The last of the anti-slop interaction work (guidelines G6, G7, G14, G15,
G17), plus booth-dev's note from S5b's gate. Every S5b promise holds: no
re-POST, serialized saves, a batch never reloads, focus survives a swap.
- Keys (G6): one rule in base.html's <head>, BoothKeys.theirs(e), called
first by the grid, the review and compare. A field or a player owns every
key but Escape (Esc still goes back from a focused player); a control
owns Space; a focused 1:1 stage that can pan owns the
arrows and Space (Chromium puts it in the Tab order); Ctrl/Meta/Alt are
the browser's. The field check lives on as BoothKeys.isEditable. Before:
an arrow on a focused video left the review, and Enter on any control
also opened the grid cursor's tile.
- The grid cursor is real focus: the tile it moves to gets tabindex=-1
(script-set, one tile at a time) and focus, without a scroll; the cursor
is an item (its data-item), and a doc closed with its ✕ is skipped; focus that
lands on a tile (S5b's fallback) makes it the cursor; Enter opens the
review only from the body, the grid or the tile, by its view?f= link;
n opens a closed doc's fold; Escape clears the cursor
and releases the tile's focus. The reticle is its focus mark (no second
ring).
- The doc bar (G7): the controls leave the <summary>. div.doc-bar holds
details.doc-fold (its summary is the label only) and div.doc-tools beside
it; the body and notes follow in div.doc-inline, hidden with a closed
fold by :has(), scripts on or off. A closed doc keeps its tools. Renders
pixel-identical to today at 1280 and 390, light and dark.
- Tile sizes (G14): a gallery tile's <img> carries width/height, the
picture as the browser draws it (EXIF 5-8 swap), read from the header
only (no decode; PNG getexif is skipped unless the header carried it),
opened O_NOFOLLOW|O_NONBLOCK, cached by the file's identity (ctime
included, so cp -p over a file is seen), in a separate
step (items.image_dims over thumbs.drawn_size) so the Desk never pays it.
Measured before: a link to tile 30 of 40 landed 44px low (3/3); after, on
its mark. content-visibility:auto, which the report proposed too, is NOT
added: a swapped-in tile has no remembered size, and a flag far down moved
the page 2929px (3/3; 0px without it).
- The rail (G15): html:has(.rail){scroll-padding-top} replaces .item's
scroll-margin-top (the two add), so a control reached by Tab stops below
the sticky rail too. Measured before: a Tab-focused flag button at 19.6px,
under the rail's bottom at 47.6px. The scripts-off fallbacks are the old
rules' numbers (132px, 217px at <=480), now pinned by a test. The height
script follows the live rail after every in-place save (it watched the
replaced node, and read 0px after one flag), and the rail's own controls
cancel the padding (a Tab between stuck group links scrolled 357px).
- Reveal names (G17): no aria-label on any reveal control; the name is the
words on it, the glyph in an aria-hidden span, the item's name as
.sr-only text ("reveal a.png" / "hide a.png"). Reveal all drops
aria-pressed (its words already say the state; r2b rules them) and its
"on" look reads the .reveal-all class on <html>. No pixel changes.
- booth-dev's note: a refused batch's forms enter `unsent` with the
refusal's words, and a later save says every standing failure's words
(each once, in order) instead of "Saved.", and every warning says the
other standing failures first, so no failure buries another. Test first:
test_a_batch_refusal_outlives_an_unrelated_save.
- Rows re-anchored to the same failure: r2b "Space on a focused review
button", r3 "C3 a held modifier" and both "C3 Space on a focused ..."
(now in BoothKeys), r2c "the stage reveal shows with scripts off", and
this contract's S3 doc-bar row and five S5b status-line rows.
Folded from the heid contract review (BEINKA, panel 4/4, thread
01M3NZJNX8D3BEYD48M9K3MV3Q): 24 flags, all prose the tests left open; the
contract states the tile/focus/cursor seam with S5b, the helper's union and
scope, the size's source and every path to none, Reveal all's name, the
refusal sentence's lifetime, and the fallback arithmetic (one test added).
Folded from the heid bug-hunt (HRÖSKVA, panel 4/4, thread
01M3P0ZPRSASFSE5K3PR4NTQP6): R1 closed docs and the cursor as an item, R2
the rail's height after a save, R3 no warning buries another, R5 the view?f=
link, R7 ctime in the size cache, R8 Escape from a player, R9 the rail's own
controls, R10 n on a closed doc. Refuted with reasons: R4 (unreachable: refused
picks re-send together), R6 (Chrome takes the same header's size with or
without the attributes; measured), R11 (by design).
From this slice's own falsifier runs: a "one row wide" row that mutated a
flex basis a non-wrapping bar just shrinks (re-aimed at the bar's flex), and
a Reveal-all "on look" read under the clicking pointer, where :hover draws
the same border (the pointer now leaves first; 3/3 proved).
Contract: as_antislop S5c.
Falsifiers: antislop.toml S5c section.
Compare's stages load their pictures through the catch-all file route, which
caught only OSError around resolve(); an embedded NUL raises ValueError. Same
class as resolve_booth's fix in f8d136a (heid bug hunt on the race fix,
hulda). The upload route's NUL-in-filename 500 is the same class and is left
to booth-dev: it is not on compare's path.
booth-dev's race note after the merge: the compare route resolved each side
in _compare_side and again in _compare_ring, then ring.index(a) raised if the
file vanished (or was relinked outside the booth) between the two; the review
did the same through cring.index(f). The compare ring is now built once and
the sides are judged by membership of it. The review re-judges its item and
scans forward for the next comparable one (usually one step, no longer a
resolve of the whole ring per render); an item no longer comparable renders
the review without a Compare control, and C does nothing.
The contract records the once-per-request rule and that the phone-width wrap
covers doc.html's bar too. r3.toml: 59 rows, four re-anchored.
Navigation was built from the review ring while the compare GET also demands
containment, so an outside symlink (which stays in the ring) was offered by
the strip, the steps, the review's Compare control and the flag landing, and
404ed on arrival. Every one is now built from the compare ring (the review
ring filtered by the same conjunction, _in_booth).
Two pre-existing gaps compare inherits, fixed at the source: resolve_booth
caught only OSError, so a NUL in the booth segment was a 500; record_view
opened its marker blocking, so a planted FIFO hung every look. Plus: the page
treats %73ide=a as side=a, and the subgrid engine floor is stated. Two
findings refuted (a chorded click mid-drag never fires pointerup, measured;
booth_items never yields an unquotable rel). r3.toml: 57 rows.
The one drift: the separator was a border on B, making B's stage 1px
narrower than A's; it is now a 1px column gap, so the stages are the same
size to the pixel. Tests now read what the contract promises instead of a
proxy: the strip's ring order, the full bakeoff sequence, 1:1 and Fit by
geometry, the 900px break from both sides, A wrapping, each form naming its
own item, a sibling-prefix symlink, both reveals, the strip's flag, the back
arrow unlinked, a one-axis picture, a focused player, two videos with no
toggle. The contract names .cmp-cap, a press on a stage, INV-4's URL-driven
picker and the redirect branch's isinstance check. r3.toml gains ten rows.
GET /b/{name}/compare with the conjunction 404 (containment AND the review
ring), both sides recorded as seen, view state (side, link) mapped from a
closed set onto every link, side-keyed regions, and back=compare in
_mark_redirect. compare.html: two stages sharing one set of rows, the strip
as picker (the side active now), linked and per-side stepping, X/L/Z/A/B/C
keys under the review's guards, synced pan by fraction with an echo guard,
per-side blur reveals, JS-off parity.
The stage machinery moves out of view.html into _stage_js.html
(BoothMode.bind, BoothStage.attach), shared by the review and compare. The
review gains a Compare control and a C key. At phone width a full top bar
wraps.
Tables: r2c's 15 stage rows re-pointed to _stage_js.html; r2b's phone
top-bar row re-anchored (the wrap made it vacuous alone); new r3.toml. The
contract records the wrap, equal stages and C on the compare page.
vh