/*
	THE PRODUCT GALLERY IS A SLIDER (2026-11-17)
	============================================

	Operator: replace the single-product interface with the design built
	on localhost — "i dont want hardcoded think".

	WHAT IT REPLACES. `single-product.css` lays WooCommerce's gallery
	out as a two-column grid of square tiles, and
	`single-product-gallery.js` pads that grid to six with
	`.sf-pdp__slot` plates so the shape does not depend on how many
	photographs a product has. Measured on product 983 before this:
	647x982, `grid-template-columns: 311.45px 311.45px`, six children —
	three real images and three empty plates.

	A slider shows one photograph at the panel's full width instead, so
	the picture is roughly four times the area it had, and the empty
	plates are not needed at all.

	**EVERY IMAGE COMES FROM THE PRODUCT.** `pdp-gallery.js` reads
	whatever WooCommerce rendered — no list, no count, no filenames
	anywhere in here. A product with one image gets no controls; a
	product with nine gets nine dots.

	IT KEEPS THE PADDING PLATES OUT RATHER THAN DELETING THEM: the
	script hides them, so `single-product-gallery.js` can stay exactly
	as it is and removing this stylesheet restores the old grid.
*/

/* ------------------------------------------------------------------ */
/*  The track                                                          */
/* ------------------------------------------------------------------ */
/*
	Scoped to `.sf-gal`, which the script adds only once it has taken
	over — so with JS off the grid behaves exactly as it does today.
	That is the fallback, and it is why nothing here is written
	against WooCommerce's class names on their own.
*/
.sf-gal{
	position:relative;
	overflow:hidden;
	border-radius:var(--sp-radius, 12px);
}

.sf-gal .sf-gal__track{
	display:flex !important;
	grid-template-columns:none !important;
	gap:0 !important;
	width:100%;
	transition:transform .62s cubic-bezier(.22,1,.36,1);
	will-change:transform;
	touch-action:pan-y;
}

/*
	`!important` ON THE SLIDE, AND IT IS EARNED. `single-product.css` is
	166KB of rules against WooCommerce's own class names, several of
	them `#` or five-class chains that set the tile's width, aspect and
	position. Matching each would mean writing selectors against another
	file's internals that break the day one is renamed. This is a
	deliberate takeover of one component, scoped to a class the script
	only adds once it has taken over.
*/
.sf-gal .sf-gal__track > .sf-gal__slide{
	position:relative !important;
	flex:none !important;
	width:100% !important;
	min-width:100% !important;
	max-width:100% !important;
	margin:0 !important;
	opacity:1 !important;
	/* their grid gives the tile an aspect; the slide keeps a square so
	   the panel height does not jump between portrait and landscape
	   photographs */
	aspect-ratio:1 / 1;
}
.sf-gal .sf-gal__track > .sf-gal__slide img{
	width:100% !important;
	height:100% !important;
	max-width:none !important;
	object-fit:contain;
	/* `contain`, not `cover`: these are studio shots of parts on white,
	   and cropping a rim to fill a square loses the rim. The panel's
	   own fill does the framing. */
}

/* the padding plates have no job in a slider */
.sf-gal .sf-pdp__slot{ display:none !important; }

/* ------------------------------------------------------------------ */
/*  Controls                                                           */
/* ------------------------------------------------------------------ */
.sf-gal__nav{
	position:absolute;
	left:16px;
	right:16px;
	bottom:16px;
	z-index:4;
	display:flex;
	align-items:center;
	justify-content:space-between;
	gap:12px;
	pointer-events:none;
}
.sf-gal__nav > *{ pointer-events:auto; }

.sf-gal__btn{
	width:44px;
	height:44px;
	display:grid;
	place-items:center;
	border:0;
	border-radius:50%;
	cursor:pointer;
	background:rgba(13,13,16,.72);
	backdrop-filter:blur(8px);
	transition:background .25s ease, opacity .25s ease;
}
.sf-gal__btn svg{ width:16px; height:16px; display:block; }
.sf-gal__btn path{
	stroke:var(--wp--preset--color--accent-1, #ec1e79);
	stroke-width:1.8;
	fill:none;
}
.sf-gal__btn:hover{ background:rgba(13,13,16,.92); }
/* disabled, not hidden: a control that vanishes shifts the row it is
   in and makes the other one jump sideways */
.sf-gal__btn[disabled]{ opacity:.3; cursor:default; }

/*
	NINETEEN DOTS DO NOT FIT A PHONE, and until the deck arrived nobody
	could see that. Measured at 375 on the product with nineteen
	photographs: the row is 308px wide inside a 234px nav, so it ran
	past the card AND pushed the next arrow 69px off the page.

	THE PANEL WAS HIDING IT. `.sf-gal` is `overflow: hidden` for the
	slider, so the over-wide control was simply cropped at the panel
	edge -- wrong, but invisible, and it reported no page overflow
	because the clip ate it. The deck has to let the fan lean out, so
	it is `overflow: visible`, and the same row then escaped all the
	way to the document. Measured both ways: deck off 0px of page
	overflow, deck on 69px, with the dots row identical at 308px.

	WRAPPING RATHER THAN SHRINKING, and no number for the count. A
	smaller dot or a tighter gap would be fitted to nineteen and wrong
	at thirty; `min-width: 0` lets the row give up its share of the nav
	and `flex-wrap` puts what is left on a second line. Three
	photographs look exactly as they did.
*/
.sf-gal__dots{
	display:flex;
	flex:0 1 auto;
	min-width:0;
	flex-wrap:wrap;
	justify-content:center;
	gap:8px;
	align-items:center;
}
.sf-gal__dot{
	width:8px;
	height:8px;
	padding:0;
	border:0;
	border-radius:50%;
	cursor:pointer;
	background:rgba(13,13,16,.3);
	transition:background .25s ease, width .25s ease;
}
/*
	The current dot is the accent AND a lozenge. On a near-white panel
	an 8px magenta dot beside two grey ones is easy to miss; the width
	change reads at a glance.
*/
.sf-gal__dot[aria-current="true"]{
	width:20px;
	border-radius:4px;
	background:var(--wp--preset--color--accent-1, #ec1e79);
}

.sf-gal.is-dragging .sf-gal__track{ transition:none; cursor:grabbing; }

@media (prefers-reduced-motion: reduce){
	.sf-gal .sf-gal__track{ transition:none; }
}

@media (max-width: 760px){
	.sf-gal__nav{ left:12px; right:12px; bottom:12px; }
	.sf-gal__btn{ width:36px; height:36px; }
}

/* ------------------------------------------------------------------ */
/*  The panel follows the scroll                                       */
/* ------------------------------------------------------------------ */
/*
	Operator, 2026-11-18, against yucca.co.za/product/kraft-pouch-bag-with-window/:
	"if you try to open to accordion, the image will stretch height too" --
	and then, asked which of two readings they meant, "i want you to follow
	exactly same 100% what yucca's website is working for the design".

	WHAT YUCCA ACTUALLY DOES, measured on its live page at 1440 rather than
	guessed at: the picture does NOT stretch. It stays SQUARE and is PINNED.
	Its row is `display:flex; align-items:stretch`, so the image column takes
	the height of the text column -- left 603x806 against right 483x806 --
	and a GSAP ScrollTrigger pin holds a 603x603 panel inside it. The spacer
	the pin leaves behind carries `padding-bottom: 204px`, which IS the
	travel: 806 less the 603 panel.

	So opening the accordion makes the TEXT column taller, the image column
	stretches to match, and the pin therefore has further to ride. That is
	the "stretch" in the report -- the COLUMN grows, not the photograph.

	OURS ALREADY DID THE FIRST HALF AND NOBODY HAD NOTICED. `.sf-pdp`
	computes `align-items: normal`, which for a flex row is stretch, so the
	gallery column has always matched the rail: measured 654x1345 against
	654x1345, and 654x1610 with all three cards open. The square panel simply
	sat at the top of it with ~750px of empty column underneath. This is the
	second half.

	WHY THIS IS NOT A ScrollTrigger PIN, which is the one place it is
	deliberately not 100% theirs. CLAUDE.md records at length what a pin cost
	this project: ScrollTrigger writes an inline pixel width onto a
	`.pin-spacer` at the moment the pin is created, nothing re-derives it,
	and the home page carried a horizontal scrollbar on every cold load
	because of it. Every band in this repo has used `position: sticky`
	since. The painted result is the same and there is no spacer to mis-size.

	SCOPED TO `:has(.sf-gal)` SO IT ONLY APPLIES TO THE SLIDER. With JS off
	the script never takes over and `single-product.css` lays the gallery out
	as its two-column grid of tiles -- which is taller than the viewport and
	has no business being stuck to anything. A browser without `:has()` gets
	that same fallback, which is the right way round.
*/
.sf-pdp__gallery{
	/*
		THEIRS PINS 20px FROM THE TOP OF THE VIEWPORT, measured by stepping
		the scroll with dispatched events (the pane fires none of its own):
		the panel rides down to top:20 at scrollY 147 and holds there until
		~351, then releases. 204px of travel, which is the spacer's figure.

		THE 20 IS MEASURED FROM THE SCREEN EDGE THERE AND FROM THE NAVBAR
		HERE, and that is the one deliberate difference. Their header is
		`position: fixed`, 80px tall, and does NOT hide on scroll -- checked
		at scrollY 0, 300 and 700, top 0 bottom 80 every time -- so the top
		60px of their pinned panel sits BEHIND it. On a near-white header
		over a near-white panel that is hard to see. Reproducing it here
		would run a dark magenta-and-carbon navbar pill across the top of
		the photograph, so the 20px of air is kept and moved below the bar.

		DERIVED FROM THE NAVBAR TOKENS, NOT HARDCODED AT 90. navbar.css
		declares all three on `:root` and is enqueued on every page, so this
		is safe to read across files -- the trap CLAUDE.md records is a
		token declared in a CONDITIONALLY loaded sheet, which resolves by
		load order and fails silently. The bar is its gutter plus its own
		height, and its height is the logo plus its two paddings:
		24 + 48 + 9 + 9 = 90, which is what the bar measures live.

		THAT SUM HOLDS ONLY WHILE THE LOGO IS WHAT SETS THE BAR'S HEIGHT.
		CLAUDE.md says it is, and the fallbacks are the present values, so a
		missing token degrades to the right number rather than to zero.
		Re-measure the bar if it ever grows a second row.
	*/
	--sf-gal-pin-air: 20px;
	--sf-gal-pin-top: calc(
		var(--sf-navbar-gutter, 24px)
		+ var(--sf-navbar-logo-height, 48px)
		+ var(--sf-navbar-pad-y, 9px) * 2
		+ var(--sf-gal-pin-air)
	);
}

.sf-pdp__gallery .wp-block-woocommerce-product-image-gallery:has(.sf-gal){
	position: sticky;
	top: var(--sf-gal-pin-top);
}

/*
	OFF ONCE THE ROW STACKS. single-product.css section 7 gives both columns
	`flex: 1 1 100%` at this width, so the gallery column is no longer the
	height of the rail -- it is its own content, the picture is the whole of
	it, and there is nothing left to travel. Sticky would be a no-op rather
	than a bug, and it is turned off so that stays true if the stack is ever
	re-pitched. 1000px is section 7's own breakpoint, not a new one.
*/
@media (max-width: 1000px){
	.sf-pdp__gallery .wp-block-woocommerce-product-image-gallery:has(.sf-gal){
		position: static;
	}
}


/* ------------------------------------------------------------------ */
/*  The fan deck                                                      */
/* ------------------------------------------------------------------ */
/*
	THE PHOTOGRAPHS AS A DECK OF CARDS (2026-11-19)
	===============================================

	Operator, after being shown nine layouts on localhost and the page
	heights each costs at nineteen photographs: "Build the deck instead".

	WHAT IT FIXES. The slider above shows one photograph and says the
	rest exist with eight small dots. The complaint was that nobody
	reads the dots -- "i think user not know there have another image".
	The deck answers that with depth rather than with a control: the
	next three photographs sit fanned out behind the top one, so the
	FIRST FRAME already shows there is more. Nothing to read, nothing
	to find.

	WHAT IT GIVES UP is any sense of HOW MANY. Three cards behind looks
	identical at four photographs and at nineteen. The dots are kept
	for exactly that reason -- they are the count, and the fan is the
	invitation.

	**IT IS A SCOPED TAKEOVER, NOT A REWRITE.** Everything above still
	governs; this only applies once the script adds `sf-gal--deck`, and
	deleting that one word in pdp-gallery.js puts the slider back with
	no CSS change. The dots, the arrows, the keyboard, the lazy-image
	forcing and the variation reset are all shared and untouched.

	WHY THE DECK IS INSET rather than filling the panel. The fan leans
	up and to the RIGHT, so at full width it would lean into the buy
	rail -- measured in the lab at 1430: a full-width deck put its
	painted edge 15px INSIDE the rail's left edge. The track is
	`--sf-gal-deck-w` of the column and anchored LEFT, so the fan opens
	into space the gallery already owns. Measured after: 50px of clear
	carbon between the furthest card and the rail.

	AND THE PANEL HAS TO STOP CLIPPING. `.sf-gal` is `overflow: hidden`
	with the radius on it, which is right for a track that slides
	through it and wrong for a fan that leans out of it -- it would cut
	the three cards off at the panel's edge, which is precisely the
	signal this layout exists to give. The radius moves onto the card,
	which is now the thing you look at.
*/
.sf-gal--deck{
	/* the fan leans out of the box, so the box cannot clip */
	overflow:visible;
	border-radius:0;
	/*
		82% of the column. Raising it costs the clearance against the
		rail; lowering it costs the photograph. Re-measure the furthest
		card's painted edge against `.sf-pdp__rail` if it moves.
	*/
	--sf-gal-deck-w:82%;
}

.sf-gal.sf-gal--deck .sf-gal__track{
	/*
		`!important` for the same reason the base rules carry it: this is
		overriding `display:flex !important` two rules up, on a class the
		script only adds once it has taken the component over.
	*/
	display:block !important;
	position:relative;
	/*
		TWO CLASSES ON `.sf-gal` AND AN `!important`, AND BOTH ARE EARNED.
		single-product.css sets `width: 100% !important` on this element
		at (0,2,0) and loads AFTER this file, so an equal-specificity
		`!important` still loses on source order. Measured before the fix:
		the track laid out full width, the fan leaned 42px INTO the buy
		rail, and the rule was present, matching and important.

		`.sf-gal.sf-gal--deck` is (0,3,0), which is the cheapest way past
		it -- both classes are on the same element and the script adds
		them together, so this cannot match something the deck does not.
	*/
	width:var(--sf-gal-deck-w) !important;
	margin-right:auto;
	/*
		THE SQUARE IS AN ASPECT-RATIO, AND THE OBVIOUS WAY DOES NOT WORK
		HERE. The usual trick for a box whose children are all absolute is
		`height: 0` plus a percentage `padding-bottom`, which resolves
		against the parent's width. single-product.css beats both: it sets
		`padding-bottom: 0` on this element at equal specificity but later
		in the cascade, AND `height: auto !important`, which nothing at
		this specificity can win at all.

		MEASURED LIVE BEFORE THE FIX: the track laid out 518px wide and
		ZERO tall, so the cards had no box to sit in and the panel
		collapsed -- with every rule present, correct and losing quietly.

		`aspect-ratio` works WITH that `height: auto` rather than against
		it: an auto height is exactly what lets a ratio supply one. So it
		needs no `!important` and picks no fight. The absolutely
		positioned cards contribute no content height, which is what
		leaves the ratio in charge.

		AND IT STILL HAD TO BE `!important`. single-product.css also sets
		`aspect-ratio: auto` on this element, at the SAME specificity
		(0,2,0) and later in the cascade, so source order handed it the
		win. Measured: the track computed `aspect-ratio: auto` with this
		rule present and matching.
	*/
	aspect-ratio:1 / 1 !important;
	transform:none !important;
	transition:none;
}

.sf-gal.sf-gal--deck .sf-gal__track > .sf-gal__slide{
	position:absolute !important;
	inset:0 !important;
	width:auto !important;
	min-width:0 !important;
	max-width:none !important;
	/* the base gives the slide a square aspect; with `inset:0` that is
	   over-constrained, and the track's padding sets the square now */
	aspect-ratio:auto;
	/*
		12px, NOT `var(--sp-radius, 12px)` -- and the difference is not
		cosmetic. single-product.css declares `--sp-radius: 2px` for the
		whole product page, so that fallback NEVER fires: the expression
		reads as twelve and paints two. CLAUDE.md records the same trap
		costing a round on the review buttons.

		The panel above keeps the token on purpose, because a slider has a
		FRAME and a frame should follow the page. A deck has none -- the
		card is the object -- so it takes the house radius the navbar, the
		hero panel and every card on the site carry, which is also what
		was approved on localhost.
	*/
	border-radius:12px;
	overflow:hidden;
	background:#fff;
	box-shadow:0 18px 40px rgba(0,0,0,.5);
	/*
		EVERY CARD IS HIDDEN BY DEFAULT, so a deck of nineteen paints four
		and not nineteen. `!important` because the base slide rule above
		carries `opacity: 1 !important` -- it is there to beat
		single-product.css, and it beats this too. Measured before the
		fix: all nineteen cards painted, stacked dead flat on each other,
		which reads as one photograph rather than as a broken deck.
	*/
	opacity:0 !important;
	pointer-events:none;
	transform-origin:50% 100%;
	transition:transform .45s cubic-bezier(.2,.8,.2,1), opacity .35s ease;
}

/*
	THE INDEX IS DISTANCE FROM THE TOP, NOT POSITION IN THE GALLERY.
	pdp-gallery.js writes `data-k` as `(i - at + n) % n`, so the live
	photograph is always 0 and the deck looks the same at every step.
	That is also why the deck WRAPS where the slider stops: a deck that
	runs out of cards at the last photograph stops being a deck.
*/
.sf-gal.sf-gal--deck .sf-gal__track > .sf-gal__slide[data-k="0"]{
	opacity:1 !important;
	pointer-events:auto;
	z-index:5;
	transform:none;
	cursor:pointer;
}
.sf-gal.sf-gal--deck .sf-gal__track > .sf-gal__slide[data-k="1"]{
	opacity:1 !important;
	z-index:4;
	transform:translate(3.5%,-3%) scale(.94) rotate(4deg);
}
.sf-gal.sf-gal--deck .sf-gal__track > .sf-gal__slide[data-k="2"]{
	opacity:.9 !important;
	z-index:3;
	transform:translate(6.5%,-5.5%) scale(.885) rotate(8deg);
}
.sf-gal.sf-gal--deck .sf-gal__track > .sf-gal__slide[data-k="3"]{
	opacity:.6 !important;
	z-index:2;
	transform:translate(9%,-7.5%) scale(.84) rotate(11.5deg);
}

/*
	THE CONTROLS FOLLOW THE CARD, NOT THE PANEL. The nav is `left:16px;
	right:16px` of `.sf-gal`, which is the full column -- so with the
	deck inset and anchored left, the next arrow would float in the
	empty gutter beside the cards rather than sitting on the
	photograph. Pulling the right edge in by what the deck gives up
	puts it back on the card.
*/
.sf-gal--deck .sf-gal__nav{
	right:calc(100% - var(--sf-gal-deck-w) + 16px);
	z-index:6;
}

@media (prefers-reduced-motion: reduce){
	.sf-gal.sf-gal--deck .sf-gal__track > .sf-gal__slide{ transition:none; }
}

@media (max-width: 760px){
	/*
		A phone column is about 335px, so 82% of it is 275 -- a real loss
		on the one screen where the photograph is already small. The fan
		is shallower here to pay for it, which is also right: the deck
		reads at a glance on a phone because it is most of the screen.
	*/
	.sf-gal--deck{ --sf-gal-deck-w:90%; }
	.sf-gal--deck .sf-gal__nav{
		right:calc(100% - var(--sf-gal-deck-w) + 12px);
	}
	.sf-gal.sf-gal--deck .sf-gal__track > .sf-gal__slide[data-k="1"]{ transform:translate(2.5%,-2%) scale(.95) rotate(3deg); }
	.sf-gal.sf-gal--deck .sf-gal__track > .sf-gal__slide[data-k="2"]{ transform:translate(4.5%,-3.5%) scale(.9) rotate(6deg); }
	.sf-gal.sf-gal--deck .sf-gal__track > .sf-gal__slide[data-k="3"]{ transform:translate(6%,-5%) scale(.86) rotate(8.5deg); }
}

/* ------------------------------------------------------------------ */
/*  The turntable                                                     */
/* ------------------------------------------------------------------ */
/*
	ONE PANEL AND A SCRUB TRACK (2026-11-19)
	========================================

	Operator, having seen the deck live: "no i want like design 8 not
	deck". Design 8 on localhost was the turntable -- the photograph
	stays put and a magenta track under it runs through every frame.

	IT IS THE SLIDER UNDERNEATH, NOT A NEW MECHANISM. The track still
	translates and the slides are still laid out in a row; all that
	changes is the control that drives it and the speed it answers at.
	So this layout is much thinner than the deck -- nothing about the
	panel, the square or the pin moves.

	**THE TRANSITION HAS TO GO.** The base track eases .62s per step,
	which is right when a dot is pressed and useless under a drag: the
	picture would arrive half a second after the knob. Instant frame
	changes ARE the turntable feel.

	THE ARROWS AND DOTS ARE HIDDEN, NOT REMOVED. The script still
	builds them and paint() still marks them, so the variation reset,
	the keyboard and the swipe all keep working unchanged -- and
	turning this layout off restores them with no JS change. The knob
	carries the count the dots were there to give.

	**IT SAYS "BROWSE", NOT "SPIN", AND THAT IS NOT A WORDING TASTE.**
	The nineteen photographs on the product this was designed against
	are THREE COLOURWAYS shot from six angles -- 01 black, 02 grey, 03
	silver, then the same three again -- read off the files, not
	assumed. Dragging them changes colour as often as angle, so a label
	promising a spin would be describing something the photographs do
	not do. Re-shoot a product as an actual rotation and the word can
	change with it.
*/
.sf-gal--spin .sf-gal__track{
	/* a scrub that eases is a scrub that lags */
	transition:none !important;
}
.sf-gal--spin .sf-gal__nav{ display:none; }

.sf-gal__scrub{
	--sf-gal-knob-w:64px;
	/*
		WOOCOMMERCE FLOATS THE GALLERY LEFT, so a sibling placed after it
		does not go underneath -- it goes BESIDE. Measured without this:
		the scrub laid out at x=664, the panel's right edge, zero pixels
		wide, because a float shortens the line boxes of what follows it
		rather than moving the box. With `clear` it lands at x=70 under
		the photograph, which is where it belongs.

		That is also why it is a SIBLING and not a child: `.sf-gal` is
		`overflow: hidden` with the radius on it, and a control inside
		it would be wrapped by the photograph's own rounded box.
	*/
	clear:both;
	display:flex;
	align-items:center;
	gap:14px;
	margin-top:18px;
	cursor:ew-resize;
	/* the page scrolls vertically under a sideways drag, so the
	   browser must not claim the gesture first */
	touch-action:none;
	user-select:none;
}
.sf-gal__scrub:focus-visible{
	outline:2px solid var(--wp--preset--color--accent-1, #ec1e79);
	outline-offset:6px;
	border-radius:8px;
}

.sf-gal__rail{
	position:relative;
	flex:1 1 auto;
	/* without this the rail refuses to shrink past its ticks and the
	   hint is pushed off the end of a phone */
	min-width:0;
	height:34px;
	display:flex;
	align-items:center;
}
.sf-gal__rail::before{
	content:"";
	position:absolute;
	left:0;
	right:0;
	height:1px;
	background:var(--sp-line, rgba(242,242,244,.16));
}
/*
	ONE TICK PER PHOTOGRAPH, so the track says how many there are
	without a number -- which is the job the dots gave up. At three
	photographs it is three marks; at fifty they close into a rule,
	which is honest in both directions.
*/
.sf-gal__ticks{
	position:absolute;
	left:0;
	right:0;
	top:0;
	bottom:0;
	display:flex;
	align-items:center;
	justify-content:space-between;
	pointer-events:none;
}
.sf-gal__ticks i{ width:1px; height:7px; background:rgba(242,242,244,.3); }

/*
	THE KNOB IS INSET BY HALF ITS OWN WIDTH AT BOTH ENDS, which is why
	the script writes `calc(f*100% + (.5 - f) * knob)` rather than a
	plain percentage: at frame one a bare 0% would hang the pill half
	off the left of the rail. The same inset is taken off the pointer
	maths, so the knob lands where the cursor is.
*/
.sf-gal__knob{
	position:absolute;
	top:50%;
	left:0;
	width:var(--sf-gal-knob-w);
	height:24px;
	border-radius:999px;
	background:var(--wp--preset--color--accent-1, #ec1e79);
	color:#fff;
	transform:translate(-50%,-50%);
	pointer-events:none;
	display:grid;
	place-items:center;
	font-style:normal;
	font-family:var(--sp-tech, "Chakra Petch", sans-serif);
	font-size:10px;
	font-weight:700;
	letter-spacing:.1em;
	white-space:nowrap;
}

.sf-gal__hint{
	font-family:var(--sp-tech, "Chakra Petch", sans-serif);
	font-size:10px;
	font-weight:600;
	letter-spacing:.14em;
	text-transform:uppercase;
	color:var(--sp-muted, #9a9a9f);
	white-space:nowrap;
}

@media (max-width: 600px){
	/*
		THE HINT GOES BEFORE THE RAIL DOES. A phone panel is about 260px
		wide; spending 110 of it on a label leaves nineteen ticks fighting
		over 130, and on a touch screen dragging is the obvious thing to do
		with a track anyway. The knob still carries the count.
	*/
	.sf-gal__hint{ display:none; }
	.sf-gal__scrub{ gap:0; margin-top:14px; }
}

/* ------------------------------------------------------------------ */
/*  The track only moves through a custom property                     */
/* ------------------------------------------------------------------ */
/*
	THE SLIDER'S TRACK WAS NEVER MOVING, AND NOTHING SAID SO
	========================================================

	single-product.css line ~2060 sets `transform: none !important` on
	`.sf-pdp__gallery .woocommerce-product-gallery__wrapper` -- part of
	the tile-grid rules, where undoing FlexSlider's inline transform is
	exactly right. **An author `!important` beats a non-important INLINE
	style**, which is what pdp-gallery.js writes, so every
	`translateX(-N00%)` the script set computed to `none`.

	Measured on the live page: inline `translateX(-1400%)` on the track,
	computed transform `none`, the live slide's own rect at x=8385 in a
	panel that ends at 664 -- and photograph ONE still painted. The dots
	and the arrows marked themselves correctly the whole time, so every
	check that read state rather than paint came back clean.

	**THE PICTURE NEVER CHANGED.** Found the moment the turntable put a
	number on screen that disagreed with what was in the panel; the deck
	hid it completely, because a deck does not translate its track.

	THE FIX IS A CUSTOM PROPERTY, NOT AN INLINE `!important`. A custom
	property carries no priority of its own -- the script sets a VALUE
	and this rule decides what is done with it -- so the two files stop
	competing rather than escalating. `:not(.sf-gal--deck)` keeps it off
	the deck, which parks its track and moves the cards instead, and the
	extra class takes this to (0,3,0) against their (0,2,0).
*/
.sf-gal:not(.sf-gal--deck) .sf-gal__track{
	transform:translateX(var(--sf-gal-x, 0%)) !important;
}
