restart (726). Where the stamp could not be written, lane-start prints the exact write-handoff --stamp re-run. Under an active restart a second line says to run it only once the restart is confirmed ready, since that write carries none of the restart's fences and exits 7 while the intent is starting. A plain launch's words are unchanged.
28. The respawn carries the register's seams (729). Beside LANES_ROUTE_WRITES, the --restart respawn line carries LANES_WORKSPACE_ROOT, LANES_REPO and LANES_FILE where each is set, quoted as the others are. lane-handoff never sets one of its own. Without them the new pane resolved the default W, and a /ctx with a register of its own failed closed. This predates 2b.
29. Under the opt-in a duplicate row is never merged in W (730). The rows are read from W, and the refusal said to merge them "by hand in" W's register and commit with lanes-edit.sh commit, which under the opt-in commits in the lane's own checkout. Now, under the opt-in:
    • where origin/<branch>, as W last fetched it, has fewer than two such rows, the duplicate is only in W's working tree, and the refusal names it a stray in W, which a person takes out;
    • where origin carries the duplicate, the merge is a hand edit of the lane's own checkout's register, published from there with LANES_LANE=<lane> LANES_ROUTE_WRITES=1 lanes-edit.sh commit.
    • The second arm is not reached with this helper, as lane openRepoTools-3's review says ("reachable only through an uncommitted duplicate in W"): canon-lane reads the published register after W is fast-forwarded and refuses a published duplicate first (Amendment 15(d)), and lane-dir and the managed-ledger fence refuse it too. A case pins that, and the code's comment says so. Its mutant (I6c) is not pinned.
30. The t018 docstrings name slice 2b's one exception (question 3). Both module docstrings and both capture() docstrings say which lines moved (the commit act's three hunks, decision 22), why, and that every other line is c45a452's. The case names are kept.
31. The register's PAUSED line says what the write did, in the skill's words (Most reads of a lane's log cannot tell a failed read from a lane with no log: lane-dir, lane-profile, lane-agent, lane-transcript, history, lane-last, binding and lane-trees answer 8 ("none") over a log git could not read (from #290) openRepoTools#303's fit (c); question 6, 982 and 990 of the review of Amendment 23 act 2, slice 3: the handoff skill locates its file with handoff-path and writes it with ONE write-handoff --whole call from the lane's own checkout, so it spells, commits in, pulls and pushes no checkout of its own (#198) openRepoTools#311; slice 3's F23 and F24). lane-handoff's 4(b) line keys on what write-handoff did:
    • handoff refreshed on a 0, LANES_NO_GIT=1 included: it claims no commit, and the step line says what was done (25, 36);
    • handoff written, NOT published (write-handoff exited <n>) for a 2 or a 6 that wrote;
    • handoff written, its publication NOT confirmed (write-handoff exited <n>) for a 3, and for any other exit after a write, which claims nothing either way;
    • handoff NOT refreshed (<why>) where nothing was written: no handoff found, handoff-path failed, write-handoff not called, or it exited n and wrote nothing.
    • Exit 6 is not one state (982, which names lane-handoff:2689). git add or commit failed and nothing is committed, or the commit is kept on HEAD and named; and since 2a's 604, an 8(d) handoff's commit or push in its own repository failed. Both scripts' 6 words now say the file is written and NOT published, point at the helper's words for whether and where a commit was made, and name the publish act only for a commit kept in the lane's register checkout. This supersedes 5's "kept on HEAD".
    • The shell suite's RV-B2 (a) assertion moves with lane-start's words (CI only). No reader parses the suffix. Most reads of a lane's log cannot tell a failed read from a lane with no log: lane-dir, lane-profile, lane-agent, lane-transcript, history, lane-last, binding and lane-trees answer 8 ("none") over a log git could not read (from #290) openRepoTools#303's fits (a) and (b) wait for its R7, which is not pushed.
32. rv2b's NITs.
    • NIT 1: where a peer moves the lane's handoff, and the row's cell with it, while this launch stamps it, the write's replay follows the cell, and 5d now reads the file the write printed, not the path handoff-path named before it.
    • NIT 2: fifteen composite asserts (S9073) are split.
    • NIT 3: W_reflog, assert_W_moved_only_by_fast_forward and a new no_auto_gc in test_a23_lane_start_read replace their copies. _published and published stay two, because the row-stamp sandbox has no origin and returns text, and the claude-current one returns bytes.
    • NIT 4 needs no change here: slice 3 spells the PAUSED subject as lane-handoff does.
    • NIT 5: WORKBENCHES_ joins both SCRUBBED_PREFIXES.
33. The battery's survivors are pinned where a case can tell them apart. m2b's 59 one-line mutants of this slice's code left 26 alive on the bench. Fourteen are now killed by new cases or assertions (LS02, LS03, LS06, LS07, LS10, LS12, LS17, LS19, LS29, LS41, LH03, LH13, LH14, LH16), and five by this round's other cases (I1a, I2a, I2b, LS45, LS49). The rest are under "Not changed".
34. A sibling lane's cell is neither written nor delivered (723, on 2a's 601). With 601 in the translation that write-handoff, handoff-path and set-restart-intent share (A54), and 5d reading only the file handoff-path names (24), a row cell naming another lane's handoff in that lane's clone of the register beside W is refused at every call site. lane-start writes nothing there, delivers none of its block and launches with no first prompt, naming why. lane-handoff writes nothing there, prints its block, and records handoff NOT refreshed (handoff-path exited 2). No 2b code changed for it; one case per call site pins it.
35. The bound file is asked before a 7 (724). Since 2a's 603 the restart fences' 2s are 7, so a pre-image a peer changed after binding comes back as 7 too. 5c now asks hs_bound_unchanged first and dies "the restart handoff changed before its owned resume stamp"; a 7 with the bound file unchanged is still "lost ownership". Both exit 2, and S3's 2 stays a note.
36. An 8(d) handoff at both call sites (722, and 721's last case). Lane openRepoTools-3 read Amendment 23 S2 as governing, and 2a's 604 commits and pushes an 8(d) handoff in its own repository, as e2037b9 did. Pinned at both call sites: the stamp and the PAUSED block are committed there by pathspec and pushed to that repository's origin, its tree clean, and the register takes no commit for them. A 0 may still commit nothing (an 8(d) handoff in no work tree, a clone of the register not beside W), so both scripts' exit-0 words now point at the helper's note "or that nothing was committed", as the skill's 0 arm does (F24). lane-start's 0 says "handoff stamped through write-handoff, which exited 0: : (its words above say what it committed and pushed, or that nothing was committed)". The three golden lines of lane-handoff's exit-0 step and the shell suite's :3822 move with them. An 8(d) stamp whose push fails there is a 6 with the helper's words (982's 8(d) case).
37. 5c's two restart dies and a why with a tab have cases (733's M1, M2 and M4). A 7 with no written line and the bound file unchanged dies "lost ownership" (M1, the battery's LS46); 35's case is M2's (LS47); a why with a tab refreshes the handoff with a space in the subject (M4, LH04).

What changed

• lane-start:
  • two ls_refresh calls (decision 2) and the rows source (3);
  • section 5's checkout and printed handoff path (4, 16);
  • 5c as one write-handoff --stamp (5, 8);
  • 5d and the restart binding read the written file (7, 17);
  • the restart-intent reads (20);
  • the recoveries' text (9) and request-handoff's env (10);
  • the header's "never refuses" paragraph, the 5c comment and the env doc (LANES_ROUTE_WRITES).
• lane-handoff:
  • section 2's file is handoff-path --create (11);
  • the write is one write-handoff --prepend, and the commit block is deleted (12, 18, 19);
  • append-line's env (13);
  • the --restart respawn line (14);
  • the header's env doc.
• Tests:
  • new tests/test_a23_lane_start.py (8 cases at first, 24 now) and tests/test_a23_lane_handoff.py (3 cases at first, 13 now), on test_lane_start_row_stamp's sandbox with the default resolution and the opt-in;
  • slice 1a's lane-start cases in tests/test_a23_lane_start_read.py follow decisions 3 and 4: W moves only by fast-forward, and the checkout-refusal cases keep "could not be made" and "exited 4";
  • tests/test_supervised_restart_regressions.py: the pipe-sized restart-intent answer case (20); the default-resolution variant (routed_restart_box, prepare_restart_routed, sharing bind_pending with prepare_restart); test_ctx_round_trip_through_lane_checkout; test_refused_stamp_with_restart_intent_still_launches; and the atomic-stamp case's rewrite (21);
  • the shell suite tests/test_lane_helpers.sh, run in CI only:
    • :3189's recovery text (9);
    • RV-B2 (a), a refused push: exit 0, the stamp committed and kept on HEAD, no carry sentence;
    • RV-B2 (b), LANES_WORKSPACE_ROOT=$WS2: not stamped, the line printed (15);
    • RV-B3 (c): no pull, the stamp pushed;
  • tests/test_repo_hygiene.py (23);
  • the two t018 goldens (22).

Functions this PR changes in lane-start and lane-handoff

This list is for lane openRepoTools-4's opensoft/openRepoTools#303 ("every read of a lane's log reports whether it happened"), which may need these two scripts. It was compiled from git diff e2037b9 HEAD by function body, and checked again against git diff 1320847 HEAD for this round.

• lane-start (56 → 56 functions):
  • added: ls_refresh, ls_register_text, ls_rows_source, hs_bound_unchanged, intent_field;
  • deleted: hs_git_net, hs_carry_note, lane_register_follow, and, with 24, hs_resolve and hs_workspace_root;
  • body changed: rows_for_lane, row_names_for_lane;
  • top-level code changed: the refreshes, the restart-intent reads (eri_*, ri_*, the binding), the stale-checkout recoveries, request-handoff's env, section 5, 5c and 5d.
  • No log read changed: rows_for_lane reads the register, not a log, and the log and lane-reconcile calls are as they were.
• lane-handoff (47 → 48 functions):
  • added: subject_text;
  • no other function's body changed: top_block, sanitize, cut_to_line, the supervisor and every log, row and lifecycle function are as they were;
  • top-level code changed: section 2's file resolution, the write (section 3), append-line's env in section 4(b), the --restart respawn line and the header.

Tests

One pytest process at a time, with every LANES_*, LANE_START_*, TMUX* and WORKBENCHES_* variable unset, a refusing gh stub first on PATH, GH_TOKEN=bogus and GIT_SSH_COMMAND=false. The bench was loaded by other lanes' runs throughout.

The module set (run at a44b9d2, the merge of 2a'; the four rows marked 4031100 run again at this head, the merge of 2a' with A61, which changes tests/test_a23_write_handoff.py and nothing else):

module	result
tests/test_a23_lane_start.py	24 passed (4031100)
tests/test_a23_lane_handoff.py	13 passed (4031100)
tests/test_supervised_restart_regressions.py -k "routed or refused_stamp or atomic_stamp or restart_owned or respawn or ctx or second_ctx or plain_handoff or handoff_publishers or bound_handoff or lost_ownership or W_behind or larger_than_a_pipe or missing_checksum or interrupted_stamp or opening_prompt or explicitly_clears_previous"	24 passed, 157 deselected
tests/test_a23_lane_start_read.py	10 passed
tests/test_lane_start_row_stamp.py	3 passed
tests/test_lane_start_claude_current.py	34 passed
tests/test_handoff_skill_80_85.py	41 passed
tests/test_triad_lane_tooling.py	5 passed, 22 skipped
tests/test_triad_lane_tooling_handoff.py	1 passed, 21 skipped
tests/test_repo_hygiene.py	239 passed, 18 skipped (4031100)
tests/test_a23_write_handoff.py (slice 2a's)	179 passed (4031100)

The triad skips are NEEDS_UPSTREAM, as before. Before the merge of 2a', hygiene's test_the_root_carries_the_line_ending_rule failed on every commit of this round for want of .gitattributes, which came with the code leg's front door (6314fc8) through 2a'. Not run on the bench, and asked of CI: the shell suite (tests/test_lane_helpers.sh, whose :3755 and :3822 move in this round), the whole of tests/test_supervised_restart_regressions.py, and the full suite.

Red first, then green, then a killed mutant, for each item of this round:

item	red	green	mutants killed
726 (e97ad3f)	at 31990c4: no "only once this restart is confirmed ready"	1 case; supervised -k "refused_stamp or atomic_stamp or restart_owned" 3 passed	I4a, LS49
729 (67e6da5)	at e97ad3f: no LANES_WORKSPACE_ROOT= in the respawn	supervised respawn/ctx list 10, test_a23_lane_handoff 4, 80_85 41	I5a, I5b
730 (2bd3ea2)	at 67e6da5: the stray case got W's words; the published pin passed	2 cases; -k "duplicate or launch_creates or fast_forwards" 4, _read 10	I6a, I6b
question 3 (fb4a7b0)	docstrings only	t018: 5 passed 22 skipped, 1 passed 21 skipped	none needed
fit (c), 982, 990 (c1f2928)	at fb4a7b0: 6 failed (the four record rows, the LANES_NO_GIT record, the stamp's exit 6)	50 passed; -k "exit_6 or no_git" 2; t018 as above	I8a, I8b, I8c, I8d, LS45
F23/F24 (439699f)	at c1f2928: the LANES_NO_GIT record and an exit-5 row	50 passed; t018 as above	I8e, I8f, I8g, I8h
rv2b's NITs (2e4d7dd)	at c1f2928: a moved handoff launched with no first prompt	test_a23_lane_start 15, _read 10, test_a23_lane_handoff 8, row_stamp 3, claude_current 34, supervised list 13	I9a
the battery (e5c3326)	pins of today's code: each shown by its mutant	test_a23_lane_start 21, test_a23_lane_handoff 10	LS02, LS03, LS06, LS07, LS10, LS12, LS17, LS19, LS29, LS41, LH03, LH13, LH14, LH16
723 (58309ba)	at e5c3326, before the merge: the sibling's file written, at both call sites	on the merge, before any 2b code	the merge's own 601; red before it
724 (615cfdd)	on the merge: "lost ownership" for a peer's change to the bound file	11 part-2 cases	P724a (also LS47, M2)
722 (5bf9bbb)	at e5c3326: the 8(d) file left uncommitted; on the merge: the old exit-0 words	11 part-2 cases	P722a, P722b
733 (c12f078)	pins: each passed before its case was added	11 part-2 cases	LS46 (M1), LH04 (M4)

Part 2's four commits were tested as one tree (11 cases passed at 16:12Z) and then made by re-running each item's edit script in order on the merge; the four commits' tree is that tested tree byte for byte.

Before this round, on the old root: run 39893128849 at 37b8806 had tests at 3183 passed, 75 skipped, 1 failed and tests-no-submodule at 2386 passed, 872 skipped, 1 failed, the failure in both being the runner image's known test_a_command_that_ignores_term_is_killed_after_the_grace[host].

Lane openRepoTools-3's answers to this PR's four questions

1. The refreshes' scope: kept as it is until the default flips (decision 2). Lane 3 also asks that, before the flip, refresh_workspace's fetch and merge take W's mutex (try once, a note if held), and that its stray scan be lock-free. The lock-free scan is 2a's 732 (A58), merged here; the mutex is for the flip.
2. The fallback to W's copy: dropped (decision 24).
3. The t018 goldens: the exception is named in the docstrings, and the case names kept (decision 30).
4. LANES_NO_GIT=1: named (decision 25), with one case per script.

Not changed

• lane-start: a reader that stops early, under set -euo pipefail, can end lane-start with no message, or take the minted-transcript branch for a session it has recorded (the shape of #280) openRepoTools#281's class beyond the restart reads. Decision 20 reads the restart intent's answer to its end in both places lane-start reads it. That was CI's 141, and the contact point lane openRepoTools-3 named for 2b. The other early-stopping pipes are left for lane-start: a reader that stops early, under set -euo pipefail, can end lane-start with no message, or take the minted-transcript branch for a session it has recorded (the shape of #280) openRepoTools#281, which is unclaimed and waits for Brett's word. They are listed so that nobody reads them as done:
  • in lane-start:
    • sed | head -n 1 on workspace.yaml;
    • ld_row's awk … exit on swapped's answer, unguarded;
    • printf | grep -qx in an elif, where a SIGPIPE reads as "not in the cell";
    • the two lrec_* reads, guarded by || x="", so a SIGPIPE loses the verdict silently;
  • in lane-handoff: its own restart and record reads.
  • handoff_digest_of's sha256sum | awk '{print $1; exit}' is safe, because its writer's one line is complete before awk can act on it.
• LANES_ROUTE_WRITES stays opt-in, and its default is not flipped.
• lane-end keeps slice 1a's register follow (slice 4, DRAFT Amendment 23 S6.2: lane-end --sweep-checkouts removes the register checkouts CLOSED lanes left under <W>-lanes/ only where nothing holds them, leaves first, each decided under its lock after a fetch, and records them in a NOTED line (#198, slice 4) openRepoTools#308). The handoff skill still pulls and commits itself until slice 3.
• top_block, the PAUSED block's bytes, the supervisor, and every log, row and lifecycle write of lane-handoff are unchanged. So are the transcript reservation, the intent's starting write and 5e in lane-start.
• The shell suite's Amendment 17 section and its stamp assertions are unchanged, and run under the LANES_REPO=$WIP seam, as the plan accepted. CI passed them at 84633df, and again with lane-handoff's change at 37b8806 (run 39893128849). Amendment 23's default-resolution behaviour is proven in the pytest modules above.

Refs opensoft/openRepoTools#198

🤖 Generated with Claude Code

• retired_word's awk … exit over the row and over the published row (731), unguarded. It is safe in practice, since each is one short line. It joins the lane-start: a reader that stops early, under set -euo pipefail, can end lane-start with no message, or take the minted-transcript branch for a session it has recorded (the shape of #280) openRepoTools#281 list above.
• Mutants not pinned (decision 33):
  • LS08 (the first refresh never runs): the helper's own reads fetch W, and the second refresh hides it;
  • LS21 (request-handoff without LANES_LANE) and LH07 (the preparing bind's digest none): they need a binding gate or restart plumbing no bench case reaches; write_lane_of routes request-handoff by its lane argument today;
  • LH11 (the file not taken from the helper's line): equivalent while handoff-path and write-handoff print the same path;
  • LS52: its fallback is gone since decision 24;
  • LS20 (the unrouted recovery's words): the shell suite's :3189 pins it, in CI only;
  • I6c: the published arm of decision 29, which this helper does not reach.
• Most reads of a lane's log cannot tell a failed read from a lane with no log: lane-dir, lane-profile, lane-agent, lane-transcript, history, lane-last, binding and lane-trees answer 8 ("none") over a log git could not read (from #290) openRepoTools#303's fits (a) and (b) wait for its R7, which is not pushed: R7's skip before the write-handoff call, and its gap note naming write-handoff --prepend. Fit (c) is decision 31.
• The kept-block notes still say to put the printed block at the head of the handoff "by hand". Lane 3's 983 asked slice 3's skill to name the write-handoff --prepend command instead; the same change in lane-handoff's notes was not part of this round.
• docs/README-lanes.md is unchanged. The manual's lines for lane-start and lane-handoff are slice 5's.

Refs opensoft/openRepoTools#198

🤖 Generated with Claude Code

[sourcery-ai]

Reviewer's Guide

Implementation

Route lane-start’s Rule 3 stamp and lane-handoff’s PAUSED block through the lane’s writable checkout with write-handoff, while keeping W fast-forward-only and preserving restart-specific refusal fences.

• lane-handoff L2525-2547, L2619-2665, L2672-2732 (+4 small edits)
• lane-start L1209-1240, L4416-4484, L4517-4549 (+22 small edits)

Tests

Add coverage for routed stamp and handoff writes, failure and recovery messaging, restart behavior, missing or moved handoffs, alternate repositories, and W’s fast-forward-only movement.

• tests/test_a23_lane_handoff.py L1-338
• tests/test_a23_lane_start.py L1-783
• tests/test_supervised_restart_regressions.py L181-233, L484-679, L1010-1068 (+5 small edits)

Update launch-read regressions and environment scrubbing to verify W moves only by fast-forward and sandbox runs do not inherit caller window variables.

• tests/test_a23_lane_start_read.py L19-24, L55-76 (+8 small edits)
• tests/test_lane_start_claude_current.py L52
• tests/test_lane_start_row_stamp.py L60

Align shell regressions for stale-row recovery and refused or successful stamp publication with the new write-handoff behavior.

• tests/test_lane_helpers.sh L3727-3740, L3776-3792, L3811-3828 (+3 small edits)

Extend hygiene checks to prohibit pulls and carry promises in both lane scripts.

• tests/test_repo_hygiene.py L3292-3296, L3401-3410

Refresh the tooling goldens and document their lane-handoff output exception, while retaining the surrounding baseline output.

• tests/fixtures/t018-handoff/single-repository-decoys.golden L34-37, L109-112
• tests/fixtures/t018/single-repository.golden L142-145
• tests/test_triad_lane_tooling.py L28-38, L1234-1237
• tests/test_triad_lane_tooling_handoff.py L32-45, L1294-1297

Update the skill regression to expect the PAUSED record phrase produced by a successful handoff write.

• tests/test_handoff_skill_80_85.py L227

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud