From PR #280's review.
improvise-s2gb: every_silent_binding_is_on_the_list walked Normal and
search mode only, so a black hole bound in FormulaPanel (the reviewer's
Z -> commit-export) passed the whole suite. The modes are now derived:
reachable_modes starts from both halves of Normal (the grid, not a key,
picks RecordsNormal, e5pf) and closes over the modes a pressed binding
lands in. Doors `declined` shuts (the three stranding prompts, the
import wizard, records editing, search now that `/` declines) fall out
with no list to maintain. A floor assertion names seven modes the walk
must reach, so it cannot pass by shrinking. furnished_to_type seeds the
mode's own minibuffer as well as the search query.
`heard` counts a SetStatus as a decline only when it says something: a
resolver answering SetStatus("") left the user seeing nothing and passed
this guard, caught only incidentally by the Ctrl-Q tests.
Merges #275's new head (improvise-4k1j) and fixes the fmt gate.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
A multi-line flow value (- name: [a, "b ... run: ... c: x"]) is one string to
go-yaml, PyYAML and Psych, but the gate counted the inner run:. Rather than
refuse one more construct, values are now an allowlist: empty, a block header,
a closed quote, a one-line flow with no quotes or nesting, or a plain scalar
not starting with an indicator. The real ci.yml needs nothing else.
Argument: context/code/tooling.md.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
App::grids() memoises the named view's layout for as long as it borrows the
App: once for a whole frame check, once in model_effects. A frame of n writes
to another view's empty records now derives 1 + 2n grids instead of about 5n.
Empty-record ClearCell on 18k cells with a formula, M4 Max: 1.44s -> 0.66s.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
renamed_away() is a Step::Declined, like still_waiting(), so the Enter
binding's welded clear-buffer no longer empties the editor the user must
Esc away from. Shared by the value and rename paths. Red first:
a_retry_whose_cell_is_gone_keeps_the_typed_text failed with Some("").
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-4k1j, from PR #275's review. hidden_row_item refused on a fold
only when the NEAREST group header was collapsed. Under two row
categories that header can be the inner one's: with Region's Coast
collapsed and Product's Kit expanded, the client composed
HideItem{Region, "Coast"} and a twin host wrote a group name into the
document, where the host refuses.
The frame cannot answer this on gh/main: FrameGroup names no level and
an inner GroupEnd reads as an outer one. So FrameEntry::Item now carries
the layout's own DataItem::collapsed flag (a fold, or under one, at any
level) and AxisFrame::folds reads it, the same bit
GridLayout::row_folds_a_group reads. FrameEntry::item builds an
unfolded one, as AxisEntry::item does.
hiding_agrees_with_the_host_on_every_row_of_nested_groups presses H on
all 25 rows of four nested grouped grids, in two windows each. It was
red on the nearest-header check at rows 4 and 5, as the review found.
Also fixes the fmt gate the review found red: fmt-workspaces.sh is a
check, not a formatter, and its diff output was misread as success.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
8phw: when A's Enter crossed B's HideItem, the LayoutMoved refusal replayed
the editor over South and settle_opened recaptured South/Cost, so the
retry wrote there. #251's Opened::Retry/Gone now covers values:
written_to() recognises SetCell/ClearCell as well as a rename, and Retry
and Gone carry the whole OpenedTarget, which the recapture must match.
A Gone value editor declines with RECORD_RENAMED_AWAY. seen stays
rename-only.
Codex P2 on #274: commit picked value/rename/create from the cell under
the cursor before it consulted the capture, so a reframe that left no
cell there sent a captured editor's Enter upstream. commit now
dispatches on opened.target(); only an uncaptured editor is dispatched by
the cursor. A records grid's measure cell is not captured as a value,
since the browser still declines it.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
The tracker closed a multi-line quote on an escaped \" or a doubled '', so a
later quote reopened it over a job's real if:. Both styles exited 0 on the real
ci.yml. The opening line is refused instead; _opens_quote's escape handling
is now load-bearing and has a case per style. Argument: context/code/tooling.md.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
`fold_frame` wrote the rejection sentence straight into `view.status_msg`,
so it was in neither `settled` nor `in_flight`, and the xd3q rebase erased
it whenever a relay queued behind a frame still in flight: an in-flight
frame's own `SetStatus` was replayed over it.
`ClientState::notify` queues the sentence as a view-only entry at the tail
of `in_flight`, where `fold_outgoing` puts a frame the host never answers.
Frames sent before it replay first, relays queue ahead of it, frames sent
after it may replace it, and it settles when everything ahead is answered.
Writing it into `settled` too was tried as a rival and fails the same way.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
OpenedCoordinate becomes #274's OpenedCell, now carrying #251's seen
(filled for a rename only) inside Opened::Target. Retry/Gone keep #251's
rename-only semantics in this commit; the value-cell retry (improvise-8phw)
follows. #221's superseded "a retry after a refusal is a force" paragraph
is dropped in favour of #251's rule.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
reopened_at opens provisionally where the landing cell is not projected
yet; ClientState::settle_declined_editor closes an untouched editor once
the host projects that cell as one i declines. Both ask one
resolve::declined_at. A touched editor is kept.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
`i`/Enter opened an editor on a NonItem cell and the refusal came only at
commit, under a minibuffer that hides the status. `EditOrDrill` now asks
`cursor_cell_id` after the formula check, as the browser already does.
Three rivals each passed the whole workspace before this: NonItem
writable in `value_cell_key`, a Record in `committable_cell_key`, and
`None` from the browser's open-editor arm. Each now fails its own test.
The commit guard is reached through `:enter-mode editing`, which opens
an editor past `EditOrDrill`.
command.md also records improvise-4daa's argument beside gk51's.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-oyaa. every_silent_binding_is_on_the_list presses every
binding the Normal and search-mode keymaps reach (parents unshadowed,
prefixes expanded) on the furnished fixture, and requires each silent
one to be on NEEDS_THE_HOST. The two earlier guards could not see an
unlisted binding that no help row names: dropping force-quit (Ctrl-Q)
from the list failed nothing. Now it fails this test.
every_binding_reaches_parents_and_prefixes checks the walk itself
against a hand-built keymap, because no silent binding in today's maps
sits in a parent or behind a prefix, so a walk that skipped either
passed the first test unchanged.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
The toast hid every status after six seconds, so a refusal the user was
still fixing vanished with its prompt open: su4u's bug on a timer. The
browser now asks the TUI's question. EditorSession moves into
improvise-protocol, ClientState records which editor a status was
written in, and RenderModel::status_stands tells the toast to hold it
until that editor closes.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
Enter in Editing/RecordsEditing is [commit-cell-edit, clear-buffer edit].
The "still waiting" answer came back from composed() as an ordinary
Local step, so the sequence ran the welded clear after it and emptied
the editor it had left open, on the value path and on #221's rename
path alike.
The two cell commits are now answered by committed(), ahead of
composed(), and still_waiting() is a Step::Declined: the existing
channel by which a step stops a sequence. Under #227 (no weld; the
commit clears at its head) the binding is one command and the waiting
return precedes that clear, so the fix fits both.
Stacked on #274.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
A job header with a trailing comment was accepted by the readability check
(comment-stripped line) and never opened (raw line), so its if: went to the
previous job and read_graph never listed it. The audit found the stripper also
cut a quoted "x # y" open, hiding a later job-level if:. Both exited 0 on the
real ci.yml. Refusing the comment makes stripped == raw on every accepted
line, so the stripper is deleted. Argument: context/code/tooling.md.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-fo0w. Three black-hole keys closed over vocabulary that
already exists, each mirroring its Cmd line for line with a lockstep case:
- enter-advance (Normal Enter) resolves as cursor arithmetic over
GridExtent; the host's bottom-right AbortChain is Ignored here, which
loses nothing because Enter is never bound inside a Sequence.
- hide-selected-row-item (H) resolves to ModelCommand::HideItem over the
outer row category and the selected tuple's first label, and refuses
on a fold with the host's own FOLD_REFUSAL, now in improvise-protocol.
- search (/) and search-navigate (n) decline out loud as a pair: n needs
the whole model, and / alone would open a box that finds nothing.
NEEDS_THE_HOST drops to exit-search-mode, force-quit and
toggle-group-under-cursor. Tests that used Enter or "Start search" as
their upstream example move to Ctrl-Q and "Exit search mode", including
shell.spec.js's not-sent case.
Also improvise-0629: the help render test now asserts every ex row is
drawn dim, not just that its description is marked.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
The advance no longer reaches the commit door on a fold or a formula's
cell, so the two older advance tests stopped seeing it (measured: the
cell_key rival passes them). Two tests seat the editor there directly.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
resolve.rs::written read its operands at Enter: the record and measure
from the cell under the cursor, and the basis from claim(), i.e. the
version current at commit. So a HideItem from another client that
renumbered the rows under an open editor sent the write to whichever
cell the cursor then named (8xq0), and a write by another client to the
record while the editor was open was silently overwritten instead of
refused RecordMoved (xz2z, 0j0f's rule on the keyboard door).
The fix reuses #221's capture rather than adding a second one.
ClientState::opened becomes OpenedCell { record, target, version,
document_generation }, where OpenedTarget is Value { measure } or
Coordinate { category }. settle_opened captures a Record cell as a
Value target, and written builds the RecordCell and the claim_as_read
basis from it. With no capture while the host has yet to answer this
client's own last write, the commit declines with the rename's "still
waiting" status.
Red first: both round-trip probes failed on gh/main 202e7eac and on
#221's head 39821d5f. Unit-test fixtures that set Editing by assignment
now open the editor through the fold, which is where the capture is
taken.
Stacked on #221 (fix/fdk1-coordinate-rename).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
pb30: gz puts the cursor on the collapsed group through landing and
place_cursor, and their column arms were untested. Three rivals passed
all 1948 tests: the arm swapped, landing returning None for columns, and
landing reading col_tuple(0). A new test uses two column levels
(Region over Month) with the cursor at (1, 7), South/May. It requires
the cursor to land on South's Q2 at (1, 5). Each rival fails it and
nothing else. The col_tuple(0) rival needs the two column levels, since
with one level the prefix is empty everywhere.
0r01 (Codex's P2 on #263): innermost_group built open_runs for the
whole axis, cloning a level stack per row, to answer for one row on
every z. open_runs is now a lazy iterator, still the one reader, and
the lookup takes .nth(idx). a_lookup_does_not_read_past_its_row gives
it an unreadable header after row 0 (level usize::MAX). The eager rival
panics with capacity overflow.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
w0j4 read any line deeper than 4 outside steps:/needs: as content, so a
needed job whose own keys sat at indent 5 or 6 (valid YAML) had its if:
and needs: missed and the gate exited 0. A deeper line is now content
only once its job has a 4-space key. tooling.md names the three valid
styles the gate refuses loudly.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
mdfc put the "any unanswered write" check inside `advanced` alone, so an
editor opened with i/a (`edit_or_drill`) still seeded from the cache the
write had made stale: `i 9 Tab Esc i 5 Enter`, typed faster than the
answer, stored 15 in the browser where the TUI stores 195.
The in-flight half now lives in `opened_on`, so no opener can forget it.
`advanced` keeps the "this frame writes" half, since the commit being
resolved is not in flight yet; dropping it fails the 3tt4/3uzg tests.
Stacked on #227 (fix/fy9o-reopened-editor-reseeded, de4a7d78).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
GridLayout::evaluate_in and column_total took any &FoldMemo, so a memo
from a layout showing Feb, handed to one hiding it, gave 5.0 where 2.0
was right. GridLayout::evaluator(model) now returns a GridEval that owns
a memo its layout built, with cell() and column_total(); the memo-taking
methods and the public fold_memo are gone. grid_body and prune_empty use
it, and the earlier rivals still fail when restated for the new API.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-uln5. The browser painted every help-overlay row plainly,
including six whose keys go upstream to a host that answers nothing:
/, n, H, z, Enter, and search mode's Enter/Esc. Every ex-command row was
painted plainly too, although `:` opens a command line the browser
cannot run. A dead key is worse than a gap: the keys typed after it land
on live bindings (improvise-koh6).
- resolve::NEEDS_THE_HOST is now the one list of names this client has
no arm for. KeyRow keeps the command it was built for.
- The browser's RenderHelpRow draws a listed row, and every ex row while
the command line is stranded, as `dim` with "(not in the browser yet)".
It marks rather than omits: both front ends keep drawing the same rows,
the feature stays findable, and web/ needs no change.
- the_command_names_that_still_need_the_host_are_these now drives that
list rather than a table of its own.
- every_help_row_the_browser_cannot_run_is_one_it_marks holds it from
the other side: silent help rows == listed rows, and no hint line
names one.
- every_advertised_command_answers_in_its_own_mode no longer restates
the black holes.
Red: a_help_row_the_browser_cannot_run_is_drawn_marked at 202e7eac.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
A pruning commit is answered with ResetCursor and an echo of its own
placement. A frame sent before that answer had already been folded
optimistically, and fold applied the relays straight onto view on top of
it. The echo is per frame, so the later placement was never re-sent: the
client sat at (1,0) while its mirror sat at (2,0), and the next write
landed on West where the TUI writes East.
#246 already queues each relay behind the head frame, so the queue holds
the right order. fold now rebuilds view as settled plus everything still
in flight whenever a relay queues, the same replay a refusal uses.
Replaying from settled, rather than re-applying what queued behind the
relay, is pinned by a ViewBack test: re-applying stepped back twice.
Stacked on #246 (fix/hv0r-jpba-relays, 790a33ba).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
#257 restored a stale message whenever the mode after a key equalled the
mode before it. Editing and RecordsEditing commit with CommitAndAdvance,
which re-opens an equal AppMode on the next cell. So the message followed
the editor from cell to cell until Esc. The reviewer's repro was R, o, 5,
Tab: "Added new record row", or o's "That record row is already here"
refusal, survived the Tab.
press now compares an EditorSession, made of the things an opening sets:
- the whole AppMode, not its variant;
- the cursor cell, which is a cell editor's target;
- the editor's seed, which EnterEditAtCursor re-seeds from the cell, so a
commit that re-opens on the same cell (Tab in the last column) is new.
A generation counter was rejected: editors are opened by more than one
effect, and a counter in each is a hand-kept list. The view was left out:
no minibuffer key can switch it and stay.
Tests:
- four press scenarios: the two records repros, Enter in a plain grid,
and Tab in the last column;
- a direct check that ItemAdd on another category is another editor.
Rivals: mode-only (#257 as pushed), no cell, no seed and discriminant
each fail a distinct step.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-9ww8: a crossing rename by another client bumps layout_version,
so the refusal is LayoutMoved, not RecordGone, and the retry recaptured
whatever record was under the cursor. refused() now pins Opened::Retry(r)
for any refused rename of r; the recapture becomes the target only if it
is r, else Gone(r). RecordMoved on the same record stays a force. The
explicit RecordGone pin is subsumed and removed.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
uv9e: ToggleGroupAtCursor { is_row: false } (gz) had no command-level
test, and a version reading col_group_for(0) passed the workspace. A new
test seats the cursor at (0, 2) with Month on the columns and requires
Q2. Reading column 0, or reading the row half of the cursor, both fail
it. The row half was already closed: the grid tests sit on row 2, and
injecting row_group_for(0) fails five of them. The browser's z is still
Upstream with no resolver arm (improvise-uln5), so it has no cursor
reading to test.
6u6s: the e265 enumeration could not see run identity. A pruner keying
runs by (level, category, group name) instead of header position
passed. The enumeration now:
- compares every level's run name and whether each survivor shares a
run with the previous one;
- adds three fixtures: a non-contiguous group, a level with no groups,
and two adjacent outer groups;
- builds each layout once per collapse subset and drives
prune_axis_items with a keep predicate per row set: 4,992 cases.
The name-keyed rival now fails at fixture 2, mask 0b101. open_runs is
the one reader of the markers. improvise-core's lib went from
0.86-0.92s to 0.17-0.23s on an Apple M4 Max under load average 39-77.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
Measured in Chromium over the worker: with keys pressed back to back,
`i 9 Tab Tab 5 Enter` stores 5 or 5095 depending on whether `5` beats the
answer; with a pause, 5095 in 5 of 5 runs. Never the stale 505.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
Quoted, anchored and multi-line needs: values, and quoted, spaced or
explicit ? keys, all skipped kb on a PR while the gate reported every
suite run. Rather than parse more YAML, every structural line inside
jobs: must take one plain form or the gate names it and fails. Also a
1be2 case at the one-column boundary, which the owner+1 rival passed.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-mdfc: an untouched commit seeded the reopened editor from the
cache even while an earlier write from this client was unanswered, so
Tab Tab at the rightmost column stored 505 where the TUI stores 5095.
`advanced` now awaits when this frame writes or any in-flight frame is
awaited. The lockstep test sends three Tabs before answering any.
improvise-4cb5: `settle_awaited_seed` also requires an editing mode, so
an Esc before the answer no longer leaves the seed in Normal's buffer.
improvise-mgis: `OpenedEditor::awaiting` loses `serde(default)`; a frame
missing it fails to parse as the missing field (protocol.md §3, 1wl0).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-pntu: FoldMemo was keyed by (key, span) while the aggregation
and covers were call arguments, so one memo across two covers answered
the first (5.0 where 2.0 was right). FoldMemo::new(agg, covers) now owns
both and the _in functions read them off it; GridLayout::fold_memo builds
the grid's. The mismatched call no longer compiles.
improvise-0jw5: a test in improvise-command counts fold evaluations across
one grid_body call (18) against a fresh-memo control (90). A per-cell or
per-total fresh memo fails it. The counter is exposed through a new
test-support feature on improvise-core, enabled only as a dev-dependency.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
#233 clears the status message at the start of every TUI keypress. In a
text-entry mode the message sits beside the prompt, and a formula that
does not parse keeps the editor open with the error. So the error
vanished on the first correcting key.
press now remembers the mode and the message. It clears the message as
before and dispatches the key. It puts the stale message back only if
the key left the user in the same text-entry mode and wrote no message
of its own. Editing keys keep the refusal on screen, a key's own message
still wins, and leaving the minibuffer (a successful commit or Esc)
drops it.
The mode check is observed, not predicted: which keys commit is the
keymap's knowledge, and the TUI loop does not copy it.
Rivals, each injected into press and restored: never restoring (#233),
restoring over a key's own message, and restoring whatever the mode each
fail a distinct step of the test.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-nacd and improvise-pr8t.
The b114 rule refused a coordinate that is a group. The invariant it
protects (improvise-gw18) is positive: every coordinate names an item.
Two writes onto a stored record fell in the gap and were accepted: a
measure that is no item at all (Bogus), and a parent group that holds no
item directly (Totals), where items_in_group is empty.
App::group_refusal becomes App::non_item_refusal. It refuses any
coordinate that is not in Model::effective_item_names for its category.
Formula targets count as _Measure items, so a write under a formula is
still refused by the Derived rule, with the true reason.
Tests:
- pr8t pins the same-name case: an item named like a group of its
category is writable. Removing the old exemption fails it.
- nacd is red at 202e7eac for both Bogus and Totals.
- A third test pins that a formula cell is refused for the formula.
Two reduce.rs fixtures wrote _Measure=Revenue without the item. They now
add it.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
On an axis with two grouped categories, runs nest: North in Upper holds
Jan in Q1, and after Q1 ends North/Mar is still in Upper. The markers
carried no level, and every reader was flat, so the nearest marker
decided. Pruning an inner header dropped its GroupEnd too, and a
surviving row changed group. The unpruned answer was lossy as well:
North/Mar was in no group while it was drawn under Upper.
The user ruled per-level membership, with z toggling the innermost group
(the one nearest the cursor). So:
- AxisEntry::GroupHeader and GroupEnd carry a level. expand_category
reads it off its prefix, which holds one item per outer category.
Membership is one open run per level, and a row's group is the
innermost one (row_group_for/col_group_for, hence z).
- prune_axis_items re-marks the survivors instead of filtering the
markers. Before each kept item it emits the markers that make the
output's open runs equal the input's, closing deeper runs innermost-out
first. So each survivor keeps exactly its runs, no header is kept over
nothing, no trailing GroupEnd is left, and when nothing is pruned the
output equals the input. This replaces the pending/close_runs_of/
run_open machinery.
- The wire: FrameGroup gains level (an index into AxisFrame::cats), and
FrameEntry::GroupEnd becomes GroupEnd { level }. groups()/group_of
read the same per-level rule, and group_runs is unchanged (h0iv).
- axis_frame carries the open run at each level above a scrolled
window, not the single nearest header, so a window opening inside
North/Q1 still knows Upper.
Tests:
- an enumeration over every funded mask x collapse subset of a two- and
a three-level fixture (about 4,000 cases), asserting that pruning
moves no row's group, that no header is drawn over nothing and no axis
ends on a marker, and that nothing-pruned is marker-for-marker equal;
- the ruling at the layout and through z;
- frame/layout agreement at every window offset;
- the nested window carry.
Rivals, each injected and restored, are in the evidence files: flat
membership, a pruner that owes no GroupEnd, ends passed through eagerly,
headers kept eagerly, nearest-only window carry, and a flat frame.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
find_modern_bash accepted any 5.x, so a bash 5.0 was pinned and
pr-reviews-test.sh then failed on an invalid option. The probe now
reads major and minor; a 5.0 case and a 5.1 control guard it.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
Edward Langley