Amendment 23 act 2, slice 2a: write-handoff writes a lane's handoff in the checkout its writes go through, handoff-path says where, publish-handoff keeps its contract on one shared act, and set-restart-intent takes the row's spellings (#198)

docs/README-lanes.md

@@ -1218,12 +1218,12 @@ else.
 |---|---|
 | 0 | done — and for a write, **confirmed**: the commit was read back on origin before the writer said `pushed` (Amendment 23 S4) |
 | 1 | environment — no register, no writer |
-| 2 | refusal — bad arguments, the object is held, an unknown alias, or `commit` refusing to publish local commits that are not register or handoff writes, from a checkout whose HEAD is off the branch, or that it cannot read (#194; it pushes none of them); and, Amendment 23 S3, every writer with nothing written, where its checkout cannot be brought to origin's tip: a fetch that failed (`LANES_NO_FETCH=1` is one, for a writer), an operation in progress or unmerged entries, HEAD off the branch, staged entries that are not the write's own, unpushed commits (the publish act excepted), a diverged branch, a pending change to a path the write commits, or a fast-forward git refused; a `LANES_REPO` naming a directory below its checkout's top level, from which git would read the write's pathspecs, before any edit; the withdrawn `--sweep`; the publish act over a stack origin has moved past (S4); and `claim` or `add-row` meeting origin moved under its own checks on each of six attempts, before any edit (S3, S4) |
+| 2 | refusal — bad arguments, the object is held, an unknown alias, or `commit` refusing to publish local commits that are not register or handoff writes, from a checkout whose HEAD is off the branch, or that it cannot read (#194; it pushes none of them); and, Amendment 23 S3, every writer with nothing written, where its checkout cannot be brought to origin's tip: a fetch that failed (`LANES_NO_FETCH=1` is one, for a writer), an operation in progress or unmerged entries, HEAD off the branch, staged entries that are not the write's own, unpushed commits (the publish act excepted), a diverged branch, a pending change to a path the write commits, or a fast-forward git refused; a `LANES_REPO` naming a directory below its checkout's top level, from which git would read the write's pathspecs, before any edit; the withdrawn `--sweep`; the publish act over a stack origin has moved past (S4); and `claim` or `add-row` meeting origin moved under its own checks on each of six attempts, before any edit (S3, S4); and `publish-handoff`'s restart fence on its bound file or pre-image (`write-handoff` answers those 7, so its 2 is never a restart fence's) |
 | 3 | **pushed, not confirmed: `<sha>`** (Amendment 23 S4). The write committed and pushed — or its push timed out, and a hung push often lands — and the read-back could not confirm the commit on origin: the fetch failed or timed out, or origin does not hold it. Never success. The commit stays on HEAD — unless another writer in the checkout **rewrote** it while its push was in flight (an unlocked writer's rebase), which the words then say: it is not on HEAD at all, and what reached origin is read by content — and the next write's S3 finds it by content either way: published, there is nothing to do; not, every write refuses until the publish act sends it or an archive (S6.3) takes it |
 | 4 | the mutex could not be taken within 60s |
 | 5 | an edit moved more than one line and was refused |
-| 6 | `git add` or `commit` failed — or the commit is **kept on HEAD and named** (sha, subject, `git show`): a sixth push refused because origin moved (Amendment 23 S4's bound; where S3 finds the sixth move before any edit, nothing is committed and it is 2), a replay's drop refused over an uncommitted change, a commit another writer made on top of this one in the checkout before the replay drops it, or under it (this write's commit is known by what it is, never by HEAD read after the fact, and the drop moves the branch only from that commit and only where it sits on origin's history, so it never takes another; `commit`'s carry form drops to its own commit's parent by design, so one under it stays, the hand edit goes back to the working tree with its copy kept, and the re-run's S3 refuses on it: 2, not 6), a commit made in the same moment as this write's that cannot be told from it (nothing is pushed), or a push (a write's, or the publish act's) refused with origin unmoved, which no replay can mend. Or another writer in the checkout **rewrote** this write's commit (a `pull --rebase` with no lock): it is not on HEAD at all, and what reached origin is read by content (`git cherry -v origin/main HEAD`). Or, where the drop's tree step was refused and another writer committed on the branch in between, the commit is named off HEAD, with `reflog -5`. Every later write from that checkout refuses until the lane publishes or archives it |
-| 7 | another act got there first and this one wrote nothing: `claim`'s `CLAIM-LOST` (another lane's claim landed first), and `set-lane-state` / `set-lane-tree`'s lifecycle fence (openRepoTools#91, below) |
+| 6 | `git add` or `commit` failed — or the commit is **kept on HEAD and named** (sha, subject, `git show`): a sixth push refused because origin moved (Amendment 23 S4's bound; where S3 finds the sixth move before any edit, nothing is committed and it is 2), a replay's drop refused over an uncommitted change, a commit another writer made on top of this one in the checkout before the replay drops it, or under it (this write's commit is known by what it is, never by HEAD read after the fact, and the drop moves the branch only from that commit and only where it sits on origin's history, so it never takes another; `commit`'s carry form drops to its own commit's parent by design, so one under it stays, the hand edit goes back to the working tree with its copy kept, and the re-run's S3 refuses on it: 2, not 6), a commit made in the same moment as this write's that cannot be told from it (nothing is pushed), a push (a write's, or the publish act's) refused with origin unmoved, which no replay can mend, or a whole-file `write-handoff` whose copy origin changed meanwhile, which no replay may overwrite (S4.4), or a `--stamp` or `--prepend` one whose handoff origin deleted meanwhile, which a replay would find nothing to write into (one moved with its row's cell, to a file under `handoffs/` that origin has, is replayed into the moved file), or an Amendment 8(d) handoff `write-handoff` wrote whose commit or push in its own repository failed (the command that finishes it is said). Or another writer in the checkout **rewrote** this write's commit (a `pull --rebase` with no lock): it is not on HEAD at all, and what reached origin is read by content (`git cherry -v origin/main HEAD`). Or, where the drop's tree step was refused and another writer committed on the branch in between, the commit is named off HEAD, with `reflog -5`. Every later write from that checkout refuses until the lane publishes or archives it |
+| 7 | another act got there first and this one wrote nothing: `claim`'s `CLAIM-LOST` (another lane's claim landed first), and `set-lane-state` / `set-lane-tree`'s lifecycle fence (openRepoTools#91, below), and `publish-handoff` / `write-handoff`'s restart fence — `write-handoff` answers its bound file and pre-image 7 too, where `publish-handoff` keeps its 2 (Amendment 23, slice 2a) |
 | 9 | `CLAIM-LOST` — issue #30's own dead-lane verdict could not be reconfirmed before a `--force` takeover's push landed: the source lane resumed, a live session now backs it up, or that could not be read at all (`claim` only). Never 7 — that code is a RIVAL's claim, and this is the same lane the takeover was granted over |
 | 10 | `lane-state`: the lane's lifecycle snapshot is there and could not be read, or is of this schema and not a whole record, so nothing can be read out of it as a snapshot (openRepoTools#91 and #115, below) — never the 8 that means it has none. It was 9 until #61 spent 9 on `claim`. `lane-trees`: the lane's `trees` directory is there and could not be listed (#157 item 4e) — never the 8 that means it recorded no tree |
 | 8 | no record — and no other meaning |