/* =============================================================
 * SUPERFLY — HOME SECTION 5 — THE DEALER WALL — looks
 * -------------------------------------------------------------
 * Built 2026-11-11 on the operator's instruction, to the stats
 * grid on dozfx.com. That section is measured value by value in
 * docs/research/2026-11-11-dozfx-stats-and-handoff.md — read it
 * before re-pitching anything here, because most of the numbers
 * below are theirs re-fitted to this page rather than guesses.
 *
 * Markup:   assets/html/dealer-wall.html (a readable copy — the
 *           live markup is in page 42's Elementor data)
 * Movement: assets/js/dealer-wall.js
 *
 * NAMESPACE. Everything is scoped under .sfdw. The prefixes already
 * taken in this folder are sf-acc, sf-bike(s), sf-cart, sf-cat(s),
 * sf-checkout, sf-dealers, sf-drawer, sf-filters, sf-hero, sf-intro,
 * sf-lights, sf-live, sf-livery, sf-login, sf-maskword, sf-mega,
 * sf-navbar, sf-net, sf-pdp, sf-pf, sf-sequence, sf-shop, sf-social,
 * sf-sort, sf-step, plus .sfbw, .sfc, .sfcg, .sfcl, .sfidx, .sfop,
 * .sfrev, .sfsr, .sfv and .sfvp — grep assets/css before adding
 * another. Note .sfdna is the REMOVED DNA band; .sfdw is not it.
 *
 * BOX SIZING stated on the root itself, so this band measures the
 * same inside Elementor and in the local harness. CLAUDE.md has the
 * full story of the 80px that goes missing otherwise.
 *
 * NO GSAP. The heading reveal, the odometer and the accent bars are
 * all CSS transitions driven by two IntersectionObservers, so this
 * band loads nothing but its own two files, creates no ScrollTrigger
 * pin and has nothing to fight Lenis. Same arrangement as .sfcg and
 * .sfbw, and the same arrangement the reference itself uses — dozfx
 * runs no animation library on that page at all.
 * ============================================================= */

.sfdw,
.sfdw *{ box-sizing: border-box }

.sfdw{
	--sfdw-accent: var(--wp--preset--color--accent-1, #ec1e79);
	--sfdw-page:   var(--wp--preset--color--base, #171717);
	--sfdw-ink:    var(--wp--preset--color--contrast, #f2f2f4);
	--sfdw-dim:    rgba(242,242,244,.64);
	--sfdw-faint:  rgba(242,242,244,.52);
	--sfdw-rule:   rgba(242,242,244,.14);

	/*
		THE CARD IS THE FOOTER'S SUPERFLY WORDMARK — operator,
		2026-11-11: "i want the card same like SUPERFLY text in footer
		.. the card is neumorpishm , then when hover will pink color
		like that footer".

		That word is in parts/footer.html as four stacked <text>
		copies, styled in footer.css:

		    .sf-pf__word-lo    #000 at .7, blur 3px, nudged +2 +2
		    .sf-pf__word-hi    #23232a,    blur 3px, nudged -2 -2
		    .sf-pf__word-base  #141418                  (the face)
		    .sf-pf__word-lit   a magenta radial, faded in on hover

		So it is RAISED, not recessed: dark edge down-right, light
		edge up-left, lit from the top-left.

		THE CARD IS THE PAGE COLOUR — operator, 2026-11-11: "then the
		background card change to this color .. #171717". That IS the
		page: it is theme.json's `base`, which is why the token points
		at var(--sfdw-page) rather than carrying the literal. Written
		that way so the cards follow the base if it ever moves again —
		it went #0a0a0a to #171717 on 2026-11-10 — and so the next
		reader cannot miss that the two are deliberately the same
		value rather than coincidentally equal.

		SO THERE IS NO FILL DIFFERENCE AT ALL NOW, and the shadows do
		the entire job. That is the purest form of the thing and it is
		what the page's own texture makes possible: the band is carbon
		weave over #171717 while the card is FLAT #171717, so the card
		reads as a smooth panel pressed out of a woven surface even
		where the light does not reach it. On a flat page it would
		disappear.

		**DO NOT PUT A TEXTURE ON THE CARD.** Repeat the weave on it
		and the only thing separating card from page is gone.

		It was #212124 for part of the day — the footer wordmark's own
		+10 +10 +13 delta applied to this page, so the face sat
		slightly above its surround the way the footer's does. That
		is the value to go back to if these ever read as too faint;
		the shadows need no change with it.

		The edges are alpha rather than the footer's flat #23232a, so
		they keep working if the face is re-pitched. Over #212124 the
		.10 white computes close to the footer's own highlight.

		THE OFFSETS ARE SCALED, THE RATIO IS NOT. The word uses ±2px
		at a 3px blur on 250px glyphs — on a 200px card those are
		invisible, so they are up to -4/-4/8 and 5/5/12, keeping the
		roughly 2:3 offset-to-blur the word has.

		THIS REPLACED /return-refund/'s PANEL, which was here for part
		of the same day: #0d0d10 with rgba(255,255,255,.024) at
		-3/-3/8 and rgba(0,0,0,.26) at 4/4/10. That is the quiet house
		card privacy, contact and return-refund share; this is the
		footer's carved one, which is a good deal stronger. **If those
		three files are ever re-tuned, this one is NOT part of that
		set any more** — it follows the footer.
	*/
	--sfdw-card: var(--sfdw-page);
	--sfdw-lift: rgba(255,255,255,.10);
	--sfdw-sink: rgba(0,0,0,.70);

	/*
		12px, THE HOUSE RADIUS. The navbar bar, the hero panel, the
		video panel and section 2's cards all carry it and were
		standardised onto it on purpose (CLAUDE.md). A card on this
		page with a different corner reads as a mistake, so this token
		exists to be read, never to be re-pitched on its own.

		DELIBERATELY NOT /return-refund/'s 8px, even though the fill
		and the shadows below are taken from it. The interior pages —
		privacy, contact, return-refund — came down from 12 to 8
		together on 2026-09-30 and agree with each other; the home
		page's own bands agree on 12. This band is on the home page,
		so it follows the home page. The two conventions are allowed
		to differ because the pages never sit side by side.
	*/
	--sfdw-radius: 12px;

	/* INTER FOR THE HEADLINE AND THE FIGURES, Chakra Petch for the
	   small tracked-out labels. That is the arrangement sections 1,
	   2 and 4 already settled on, and this band is between two of
	   them — a fifth family here would read as a seam.

	   The figures are 700 where everything else is 600. Inter at
	   600 is right for a heading and reads light as a 50px numeral,
	   and the figures are the subject of this band. Raise this
	   before raising the size if they ever read as weak. */
	--sfdw-display: "Inter", system-ui, -apple-system, sans-serif;
	--sfdw-tech:    "Chakra Petch", ui-monospace, monospace;

	/* THE GUTTER IS SECTION 2's AND SECTION 4's, 80px at the top end.
	   It is deliberately NOT the navbar logo's 44 — the two bands
	   either side of this one sit at 80 and an 80/44/80 run reads as
	   a mistake. Do not "align" it without asking. */
	--sfdw-gutter: var(--sf-gutter);

	/*
		THE FIXED SOCIAL RAIL, and this band has to clear it in two
		different places because it has two kinds of content.

		.sfsr is 46px wide (40 below 782px) on the left edge of every
		page at z-index 90, vertically centred, so as you scroll every
		card in the left column passes behind it. 58 is the house
		clearance — .sfidx uses it for its row numbers, .sfcg for its
		heading, .sfbw for its brand words.

		--sfdw-rail-clear is for the HEAD, whose text sits directly on
		the section's content edge. A max() rather than a media query,
		so it is 0 on a desktop and disappears on its own if the gutter
		ever passes 58.

		IT IS NOT ON THE SECTION ANY MORE, AND THAT WAS A REAL BUG.
		It used to be folded into `padding-inline`, which sets BOTH
		sides — so a phone paid 38px of rail clearance on the RIGHT,
		where there is no rail. Measured at 375: the cards came out
		124px wide against 161 once it was taken off. Harmless while
		the cells were bare text and the air just read as generous;
		the moment they became boxes it was 76px of lost card.
	*/
	--sfdw-rail-clear: max(0px, 58px - var(--sfdw-gutter));

	/*
		--sfdw-card-indent is the same job for the CARDS, and it is a
		different sum: their text starts at the card's own padding, not
		at the section's content edge, so they need less. It clears the
		rail's real width plus 12px of air, and the max() makes it a
		no-op the moment the normal padding already does.

		Measured left-column text positions: 52 at 375 and at 768, 63
		at 1024, 88 at 1440 — against a rail ending at 40 / 40 / 46 / 46.
	*/
	--sfdw-card-pad:    clamp(18px, 1.9vw, 28px);
	--sfdw-rail-w:      46px;
	--sfdw-card-indent: max(
		var(--sfdw-card-pad),
		calc(var(--sfdw-rail-w) + 12px - var(--sfdw-gutter))
	);

	/* The figure. It was the reference's own clamp(40px, 3.7vw, 62px)
	   and went up on the operator's instruction, 2026-11-11 — "make
	   it the number more bigger". 69px at 1440, against 53 before.
	   The arithmetic that caps it is beside .sfdw__v. */
	--sfdw-v: clamp(46px, 4.8vw, 80px);

	position: relative;
	/*
		THE BOTTOM IS SHORTER THAN THE TOP — operator, 2026-11-12:
		"reduce more a space .. maybe margin bottom for second last
		section and space margin top for last section".

		It was the house clamp(72px, 7.5vw, 128px) on both edges. The
		band below, `.sfsg`, carried the same on ITS top, so the two
		simply added up: **216px at 1440** between this band's caption
		row and that band's eyebrow, both of which are 10-11px labels.
		Air between two bands is the sum of their two paddings, and
		that is easy to miss when you are reading one file.

		THE TOP KEEPS THE HOUSE VALUE. The bike ribbon above is a band
		like any other and that join has to match every other join on
		the page.

		**THIS IS HALF OF A PAIR.** `singapore.css` carries the same
		value on its top, written the same way and pointing back here.
		Change one and check the other, or the gap goes lopsided — and
		if a band is ever inserted between them, both revert to the
		house value.
	*/
	padding-top: clamp(72px, 7.5vw, 128px);
	padding-bottom: clamp(40px, 3.6vw, 64px);
	padding-inline: var(--sfdw-gutter);   /* symmetric — see --sfdw-rail-clear */

	/* ONE SURFACE FOR THE HOME PAGE. Both diagonals are needed — the
	   light one is the sheen on the tow, the dark one the shadow under
	   it, and with one alone it reads as corduroy. CLAUDE.md records
	   that the weave is deliberately louder since the base went to
	   #171717; do not re-tune it here alone. */
	background-color: var(--sfdw-page);
	background-image:
		repeating-linear-gradient( 45deg, rgba(255,255,255,.025) 0 2px, transparent 2px 4px),
		repeating-linear-gradient(-45deg, rgba(0,0,0,.35)        0 2px, transparent 2px 4px);

	color: var(--sfdw-ink);
	font-family: var(--sfdw-display);
}


/* -------------------------------------------------------------
 * 1. THE HEAD
 * ----------------------------------------------------------- */

/*
	ONE COLUMN. The head was a 7fr/5fr grid with a "See the dealer
	map" link in the right-hand track; the operator had that link
	removed on 2026-11-11, because the eighth card's "See more
	dealers" button already goes to the same page and two routes out
	of one band to one page is one too many.

	The grid went with it rather than being left holding an empty
	track — a one-child grid with a 5fr column beside it is a
	headline capped at 58% of the band for no reason any longer
	visible in the file. The .sfdw__link rules are in git.
*/
.sfdw__head{
	margin: 0 0 clamp(40px, 4.4vw, 72px);

	/* The head's text sits on the section's content edge, so it is
	   this that carries the full house clearance. The cards have
	   their own, smaller, sum — see --sfdw-card-indent. */
	padding-left: var(--sfdw-rail-clear);
}

.sfdw__eyebrow{
	margin: 0 0 clamp(16px, 1.6vw, 24px);
	font-family: var(--sfdw-tech);
	font-size: 11px;
	font-weight: 500;
	letter-spacing: .16em;
	text-transform: uppercase;
	color: var(--sfdw-accent);
}

/*
	THE HEADLINE IS ONE NOTCH UNDER SECTION 2's, ON PURPOSE.
	Section 2 carries the page's big statement at 72px (sohub's own
	ladder) and section 4's title is deliberately small because it
	sits over 128px brand words. Nothing here is bigger than a 50px
	figure, so this can be the middle of the three without competing
	with either: 46px at 1353, 49 at 1440.

	MIXED CASE, and the case is never changed alone. The whole home
	page went mixed case on 2026-11-10 and only the eyebrow and the
	PLACEHOLDER tag are uppercase now, both by role. There is no
	text-transform in this file and there must not be one.
*/
/* SECTION 2's TITLE TYPE, AND UPPERCASE WITH IT (operator,
   2026-11-12: "for last section homepage i want that tile sameeee
   title font style and uppercase like section 2"). So the weight,
   the leading, the tracking and the case are `.sfcg__title`'s, value
   for value — and this band now carries the same literals the other
   four do. **Change all five files or none**; `.sfop__title` carries
   the note on why they are literals rather than a shared token.

   **THE SIZE IS SECTION 2's LADDER TOO, AS OF THE SAME DAY**
   (operator: "Match the size too"). It was `clamp(28px, 3.4vw, 52px)`
   — 49px at 1440 — and the gap was the one thing in this band argued
   rather than copied: this band's own odometer figures run to 69px,
   so a 72px headline over them is two things of a size. That was put
   to the operator and the answer was to match. So the five files now
   share the SIZE as well as the style, and **a change to any of them
   is a change to all five.**

   THE LADDER IS STEPPED, NOT A CLAMP, because section 2's is — the
   steps and their breakpoints are sohub.digital's own, and leading
   and tracking are ratios so only the size moves. See
   category-grid.css, which carries the measured line-shape table.

   **A SIZE CHANGE RESHAPES A HEADLINE**, which is why that table
   exists: it moves where the words break. This band's two lines are
   written as two spans, so they cannot re-break — but they CAN
   overflow the measure, and at 1900 the step is 112px. Checked after
   the change at 375, 1024, 1440 and 1900. */
.sfdw__title{
	margin: 0;
	font-family: var(--sfdw-display);
	font-size: 30px;
	font-weight: 700;
	line-height: 1;
	letter-spacing: -.075em;
	text-transform: uppercase;
	color: var(--sfdw-ink);
}

@media (min-width:  768px){ .sfdw__title{ font-size:  48px } }
@media (min-width: 1024px){ .sfdw__title{ font-size:  60px } }
@media (min-width: 1280px){ .sfdw__title{ font-size:  72px } }
@media (min-width: 1900px){ .sfdw__title{ font-size: 112px } }

/* THE SHEEN, PORTED FROM `.sfcg__word` — the same ink -> magenta ->
   ink gradient clipped to the glyphs and looping for ever, so it
   reads as a light passing through the word rather than as a second
   colour. Asked for by name: "have effect at text STATE like effect
   at text PERFORMANCE".

   TWO THINGS IN IT ARE LOAD-BEARING and both are recorded elsewhere
   in this project from other bands:

     · `background-clip: text` must be written as BOTH longhands or
       Blink paints a solid magenta band with invisible type inside
       it. The `-webkit-` one is not optional;
     · the negative-margin / positive-padding pair gives the last
       glyph the room the background box would otherwise clip.

   IT RIDES ON `.sfdw__wi`, THE WORD THE SPLITTER ALREADY BUILDS, so
   the sheen word takes part in the reveal like every other word —
   see the note in dealer-wall.js for why it cannot be a span in the
   markup. */
.sfdw__wi.is-sheen{
	margin-right: -.12em;
	padding-right: .12em;

	color: transparent;
	background-image: linear-gradient(90deg,
		var(--sfdw-ink)    0%,
		var(--sfdw-accent) 48%,
		var(--sfdw-ink)   100%);
	background-size: 220% 220%;
	-webkit-background-clip: text;
	background-clip: text;
	animation: sfdwSheen 5s ease infinite;
}

@keyframes sfdwSheen{
	0%   { background-position: 10% 0 }
	50%  { background-position: 91% 100% }
	100% { background-position: 10% 0 }
}

/* The loop is decoration, so it stops where motion is unwelcome —
   the word keeps the gradient, it simply holds still. */
@media (prefers-reduced-motion: reduce){
	.sfdw__wi.is-sheen{ animation: none }
}

.sfdw__title .sfdw__line{ display: block }

/*
	THE WORD REVEAL, PORTED FROM THE REFERENCE AS-IS.

	Each word sits in its own overflow:hidden mask and rises from
	140% with a 7deg rotation and an 8px blur. The rotation and the
	blur are what make it read as type arriving rather than as a list
	sliding, and they are the half most ports leave out.

	THE PADDING / MARGIN PAIR ON THE MASK IS LOAD-BEARING. Without it
	the box shaves the descenders off a word sitting still. navbar.css
	and bike-wall.css both record the same trick; the values here are
	the reference's own.
*/
.sfdw__wd{
	display: inline-block;
	overflow: hidden;
	vertical-align: top;
	padding: .18em .1em .24em;
	margin:  -.18em -.1em -.24em;
}

.sfdw__wi{
	display: inline-block;
	transform-origin: 0 100%;
	transition:
		transform 1.05s cubic-bezier(.16,1,.3,1),
		opacity    .8s cubic-bezier(.22,1,.36,1),
		filter     .9s cubic-bezier(.22,1,.36,1);
	transition-delay: calc(var(--i, 0) * 60ms);
}

/*
	THE FAILURE MODE MUST BE "ALREADY REVEALED". The start state is
	scoped to .sfdw--js, which dealer-wall.js adds to the section as
	its first act — so with no JS, or with the script blocked, the
	words are simply there. Reduced motion is handled at the foot of
	this file by clearing the same rule.
*/
.sfdw--js:not(.is-in) .sfdw__wi{
	transform: translate3d(0, 140%, 0) rotate(7deg);
	opacity: 0;
	filter: blur(8px);
}

/* .sfdw__link lived here — the head's "See the dealer map", removed
   2026-11-11. Deleted rather than left dormant: it had no consumer
   the moment the markup went, and this file already carries enough
   that is kept on purpose. It is in git. */


/* -------------------------------------------------------------
 * 2. THE GRID
 * ----------------------------------------------------------- */

.sfdw__grid{
	display: grid;
	grid-template-columns: repeat(4, minmax(0, 1fr));
	gap: clamp(14px, 1.5vw, 22px);
	margin: 0;
	padding: 0;
	list-style: none;
}

/*
	EVERY CELL IS A REAL CARD — operator, 2026-11-11: "i want you make
	it the card for each states , and remove line also please redesign
	the card".

	THE DOTTED RULE AND ITS MAGENTA STUB ARE GONE. They were the
	reference's whole visual signature — every divider on dozfx.com is
	a repeating radial-gradient of dots with a short solid accent
	segment at the left end — and a rule plus a box is one device too
	many: the card's own edge already does the separating. Both are in
	git if that decision is ever revisited, and the study at
	docs/research/2026-11-11-dozfx-stats-and-handoff.md has the values.

	WHAT THE BOX HAD TO BUY BACK. The rule was also the card's
	ENTRANCE — a 56px bar drawing across it on the count. With it gone
	the figures would simply have appeared, so .sfdw__card below rises
	and fades on the same --d instead. The stagger the operator asked
	to keep is unchanged; only what carries it has moved.

	THE GAP CAME DOWN HARD, 56-96px to 14-22px. Those were the gaps
	between bare columns of text, where the air IS the separation.
	Boxed cards separate themselves, and at the old row gap the two
	rows read as two unrelated groups.
*/
.sfdw__card{
	position: relative;
	display: flex;
	flex-direction: column;
	gap: 14px;
	padding: var(--sfdw-card-pad);
	padding-left: var(--sfdw-card-indent);

	/*
		GUNMETAL GREY, the operator's own colour (2026-11-11). It is
		NOT in theme.json — the palette's two panel colours are
		accent-2 #131317 and accent-3 #15151a, both near-black, and a
		card in either would be invisible against the #171717 page.
		So it is a literal here, declared once as a token.

		There IS a "gunmetal" in hero.css, --sf-hero-metal #9aa0aa.
		That is a light silver for icon strokes and has nothing to do
		with this; do not consolidate the two.
	*/
	background-color: var(--sfdw-card);
	border: 0;
	border-radius: var(--sfdw-radius);

	/* Raised: light edge up-left, dark edge down-right, no border —
	   the light IS the edge, exactly as the footer's word has it. */
	box-shadow:
		-4px -4px  8px var(--sfdw-lift),
		 5px  5px 12px var(--sfdw-sink);

	/*
		THE EMBOSS DOES NOT CHANGE ON HOVER, because the footer's
		word does not either — it keeps its carving and only lights
		up. A card that deepens its own shadow AND lights pink is two
		effects where the reference has one.
	*/
	transition:
		transform .9s cubic-bezier(.16,1,.3,1),
		opacity   .7s cubic-bezier(.22,1,.36,1);
	transition-delay: calc(var(--d, 0ms) + .1s);

	/* the pink light rides on ::after and must be clipped to the
	   card's own corner */
	overflow: hidden;
	isolation: isolate;
}

/*
	THE PINK LIGHT, PORTED FROM THE FOOTER'S LIT COPY.

	footer.html fills a second copy of the word with a radial
	gradient — #ec1e79 at .90, .32 at 45%, 0 at the edge — and
	footer.js eases its centre toward the cursor, so the letters catch
	a light that trails the pointer. dealer-wall.js does the same here
	with --sfdw-mx / --sfdw-my.

	THE ALPHA IS MUCH LOWER THAN THE FOOTER'S AND HAS TO BE. There the
	gradient is clipped to the glyphs, so .90 magenta covers maybe a
	third of the band and reads as light caught on an edge. A card is
	a filled rectangle: the same .90 is a pink blob the size of the
	card. .26 falling to .09 puts about the same amount of magenta on
	screen.

	The 215px radius IS the footer's own r, unscaled — it is close
	enough to a card's own size to read the same way, and a radius
	tied to the card would change the feel between the four-column and
	the one-column layouts.
*/
.sfdw__card::after{
	content: "";
	position: absolute;
	inset: 0;
	pointer-events: none;
	opacity: 0;
	transition: opacity .35s ease;   /* the footer's own fade */
	background: radial-gradient(
		circle 215px at var(--sfdw-mx, 50%) var(--sfdw-my, 50%),
		rgba(236,30,121,.26) 0%,
		rgba(236,30,121,.09) 45%,
		rgba(236,30,121,0)  100%
	);
}

.sfdw__card.is-lit::after{ opacity: 1 }

/*
	The card's own entrance, on the same beat as its figure. Small
	travel on purpose — a boxed card sliding a long way reads as a
	carousel, and the odometer is meant to be the thing you watch.
*/
.sfdw--js .sfdw__card:not(.is-count){
	opacity: 0;
	transform: translate3d(0, 18px, 0);
}

/*
	NO :hover RULE ON THE CARD. The light is .is-lit, set by
	dealer-wall.js from real pointer events, because the gradient's
	centre has to follow the cursor and CSS alone cannot know where it
	is. A :hover fallback would fire with the centre stuck at 50% 50%,
	which is a worse effect than none — so a visitor on a device with
	no fine pointer simply gets the carved card, as the footer's word
	does at the same widths.
*/

/*
	--d IS THE CARD'S WHOLE ENTRANCE, IN ONE NUMBER. The accent bar,
	the odometer roll and the unit all read it, so a card arrives as
	one event rather than three that happen to agree.

	IT RESTARTS ON EACH ROW, AND THE ROW LENGTH IS THE COLUMN COUNT.
	The reference computes it from the card index with a modulo of 4
	and leaves it at that, so on their own two-column phone layout the
	fourth card waits 420ms behind a card it is no longer beside. The
	nth-child pair below re-deals it at the breakpoint instead.

	IT IS IN CSS, NOT IN THE MARKUP, AND IT HAS TO BE. An inline
	style:"--d:420ms" beats any stylesheet rule, media query or not —
	so the typed version of this could not be corrected for the narrow
	layout at all. dealer-wall.js therefore reads it off getComputedStyle,
	never off element.style, which returns nothing for a value that
	arrives this way.
*/
.sfdw__card:nth-child(4n+1){ --d:   0ms }
.sfdw__card:nth-child(4n+2){ --d: 140ms }
.sfdw__card:nth-child(4n+3){ --d: 280ms }
.sfdw__card:nth-child(4n+4){ --d: 420ms }

.sfdw__n{
	font-family: var(--sfdw-tech);
	font-size: 10px;
	font-weight: 500;
	letter-spacing: .14em;
	font-variant-numeric: tabular-nums;
	color: var(--sfdw-faint);
}

/*
	PINK AND BIGGER — operator, 2026-11-11: "for the number dealer
	change to pink color and make it the number more bigger".

	It was Inter 700 at the reference's own clamp(40px, 3.7vw, 62px)
	in ink. Now accent-1 at clamp(46px, 4.8vw, 80px) — 69px at 1440,
	against 53 before.

	THE SIZE IS CAPPED BY THE CARD, NOT BY TASTE. Four columns of a
	1430px band at a 20px gap leave each card about 342px wide, less
	2 x 28px of padding = 286px of content. "157" at this ramp measures
	about 118px, so there is room; a four-digit figure would not fit
	and the ramp would have to come down with it. Redo that sum before
	raising it again.
*/
.sfdw__v{
	margin: 0;
	display: flex;
	align-items: baseline;
	font-family: var(--sfdw-display);
	font-size: var(--sfdw-v);
	font-weight: 700;
	line-height: .9;
	letter-spacing: -.035em;
	color: var(--sfdw-accent);
}

/*
	500, NOT 400, AND THAT IS NOT A TASTE CALL. The shared
	`superfly-inter` handle requests wght 500;600;700 and is registered
	byte-identically from three enqueues — asking for 400 here would
	not add it, it would hand the browser a synthesised light cut of
	the 500 outline. Widening that URL is a decision for the whole
	page, not for one paragraph. CLAUDE.md records the same trap on
	the navbar's own font handle from the heavy end.
*/
.sfdw__t{
	margin: 0;
	max-width: 16em;
	font-family: var(--sfdw-display);
	font-size: 15px;
	font-weight: 500;
	line-height: 1.4;
	color: var(--sfdw-dim);
}

/* The state is the subject of the card; the towns are the detail. */
.sfdw__st{
	font-weight: 600;
	color: var(--sfdw-ink);
}

.sfdw__card:hover .sfdw__v{
	text-shadow: 0 0 30px rgba(236,30,121,.35);
}


/* -------------------------------------------------------------
 * 3. THE ODOMETER
 * -------------------------------------------------------------
 * Built by dealer-wall.js. Each digit is a vertical strip of <i>
 * elements that slides up by exactly --k line heights.
 *
 * Three things here are load-bearing and all three are easy to
 * "tidy" into a bug:
 *
 *  - THE GHOST DIGIT. .sfdw__g is a hidden copy of the final digit
 *    that gives the column its real width. The strip is absolute,
 *    so without it the column is zero-wide and a 1 and an 8 take
 *    the same space.
 *  - clip-path: inset(0 -.35em) RATHER THAN overflow: hidden. It
 *    clips vertically but leaves horizontal room, so the negative
 *    tracking and the hover glow are not sheared off.
 *  - --lh MUST EQUAL THE FIGURE'S OWN line-height (.9 above), or
 *    every digit lands a fraction off its baseline.
 * ----------------------------------------------------------- */

.sfdw__od{ display: inline-flex; --lh: .9em }

/*
	The strip is forty-odd loose digits as far as the accessibility
	tree is concerned, so the script hides it and this carries the
	real figure. Defined here rather than leaning on .sr-only or
	WordPress's .screen-reader-text: neither is guaranteed in the
	local harness, and a figure that renders twice is a worse bug
	than one that is not announced.
*/
.sfdw__sr{
	position: absolute;
	width: 1px;
	height: 1px;
	padding: 0;
	margin: -1px;
	overflow: hidden;
	clip-path: inset(50%);
	white-space: nowrap;
	border: 0;
}

.sfdw__d{
	position: relative;
	display: inline-block;
	clip-path: inset(0 -.35em);
}

.sfdw__g{ visibility: hidden }

.sfdw__s{
	position: absolute;
	left: 50%;
	top: 0;
	display: flex;
	flex-direction: column;
	transform: translate3d(-50%, 0, 0);
	will-change: transform;
}

.sfdw__s i{
	display: block;
	height: var(--lh);
	line-height: var(--lh);
	font-style: normal;
	text-align: center;
}

/*
	--j is the digit's index from the LEFT, so each column runs
	LONGER and starts LATER than the one before it. Combined with the
	strip length (the script gives column j exactly j+2 full
	revolutions) the number resolves left to right and the last digit
	is still moving when the first has stopped.

	THE BLUR IS WHAT SELLS THE SPEED. Without it the roll reads as a
	list sliding past a window. It peaks a tenth of the way in and is
	gone by the end, so nothing is ever blurred at rest.

	--d IS INHERITED FROM THE CARD, not copied onto the strip by the
	script. Custom properties inherit, so the strip picks up whichever
	value the breakpoint is currently handing its card — copy it in JS
	instead and it goes stale the moment the window crosses 1023px.
	The reference copies it and has exactly that bug.
*/
.sfdw__od.is-on .sfdw__s{
	transform: translate3d(-50%, calc(var(--k) * -1 * var(--lh)), 0);
	transition: transform
		calc(1.5s + var(--j) * .25s)
		cubic-bezier(.12,.82,.16,1)
		calc(var(--d, 0ms) + var(--j) * 70ms);
	animation: sfdwBlur
		calc(1.5s + var(--j) * .25s)
		ease-out
		calc(var(--d, 0ms) + var(--j) * 70ms) both;
}

@keyframes sfdwBlur{
	0%   { filter: blur(0) }
	12%  { filter: blur(2.4px) }
	70%  { filter: blur(.4px) }
	to   { filter: blur(0) }
}

/*
	The unit is held back and brought in after the roll has landed.

	NOTHING USES IT TODAY, and that is deliberate rather than leftover.
	The reference's figures are all approximations and carry a "+"; a
	dealer count is exact, so printing one here would be a claim the
	data does not support. The rule stays because the moment a figure
	does need a "+", a "/7" or an "h", it is one span in the markup.
*/
.sfdw__u{
	display: inline-block;
	margin-left: .04em;
	font-size: .72em;
	color: var(--sfdw-accent);
	opacity: 0;
	transform: translate3d(0, .2em, 0);
}
.sfdw__card.is-count .sfdw__u{
	opacity: 1;
	transform: none;
	transition:
		opacity   .6s cubic-bezier(.22,1,.36,1) calc(1.25s + var(--d, 0ms)),
		transform .9s cubic-bezier(.16,1,.3,1)  calc(1.25s + var(--d, 0ms));
}


/* -------------------------------------------------------------
 * 4. THE EIGHTH CELL — the way out, marked by light
 * -------------------------------------------------------------
 * THE TILE IS MARKED BY A MAGENTA BLOOM, NOT BY A MAGENTA FILL.
 * Chosen by the operator on 2026-11-11 from five variants built
 * side by side; the page that offered them is
 * docs/home-concepts/dealer-note-variants.html, which still holds
 * the other four as scoped overrides if any of them is ever wanted.
 *
 * THE ROUTE TO HERE IS WORTH KNOWING, because it crossed its own
 * tracks twice and the file should not pretend otherwise:
 *
 *   1. It was a bordered note card over the page colour with a
 *      bloom in its corner — the dozfx reference's own arrangement.
 *   2. "make it the card match with new card" made it identical to
 *      the other seven, and the note written then said that if it
 *      ever needed to stand out, the HAIRLINE should go magenta
 *      rather than the fill.
 *   3. "background change color pink" made the whole fill magenta,
 *      overruling that.
 *   4. "did you have another design" produced the five, and this is
 *      the one picked: the fill goes back to the house card and the
 *      magenta returns as light.
 *
 * So it is close to where it started, and that is fine — what the
 * loop settled is WHY: the eighth cell is not a peer of the other
 * seven (they are figures, it is the way out) so it has to be
 * marked, but marking it by fill takes it out of the grid
 * altogether. Light marks it and keeps it a member.
 *
 * IT IS ALSO THE SAME IDEA AS THE REST OF THE BAND. The cards are
 * the footer wordmark's emboss and the hover is the footer's lit
 * copy; this is a third use of the same light. Nothing here is a
 * new device.
 *
 * THE EMBOSS AND THE FILL ARE UNCHANGED from .sfdw__card — only a
 * background-image is added — so this is the same card with a light
 * on it rather than a different object. Do not re-pitch the shadows
 * for this one alone.
 * ----------------------------------------------------------- */

.sfdw__card--note{
	justify-content: space-between;
	gap: 20px;

	/*
		From the top-right corner, because that is the corner the
		emboss does NOT light — the carving is lit from the top-left,
		so a bloom there would double up on it and read as a blown
		highlight rather than as a separate light.
	*/
	background-image: radial-gradient(120% 110% at 88% 0%,
		rgba(236,30,121,.55)  0%,
		rgba(236,30,121,.16) 45%,
		rgba(236,30,121,0)   72%);
}

/*
	NO ::after OVERRIDE. The hover light is the plain magenta radial
	every card gets, and on this tile it simply arrives on top of a
	surface that already carries some — which reads as the same light
	moving rather than as a second effect.
*/

.sfdw__q{
	margin: 0;
	font-family: var(--sfdw-display);
	font-size: clamp(18px, 1.5vw, 23px);
	font-weight: 600;
	line-height: 1.15;
	letter-spacing: -.015em;
	color: #fff;
}

/* The same two-tone the dark cards use: the figure in magenta, the
   sentence in near-white. On the magenta-FILL version this had to
   inverted to near-black; with the fill gone it goes back. */
.sfdw__q em{
	font-style: normal;
	color: var(--sfdw-accent);
}

/*
	THE PILL IS MAGENTA AGAIN. It went near-black while the card was
	a magenta fill — a contrast decision, not a taste one: a white
	pill with a magenta label puts 13px text at about 4.2:1, under
	the 4.5:1 AA wants at that size. On this dark surface the magenta
	pill with white text measures about 4.2:1 as well, which is the
	same arithmetic — SO KEEP THE LABEL AT 13px/600 OR LARGER, and
	do not set anything smaller in magenta on this tile.
*/
.sfdw__cta{
	align-self: start;
	display: inline-flex;
	align-items: center;
	gap: 10px;
	padding: 11px 18px;
	border-radius: 999px;
	background: var(--sfdw-accent);
	font-family: var(--sfdw-display);
	font-size: 13px;
	font-weight: 600;
	line-height: 1;
	letter-spacing: .01em;
	color: #fff;
	text-decoration: none;
	transition: transform .4s cubic-bezier(.16,1,.3,1), box-shadow .4s ease;
}
.sfdw__cta:hover{
	transform: translateY(-2px);
	box-shadow: 0 10px 28px -10px rgba(236,30,121,.9);
	color: #fff;
}
.sfdw__cta span[aria-hidden]{
	transition: transform .4s cubic-bezier(.16,1,.3,1);
}
.sfdw__cta:hover span[aria-hidden]{ transform: translate(3px, -3px) }


/* -------------------------------------------------------------
 * 5. THE CAPTION — under the grid, outside it
 * ----------------------------------------------------------- */

.sfdw__caption{
	display: flex;
	justify-content: space-between;
	gap: 16px;
	margin: clamp(32px, 3.4vw, 56px) 0 0;
	font-family: var(--sfdw-tech);
	font-size: 10px;
	font-weight: 500;
	letter-spacing: .14em;
	text-transform: uppercase;
	color: var(--sfdw-faint);
}


/* -------------------------------------------------------------
 * 6. NARROWER
 * -------------------------------------------------------------
 * The reference goes to two columns at 1024 and stays there. Four
 * columns of a 69px figure plus a 16em sentence do not survive a
 * tablet, and one column makes the band eight screens tall.
 * ----------------------------------------------------------- */

@media (max-width: 1023px){
	.sfdw__grid{ grid-template-columns: repeat(2, minmax(0, 1fr)) }
	.sfdw__card--note{ grid-column: 1 / -1 }

	/* Two columns, so the delay restarts every two cards. See the
	   note beside the 4n rules — this is the whole reason --d is not
	   written into the markup. */
	.sfdw__card:nth-child(2n+1){ --d:   0ms }
	.sfdw__card:nth-child(2n+2){ --d: 140ms }
}

/*
	ONE COLUMN BELOW 600px, AND THE REASON IS THE BOX, NOT THE WIDTH.
	The reference stays at two columns all the way down — but its
	cells are bare text, so a narrow column just wraps. A boxed card
	spends its width twice over, on its own padding and on the rail
	indent: measured at 375 in two columns, each card was 161px wide
	and left 111px of content, which put "shops across Johor, from
	Johor Bahru to Segamat" on five lines under a 46px figure.

	One column gives 335px wide and 285px of content — and the band
	is no taller for it, because each card is now short.
*/
@media (max-width: 599px){
	.sfdw__grid{ grid-template-columns: minmax(0, 1fr) }

	/*
		One column, so every card is its own row and there is nothing
		to stagger — each one has its own observer and arrives on its
		own scroll position.

		:nth-child(n) IS NOT DECORATION. The rules it has to beat are
		.sfdw__card:nth-child(2n+1) at 0,2,0; a plain .sfdw__card here
		is 0,1,0 and would lose however late it appears in the file.
	*/
	.sfdw__card:nth-child(n){ --d: 0ms }

	.sfdw__card{ gap: 10px }
	.sfdw__t{ font-size: 14px; max-width: none }
	.sfdw__caption{
		flex-direction: column;
		gap: 6px;
	}
}


/* -------------------------------------------------------------
 * 7. REDUCED MOTION
 * -------------------------------------------------------------
 * Everything lands at its end state. The odometer strip is left in
 * place and the script sets --k to 0 instead, so the correct digit
 * is the one on screen — see the note in dealer-wall.js.
 * ----------------------------------------------------------- */

@media (prefers-reduced-motion: reduce){
	.sfdw--js:not(.is-in) .sfdw__wi{
		transform: none;
		opacity: 1;
		filter: none;
	}
	.sfdw__wi{ transition: none }

	/* The cards' own entrance replaced the accent bar's when the rule
	   was removed, so it is this that has to be cleared now. */
	.sfdw--js .sfdw__card:not(.is-count),
	.sfdw__card{
		opacity: 1;
		transform: none;
		transition: none;
	}

	/* The pink light still appears on hover — it is a colour change,
	   not motion, and the gradient's centre moving with the cursor is
	   the pointer's own movement rather than an animation. Only the
	   fade is dropped. */
	.sfdw__card::after{ transition: none }

	.sfdw__u{ opacity: 1; transform: none }
	.sfdw__od.is-on .sfdw__s{ transition: none; animation: none }
	.sfdw__cta,
	.sfdw__cta span[aria-hidden]{ transition: none }
}
