socket.rs: EditTarget takes the union (document_generation, cell, version,
opened); typed_target_frame keeps #205's untouched early return, then claims
as read at the captured generation. Evidence appended after gh/main's.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
GridLayout::column_total asks Model::evaluate_spanning, whose Span adds to
#203's fold the categories a key omits. A formula is evaluated over the
pooled rows, a stored measure folds its records under its own aggregator,
text is unanimous or mixed, an error propagates, and _Measure on rows
leaves no strip (the user's rulings). Hidden rows stay out.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-7fl8: on the one-tick-per-read clock --wait 5 polls exactly 6
times, but the test asserted at least 5. So a loop capped at 5 polls, or
a deadline of now + WAIT - 1, passed at 25/25. It now asserts
polls = WAIT + 1, and both rivals fail it.
With the deadline check deleted, the suite used to hang at the first
real-clock timeout case. Every pr-reviews call now runs beside a
/bin/sleep watchdog (wait -n -p), and loses to it as exit 124. The same
rival now finishes in 93s with four FAILs.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
landed_seed parses committed text with typed(), which puts
i64.trunc_sat_f64_s in the client bundle; the bundled wasm-opt rejects it
unless the feature is named, as the worker bundle already does.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
The reviewer's probe_fold_hidden_measure, over its three rows (hidden,
bogus, no measure). The b114 group rule refuses all three; removing it
fails this test and the b114 test and nothing else.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
App::group_refusal asks the document, not the layout, whether any
coordinate names a group rather than an item, and runs first in
unwritable: a measure-half fold hides from both the stored arm and the
projection. Rival (rule removed) fails the new test and nothing else.
command.md and the ciw8 fold evidence now say the aggregate arm guards
only the None-axis case.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
#197 changed project::cell_id to take &Model; aggregate_refusal and the two
sjdb tests now pass the host model. The text merge was clean there and did
not compile (E0308) until they did.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-cq3d: nothing in production cleared status_msg, and draw_status
prints it instead of the derived hint, so after the first message of a
session the TUI never drew the key hint again, in any mode.
run_tui now sends each key through draw.rs::press, which applies
SetStatus("") and then dispatches: a message a key produces is drawn after
that key and expires at the next one, vim-style, so an error is never
wiped before it is drawn. The expiry lives in the TUI front end rather than
App::dispatch_key, which the lockstep harness shares with the browser. The
browser already shows messages transiently (a TOAST_MS toast) beside an
always-drawn #hint. Reasoning: context/code/tui.md.
The test drives press: a stale message must give way to TileSelect's hint,
a message must survive its own key, and give way at the next. Red at base
at the first step; clearing after dispatch fails the second.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
Edward Langley