3 Commits
Author SHA1 Message Date
rzen 2090d742b9 Build the trash quasi-lane
Deletion becomes a two-stage, Finder-style story. File > Delete and
plain Backspace tombstone the live selection; View > Show Trash (no
chord — Shift-Cmd-T stays with the system's tab bar) reveals the
quasi-lane: trailing, one fixed width unit consumed only while shown,
hatched dimmed header, count badge, no new-card button, exempt from
resize and reorder alike. Its contents are a pure view over the
snapshot — the deterministic sort (deleted newest first, folder-name
ties, unparseable stamps oldest) interleaves card rows with a
tombstoned lane's single entry, whose count names what Put Back
returns; the ancestor walk is absolute, so an own-flag card beneath a
tombstoned lane has no row and recovery is deliberately two steps.
Put Back twins Delete on Cmd-Backspace with validation enabling
exactly one; restore fidelity is byte-perfect because nothing ever
moved. Delete Immediately confirms exactly where loss is real (every
board is mode-none today; the predicate names the git carve-out for
m7), Empty Trash always confirms with the true whole-board count,
and dragging a tombstoned card onto a live lane restores it there —
positional drops and cross-board locality arrive with m5's machinery.
The banner's delete phrasing drops "move to the trash" per the naming
constraint: board deletion says Delete, "Move to Trash" stays
reserved for the system Trash. 47 new tests.

Claude-Session: https://claude.ai/code/session_01SR4XGjmBE16ZUYWpfFHXwY
2026-07-27 15:38:34 -04:00
rzen bea6d02d1d Realign read-side rules — width range coercion, finite order, symlink pins
The design corpus ratified that ranges are part of a sensible reading:
an exact-integer width below 1 now coerces to 1 read-side (bytes
untouched) instead of reading as malformed — the width division must
never see a zero or negative unit — while a non-finite order (.nan,
.inf) is now the same loud malformed-order rejection as a non-numeric
one, guarded at the single point where the double arrives so loader
and Writer inherit it together. The symlink-never-traversed rule
turned out to be already enforced (the loader has filtered symlinks
ahead of the directory check since the first commit); it and the
copy-preserves-the-link-verbatim behavior are now pinned by tests,
alongside the two hostile shapes the corpus names (width: 0,
order: .nan). Five new tests.

Claude-Session: https://claude.ai/code/session_01SR4XGjmBE16ZUYWpfFHXwY
2026-07-27 14:21:03 -04:00
rzen ff3ba298f0 Build the full-visibility lane layout
The lane strip replaces the placeholder board: window width divides
across the lanes' width units (no horizontal scroll, no minimum width,
degenerate compression accepted), a lane of n units flowing its cards
into n round-robin masonry columns via a measurement-cached Layout.
Width has two deliberately opposite controls, both landing here: the
right-edge drag (ported verbatim from the pathfinder's ColumnResize)
freezes the 1x standard at drag start, snaps between integer widths
with the asymmetric shadow-leads tick and 10pt re-entry, grows the
window one standard width per snap so siblings keep their exact
pixels, and rubber-bands at the screen's visible frame — uncapped
otherwise; the Increase/Decrease Lane Width items (new Board menu,
Cmd-Opt-arrows) are the stepper's keyboard face and re-divide the
existing window width instead, never touching the window. Width
changes write through the new .resize WriteOperation ("Couldn't
resize…" in the banner vocabulary, which grows with the surfaces by
design); malformed width values render as one unit and stay untouched
on disk. 27 new tests port the pathfinder's resize-math suite onto
the uncapped range and pin the write path's fidelity.

Claude-Session: https://claude.ai/code/session_01SR4XGjmBE16ZUYWpfFHXwY
2026-07-27 08:16:14 -04:00