/* =============================================================
 * SUPERFLY — PRODUCT ARCHIVE (the shop and every category page)
 *
 * Shape: revamp/category-b.html, itself a measured reconstruction
 * of lupinelights.com/en/collections/bike-lighting.
 *
 *   ┌──────────────────────────────────────────────────────────┐
 *   │                                                          │
 *   │                    W H E E L S            160px          │
 *   │          the category description, centred               │
 *   │                                                          │
 *   │  FILTERS +        3 PRODUCTS         [ SORTING      ▾ ]  │
 *   │  ┌ colour ─┬─ fits ─┬─ price ─┬─ availability ┐ (folded) │
 *   │  ┌────────┐   ┌────────┐   ┌────────┐                    │
 *   │  │ white  │   │ white  │   │ white  │  4/5 tiles         │
 *   │  └────────┘   └────────┘   └────────┘                    │
 *   │  PROJECT X    PROJECT X    PROJECT X     40px            │
 *   │  RM574.00     RM448.00     RM433.00      muted           │
 *   └──────────────────────────────────────────────────────────┘
 *
 * THIS REPLACED the two-column version (a sticky 470px filter
 * rail beside the grid, revamp/category-a.html). The operator
 * picked B. The old file is in git, not in a comment.
 *
 * ---- THE MEASUREMENTS ----
 * Read off the reference at 1440px, where the container is 1440
 * with 48px padding, so the content is 1344.
 *
 *   container    padding 0 48, max-width 2000
 *   head         centred, margin 48 0, gap 8
 *   title        160px / 144px uppercase, padding-top 98
 *   description  24px / 28.8, centred, margin-bottom 48
 *   bar          flex space-between, height 32, margin-bottom 24
 *   grid         3 columns, gap 98px ROW / 48px COLUMN
 *   card         flex column gap 24
 *   card image   aspect-ratio 4/5, radius 2
 *   card content gap 8
 *   card title   40px / 40, letter-spacing .8, uppercase
 *   excerpt      italic 18px / 23.76
 *   price        tech 14px / 16.1, ls .56, colour #878787
 *
 * THE PRICE IS MUTED, NOT ACCENT. That is deliberate in the
 * original and it is the opposite of the product page: on a
 * collection the product NAME carries the row, not the number.
 *
 * The 98px row gap is not a typo. It is what stops a wall of
 * cards reading as a spreadsheet.
 *
 * ---- WHY THIS FILE EXISTS RATHER THAN MORE woocommerce.css ----
 * woocommerce.css skins everything WooCommerce draws, including
 * cart and account. This changes archives only, loads after it,
 * and is enqueued only on is_shop() || is_product_taxonomy() — so
 * a mistake here cannot reach the cart, and reverting is one
 * enqueue line. It overrides woocommerce.css section 9, which
 * draws cards as bordered panels; those rules stay for the shop's
 * other surfaces.
 *
 * ---- MARKUP IS WOOCOMMERCE'S OWN ----
 * WooCommerce 11.1.1 renders archives with blocks, so the class
 * names below are theirs and were read off the live page, not
 * guessed. If this page ever goes plain after a WooCommerce
 * update, check those names first. The `sf-shop*` classes are
 * ours, set in templates/archive-product.html.
 *
 * Type is the house trio and the accent is SUPERFLY's #ec1e79 —
 * the same in revamp/category-b.html since 2026-09-29, so the
 * concept and this file cannot drift apart on colour.
 *
 * Enqueue: functions.php — superfly_shop_archive_assets()
 * Fold:    assets/js/shop-archive.js
 * Concept: revamp/category-b.html
 * ============================================================= */

.sf-shop {
	--sa-bg:     #101014;
	--sa-text:   #f2f2f4;
	--sa-muted:  #878787;
	--sa-faint:  rgba(242, 242, 244, 0.34);
	--sa-line:   rgba(242, 242, 244, 0.14);
	--sa-hair:   rgba(242, 242, 244, 0.08);
	--sa-accent: #ec1e79;
	--sa-tile:   #ffffff;              /* unused since 2026-10-05 — see --sa-well */
	/* THE CARD HAS A FACE (operator, 2026-10-22, pointing at the Filters
	   button: "the background is color this ... no need transparent for the
	   card").

	   #16161a is that button's plate, set at line ~426 of this file, and the
	   note there works through why it is that value: #0d0d10 on a #101014
	   page is three levels apart, under what a screen resolves, so the
	   control was invisible. #16161a is twelve levels and reads as a surface
	   without needing a line drawn round it.

	   THE HISTORY THIS REVERSES, so nobody treats it as a slip:
	     #0d0d10     a face three levels off the page -- invisible
	     transparent the page itself, 2026-10-08, when the photograph was
	                 full bleed and WAS the card. Its note argued the weave
	                 should run through unbroken.
	     #16161a     this. The photograph is inset now and no longer
	                 describes the card, so something has to.

	   The weave objection from 2026-10-08 is real and is accepted: a flat
	   fill does stop the carbon texture at the card edge. That is the point
	   here rather than a cost -- the card is meant to read as a panel ON the
	   weave, the way the Filters button does.

	   STILL NO BORDER. A hairline was added and removed once already and the
	   note at section 5 says not to put it back without asking. The Filters
	   button pairs this face with a 1px rgba(242,242,244,.14) ring; if the
	   card is ever wanted to match it exactly, that ring is one line in the
	   card's box-shadow -- but it is a separate decision from the colour,
	   and only the colour was asked for. */
	/* LIGHT (operator, 2026-10-22: "i want the interface exactly 100% same
	   with this"). Measured off the reference card on
	   logitechg.com/en-gb/shop/c/microphones at a 1440 viewport, not picked
	   by eye: the card is rgb(234,235,237) with 24px padding and a 12px
	   radius, and its type is rgb(27,27,27) over rgb(48,49,51).

	   The page around it stays near-black. That is the reference too -- its
	   page is dark and the cards are light panels on it.

	   The sequence for this one value is now #0d0d10 -> transparent ->
	   #16161a -> #EAEBED, each on a request. The three dark values and the
	   reasoning behind them are in the history above; none of it applies to
	   a light card, where there is no question of a surface being too close
	   to the page to see. */
	/* BACK TO THE DARK GREY (operator, 2026-10-22, same day: "change the
	   background to grey like previously you put").

	   #16161a is the Filters button's plate -- set at line ~426 of this file,
	   where the note works through why that value and not a darker one. The
	   card wore it for a few hours this afternoon before the light reference
	   colour replaced it, and this puts it back.

	   ONLY THE SURFACE AND ITS INK REVERT. Everything the card learned from
	   logitechg.com stays: the 24px padding, the 12px radius, the 4:3 photo
	   box at 47.2 from the top, the 12 / 24 / 12 gaps, the swatch row, the
	   description under the name, the badge geometry. Those are measurements,
	   not colours, and they do not care which way round the card is.

	   The full sequence for this one value is now
	       #0d0d10 -> transparent -> #16161a -> #EAEBED -> #16161a
	   and every step was asked for. The reasoning behind the three dark
	   values is in the history above.

	   THE INK IS ROLE-MAPPED, NOT INVERTED. The reference sets its name in
	   near-black and its description, price and CTA one step lighter; these
	   are the same two roles in this site's own palette, which is what the
	   card used before the light detour. */
	/* THE CARD IS THE RAIL'S CARD (operator, 2026-11-16: "the product card
	   color follow the card filter color").

	   #16161a -> the page colour, and it HAD to take the rail's two shadows
	   with it. The rail's console has no fill of its own: it is #101014 on a
	   #101014 woven page, carved out by light alone, so copying the colour
	   without the emboss leaves a card with no face, no border and no shadow
	   — nothing at all. dealer-wall.css argues the same pair for the same
	   reason, and that is why neither carries a hairline: the light IS the
	   edge, and a border flattens it into a rectangle with a glow.

	   What it buys: the weave stops at the card's edge, so the card reads as
	   a panel pressed out of the page rather than a tile laid on it, and the
	   rail and the grid are one material.

	   The full sequence for this one value is now
	       #0d0d10 -> transparent -> #16161a -> #EAEBED -> #16161a -> page
	   and every step was asked for. */
	--sa-card:   var(--sa-bg);

	/* THE EMBOSS, AND IT IS THE INFORMATION PAGES' SET NOW (operator,
	   2026-11-16: "i think the shadow is too much ... open contact-us and
	   see the shadow for this box, i want like that shadow").

	   Read off the live /contact-us/ rather than out of a comment —
	   .superfly-contact-col-form and -col-detail both compute:

	       rgba(255,255,255,.024) -3px -3px  8px
	       rgba(0,0,0,.26)         4px  4px 10px

	   so privacy, contact, return-refund, the rail and the product card
	   now carry one set between them. CHANGE THEM TOGETHER.

	   IT WAS .10 / .70 AT -4/-4/8 AND 5/5/12 — section 5's values from the
	   home page's dealer wall, which is where the rail's look came from on
	   2026-11-16. Those are pitched for a card on the carbon weave with no
	   face of its own; these are pitched for a panel that is already three
	   units darker than its page. We have the first situation and are now
	   using the second's numbers, which is a deliberate trade for matching
	   the information pages and was asked for in those terms.

	   WHAT THAT COSTS, because it is the thing to look at first if the
	   cards ever read as flat: /contact-us/ gives its panels rgb(13,13,16)
	   on a rgb(16,16,20) page, so the FACE does half the work and the
	   light only has to finish the job. Our card is the page colour
	   exactly, so the light is doing all of it. The weave stopping at the
	   card's edge is what carries the difference. If that is not enough,
	   the fix is a face a few units off the page — NOT a heavier shadow,
	   which is what was just asked to stop. */
	--sa-lift: rgba(255, 255, 255, 0.024);
	--sa-sink: rgba(0, 0, 0, 0.26);

	--sa-card-ink:  var(--sa-text);     /* the name, and the heart */
	--sa-card-ink2: var(--sa-muted);    /* the description, price and CTA */
	--sa-well:   #0f0f0f;              /* the reference's product panel, measured */
	--sa-radius: 2px;

	/* THE PHOTO'S INSET inside the card. 0 -- full bleed.

	   THREE THINGS HAVE BEEN TRIED HERE. Read this before touching it:

	     6px  a frame of the card's own face. The face was #0d0d10, lighter
	          than the page, so the frame read as a bordered panel.
	     0px  full bleed, no frame -- but then the PHOTOGRAPH's edge is the
	          card's edge, and these are light grey studio backdrops. A
	          bright rectangle on a near-black page reads as a border just
	          as much as a hairline does.
	     10px frame again, this time transparent so the page showed through.
	          Tried 2026-10-08 and it is WORSE: the bright rectangle is
	          still bright, it just floats in a gutter, so the card gains an
	          edge AND loses its bleed. Do not try this a third time.

	   What actually works is leaving the photo full bleed and FEATHERING
	   it -- see the mask on the image wrapper below. The photo stops being
	   a rectangle at all, so there is no edge to mistake for a border, and
	   the card is described by its shadow. */
	--sa-inset: 0px;

	--sa-title: "Inter", system-ui, -apple-system, sans-serif;
	--sa-body:  "Inter", system-ui, -apple-system, sans-serif;
	--sa-tech:  "Chakra Petch", "Archivo", sans-serif;

	--sa-step: 48px;

	/* THE SIDE GUTTER IS /shop-by-category/'s, NOT --sa-step (operator,
	   2026-11-16: "is it have padding left, right, margin top, bottom same
	   like this page", pointing at /shop-by-category/). It was a flat 48px
	   here against that page's clamp, so the two shop entry points started
	   their content on different edges one click apart.

	   IT IS A SEPARATE TOKEN FROM --sa-step ON PURPOSE. --sa-step is also
	   the gap between the filter columns and the space above the pagination;
	   folding the two together would re-pitch those as well.

	   TWO LITERALS, NOT A var() READ FROM category-index.css. Both files load
	   on their own pages, and a custom property substitutes at computed-value
	   time on the element where it is DECLARED — a :root token crossing files
	   resolves by load order and fails silently. CHANGE BOTH OR NEITHER:
	   this is --sfidx-gutter. */
		/* SUPERSEDED THE SAME DAY by the FAQ geometry above, which is the
	   current state. Kept because it is the one record of what the
	   80px flat ramp measured, and because the two-literals / four-pages
	   argument in it still holds.

	   RAISED 2026-11-16 (operator: "i want you add padding for shop by
	   bike , shop by category , product category , dengan brand page",
	   then 80px from three measured options).

	   Sides went clamp(20px, 3.1vw, 44px) -> clamp(24px, 4.2vw, 80px)
	   and the block padding clamp(40px, 4.4vw, 72px) ->
	   clamp(48px, 5.6vw, 96px). Measured: 24/48 at 375, 43/57 at 1024,
	   60/81 at 1440, 80/96 at 1900.

	   80 IS THE HOME PAGE'S OWN WIDEST BAND. .sfcg carries
	   clamp(20px, 4.2vw, 80px) — sohub's xl:px-20 — so the shop pages
	   now sit on the same line as the category grid rather than on the
	   navbar logo's 44.

	   THE FOUR PAGES MOVE TOGETHER BY CONSTRUCTION, which is the whole
	   reason this is two literals and not four: Shop by Bike and Shop
	   by Category are both .sfidx, and every product-category and brand
	   archive is .sf-shop. The two files were matched value for value
	   earlier today and must stay matched — CHANGE BOTH OR NEITHER.

	   WHAT IT COSTS: the measure on both is 1560 minus twice the
	   gutter, so the content narrows by twice any increase — about 33px
	   at 1440 and 72px at 1900, off a three-column grid. That was put
	   to the operator with the options and taken.

	   AND THE RAIL CLEARANCE RETIRES ITSELF. Both files indent their
	   headings by max(0px, 58px - gutter) to clear the fixed social
	   rail; above a 58px gutter that computes 0 on its own, so from
	   about 1381px up the headings sit flush on the gutter instead of
	   14px inside it. That is the max() doing what it was written to
	   do, not a regression. */
	/* THE SIDES ARE THE FAQ PAGE'S, EXACTLY (operator, 2026-11-16:
	   "added moreee padding , you can open FAQ page , i want like that
	   padding").

	   It is NOT a bigger gutter. The FAQ page — and contact, privacy,
	   terms and About Us with it — is a 1380px box CENTRED in the
	   window with a 40px gutter inside it, so what you read as padding
	   is the centring plus the gutter, and it grows with the screen
	   while the measure stays put. The 80px gutter this replaces was a
	   flat ramp that could never do that.

	   Measured text edge, this against the 80px version it replaces:

	        375   24   (was 24)
	       1024   51   (was 43)
	       1440   65   (was 60)
	       1780  235   (was 179)   <- the operator's own screen
	       1900  295   (was 244)

	   SO THE SHOP IS BACK ON THE HOUSE MEASURE. CLAUDE.md records it
	   leaving --sf-measure earlier today, when the archives were lined
	   up with /shop-by-category/ at 1560. Both families move together
	   here, so those two stay matched with each other AND rejoin the
	   information pages. Change --sf-measure in style.css and all of
	   them follow.

	   TWO DELIBERATE DEVIATIONS FROM THE FAQ, both because the ask was
	   MORE padding and matching it exactly would have given less:

	     - the floor is 24px, not the FAQ's 20, so a phone does not go
	       backwards by 4px;
	     - the TOP padding is left at clamp(48px, 5.6vw, 96px), which is
	       already more than the FAQ's clamp(34px, 5vw, 64px). Only the
	       bottom takes the FAQ's own clamp(60px, 9vw, 120px), which is
	       more than the 96 it replaces.

	   --sf-measure-box IS DECLARED IN style.css, on :root, and that is
	   safe to read from here only because that sheet loads on every
	   page — the cross-file token trap this project records applies to
	   a token declared in a CONDITIONALLY loaded stylesheet. faq.css
	   and dealer-map.css already read it the same way. */
	--sa-gutter: var(--sf-gutter);

	box-sizing: border-box;
	max-width: 2000px;
	margin: 0 auto;
	/* THE BOTTOM PADDING IS NOT OPTIONAL (operator, 2026-10-05).
	   This was `padding: 0 var(--sa-step)` — sides only — so the last row of
	   cards (or the pagination) ended flush against the top of the footer
	   with nothing between them. Measured: 0px from the grid's bottom edge
	   to the footer.

	   It is NOT --sa-step, because that drops to 20px on a phone, which is
	   where the collision looked worst. A clamp keeps real air at every
	   width.

	   IT IS /shop-by-category/'s clamp SINCE 2026-11-16, top and bottom,
	   which is --sfidx's own padding-block. It was clamp(56px, 8vw, 104px)
	   on the bottom only — 104px against that page's 60 at 1353px — and
	   zero on top, the top air being carried by the masthead's margin
	   instead. See the masthead note below for why that margin is gone.

	   A REAL padding-top IS WHAT STOPS THE COLLAPSE, and that is the point
	   of doing it here rather than on the head: with no top padding or
	   border, a margin on the first child collapses straight out through
	   this box and lands above it. Now the section's own box starts at the
	   header's flow edge and the padding is spent inside it, exactly as
	   .sfidx does it. */
	padding: clamp(48px, 5.6vw, 96px) var(--sa-gutter) clamp(60px, 9vw, 120px);
}

/* THE CONTENT COLUMN IS CAPPED AT 1180px (operator, 2026-10-05).

   The section itself stays wide — it still carries the page background
   out to the edges — but everything inside it now stops at 1180 and
   centres, which is the measure About Us, FAQ, privacy, terms and contact
   all use. The shop finally lines up with them.

   WHY IT MATTERED HERE: without the cap the grid grew with the window. On
   a 1522px screen it reached 1416px and each card became 341px, where the
   design is drawn around roughly 280. Four columns in 1180 with an 18px
   gap gives (1180 - 54) / 4 = 281.5px, which is the card the operator has
   been asking for.

   Applied to the children rather than to .sf-shop, because .sf-shop paints
   the background and must stay full width. */
/* .sf-shop__body REPLACED .sf-shop__filters AND .sf-shop__grid HERE on
   2026-11-16, and leaving either of them in would have been the bug.

   Those two are INSIDE the two-column wrapper now. A max-width on a grid
   item is a cap on the item, not on the row — so the rail would have been
   capped at the full measure, the grid capped again inside its own track,
   and the two columns would have disagreed about where the page ends. The
   measure belongs to the one box that spans both. */
.sf-shop__head,
.sf-shop__bar,
.sf-shop__body {
	/* /shop-by-category/'s measure since 2026-11-16, chosen by the operator
	   when the gutters were matched, and --sf-measure-box since the FAQ
	   match later the same day: --sfidx-wrap is that same token INCLUDING its own
	   gutter, and that wrap is border-box, so the content inside it is
	   1560 - 2 x gutter. Written that way here because these children sit
	   inside this section's PADDING box, so they must subtract the same
	   gutter to land on the same edge.

	   It only binds above about 1560px of client width; below that the gutter
	   is what sets the measure and the two pages agree by construction.

	   WAS var(--sf-measure) (1300px), the house measure About Us, FAQ,
	   privacy, terms, contact, cart and checkout all share. The shop is
	   deliberately wider than those now — it was the only way to line the
	   two shop pages up on a wide screen, where 1300 centred inside a 44px
	   gutter put this page's title 90px further right than the other's.
	   Measured at 1782px: 240 before, 150 after, against 150 there. */
	/* NO CAP SINCE 2026-11-17 (operator: "Yes, drop it"). This page runs
	   FULL BLEED inside --sf-gutter, so its edge is section 2's edge at every
	   width -- measured 75px against section 2's 75 at a 1770px client, where
	   the cap used to put it at 235.

	   IT WAS calc(var(--sf-measure-box) - 2 * var(--sa-gutter)), i.e. 1300 of
	   content, the house measure the information pages still share. Putting
	   that one line back restores it, and /shop-by-category/ MUST come back
	   with it -- --sfidx-wrap in category-index.css. The two shop pages are
	   only ever right together; the whole point of the 2026-11-16 round was
	   lining them up with each other.

	   WHY THE GRID PAGES AND NOT THE PROSE PAGES. About Us, FAQ, privacy,
	   terms, contact, cart and checkout keep the cap: a paragraph at this
	   width runs to ~1620px lines, which is the one thing a measure exists
	   to prevent. A grid of cards has no such limit -- the cards simply get
	   bigger, and the extra room goes to the photographs. */
	max-width: none;
	margin-left: auto;
	margin-right: auto;
}

.sf-shop *,
.sf-shop *::before,
.sf-shop *::after {
	box-sizing: border-box;
}


/* -------------------------------------------------------------
 * 0a. WOOCOMMERCE MAKES <body> A SCROLL CONTAINER ON EVERY ARCHIVE,
 *     AND IT PUTS THE NAVBAR 32px TOO LOW FOR LOGGED-IN USERS
 * -------------------------------------------------------------
 * (operator, 2026-11-16: "why product category navbar too bottom?",
 * with two screenshots taken while logged in.)
 *
 * The Product Filters block resets its overlay lock like this, in
 * plugins/woocommerce/assets/client/blocks/woocommerce/product-filters.js:
 *
 *     scrollLimit: () => {
 *       document.body.style.overflow = isOverlayOpened ? "hidden" : "auto";
 *     }
 *
 * So every archive load leaves `<body style="overflow: auto">` — it should
 * clear the property, not set a value. That makes <body> a SCROLL
 * CONTAINER, and `position: sticky` resolves its offsets against the
 * nearest scrollport.
 *
 * WHY THAT MOVES THE NAVBAR. navbar.css pins the header with
 * `top: var(--wp-admin--admin-bar--height, 0px)` so it sits below the
 * admin bar once you scroll. Logged OUT that variable is 0 and the body
 * box starts at the top of the viewport, so the two agree and nothing
 * shows. Logged IN, WordPress adds `html { margin-top: 32px }`, so the
 * body box now starts 32px down — and the sticky threshold becomes that
 * 32 PLUS the 32 of the variable. Measured on /product-category/brake-systems/
 * with the admin bar: header top 64 where it should be 32, which is the
 * whole of the gap the operator photographed. /shop-by-category/ has no
 * filters block, so its body stays `visible` and its header sits at 32.
 *
 * ONLY ADMINS EVER SAW IT. A customer is logged out, html carries no
 * margin, and the band itself was always in the right place — measured
 * 98 on both pages either way. It is still worth fixing: the operator
 * looks at this site logged in all day.
 *
 * THE SELECTOR MATCHES THE VALUE, NOT THE ELEMENT, AND THAT IS THE POINT.
 * `body { overflow: visible !important }` would also defeat a real scroll
 * lock — the cart drawer sets `overflow: hidden` inline (cart-drawer.js),
 * and so does this same WooCommerce function while its overlay is open.
 * Matching the literal `overflow: auto` neutralises the stray reset and
 * leaves every `hidden` lock working. Verified both ways: with the rule,
 * the header measures 32 and an inline `hidden` still computes `hidden`.
 *
 * It depends on how Chrome serialises an inline style ("overflow: auto;",
 * with the space). If WooCommerce ever clears the property properly this
 * rule stops matching and can go. */
body[style*="overflow: auto"] {
	overflow: visible !important;
}


/* -------------------------------------------------------------
 * 1. THE MASTHEAD
 * ------------------------------------------------------------- */
/* LEFT, NOT CENTRED (operator, 2026-10-05).
   It was a centred masthead sitting above a left-aligned grid, so the page
   had two different left edges and the title had nothing holding it down
   once the breadcrumb was removed. Everything now starts on the same edge
   as the first card, and a hairline closes the header off. */
.sf-shop__head {
	display: flex;
	flex-direction: column;
	align-items: flex-start;
	gap: 10px;

	/* THE FIRST CARD LINES UP WITH /shop-by-category/'s FIRST ROW, and the
	   margin is what pays for it (operator, 2026-11-16: "line them up").

	   THIS REVERSES THE EARLIER PASS THE SAME DAY, deliberately, and the
	   two cannot both be had. That pass gave this element .sfidx__head's own
	   clamp(44px, 6vw, 92px) so the GAP UNDER THE TITLE matched; the content
	   below it then started about 35px lower anyway, because an archive runs
	   title -> count -> hairline -> FILTERS BAR -> cards where that page runs
	   title -> lede -> rows, and the bar block is 62px the other page does
	   not spend. Asked which of the two to match, the operator chose the
	   content. Delete the `- 35px` term to go back to matching the gap.

	   THE CONSTANT IS MEASURED, NOT DERIVED, and it folds in two things:
	   the bar plus the space under it MINUS the amount this masthead is
	   SHORTER than that page's, whose lede runs two or three lines.

	   So it tracks /shop-by-category/'s COPY **and this page's own
	   furniture**. Rewrite that lede to a different number of lines, or
	   move something in or out of the masthead, and it drifts; re-measure
	   both pages rather than trusting the number.

	   IT WAS 35 UNTIL 2026-11-16, when the results count moved out of the
	   masthead and down into the bar. That took ~11px off this page's
	   header, so the grid rose by the same amount and the two pages fell
	   out of line — measured 407 against 418 at 1440. 24 puts them back.
	   Nothing about the rule was wrong; the thing it measures changed.

	   Measured after: 418.3 against 418.5 at 1440.

	   THE 12px FLOOR IS LOAD-BEARING ON A PHONE. At 375 the whole gap is
	   44px and the bar block alone is 62, so the subtraction goes negative
	   and the hairline would sit on the Filters button. The floor costs the
	   1px of the three readings above and keeps the rule off the control. */
	margin: 0 auto max(12px, calc(clamp(44px, 6vw, 92px) - 24px));   /* auto, so it centres with the grid */
	padding-bottom: clamp(14px, 1.6vw, 20px);
	border-bottom: 1px solid var(--sa-line, rgba(245, 245, 245, 0.10));
	text-align: left;
}

/* THE MASTHEAD'S TOP MARGIN HAD NEVER APPLIED, AND THE TITLE SAT UNDER THE
   NAVBAR (2026-11-13, operator: "the title overlap with navbar", pointing at
   /product-category/wheels/).

   WordPress emits this inline in <head>, after every enqueued stylesheet:

       :root :where(.is-layout-flow) > :first-child { margin-block-start: 0 }

   `:where()` contributes nothing, but `:root` and `:first-child` are one
   class each, so that rule is 0,2,0 against `.sf-shop__head`'s 0,1,0 — and
   it is a LONGHAND, so it silently zeroed the `margin-top` half of the
   shorthand above while `margin-bottom` kept applying. Measured live before
   this rule existed: `margin-top` computed 0px and the head sat at y=85.

   WHY THAT IS A COLLISION AND NOT JUST A TIGHT GAP. The navbar is a floating
   pill: `<header>` reserves 66px of flow, and navbar.css then pushes the bar
   itself down by `--sf-navbar-float-gap` (24px here, 20px at 599px and
   under), so the bar PAINTS to y=90 while the flow it reserves ends at 66.
   Anything that starts at the header's flow edge is 24px underneath it. The
   first line of "KNOW YOUR <CATEGORY>." ran behind the glass.

   It is every archive, not just one category — /shop/ and the search results
   share this element and measured identically.

   TWO THINGS HAD TO MOVE, and only the first is the bug. `body.woocommerce`
   plus the two classes is 0,3,0, which clears the inline rule whatever the
   source order — that alone restores the clamp. The overhang is then ADDED
   to it, because the clamp was pitched on 2026-10-05 against a navbar that
   was still a full-bleed band flush with its own flow; the pill arrived on
   2026-11-10 and quietly took 24 of its 44px. Restoring the clamp on its
   own measured 20px of air, which is not what it was drawn to mean.

   THE OVERHANG IS A LITERAL, not `var(--sf-navbar-float-gap)`, for the
   reason this project already records for `--sfop-radius` against the
   navbar's own: a token declared in one stylesheet and read from another
   resolves by load order and fails silently. CHANGE IT WITH navbar.css OR
   NEITHER — it is that file's `--sf-navbar-gutter`, which the float gap
   reads, and which steps to 20px in navbar.css's own 599px block.

   Measured after, at 1440x900: head top 134, bar bottom 90, 44px of air.

   NOTE THE MARGIN COLLAPSE, or the arithmetic here will not reproduce.
   `<main>` carries its own 19.2px block margin and `.sf-shop` has no top
   padding or border, so the two COLLAPSE rather than add: the head sits at
   the header's flow edge plus max(19.2px, this), never plus the sum. Any
   value under 19.2px would therefore change nothing at all.

   DO NOT fold this back into the shorthand above. It loses again the moment
   it does, with no error and no visible difference until a page is looked
   at from the very top of the scroll.

   SUPERSEDED 2026-11-16 - THE SECTION OWN padding-top CARRIES THIS NOW, and
   it is the same clamp /shop-by-category/ uses. The whole margin went, the
   inline rule it had to out-specify is no longer in play, and the navbar
   clearance comes out of a padding that cannot collapse. Everything above
   still explains why a margin here was the wrong tool; keep it.

   WHAT REPLACED IT IS THE RULE BELOW, and it is the other half of the same
   collapse. The main element on an archive is has-global-padding plus
   is-layout-constrained and carries a 19.2px block margin of its own; on
   /shop-by-category/ it is a plainer group with none, so that page section
   starts exactly at the header flow edge (measured: 66) and this one
   started 19.2px lower. Zeroed, the two boxes begin at the same pixel and
   their identical padding does the rest.

   SCOPED WITH :has() TO THIS SECTION so no other archive layout is touched.
   A browser without :has() simply keeps the 19.2px, which is a 19px
   difference between two pages rather than a broken one. */
body.woocommerce main:has(> .sf-shop),
body.woocommerce-page main:has(> .sf-shop) {
	margin-top: 0;
}

/* THE "SHOP BY CATEGORY" EYEBROW WAS REMOVED (operator, 2026-10-05).
   This header opens on the title instead. The rules are in git if it is
   wanted back; they drew the pink dash as the pseudo-element's own
   background, since a pseudo-element cannot carry a second one. */

/* The category name is the largest thing on the page — that is the
 * whole idea of this layout.
 *
 * woocommerce.css section 10 screen-reader-hides every archive title
 * site-wide, on the grounds that the breadcrumb already says it. Here
 * the title IS the page, so it is put back — but only inside .sf-shop,
 * so nothing else on the site changes. The selector has to out-specify
 * `body.woocommerce h1.wp-block-query-title` (0,2,2), hence the body
 * prefix and the h1. */
body.woocommerce .sf-shop__head h1.wp-block-query-title,
body.woocommerce-page .sf-shop__head h1.wp-block-query-title {
	position: static;
	width: auto;
	height: auto;
	overflow: visible;
	clip: auto;
	clip-path: none;
	white-space: normal;
	margin: 0;
	padding: 0;
	max-width: 100%;
	font-family: var(--sa-title);
	/* Smaller than the other information pages on purpose (2026-10-05).
	   Those pages ARE their heading; this one is a label on a grid of
	   products, and at the house size it measured 104px and was the loudest
	   thing on the screen by a wide margin. The reference's 160px is kept in
	   revamp/category-b.html if the argument is ever reopened.
	   Was clamp(32px, 6vw, 68px). */

	/* IT IS /shop-by-category/'s TYPE SINCE 2026-11-16 (operator: "follow
	   shop by category size"), which is .sfidx__title value for value. The
	   face was already shared — both are Archivo Black, uppercase, 400 —
	   so three numbers moved:

	       font-size       clamp(30px, 4.6vw, 58px) -> clamp(28px, 4.4vw, 60px)
	       line-height     0.94  ->  1.04
	       letter-spacing  -0.028em  ->  -0.02em

	   THE SIZE WAS NEVER THE VISIBLE DIFFERENCE. The two ramps were within
	   2px at every width and they CROSSED: this title was the larger one
	   below about 1300px and the smaller one above it (30 against 28 at 375,
	   58 against 60 at 1440). What the eye reads is the leading — at 58px
	   the two lines sat 54.5px apart against the other page's 60.3, so
	   "KNOW YOUR / BRAKES." was visibly tighter than "KNOW YOUR PART. /
	   SHOP BY CATEGORY.". Only "size" was asked for; all three moved
	   because matching one of them alone leaves the pair still looking
	   different, which is what the ask was about.

	   THE COST: the title block grows about 6px on two lines and the grid
	   below it follows. The note above still stands — this is a label on a
	   page of products, not the page itself — so if it ever reads as too
	   loud again, come down from HERE rather than back to the old values,
	   and change category-index.css with it or the two drift apart again.

	   FIVE LITERALS AGAINST .sfidx__title, not a shared token: a custom
	   property substitutes at computed-value time on the element where it
	   is DECLARED, so a :root token crossing stylesheets resolves by load
	   order and fails silently. CHANGE BOTH FILES OR NEITHER. */
	/* SECTION 2's TYPE, EXCEPT FOR THE SIZE (operator, 2026-11-16:
	   "please make it the title font and style same like homepage title
	   section 2", then "everything but the size" from two measured
	   options).

	   Family, weight, leading, tracking and case are .sfcg__title's,
	   value for value. The SIZE is deliberately NOT: section 2's ladder
	   tops at 112px and this title is capped at 60, which is the
	   2026-10-05 decision that this is a LABEL over a grid of products
	   rather than the page itself. Measured at 1900 the full ladder
	   would have nearly doubled it.

	   Inter is narrower and far lighter than the Archivo Black this
	   replaces, so the heading reads quieter at the same size even at
	   700 — which is why keeping the cap costs nothing here.

	   CHANGE category-index.css WITH IT. .sfidx__title carries the same
	   four values as literals, for the usual reason: a :root token
	   declared in one stylesheet and read from another resolves by load
	   order and fails silently. */
	/* THE SIZE IS SECTION 2's TOO NOW (operator, 2026-11-16: "i want the
	   font size is same like homepage 112px right for the title section
	   2 in homepage ?").

	   This REVERSES the "everything but the size" answer given an hour
	   earlier, and with it the 2026-10-05 ruling that this heading is a
	   label over a grid rather than the page itself. The cap was 60px.

	   It is a STEPPED LADDER, not a clamp, because that is what
	   .sfcg__title is — 30 / 48 / 60 / 72 / 112 at 768 / 1024 / 1280 /
	   1900 — and the point is that the two compute identically at every
	   width. A clamp fitted to the same endpoints would not: it ramps
	   between them where the ladder jumps.

	   SO ALL SIX VALUES ARE SHARED AGAIN: family, weight, size, leading,
	   tracking, case. CHANGE category-index.css AND category-grid.css
	   WITH THIS FILE, or they drift — the values are literals in each
	   for the usual reason, that a :root token declared in one
	   stylesheet and read from another resolves by load order and fails
	   silently.

	   WHAT IT COSTS, measured: nothing below 1024, where the old clamp
	   was already at or above these steps. Above it the heading nearly
	   doubles — 60 to 72 at 1440 and 60 to 112 at 1900 — and the grid
	   below it moves down by the difference on a two-line title. That
	   was the argument against it in October; it was asked for anyway,
	   with the numbers in front of the operator. */
	font-size: 30px;
	font-weight: 700;
	line-height: 1;
	letter-spacing: -0.075em;
	text-transform: uppercase;
	text-align: left;
	color: var(--sa-text);
}

@media (min-width:  768px){ body.woocommerce-page .sf-shop__head h1.wp-block-query-title{ font-size:  48px } }
@media (min-width: 1024px){ body.woocommerce-page .sf-shop__head h1.wp-block-query-title{ font-size:  60px } }
@media (min-width: 1280px){ body.woocommerce-page .sf-shop__head h1.wp-block-query-title{ font-size:  72px } }
@media (min-width: 1900px){ body.woocommerce-page .sf-shop__head h1.wp-block-query-title{ font-size: 100px } }

/* ---- "KNOW YOUR <CATEGORY>." (operator, 2026-10-05) ----
   Only the category name comes from WooCommerce. "Know your" is a constant
   and the full stop is punctuation, so both are drawn in CSS and the name
   itself is the only live text. That keeps it automatic: a new category
   gets the treatment with nothing to type.

   CATEGORY ARCHIVES ONLY. /shop/ has the title "Shop", and "Know your
   shop." is nonsense, so the scope is body.tax-product_cat and the shop
   page keeps a plain title.

   WORTH KNOWING: the second line is the category name exactly as it is
   written in WooCommerce, so it inherits whatever is there. "Suspensions"
   gives "KNOW YOUR SUSPENSIONS." and "Yamaha" gives "KNOW YOUR YAMAHA.",
   both of which read well — but the placeholder "Other" category gives
   "KNOW YOUR OTHER.", which does not. Rename that category and the line
   fixes itself. */
body.tax-product_cat .sf-shop__head h1.wp-block-query-title::before {
	content: "Know your";
	display: block;
	color: var(--sa-text);
}

/* THE MAGENTA HALF CARRIES SECTION 2 SHEEN (operator, 2026-11-16:
   "also the color pink text in title please match with effect
   PERFORMANCE text").

   IT GOES ON THE h1, NOT ON A SPAN, because there is no span: the
   white overline is a ::before and the full stop an ::after, and
   only the h1 own text is magenta. background-clip: text clips the
   gradient to EVERY glyph in the element, the ::before included —
   but that pseudo sets its own opaque colour and paints over it, so
   only the magenta line shows the sweep. The ::after inherits
   transparent and is swept with the word, which is right: the full
   stop was magenta before.

   DO NOT give the ::before a transparent colour for consistency —
   it would immediately start showing the gradient and the overline
   would stop being white.

   width: fit-content is NOT set here, unlike the information pages:
   this element holds both lines, so shrink-wrapping it would still
   span the wider of the two and buys nothing. */
body.tax-product_cat .sf-shop__head h1.wp-block-query-title {
	color: transparent;
	background-image: linear-gradient(90deg,
		var(--sa-text)    0%,
		var(--sa-accent) 48%,
		var(--sa-text)   100%);
	background-size: 220% 220%;
	-webkit-background-clip: text;
	background-clip: text;
	animation: sfSheen 5s ease infinite;

	/* the last glyph would otherwise be clipped by the background box */
	margin-right: -.12em;
	padding-right: .12em;
}

body.tax-product_cat .sf-shop__head h1.wp-block-query-title::after {
	content: ".";
}

/* ---- "FITS YOUR <BRAND>." (operator, 2026-11-16: "why the title for
   brand not same title like product category .. what title that
   suitable ?", then "Fits your" from three measured options) ----

   The brand archives had NO overline at all: /brand/yamaha/ read a bare
   white "YAMAHA" where every other heading on the site is a white label
   over a magenta subject. That is the inconsistency, and it was only
   ever an omission — these rules are the category block above with one
   body class and one word changed.

   "Fits your" rather than "Built for" or "Parts for your": it is the
   closest in rhythm to "Know your", so the two archive types read as a
   pair, and it is two words like the other two overlines.

   AUTOMATIC, like the category one. Only the brand name comes from
   WooCommerce; the overline and the full stop are drawn here, so a
   brand added in wp-admin gets the treatment with nothing to type.
   Checked against all six: Honda, Yamaha, Suzuki, SYM, Zontes and
   Modenas all read correctly. A brand whose name is a model rather
   than a make would not — see the "KNOW YOUR OTHER." warning above for
   the same trap on the category side. */
body.tax-product_brand .sf-shop__head h1.wp-block-query-title::before {
	content: "Fits your";
	display: block;
	color: var(--sa-text);
}

body.tax-product_brand .sf-shop__head h1.wp-block-query-title {
	color: transparent;
	background-image: linear-gradient(90deg,
		var(--sa-text)    0%,
		var(--sa-accent) 48%,
		var(--sa-text)   100%);
	background-size: 220% 220%;
	-webkit-background-clip: text;
	background-clip: text;
	animation: sfSheen 5s ease infinite;

	/* the last glyph would otherwise be clipped by the background box */
	margin-right: -.12em;
	padding-right: .12em;
}

body.tax-product_brand .sf-shop__head h1.wp-block-query-title::after {
	content: ".";
}

/* ---- /shop/ ----
   Built exactly like a category heading so the two match: a white first line
   drawn in CSS, the page's own title in pink beneath it, and a full stop.

       Every part        <- drawn here, white
       WE MAKE.          <- page 7's title, pink

   Page 7 is titled just "We make" for that reason. The browser tab and
   search results would then read "We make", which is why functions.php
   carries a document_title_parts filter putting the full phrase back.

   It gets no "Know your": that belongs to the category pages, and the shop
   has no category of its own. */
/* :not(.search) ADDED 2026-10-07. A product search carries BOTH
   `post-type-archive-product` AND `search` on <body>, so this fired there
   too and the results page read "Every part / Search results for: ...".
   The overline is the shop page's own title, not a decoration for any list
   of products. */
body.post-type-archive-product:not(.search) .sf-shop__head h1.wp-block-query-title::before {
	content: "Every part";
	display: block;
	color: var(--sa-text);
}

body.post-type-archive-product .sf-shop__head h1.wp-block-query-title {
	color: transparent;
	background-image: linear-gradient(90deg,
		var(--sa-text)    0%,
		var(--sa-accent) 48%,
		var(--sa-text)   100%);
	background-size: 220% 220%;
	-webkit-background-clip: text;
	background-clip: text;
	animation: sfSheen 5s ease infinite;

	/* the last glyph would otherwise be clipped by the background box */
	margin-right: -.12em;
	padding-right: .12em;
}

body.post-type-archive-product .sf-shop__head h1.wp-block-query-title::after {
	content: ".";
	color: var(--sa-accent);
}


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

/* ---- SEARCH RESULTS (operator, 2026-10-07: "can you make it match with
   others title ... cause its pink color for all text") ----

   The whole sentence was coming out of wp:query-title as one string, so the
   whole sentence went magenta: "SEARCH RESULTS FOR: "SUPERFLY PROJECT X""
   at 58px, two lines of solid pink. Every other heading on the site is a
   white label over a magenta subject - "Know your / BRAKES.", "Every part /
   WE MAKE." - and this was the only one not built that way.

   The template now passes showPrefix:false, so the <h1> holds just the term
   and the label is drawn here like the other two.

   SMALLER THAN THE OTHER TWO ON PURPOSE. "Brakes" is one word the operator
   chose; a search term is whatever a customer types, and at 58px a phrase
   like "superfly project x 4 spoke" ran four lines and pushed the results
   off the screen. 44px caps it. */
body.search .sf-shop__head h1.wp-block-query-title::before {
	content: "Search results for";
	display: block;
	color: var(--sa-text);
}

body.search .sf-shop__head h1.wp-block-query-title {
	font-size: clamp(26px, 3.6vw, 44px);
	color: var(--sa-accent);
}

body.search .sf-shop__head h1.wp-block-query-title::after {
	content: ".";
	color: var(--sa-accent);
}

/* The real WooCommerce field, edited per category in Products →
 * Categories. Every category is empty today, so this renders nothing
 * and the header is title-only until someone writes them. */
.sf-shop__head .wp-block-term-description {
	margin: 0;
	/* JUSTIFIED (operator, 2026-10-05). Both edges line up, which sits well
	   under a two-line title and against the ruled grid below.
	   The max-width matters more now than it did: justified text in a very
	   wide column opens "rivers" of white space between the words, so the
	   measure is held to 56ch. */
	text-align: justify;
	max-width: 56ch;
	font-family: var(--sa-body);
	font-style: normal;
	font-size: 15px;
	line-height: 1.65;
	color: var(--sa-muted);
}

.sf-shop__head .wp-block-term-description p {
	margin: 0;
}

/* nothing written yet → no empty 48px hole under the title */
.sf-shop__head .wp-block-term-description:empty {
	display: none;
}


/* -------------------------------------------------------------
 * 2. THE BAR
 * ------------------------------------------------------------- */
.sf-shop__bar {
	box-sizing: border-box;
	align-items: center;
	gap: 16px;
	min-height: 32px;
	margin: 0 auto 24px;                       /* auto, so it centres with the grid */

	/* THE BAR IS THE CARD (operator, 2026-11-16: "i mean the card like
	   this", with a shot of one wide panel holding the count at one end
	   and the sorting at the other).

	   It is the house card, read from the same three tokens the filter
	   rail and the product cards read, so all three follow each other
	   if any is ever re-pitched:

	       background: var(--sa-card);            (that is --sa-bg)
	       box-shadow: -3px -3px 8px var(--sa-lift),
	                    4px 4px 10px var(--sa-sink);

	   NO BORDER, radius 12 — the card conventions this file already
	   argues: on a carved card the light IS the edge, and a hairline
	   flattens it into a rectangle with a glow.

	   THE PADDING IS WHY box-sizing IS STATED. This element carries the
	   shared measure (see section 1), and the head and the grid are on
	   that same measure with no padding of their own. Border-box keeps
	   the CARD on the measure and insets its text by 20px, which is the
	   reference's arrangement — the panel's edges line up with the
	   masthead and the grid, and only the type inside it comes in.
	   Content-box would push the card 40px past both. */
	padding: 14px 20px;
	border-radius: 12px;
	background: var(--sa-card);
	box-shadow: -3px -3px 8px var(--sa-lift), 4px 4px 10px var(--sa-sink);

	/* IT IS A BAND NOW, NOT A FLOATING CONTROL (operator, 2026-11-16:
	   "default sorting maybe can full width and put showing result is
	   same row .. but i dont want same like durian bikers").

	   The count came back down from the masthead, so this row runs the
	   full measure with the count at one end and the sorting at the
	   other.

	   AND IT CARRIES NO RULE OF ITS OWN (operator, 2026-11-16: "if you
	   put two lines, its like redundant"). It had a border-bottom for a
	   few minutes, which put a second hairline 15px under the masthead's
	   — two rules with one row of small type trapped between them. The
	   masthead's is the single line, directly above this card.

	   WHAT IS OURS RATHER THAN THEIRS: the reference is a white panel
	   with a 1px border and a boxed select inside it. Here the panel is
	   carved out of the page rather than drawn on it, and the sorting
	   is type — "Sort by" quiet in front of the value, going magenta
	   when you open it — so the card holds one control and not two
	   boxes nested inside each other. */
}

/* -------------------------------------------------------------
 * 2b. THE GRID STARTS ON THE RAIL'S LINE
 * -------------------------------------------------------------
 * (operator, 2026-11-16: "why not same row", with a shot of the rail's
 * panel sitting 24px above the first row of cards.)
 *
 * THE OFFSET IS WOOCOMMERCE'S OWN GUARD FAILING, NOT OUR LAYOUT. The
 * product collection renders a notices container before the product
 * list:
 *
 *     .sf-shop__grid
 *       > div.wc-block-components-notices      <- first child
 *       > ul.wc-block-product-template         <- SECOND child
 *
 * and WordPress's flow layout gives every child but the first
 * `margin-block-start: 24px`. WooCommerce knows this and ships the
 * cure in product-collection-style.css:
 *
 *     .wc-block-components-notices:not(:has(*)) + .wc-block-product-template
 *       { margin-block-start: 0 }
 *
 * IT NEVER FIRES. Their guard tests for a container with NO element
 * children, and the Interactivity API version of this block always puts
 * a <template> inside it — the banner markup that `data-wp-each` clones
 * when a notice actually exists. A <template> is an element, so
 * `:has(*)` matches, `:not(:has(*))` fails, and every archive on the
 * site carries 24px of dead space above its grid.
 *
 * NOTHING IS VISIBLE IN IT, which is why this survived: a <template>
 * renders nothing at all, so the symptom is a 24px gap with no element
 * in it to inspect and no rule of ours to blame.
 *
 * SO THE TEST IS THE BANNER, NOT THE CONTAINER'S CHILDREN. A real
 * notice is cloned in as a sibling of the template, carrying
 * .wc-block-components-notice-banner — so this zeroes the margin only
 * while there is no banner, and a genuine "added to cart" notice still
 * gets its 24px of air. Do not "simplify" it to :not(:has(*)): that is
 * exactly the rule that does not work.
 *
 * A browser without :has() drops the rule and keeps the 24px, which is
 * today's behaviour — the right way for this to fail.
 * ------------------------------------------------------------- */
.sf-shop__grid > .wc-block-components-notices:not(:has(.wc-block-components-notice-banner))
	+ .wc-block-product-template {
	margin-block-start: 0;
}


/* the sort control takes the right-hand end and may shrink; the Filters
   button never does, so it cannot be squeezed off a narrow screen */
/* WooCommerce nests the control twice: div.wc-block-catalog-sorting >
   form.woocommerce-ordering > select. The DIV is the bar's flex item, so
   that is what has to be sized — targeting the form alone left the wrapper
   at its own height and the two controls 8px out of line. */
/* ALL THREE SHRINK TO THE WORDS NOW. This was `0 1 260px` — a width
   for a boxed control — and it left the sorting 76px short of the
   bar's right edge once the box came off: WooCommerce's own wrapper
   kept the 260 and the type sat left-aligned inside it. Measured at
   1440: sort ending on 1310 against a bar ending on 1386.

   It has to be all three, because WooCommerce nests the control twice
   (div.wc-block-catalog-sorting > form.woocommerce-ordering > select)
   and any one of them left at a fixed basis reintroduces the gap. */
.sf-shop__bar .wc-block-catalog-sorting,
.sf-shop__bar .sf-sort,
.sf-shop__bar form.woocommerce-ordering {
	flex: 0 0 auto;
	min-width: 0;
	margin: 0;
	justify-content: flex-end;
}

.sf-shop__bar .wc-block-catalog-sorting,
.sf-shop__bar form.woocommerce-ordering {
	display: flex;
	align-items: center;
}

/* IT HUGS THE RIGHT NOW rather than filling the track. It was
   `1 1 auto` while it was a boxed control that wanted a width; as a
   line of type it should be as wide as its words and no wider, so the
   count keeps the rest of the row. */
.sf-shop__bar .sf-sort {
	flex: 0 0 auto;
}

.sf-shop__bar .sf-shop__filters-toggle {
	flex: none;
}

/* ---- THE TWO BAR CONTROLS ARE ONE FAMILY (operator: "the design is not
   standardized", 2026-10-05) ----
   Before this, Filters was bare text with no box and the sorting control was
   a 32px hairline-bordered select. Two different objects doing the same kind
   of job, side by side on the same line.

   They are now the same plate as the pagination numbers: 38px tall, 6px
   radius, the card's neumorphic light-and-shadow, Chakra Petch at 10px
   uppercase. The archive ends up with ONE control shape — cards, pagination
   and bar all drawn the same way.

   The shared properties live here; only the select adds a chevron. */
.sf-shop__filters-toggle,
.sf-shop__bar select.orderby {
	display: inline-flex;
	align-items: center;
	height: 38px;
	padding: 0 16px;
	border: 0;
	border-radius: 6px;
	/* THE BOX HAS TO BE DRAWN WITH LIGHT (operator: "the box tak nampak
	   pun", 2026-10-05). The same arithmetic as the product card: the face
	   is #0d0d10 on a #101014 page — three levels apart — and the
	   neumorphic sink adds about three more. Both are under what a screen
	   resolves, so the plate was invisible and the two controls read as
	   floating text.

	   Two changes: the face is lifted to #16161a, and a 1px ring is added.
	   rgba(242,242,244,.14) over the page renders about rgb(42,42,42) — a
	   32-level step. The shadow is kept underneath; it does no harm. */
	background-color: #16161a;
	box-shadow: 0 0 0 1px rgba(242, 242, 244, 0.14),
	            -2px -2px 6px rgba(255, 255, 255, 0.022),
	             3px 3px 8px rgba(0, 0, 0, 0.26);
	color: var(--sa-text);
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 600;
	line-height: 12px;
	letter-spacing: 0.14em;
	text-transform: uppercase;
	cursor: pointer;
	transition: color 0.22s ease, box-shadow 0.22s ease,
	            background-color 0.22s ease;
}

.sf-shop__filters-toggle:hover,
.sf-shop__bar select.orderby:hover {
	color: var(--sa-accent);
	background-color: #1b1b20;
	box-shadow: 0 0 0 1px var(--sa-accent),
	            -3px -3px 8px rgba(255, 255, 255, 0.022),
	             4px 4px 10px rgba(0, 0, 0, 0.26);
}

.sf-shop__filters-toggle {
	gap: 10px;
}

.sf-shop__filters-toggle i {
	font-style: normal;
	color: var(--sa-muted);
	transition: color 0.22s ease, transform 0.22s ease;
}

/* open: pressed in. The GLYPH is swapped by shop-archive.js (+ to x), so
   there is deliberately no rotation here — an earlier version rotated 45deg
   as well, which turned the script's x back into a + on open. */
.sf-shop__filters-toggle[aria-expanded="true"] {
	color: var(--sa-accent);
	box-shadow: inset 2px 2px 6px rgba(0, 0, 0, 0.34),
	            inset -1px -1px 3px rgba(255, 255, 255, 0.02);
}

.sf-shop__filters-toggle[aria-expanded="true"] i {
	color: var(--sa-accent);
}

/* The count lives in the MASTHEAD now, not the bar (2026-10-05). The bar
   selector is kept alongside it so the rule still applies if it is ever
   moved back, and so a cached page served mid-change does not lose it. */
.sf-shop__bar .wc-block-product-results-count,
.sf-shop__head .wc-block-product-results-count {
	margin: 0;
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 500;
	line-height: 12px;
	letter-spacing: 0.1em;
	text-transform: uppercase;
	color: var(--sa-muted);
}

/* a little air between the description and the count under it */
.sf-shop__head .wc-block-product-results-count {
	margin-top: 4px;
}

/* ---- THE SORT LISTBOX ----
   Built by shop-archive.js in front of WooCommerce's <select>. The native
   select's OPEN LIST is drawn by the operating system and no stylesheet can
   reach it — the white panel and blue highlight the operator reported are
   Windows, not this theme. This is the only way to restyle it.

   The button shares every rule with .sf-shop__filters-toggle above, so the
   bar still reads as one family. */
.sf-sort {
	position: relative;
}

/* hidden only once the script has replaced it: the class is added by the
   script, so with no JS the native select stays visible and usable */
.sf-shop__bar select.orderby.sf-sort__native {
	position: absolute;
	width: 1px;
	height: 1px;
	padding: 0;
	margin: -1px;
	overflow: hidden;
	clip-path: inset(50%);
	pointer-events: none;
}

/* THE PLATE AND THE RING ARE BACK (operator, 2026-11-16: "why not like
   the button before this").

   They came off earlier the same day, on the reasoning that a boxed
   select beside a line of tracked-out text is Durian Bikers'
   arrangement and the thing to avoid. The operator looked at it
   unboxed and wanted the button back — so the difference from theirs
   is carried by the type and the magenta instead, which it already was.

   WHAT IS KEPT FROM THE UNBOXED PASS: the "Sort by" label in front of
   the value, and the shortened option text behind it — "Sort by
   Default" rather than "Default sorting". Both are improvements the
   box does not undo. Drop the ::before below to go back to the value
   alone.

   IT SIZES TO ITS WORDS NOW rather than filling a 260px track; that
   part of the band change stays, so the button sits on the measure's
   right edge instead of floating inside a wider box.

   AND IT HAS NO BOX OF ITS OWN, because the BAR is the card now
   (operator, 2026-11-16: "i mean the card like this"). It went through
   a #16161a plate with a ring and then the house card before the
   screenshot arrived; both put a box inside a box, which is the one
   thing the reference does not do.

   So the plate, the ring, the radius and the emboss are all off and
   this is type and a chevron again, sitting inside .sf-shop__bar's
   padding. The three card tokens live on the BAR — see section 2. */
.sf-sort__btn {
	display: inline-flex;
	align-items: center;
	justify-content: flex-end;
	gap: 10px;
	width: auto;
	height: auto;
	padding: 0;
	border: 0;
	border-radius: 0;
	background: none;
	box-shadow: none;
	color: var(--sa-text);
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 600;
	line-height: 12px;
	letter-spacing: 0.14em;
	text-transform: uppercase;
	cursor: pointer;
	transition: color 0.22s ease, box-shadow 0.22s ease,
	            background-color 0.22s ease;
}

/* THE QUIET LABEL IN FRONT. A ::before rather than markup, so the
   script that builds this button needs no change — and so it cannot be
   read as part of the current value by a screen reader, which gets the
   button's own aria-label instead. */
.sf-sort__btn::before {
	content: "Sort by";
	color: var(--sa-muted);
	font-weight: 500;
	letter-spacing: 0.12em;
	transition: color 0.22s ease;
}

.sf-sort__btn:hover::before,
.sf-sort__btn[aria-expanded="true"]::before {
	color: var(--sa-muted);
}

/* Hover and open both just take the value to the accent. There is no
   plate left to deepen or press in — the card is the bar, and a card
   that moved under one of its own children would read as the whole row
   reacting to the sorting. */
.sf-sort__btn:hover,
.sf-sort__btn[aria-expanded="true"] {
	color: var(--sa-accent);
	background: none;
	box-shadow: none;
}

/* the chevron, drawn; it turns over when the list is open */
.sf-sort__btn::after {
	content: "";
	flex: none;
	width: 9px;
	height: 9px;
	background: currentColor;
	transition: transform 0.22s ease;
	-webkit-mask: var(--sf-chev-d) center / contain no-repeat;
	        mask: var(--sf-chev-d) center / contain no-repeat;
	--sf-chev-d: url('data:image/svg+xml;utf8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="black" stroke-width="3" stroke-linecap="round" stroke-linejoin="round"><path d="M5 9l7 7 7-7"/></svg>');
}

.sf-sort__btn[aria-expanded="true"]::after {
	transform: rotate(180deg);
}

/* THE LABEL TRUNCATES RATHER THAN STRETCHING THE BAR. "Sort by price: high
   to low" is 26 characters; left to grow it pushed the Filters button off a
   375px screen. The full text is in the list and announced by the button's
   aria-label. */
.sf-sort__label {
	overflow: hidden;
	text-overflow: ellipsis;
	white-space: nowrap;
}

.sf-sort__list {
	position: absolute;
	z-index: 20;
	top: calc(100% + 8px);
	right: 0;
	min-width: 100%;
	max-width: min(320px, calc(100vw - 40px));
	max-height: min(60vh, 340px);
	overflow-y: auto;
	margin: 0;
	padding: 6px;
	list-style: none;
	border-radius: 8px;
	background: #121216;
	box-shadow: 0 0 0 1px rgba(245, 245, 245, 0.10),
	            0 18px 40px rgba(0, 0, 0, 0.72);
}

.sf-sort__list[hidden] {
	display: none;
}

.sf-sort__opt {
	padding: 10px 12px;
	border-radius: 5px;
	cursor: pointer;
	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 600;
	letter-spacing: 0.1em;
	text-transform: uppercase;
	color: var(--sa-muted);
	white-space: nowrap;
	transition: color 0.16s ease, background-color 0.16s ease;
}

.sf-sort__opt:hover,
.sf-sort__opt:focus {
	outline: none;
	background: rgba(245, 245, 245, 0.06);
	color: var(--sa-text);
}

/* the one in force: magenta, and marked with a bar rather than colour
   alone, so it still reads without the hue */
.sf-sort__opt.is-current {
	position: relative;
	padding-left: 20px;
	background: rgba(236, 30, 121, 0.10);
	color: var(--sa-accent);
}

.sf-sort__opt.is-current::before {
	content: "";
	position: absolute;
	left: 9px;
	top: 50%;
	width: 3px;
	height: 12px;
	margin-top: -6px;
	border-radius: 2px;
	background: var(--sa-accent);
}

/* only what is particular to the native select: room for the chevron, and
   the chevron itself, drawn rather than shipped as an asset */
.sf-shop__bar select.orderby {
	padding-right: 34px;
	appearance: none;
	-webkit-appearance: none;
	background-image:
		linear-gradient(45deg, transparent 50%, var(--sa-muted) 50%),
		linear-gradient(135deg, var(--sa-muted) 50%, transparent 50%);
	background-position: calc(100% - 18px) 17px, calc(100% - 13px) 17px;
	background-size: 5px 5px, 5px 5px;
	background-repeat: no-repeat;
}

/* the native dropdown list: Chakra Petch ships 500 and up, so name a
   weight it has or this falls back to Archivo */
.sf-shop__bar select.orderby option {
	background: #111114;
	color: #ffffff;
	font-weight: 500;
}


/* -------------------------------------------------------------
 * 3b. THE FILTER SHEET (operator, 2026-10-05)
 *     ** READ BY NOTHING SINCE 2026-11-16 **
 * -------------------------------------------------------------
 * The sheet was replaced by the left rail at section 13, and
 * shop-archive.js no longer builds the element these rules style.
 * Every selector below matches zero nodes on every page.
 *
 * KEPT, NOT DELETED, for one reason: it is ~340 lines of settled
 * decisions and it is the only record of how the filter blocks
 * were made to look like this site's. Section 13 starts from it.
 *
 * ITS THREE ELEMENT CLASSES WERE RENAMED IN THE RAIL — sf-rail__head,
 * sf-rail__body, sf-rail__mark — and that is NOT cosmetic. The rules
 * below are mostly bare class selectors (.sf-filters__head, not
 * .sf-filters .sf-filters__head), so reusing the names would have had
 * the dead sheet silently styling the live rail from 2500 lines away.
 * If this block is ever deleted, leave the rail's names alone anyway.
 *
 * A full-screen panel over the page, after the reference the
 * operator sent: dark sheet, a centred column of sections, Clear
 * all / Apply at the foot. The type is this site's — Archivo
 * Black for the section headings, Chakra Petch for the controls
 * — and the accent is the brand magenta, not the reference's.
 *
 * Built by shop-archive.js, which MOVES WooCommerce's own filter
 * blocks inside it. Nothing here reimplements a filter.
 * ------------------------------------------------------------- */
/* BORDER-BOX INSIDE THE SHEET TOO. The archive sets it with `.sf-shop *`,
   and the script moves the filter panel OUT of .sf-shop — so inside the sheet
   everything reverted to content-box. The amount inputs measured 312px inside
   a 292px column: `width: 100%` plus 20px of padding, overflowing by exactly
   their padding. Same cause as the --sa-* tokens needing to be redeclared
   below. */
.sf-filters,
.sf-filters *,
.sf-filters *::before,
.sf-filters *::after {
	box-sizing: border-box;
}

html.sf-filters-open {
	overflow: hidden;              /* the page must not scroll behind the sheet */

	/* AND THE RESERVED SCROLLBAR GUTTER HAS TO GO WITH IT.
	   The theme sets `scrollbar-gutter: stable` on <html>, which keeps the
	   10px gutter reserved even once `overflow: hidden` has taken the
	   scrollbar away. <html> therefore stays 1430px in a 1440px window, and
	   a `position: fixed; inset: 0` sheet resolves against THAT — so the
	   sheet sat 10px narrow and flush left, leaving a strip down the right
	   and putting everything inside it 10px left of the window's centre.
	   Measured: html 1430, innerWidth 1440, column 395 either side of the
	   sheet but 395/405 against the window.

	   The content inside the sheet was centred correctly the whole time; it
	   was the sheet that was off. */
	scrollbar-gutter: auto;
}

.sf-filters {
	/* THE TOKENS ARE REDECLARED HERE, and they have to be. Every --sa-*
	   variable is defined on .sf-shop, and the script MOVES the filter panel
	   out of .sf-shop and into this sheet at body level — so inside the sheet
	   those variables resolve to nothing. The checkbox borders came out
	   rgba(0,0,0,0) and vanished. Keep this list in step with .sf-shop. */
	--sa-text:   #f2f2f4;
	--sa-muted:  #878787;
	--sa-faint:  rgba(242, 242, 244, 0.34);
	--sa-line:   rgba(242, 242, 244, 0.14);
	--sa-hair:   rgba(242, 242, 244, 0.08);
	--sa-accent: #ec1e79;
	--sa-card:   #0d0d10;
	--sa-radius: 2px;
	--sa-title: "Inter", system-ui, -apple-system, sans-serif;
	--sa-body:  "Inter", system-ui, -apple-system, sans-serif;
	--sa-tech:  "Chakra Petch", "Archivo", sans-serif;

	position: fixed;
	inset: 0;
	z-index: 1000;
	display: flex;
	flex-direction: column;
	/* THE PAGE SHOWS THROUGH, BLURRED (operator, 2026-10-05). Not an opaque
	   sheet: the archive stays visible behind it, so this reads as a layer
	   over the shop rather than a separate page.

	   The tint is carried AS WELL AS the blur, because backdrop-filter is not
	   universal — without a background colour the text would sit on the raw
	   page wherever the filter is unsupported. .82 is dark enough to read
	   against the white studio shots that fill the cards behind it. */
	background: rgba(8, 8, 10, 0.82);
	-webkit-backdrop-filter: blur(20px) saturate(120%);
	        backdrop-filter: blur(20px) saturate(120%);
	color: var(--sa-text, #f2f2f4);
}

.sf-filters[hidden] {
	display: none;
}

/* ---- the bar across the top ---- */
.sf-filters__bar {
	display: flex;
	align-items: center;
	justify-content: space-between;
	gap: 16px;
	flex: none;
	padding: 20px clamp(20px, 4vw, 48px);
	border-bottom: 1px solid rgba(242, 242, 244, 0.08);
}

.sf-filters__eyebrow {
	font-family: var(--sa-tech, "Chakra Petch", sans-serif);
	font-size: 11px;
	font-weight: 700;
	letter-spacing: 0.2em;
	text-transform: uppercase;
	color: var(--sa-muted, #878787);
}

.sf-filters__close {
	display: inline-flex;
	align-items: center;
	gap: 10px;
	height: 38px;
	padding: 0 16px;
	border: 1px solid var(--sa-accent, #ec1e79);
	border-radius: 6px;
	background: transparent;
	cursor: pointer;
	font-family: var(--sa-tech, "Chakra Petch", sans-serif);
	font-size: 10px;
	font-weight: 700;
	letter-spacing: 0.16em;
	text-transform: uppercase;
	color: var(--sa-text, #f2f2f4);
	transition: background-color 0.22s ease, color 0.22s ease;
}

.sf-filters__close:hover {
	background: var(--sa-accent, #ec1e79);
	color: #fff;
}

.sf-filters__close i {
	font-style: normal;
	font-size: 15px;
	line-height: 1;
}

/* ---- the middle ----
   It can still scroll if every section is opened at once, but with them
   folded nothing has to — which is the point of the fold. */
.sf-filters__scroll {
	flex: 1 1 auto;
	overflow-y: auto;
	-webkit-overflow-scrolling: touch;
	padding: clamp(16px, 3vw, 32px) clamp(20px, 4vw, 48px);

	/* THE BLOCK SITS IN THE MIDDLE OF THE SCREEN (operator, 2026-10-05),
	   rather than stacked at the top with the rest of the sheet empty.

	   `safe center` is the important half. Plain `center` centres the column
	   even when it is TALLER than the sheet, and the overflow then goes off
	   the top where it cannot be scrolled back to — the classic flex-centring
	   trap. `safe` falls back to flex-start the moment it would overflow, so
	   opening both sections on a short screen still scrolls normally. The
	   plain value is left above it for browsers that do not know `safe`. */
	display: flex;
	justify-content: center;
	align-items: center;
	align-items: safe center;
}

/* ---- a section heading, which is also its control ---- */
.sf-filters__head {
	display: flex;
	align-items: center;
	gap: 14px;
	width: 100%;
	padding: 0;
	border: 0;
	background: none;
	cursor: pointer;
	text-align: left;
	color: inherit;
}

/* the + on the LEFT, as the reference draws it */
.sf-filters__head::before {
	content: "+";
	flex: none;
	font-family: var(--sa-tech);
	font-size: 15px;
	line-height: 1;
	color: var(--sa-accent);
}

/* and the + / x on the right, pushed out by the heading between them */
.sf-filters__head .sf-filters__mark {
	flex: none;
	margin-left: auto;
	font-style: normal;
	font-size: 15px;
	line-height: 1;
	color: var(--sa-muted);
	transition: color 0.18s ease;
}

.sf-filters__head:hover .sf-filters__mark,
.sf-filters__head[aria-expanded="true"] .sf-filters__mark {
	color: var(--sa-text);
}

.sf-filters__head h3,
.sf-filters__head .sf-filters__h,
.sf-filters__head .wp-block-group {
	margin: 0;
}

.sf-filters__body {
	padding-top: 18px;
}

.sf-filters__body[hidden] {
	display: none;
}

.sf-filters__col {
	width: 100%;
	max-width: 640px;
	margin: 0 auto;
}

/* the moved WooCommerce panel sheds the fold's own chrome */
.sf-filters .sf-shop__filters {
	margin: 0;
	padding: 0;
	border: 0;
}

/* inside the sheet the groups stack down the column, each with a rule
   above it — the reference's shape */
.sf-filters .wc-block-product-filters {
	display: block;
}

/* THE `.sf-shop__filters` IN THE MIDDLE IS LOAD-BEARING. The panel keeps
   that class when the script moves it in here, and section 4 caps the price
   group at `width: min(420px, 100%)` for the no-script inline layout. Both
   selectors were (0,2,0), the inline one sits later in this file, so it won
   and the Price section came out 420px inside a 640px column while Category
   filled it — visibly ragged, and only at desktop width, since at 375px both
   are the same 335px anyway. This selector is (0,3,0) and settles it. */
.sf-filters .sf-shop__filters .wp-block-woocommerce-product-filter-taxonomy,
.sf-filters .sf-shop__filters .wp-block-woocommerce-product-filter-price,
.sf-filters .wp-block-woocommerce-product-filter-taxonomy,
.sf-filters .wp-block-woocommerce-product-filter-price {
	width: 100%;
	max-width: none;
	padding: 20px 0;
	border-top: 1px solid rgba(242, 242, 244, 0.08);
}

/* The ACTIVE-FILTER CHIPS are not wanted in the sheet: the controls that
   set them are right there, and Clear all is at the foot. Left in, the block
   rendered as an empty bordered strip above Category — the stray rule and
   60px of nothing at the top of the panel. */
.sf-filters .wp-block-woocommerce-product-filter-active {
	display: none;
}

/* no rule above the first group */
.sf-filters .wp-block-woocommerce-product-filter-taxonomy {
	border-top: 0;
	padding-top: 0;
}

.sf-filters h3.wp-block-heading,
.sf-filters__h {
	margin: 0;
	font-family: var(--sa-title, "Inter", sans-serif);
	font-size: 15px;
	/* 700, NOT 400. This was Archivo Black, which has ONE weight and is
	   black at 400; Inter at 400 in the same place reads as a caption.
	   category-grid.css records the same correction from the same cause
	   when section 2 changed face. */
	font-weight: 700;
	line-height: 20px;
	letter-spacing: 0.1em;
	text-transform: uppercase;
	color: var(--sa-text, #f2f2f4);
}

/* ---- the category CHIPS ----
   wp:woocommerce/product-filter-chips in the template. Bordered pills that
   wrap two or three to a row with WooCommerce's own count beside the name,
   as the reference draws them. Four categories cost one line this way
   instead of four, which is most of what keeps the sheet off a scrollbar. */
.sf-filters .wc-block-product-filter-chips__items,
.sf-filters .wp-block-woocommerce-product-filter-chips {
	display: flex;
	flex-wrap: wrap;
	justify-content: center;      /* the row of chips is centred, not ranged left */
	gap: 10px;
}

.sf-filters .wc-block-product-filter-chips__item {
	display: inline-flex;
	align-items: center;
	gap: 7px;
	padding: 9px 14px;
	border: 1px solid rgba(242, 242, 244, 0.22);
	border-radius: 5px;
	background: transparent;
	cursor: pointer;
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 700;
	letter-spacing: 0.14em;
	text-transform: uppercase;
	color: var(--sa-text);
	transition: border-color 0.2s ease, background-color 0.2s ease,
	            color 0.2s ease;
}

.sf-filters .wc-block-product-filter-chips__item:hover {
	border-color: var(--sa-accent);
}

.sf-filters .wc-block-product-filter-chips__item[aria-checked="true"],
.sf-filters .wc-block-product-filter-chips__item.is-selected {
	border-color: var(--sa-accent);
	background: var(--sa-accent);
	color: #fff;
}

.sf-filters .wc-block-product-filter-chips__item-count,
.sf-filters .wc-block-product-filter-chips__count {
	font-weight: 500;
	color: var(--sa-faint);
}

/* the category labels match the Sort list rather than body copy, so the two
   lists in the sheet read as one set */
.sf-filters .sf-shop__filters .wc-block-product-filter-checkbox-list__text,
.sf-filters .wc-block-product-filter-checkbox-list__text {
	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 600;
	letter-spacing: 0.14em;
	text-transform: uppercase;
}

.sf-filters .wc-block-product-filter-checkbox-list__items {
	gap: 14px;
}

/* ---- Sort is NOT in this sheet (operator, 2026-10-05) ----
   A radio list lived here for one revision. It was removed because the bar
   already carries a sort control — see "THE SORT LISTBOX" — and two controls
   driving the same <select> is one more than anyone needs. The rules went
   with it rather than being left behind as dead weight. */

/* ---- the foot ---- */
.sf-filters__foot {
	display: flex;
	justify-content: center;
	gap: 14px;
	flex: none;
	padding: 18px clamp(20px, 4vw, 48px);
	border-top: 1px solid rgba(242, 242, 244, 0.08);
}

.sf-filters__clear,
.sf-filters__apply {
	flex: 0 1 220px;
	height: 50px;
	border: 0;
	border-radius: 6px;
	cursor: pointer;
	font-family: var(--sa-tech, "Chakra Petch", sans-serif);
	font-size: 11px;
	font-weight: 700;
	letter-spacing: 0.18em;
	text-transform: uppercase;
	transition: background-color 0.22s ease, color 0.22s ease,
	            border-color 0.22s ease;
}

.sf-filters__clear {
	border: 1px solid rgba(242, 242, 244, 0.22);
	background: transparent;
	color: var(--sa-text, #f2f2f4);
}

.sf-filters__clear:hover {
	border-color: var(--sa-text, #f2f2f4);
}

.sf-filters__apply {
	background: var(--sa-accent, #ec1e79);
	color: #fff;
}

.sf-filters__apply:hover {
	background: #ff2e8c;
}

/* the foot sits in the middle column too, not edge to edge on a big screen */
@media (min-width: 720px) {
	.sf-filters__foot {
		max-width: calc(640px + 2 * clamp(20px, 4vw, 48px));
		margin: 0 auto;
		width: 100%;
	}
}


/* -------------------------------------------------------------
 * 3. THE FILTERS, FOLDED
 *
 * Closed by the JS on load, not here, so with scripts blocked the
 * filters are simply open and usable rather than shut behind a
 * dead button. See assets/js/shop-archive.js.
 * ------------------------------------------------------------- */
.sf-shop__filters {
	margin: 0 0 var(--sa-step);
	padding: 32px 0;
	border-top: 1px solid var(--sa-hair);
	border-bottom: 1px solid var(--sa-hair);
}

.sf-shop__filters[hidden] {
	display: none;
}

/* ONE FILTER, NOT FOUR (operator, 2026-10-05: colour, fits and availability
   removed; price is the only one wanted). A four-column grid with a single
   child left the price squeezed into a quarter of the width with three empty
   columns beside it, so the track was about 250px and unusable. It is a flex
   row now: the price takes a comfortable measure and anything added later
   sits beside it without a grid to re-count. */
.sf-shop__filters .wc-block-product-filters {
	display: flex;
	/* ROW, SAID OUT LOUD. WooCommerce sets `flex-direction: column` on this
	   container for its drawer. Leaving it meant flex-basis was read as a
	   HEIGHT, so `flex: 1 1 320px` on the price group below — intended as a
	   width — produced a 320px-tall box with 198px of empty space under the
	   slider. Measured before this line: container height 320px, price group
	   height 320px, content about 120px. */
	flex-direction: row;
	flex-wrap: wrap;
	gap: var(--sa-step);
	align-items: flex-start;
}

/* sized by width, not by flex-basis, so the direction above cannot bite
   again if WooCommerce changes it */
.sf-shop__filters .wp-block-woocommerce-product-filter-price {
	flex: 0 1 auto;
	width: min(420px, 100%);
}

/* WooCommerce's own mobile drawer button and overlay chrome: this
   layout has its own toggle, so the plugin's is not wanted. */
.sf-shop__filters .wc-block-product-filters__open-overlay,
.sf-shop__filters .wc-block-product-filters__overlay-header {
	display: none;
}

/* WooCommerce wraps the filters in a DRAWER below its own breakpoint, and
   this layout has its own fold, so every wrapper is dissolved with
   `display: contents` and the filters flow inline at any width.

   __overlay-dialog WAS MISSING FROM THIS LIST and it is the one that
   matters: it is `position: absolute; inset: 0` with a transform, so on a
   phone the whole panel was lifted out of the page. The fold opened to a
   66px-tall empty box while the price filter was laid out 900px further
   down inside the drawer. Measured before this line: panel 66px high,
   price group 936px high at top 919. Add any new wrapper WooCommerce
   introduces to this list. */
.sf-shop__filters .wc-block-product-filters__overlay,
.sf-shop__filters .wc-block-product-filters__overlay-wrapper,
.sf-shop__filters .wc-block-product-filters__overlay-dialog,
.sf-shop__filters .wc-block-product-filters__overlay-content {
	/* !important IS REQUIRED, not laziness. WooCommerce ships
	       :where(.wc-block-product-filters) .…__overlay-dialog {
	           display: flex !important;
	       }
	   and `:where()` contributes ZERO specificity, so that rule is (0,1,0)
	   — this one at (0,2,0) already outranks it and still lost, because
	   !important beats specificity outright. Checked with
	   getPropertyPriority('display') rather than assumed. */
	display: contents !important;
}

/* the drawer's own header and footer go with it */
.sf-shop__filters .wc-block-product-filters__overlay-header,
.sf-shop__filters .wc-block-product-filters__overlay-footer {
	display: none;
}

/* 16px under the heading. With four groups stacked in one column on a phone
   this was too tight and "Matte black" ran straight into "FITS" — visible in
   the operator's screenshot. Only price is left, but the spacing is kept
   honest in case a group comes back. */
.sf-shop__filters h3.wp-block-heading {
	margin: 0 0 16px;
	font-family: var(--sa-title);
	font-size: 14px;
	/* 700, NOT 400. This was Archivo Black, which has ONE weight and is
	   black at 400; Inter at 400 in the same place reads as a caption.
	   category-grid.css records the same correction from the same cause
	   when section 2 changed face. */
	font-weight: 700;
	line-height: 18px;
	letter-spacing: 0.1em;
	text-transform: uppercase;
	color: var(--sa-text);
}

/* ---- checkbox lists (Colour, Fits, Availability) ---- */
.sf-shop__filters .wc-block-product-filter-checkbox-list__items {
	display: flex;
	flex-direction: column;
	gap: 12px;
}

.sf-shop__filters .wc-block-product-filter-checkbox-list__label {
	display: flex;
	align-items: center;
	gap: 12px;
	cursor: pointer;
}

.sf-shop__filters .wc-block-product-filter-checkbox-list__text {
	font-family: var(--sa-body);
	font-size: 14px;
	line-height: 1.32;
	letter-spacing: 0.24px;
	color: var(--sa-muted);
	transition: color 0.18s ease;
}

.sf-shop__filters .wc-block-product-filter-checkbox-list__label:hover
	.wc-block-product-filter-checkbox-list__text {
	color: var(--sa-text);
}

/* The tick is an <svg> mark layered over a real <input>, so both get
   restyled rather than the input hidden — hiding it takes the keyboard
   with it. */
.sf-shop__filters .wc-block-product-filter-checkbox-list__input-wrapper {
	position: relative;
	flex: none;
	width: 18px;
	height: 18px;
}

/* THE TYPE SELECTOR IS LOAD-BEARING. WooCommerce ships
       input[type="checkbox"].wc-block-product-filter-checkbox-list__input {
           border: 1px solid var(--wc-...-border, transparent);
           width: 1em;
       }
   at (0,2,1), which outranks a plain `.sf-shop__filters .__input` at
   (0,2,0) - so this rule was losing and the boxes were drawn with a
   TRANSPARENT border at 1em. They have been invisible on the archive since
   the filters shipped; it only became obvious once they were the only thing
   in the sheet. Measured: border-color rgba(0,0,0,0), width 16.8px. */
.sf-shop__filters,
.sf-filters {
	--wc-product-filter-checkbox-list-option-element-border: rgba(242, 242, 244, 0.30);
}

.sf-shop__filters input[type="checkbox"].wc-block-product-filter-checkbox-list__input,
.sf-filters input[type="checkbox"].wc-block-product-filter-checkbox-list__input {
	appearance: none;
	-webkit-appearance: none;
	margin: 0;
	width: 18px;
	height: 18px;
	border: 1px solid rgba(242, 242, 244, 0.30);
	border-radius: var(--sa-radius);
	background: transparent;
	cursor: pointer;
	transition: border-color 0.18s ease, background-color 0.18s ease;
}

.sf-shop__filters input[type="checkbox"].wc-block-product-filter-checkbox-list__input:checked,
.sf-filters input[type="checkbox"].wc-block-product-filter-checkbox-list__input:checked {
	border-color: var(--sa-accent);
	background: var(--sa-accent);
}

.sf-shop__filters input[type="checkbox"].wc-block-product-filter-checkbox-list__input:hover,
.sf-filters input[type="checkbox"].wc-block-product-filter-checkbox-list__input:hover {
	border-color: var(--sa-accent);
}

.sf-shop__filters .wc-block-product-filter-checkbox-list__input:checked {
	border-color: var(--sa-accent);
	background: var(--sa-accent);
}

.sf-shop__filters .wc-block-product-filter-checkbox-list__input:focus-visible {
	outline: 2px solid var(--sa-accent);
	outline-offset: 2px;
}

.sf-shop__filters .wc-block-product-filter-checkbox-list__mark {
	position: absolute;
	inset: 0;
	margin: auto;
	width: 12px;
	height: 12px;
	pointer-events: none;
	color: #ffffff;
	opacity: 0;
	transition: opacity 0.18s ease;
}

.sf-shop__filters .wc-block-product-filter-checkbox-list__input:checked
	+ .wc-block-product-filter-checkbox-list__mark {
	opacity: 1;
}

.sf-shop__filters .wc-block-product-filter-checkbox-list__count {
	margin-left: auto;
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 500;
	letter-spacing: 0.1em;
	color: var(--sa-faint);
}

/* ---- PRICE ----
 * WooCommerce's markup, for reference, because the selectors are opaque:
 *   .__range            the rail. Carries --low / --high as PERCENTAGES of
 *                       the span, which is what lets the fill below be drawn
 *                       in CSS with no JavaScript of ours.
 *     .range-bar        the filled portion
 *     input.min/.max    two overlaid range inputs, height 0; only their
 *                       thumbs are visible, so the thumb is styled and the
 *                       track is not
 *   .__left / .__right  the two amount boxes
 *
 * It shipped as a bare white line with two white dots and two filled grey
 * rectangles — WooCommerce's defaults showing through. Rebuilt here in the
 * brand's own language (operator, 2026-10-05).
 * ------------------------------------------------------------- */
.sf-filters .wc-block-product-filter-price-slider__range,
.sf-shop__filters .wc-block-product-filter-price-slider__range {
	position: relative;
	height: 4px;
	border-radius: 999px;
	background: rgba(242, 242, 244, 0.13);
}

/* the selected span, drawn from the two custom properties */
.sf-filters .wc-block-product-filter-price-slider__range .range-bar,
.sf-shop__filters .wc-block-product-filter-price-slider__range .range-bar {
	position: absolute;
	top: 0;
	bottom: 0;
	left: var(--low, 0%);
	right: calc(100% - var(--high, 100%));
	width: auto;
	border-radius: 999px;
	background: var(--sa-accent);
	box-shadow: 0 0 12px rgba(236, 30, 121, 0.45);
}

/* the two inputs are invisible rails; only the thumbs show */
.sf-filters .wc-block-product-filter-price-slider__range input[type="range"],
.sf-shop__filters .wc-block-product-filter-price-slider__range input[type="range"] {
	appearance: none;
	-webkit-appearance: none;
	background: transparent;
	pointer-events: none;        /* the thumb takes the pointer back below */
}

.sf-filters .wc-block-product-filter-price-slider__range input[type="range"]::-webkit-slider-thumb,
.sf-shop__filters .wc-block-product-filter-price-slider__range input[type="range"]::-webkit-slider-thumb {
	-webkit-appearance: none;
	pointer-events: auto;
	width: 16px;
	height: 16px;
	border: 2px solid var(--sa-accent);
	border-radius: 50%;
	background: #0d0d10;
	box-shadow: 0 2px 6px rgba(0, 0, 0, 0.6);
	cursor: grab;
	transition: transform 0.16s ease, box-shadow 0.16s ease;
}

.sf-filters .wc-block-product-filter-price-slider__range input[type="range"]::-moz-range-thumb,
.sf-shop__filters .wc-block-product-filter-price-slider__range input[type="range"]::-moz-range-thumb {
	pointer-events: auto;
	width: 16px;
	height: 16px;
	border: 2px solid var(--sa-accent);
	border-radius: 50%;
	background: #0d0d10;
	box-shadow: 0 2px 6px rgba(0, 0, 0, 0.6);
	cursor: grab;
}

.sf-filters .wc-block-product-filter-price-slider__range input[type="range"]:hover::-webkit-slider-thumb,
.sf-filters .wc-block-product-filter-price-slider__range input[type="range"]:focus-visible::-webkit-slider-thumb {
	transform: scale(1.18);
	box-shadow: 0 0 0 5px rgba(236, 30, 121, 0.18),
	            0 2px 6px rgba(0, 0, 0, 0.6);
}

/* ---- the two amounts ----
   Labelled, because "RM17" and "RM627" side by side say nothing about which
   end is which until you have already dragged something. */
.sf-filters .wc-block-product-filter-price-slider__left,
.sf-filters .wc-block-product-filter-price-slider__right {
	position: relative;
	padding-top: 18px;
}

.sf-filters .wc-block-product-filter-price-slider__left::before,
.sf-filters .wc-block-product-filter-price-slider__right::before {
	position: absolute;
	top: 0;
	left: 2px;
	font-family: var(--sa-tech);
	font-size: 9px;
	font-weight: 700;
	letter-spacing: 0.2em;
	text-transform: uppercase;
	color: var(--sa-faint);
}

.sf-filters .wc-block-product-filter-price-slider__left::before  { content: "From"; }
.sf-filters .wc-block-product-filter-price-slider__right::before { content: "To"; }

.sf-filters .wc-block-product-filter-price-slider__amount,
.sf-filters input.wc-block-product-filter-price-slider__amount,
.sf-filters .wc-block-product-filter-price-slider__content input[type="text"],
.sf-filters .wc-block-product-filter-price-slider__content input[type="number"],
.sf-shop__filters .wc-block-product-filter-price-slider__content input[type="text"] {
	/* width: 100% matters — the inputs measured 332px inside 308px parents
	   and were overflowing their own column. */
	width: 100%;
	min-width: 0;
	height: 34px;
	padding: 0 10px;
	border: 0;
	border-bottom: 1px solid rgba(242, 242, 244, 0.18);
	border-radius: 0;
	background: transparent;
	box-shadow: none;
	color: var(--sa-text);
	font-family: var(--sa-tech);
	font-size: 13px;
	font-weight: 700;
	letter-spacing: 0.06em;
	text-align: left;
	appearance: textfield;
	-moz-appearance: textfield;
	transition: border-color 0.2s ease;
}

.sf-filters .wc-block-product-filter-price-slider__content input:hover {
	border-bottom-color: rgba(242, 242, 244, 0.38);
}

.sf-filters .wc-block-product-filter-price-slider__content input:focus {
	outline: none;
	border-bottom-color: var(--sa-accent);
}

/* the two columns sit under a full-width rail */
.sf-filters .wc-block-product-filter-price-slider__content,
.sf-shop__filters .wc-block-product-filter-price-slider__content {
	display: grid;
	/* repeat(2, minmax(0,1fr)) and an explicit 0 for any implicit track: a
	   third 292/292/0 track was appearing, and with the column-gap counted
	   after the second column the right-hand amount finished 28px short of
	   the rail above it. Visible as the two fields sitting off-centre under
	   a full-width slider. */
	grid-template-columns: repeat(2, minmax(0, 1fr));
	grid-auto-columns: 0;
	column-gap: 28px;
	row-gap: 26px;
	align-items: center;
	margin-top: 4px;
}

.sf-filters .wc-block-product-filter-price-slider__range {
	/* explicit line numbers rather than 1 / -1, which was part of what
	   produced the phantom track */
	grid-column: 1 / 3;
	width: 100%;
}

.sf-filters .wc-block-product-filter-price-slider__left  { grid-column: 1; min-width: 0; }
.sf-filters .wc-block-product-filter-price-slider__right { grid-column: 2; min-width: 0; }

/* WooCommerce leaves an empty node in this wrapper, which auto-placement was
   turning into a third (0px) column. Pinning it to row 1 keeps the track
   count at two. */
.sf-filters .wc-block-product-filter-price-slider__content > :not(.wc-block-product-filter-price-slider__range):not(.wc-block-product-filter-price-slider__left):not(.wc-block-product-filter-price-slider__right) {
	display: none;
}

/* the group's height is the drawer's, not ours */
.sf-filters .wp-block-woocommerce-product-filter-price,
.sf-filters .wp-block-woocommerce-product-filter-price-slider {
	height: auto !important;
	min-height: 0 !important;
}

/* ---- active chips + clear ---- */
.sf-shop__filters .wc-block-product-filter-removable-chips__item {
	border: 1px solid var(--sa-line);
	border-radius: var(--sa-radius);
	background: transparent;
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 500;
	letter-spacing: 0.1em;
	text-transform: uppercase;
	color: var(--sa-text);
}

.sf-shop__filters .wc-block-product-filter-clear-button {
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 500;
	letter-spacing: 0.1em;
	text-transform: uppercase;
	color: var(--sa-accent);
	text-decoration: none;
}


/* -------------------------------------------------------------
 * 4. THE GRID
 * ------------------------------------------------------------- */
/* FIVE ACROSS, RULED (operator, 2026-10-05 — concept revamp/category-grid-c.html).
   Was 3 columns with a 98px row gap, taken from the Lupine reference.

   The operator's note was that the picture was too big: at 3 across the
   image tile measured 413×516 and one product filled a third of the
   screen. Five across takes the tile to about 172px square.

   THE ROW GAP IS NOW ZERO AND THE ROWS ARE RULED INSTEAD. At 98px the
   gap WAS the separator — that is what stopped a big grid reading as a
   spreadsheet. A 98px gap between small cards would leave them stranded,
   so the job passes to a hairline, and the page reads as a parts
   catalogue rather than a lookbook. The rule is drawn on each CELL (see
   below), never as a gap, so it survives a short last row. */
.sf-shop__grid ul.wc-block-product-template {
	display: grid;
	/* THREE ACROSS (operator, 2026-10-22: "look at that website, the card is
	   big"). The count has now gone five -> four -> three, each time for the
	   same reason, and this is the step that matches the reference:
	   logitechg.com runs three 416px cards, and three inside the 1180px
	   column gives about 381px.

	   THE COLUMN CAP IS THE NEXT LEVER, not the count. 1180px is an operator
	   decision from 2026-10-05 and is left alone here; the reference's grid
	   is 1425 wide inside a 1568 max, so raising the cap is what would take
	   the card the rest of the way to 416. Raise that before going to two
	   across -- two is a comparison, not a catalogue.

	   FOUR ACROSS AGAIN (operator, 2026-11-18: "one row have 4 instead of 3
	   .. so need to make it the card product is small"). So the count has
	   gone five -> four -> three -> four, and this time it is the card width
	   that was asked for rather than the count for its own sake.

	   THE CARD IS SMALLER BY THE SAME AMOUNT AT EVERY WIDTH -- there is no
	   second number to set. The track is 1fr, so the gap is unchanged and
	   each card gives up a quarter of itself: measured 302 -> 222 at a 1430
	   client and 406 -> 302 at 1780. Note the operator's own ~1780 screen
	   now shows the card that 1430 used to, which is the one place this
	   reads as a size change rather than as one more column.

	   IT STEPS BACK TO THREE AT 1280 (see section 8) and that step is not
	   optional: the grid lives in the 1fr track beside a 320px rail, so it
	   only ever has about (viewport - 2 x gutter - 368) to spend. At 1280
	   four cards would be 187px, against the ~280 this card is drawn
	   around. The uncapped page makes that arithmetic viewport-wide --
	   re-do it against the RAIL, not the window, before moving either
	   count. */
	grid-template-columns: repeat(4, minmax(0, 1fr));

	/* EVEN GAPS NOW (operator, 2026-10-05). A big row gap was borrowed from
	   the reference, where the products sit loose on the page and the gap is
	   the only thing separating the rows. Here each product has its own card,
	   so the card edge does that job and a tall row gap just pulls the grid
	   apart. */
	column-gap: clamp(12px, 1.4vw, 18px);
	row-gap: clamp(12px, 1.4vw, 18px);
	margin: 0;
	padding: 0;
	list-style: none;
	counter-reset: sf-item;
}


/* -------------------------------------------------------------
 * 5. ONE CARD
 *
 * The override of woocommerce.css section 9, which draws the card
 * as a bordered panel that lifts on hover.
 *
 * ONE CARD PER PRODUCT, PRICE INSIDE IT (operator, 2026-10-05).
 * The card is this site's own: flat face, NO border, 8px radius, one
 * light shadow top-left and one dark bottom-right. The edge is made of
 * light, not a line — a hairline on top of those shadows reads as a card
 * wearing two edges. Same values as the contact page and the tile board,
 * so the shop matches the rest of the site.
 *
 * THERE IS NO SEPARATE PANEL BEHIND THE PHOTO. An earlier version put a
 * near-black tile (#0f0f0f) there, copied from the reference, and on a
 * dark page it read as a black border drawn around every product. The
 * photo now sits straight on the card face.
 * ------------------------------------------------------------- */
.sf-shop__grid li.wc-block-product {
	display: flex;
	flex-direction: column;
	/* NO FLEX GAP (2026-10-22). It was 8px, which was the card's whole
	   vertical rhythm while the parts were photo / category / name / price.
	   Every gap in the card is now a measured value off the reference --
	   47.2 above the photo, 12 under it, 24 under the swatches, 12 under the
	   name -- and an 8px flex gap was being ADDED to each of them. Measured
	   before this line: 20 where 12 was intended, 32 where 24 was. The
	   margins carry the spacing; this carries none. */
	gap: 0;
	position: relative;
	/* NEUMORPHIC, NO HAIRLINE (operator, 2026-10-05, later the same day).

	   This briefly carried `border: 1px solid rgba(245,245,245,.22)` after a
	   request for "a visible edge". The operator then asked for the card to
	   follow the site's neumorphic language, which is defined by light and
	   shadow and has NO border -- privacy.css, return-refund.css and
	   contact.css all draw their cards that way. A hairline on top of those
	   two shadows reads as a card wearing two edges, which is what it looked
	   like. Do not put the border back without asking: it has now been
	   added and removed once.

	   The values below are the site's shared set, not approximations:
	       rest   -3px -3px  8px lift,  4px  4px 10px sink
	       hover  -4px -4px  9px lift,  5px  5px 12px sink
	       lift   rgba(255,255,255,.022)   sink  rgba(0,0,0,.26)
	   Change them here and in those three files together, or the shop stops
	   matching the information pages.

	   PADDING IS ZERO AT THE TOP AND SIDES so the photograph can run to
	   the card's edge. The words underneath carry their own side padding
	   instead - see the rules further down. overflow:hidden is what makes
	   the photo's top corners follow the card's radius. */
	/* THE PICTURE FILLS THE WHOLE CARD (operator, 2026-10-05).

	   It used to stop about 70% down, with the words on the card below it.
	   Now the photograph covers the card corner to corner and the words sit
	   ON it, over a gradient dark enough to read against a white studio
	   shot as well as a dark poster.

	   The card is square, matching the photographs, so nothing is cropped.
	   justify-content:flex-end pushes the words to the foot; the picture is
	   taken out of the flow and pinned behind them further down. */
	/* TALLER (operator, 2026-10-05): 4/5 portrait rather than square, so
	   282 wide becomes 352 tall instead of 282. See the note on the image
	   below about what filling a taller card costs. */
	/* THE WORDS CAME OFF THE PHOTOGRAPH (operator, 2026-10-22).

	   From 2026-10-05 the picture filled the card corner to corner and the
	   name and price sat ON it over a gradient. The card was rebuilt to the
	   anatomy of logitechg.com/en-gb/shop/c/microphones -- badges in the top
	   corners, photo, then name and price on the card's own face below it.
	   See docs/research/2026-10-22-logitechg-category-page.md.

	   So the CARD no longer carries a ratio and no longer pushes its content
	   to the foot: the photo box carries 1/1 (see the image rule below) and
	   the words follow it in normal flow. The scrim on ::after is switched
	   off with them -- the note there always said it could go once the text
	   moved off the picture, and this is that.

	   The card face, its lack of border and its lack of shadow are UNCHANGED.
	   Everything that long sequence of decisions settled still holds; only
	   the arrangement inside the card is different. */
	/* 24px ALL ROUND, AND NOTHING INSIDE CARRIES ITS OWN GUTTER (2026-10-22).
	   Measured: 416.3 wide, 527.4 tall, padding 24, radius 12. Every child
	   used to carry `padding: 0 14px` of its own because the photograph ran
	   to the card's edge and the words had to inset themselves. The card
	   pads itself now, so those are gone -- if you put one back the text
	   will sit 14px right of the photograph. */
	padding: 24px;
	border: 0;
	/* 12px, measured off the reference card (was 8). The photo box inside
	   carries the same 12. */
	border-radius: 12px;
	overflow: hidden;
	background: var(--sa-card, #0d0d10);
	color: var(--sa-card-ink);
	/* NEUMORPHIC AGAIN, NOW THAT THE CARD HAS A FACE OF ITS OWN
	   (operator chose the inset design, 2026-10-05).

	   The history matters, because it is easy to undo by accident:
	     1. neumorphic set  -> invisible. The page is rgb(10,10,10) and the
	        sink at rgba(0,0,0,.26) renders rgb(7,7,7): three levels.
	     2. sink raised to .55 -> five levels. Still invisible; black has
	        only ten levels of headroom on a near-black page.
	     3. a 1px light ring -> 24 levels, clearly visible, but a hairline
	        rather than the neumorphic look that was asked for.
	     4. THIS: the photo is inset by --sa-inset, so a frame of the card's
	        own #0d0d10 face shows all round it.
	   Step 4 is the one that actually fixes it, and not by making the
	   shadow stronger. Neumorphism needs the panel and the page to be the
	   same colour; a full-bleed photograph is not that, a 6px dark frame
	   is. The card now defines itself the way privacy.css and
	   return-refund.css do, and uses their exact values.

	   So the ring is gone and these are the site's shared numbers again:
	       rest   -3px -3px  8px lift,  4px  4px 10px sink
	       hover  -4px -4px  9px lift,  5px  5px 12px sink
	   Change them here and in those two files together. */
	/* STRONGER THAN THE INFORMATION PAGES, DELIBERATELY (2026-10-08).
	   Those panels sit on their own lighter backgrounds; a product card sits
	   on rgb(10,10,10), where a black sink has ten levels of headroom and the
	   shared rgba(0,0,0,.26) spends three of them. With the card face now the
	   same colour as the page there is nothing else left to define the card,
	   so the light side does the work: 6% white instead of 2.2%. */
	/* SOFT AND DIFFUSE, NOT STRONG (operator, 2026-10-08: "reduce a shadow
	   for all the card .. i want smooth shadow like second image").

	   This was briefly -5/-5/14 at 6% white over 6/6/16 at 55% black, pushed
	   that hard earlier the same day because the card had just lost its face
	   and its border and there was nothing else describing it. Beside a white
	   rim photograph it read as a hard dark rim rather than a shadow.

	   Smooth comes from BLUR, not from opacity: the sink is now wider (28px)
	   and much weaker (38%), with a small vertical bias so the card reads as
	   lifted off the page rather than lit from one corner. The white side
	   stays, but barely -- just enough to keep the top-left from going dead.

	   Still stronger than the information pages' shared set, which is right:
	   those panels sit on their own lighter backgrounds and these sit on
	   rgb(10,10,10). Do not copy these numbers back to contact/privacy. */
	/* NO SHADOW AT ALL (operator, 2026-10-08: "remove the shadow").

	   The end of a long walk, and the whole route is worth knowing before
	   anyone adds one back:
	     border          -> removed, the site's language is not edges
	     neumorphic set  -> invisible on rgb(10,10,10), three levels
	     strong shadow   -> visible, but a hard rim beside a white photo
	     soft wide shadow-> better, still unwanted
	     none            -> this

	   Nothing now describes the card: no border, no face (it is
	   transparent), no shadow. What separates one product from the next is
	   the feathered photograph and the grid's own gap, and that is the
	   intended result rather than an oversight. The cards are images on a
	   page, not panels.

	   If a shadow is ever wanted back, it belongs here, and read the note on
	   the feather first -- the two have to be judged together, because a
	   shadow outside a photo that fades to nothing draws a rim around empty
	   space. */
	/* A SHADOW IS BACK (operator, 2026-11-16), AND THE NOTE ABOVE IS WHY
	   THIS NEEDS READING RATHER THAN SKIMMING.

	   "remove the shadow" (2026-10-08) was the end of the walk recorded
	   above, and this reverses it. It is not an oversight — the operator
	   asked for the product card to take the filter rail's card, saw it
	   rendered that way on the localhost mock first, and confirmed. Flagged
	   to them in those terms when it went in.

	   TWO THINGS HAVE CHANGED SINCE THAT INSTRUCTION, and both are what make
	   this safe rather than a straight contradiction:

	     - THE WARNING'S PRECONDITION IS GONE. "a shadow outside a photo that
	       fades to nothing draws a rim around empty space" was about the
	       FEATHER MASK, and that mask was itself removed later the same day
	       (see the elimination list further down this file). The photo has
	       hard edges and 8px corners now, inside the card's own 24px
	       padding, so there is no fading edge for a shadow to rim.
	     - THE CARD HAD NO FACE THEN. It was transparent on 2026-10-08, which
	       is the state the whole list was eliminating options against. It
	       has had a face since 2026-10-22.

	   The values are the RAIL's, which are section 5's on the home page —
	   dealer-wall.css. Not the information pages' lighter set: those panels
	   sit on their own backgrounds, these sit on the weave. Change them with
	   the rail or the two stop matching, which is the whole point of the
	   request. */
	box-shadow: -3px -3px 8px var(--sa-lift), 4px 4px 10px var(--sa-sink);
	counter-increment: sf-item;
	transition: box-shadow 0.35s cubic-bezier(0.22, 0.61, 0.36, 1),
	            transform 0.35s cubic-bezier(0.22, 0.61, 0.36, 1);
}

.sf-shop__grid li.wc-block-product:hover {
	/* The shadow step is the site's shared hover; the 3px rise is the shop's
	   own and has no equivalent on the information pages, which do not move.
	   A product card is a target, so it answers the pointer. */
	transform: translateY(-3px);
	/* THE EMBOSS DOES NOT CHANGE ON HOVER, and this line is load-bearing
	   rather than leftover. It read `box-shadow: none` while the card had no
	   shadow at rest, which was a no-op; with the rest state carrying the
	   emboss (2026-11-16) the same line would WIPE it the moment the pointer
	   arrived and the card would flatten into the page.

	   Restating it rather than deleting the declaration keeps the transition
	   honest — and holding it steady is the rail's own rule, and the footer
	   wordmark's it came from: the carving stays, only the lift answers. */
	box-shadow: -3px -3px 8px var(--sa-lift), 4px 4px 10px var(--sa-sink);
}

/* THE INDEX NUMBER — 01, 02, 03 … — drawn by CSS, not by markup.
   WooCommerce renders this archive with blocks and there is no filter that
   adds an element per card without a template override, so the number is a
   counter on the cell. It costs no markup, it renumbers itself on any
   filter or page change, and it cannot fall out of step with what is on
   screen.

   It sits INSIDE the card now, in the top-left corner above the photo,
   where it reads as a catalogue reference rather than as a label floating
   loose over a picture. */
.sf-shop__grid li.wc-block-product::before {
	content: counter( sf-item, decimal-leading-zero );
	/* A BADGE, NOT LOOSE TYPE (operator, 2026-10-22: "keep number but make
	   it the number badge same like sale badge").

	   It was bare text with a drop shadow, which is what you reach for when
	   something has to survive landing on an unknown photograph. A plate
	   does the same job better and -- the point of the change -- makes the
	   number and the sale badge one family, the way the reference's NEW and
	   SALE pills are.

	   THE WIDTH IS FIXED ON PURPOSE. The sale badge below is positioned by a
	   hard left offset, because WooCommerce renders it as its own absolutely
	   positioned element and there is no row for the two to sit in. That
	   offset is only reliable while this pill's width is. 34px holds two
	   digits at this size with room to spare; the grid pages at 12, so the
	   counter never reaches three. If either ever changes, the two numbers
	   move together -- they are --sf-badge-w and the `left` below. */
	/* MEASURED OFF THE REFERENCE BADGE (2026-10-22): a WHITE plate, 12 / 700,
	   padding 8px 12px, radius 4, line-height 13.8 -- which makes it 29.8
	   tall -- at the card's own padding, 24 from the top and 24 from the
	   left. Normal tracking, not this site's control tracking.

	   THE WIDTH IS STILL FIXED, for the same reason as before: the sale badge
	   beside it is positioned by a hard left offset, because WooCommerce
	   renders it as its own absolutely positioned element and there is no row
	   for the two to share. 40px holds two digits at 12px with room, and the
	   grid pages at 12 so the counter never reaches three. Change this and
	   the sale badge's `left` together. */
	--sf-badge-w: 40px;
	position: absolute;
	top: 24px;
	left: 24px;
	z-index: 3;
	display: inline-flex;
	align-items: center;
	justify-content: center;
	min-width: var(--sf-badge-w);
	height: 29.8px;
	padding: 0 12px;
	border-radius: 4px;
	/* The reference's badge is a WHITE plate on a light grey card -- a step
	   lighter than the surface, so it reads as a label rather than a stamp.
	   On a #16161a card the step in the same direction is nearly black, not
	   white: a white pill here is the brightest thing on the page and fights
	   the photograph. So the geometry and the type are the reference's and
	   only the two colours are flipped. */
	background: rgba(8, 8, 9, 0.9);
	font-family: var(--sa-body);
	font-size: 12px;
	font-weight: 700;
	line-height: 13.8px;
	letter-spacing: normal;
	color: var(--sa-card-ink);
	transition: color 0.3s ease, background-color 0.3s ease;
	pointer-events: none;
}

.sf-shop__grid li.wc-block-product:hover::before {
	background: var(--sa-accent);
	color: #fff;
}

/* ---- the colour swatches: INERT (2026-10-22) ----
   NOTHING RENDERS THESE TODAY. superfly_swatches_on_image() still exists in
   functions.php but its add_filter line is commented out, on the operator's
   instruction the same day the row was added ("remove the color"). These
   rules are kept whole, and so is the arithmetic in the photo-margin note
   above, so switching the row back on is one line of PHP and needs nothing
   here.

   What it drew, for whoever turns it back on: a row read from the pa_colour
   attribute, between the photograph and the name. The reference puts this row between the photograph
   and the name, with the selected colour drawn larger than the rest.

   THE ROW IS RENDERED EVEN WHEN IT IS EMPTY, deliberately. Only 33 of 104
   products carry a colour attribute; if the row collapsed on the other 71
   their names would sit 28px higher than their neighbours' and every grid row
   would be ragged. An empty row costs one line and keeps every card
   identical, which is what this geometry depends on.

   There is no selection to show on an archive, so the FIRST colour is drawn
   as the selected one. The row is a sample of what the product comes in, not
   a control -- the card is a link, and the span is aria-hidden. */
.sf-shop__grid li.wc-block-product .sf-shop__swatches {
	display: flex;
	align-items: center;
	gap: 12px;
	height: 28px;       /* the reference's holder: a 16px dot, 6px either side */
	margin: 0 0 24px;   /* its gap between this row and the name */
}

.sf-shop__grid li.wc-block-product .sf-shop__swatch {
	display: inline-flex;
	flex: none;
	width: 12px;
	height: 12px;
	border-radius: 50%;
	/* A hairline so the dark finishes -- Black, Midnight Black, Prism Black --
	   still read as dots on a #16161a card instead of vanishing into it.
	   The ring was dark while the card was light and had the same job in the
	   other direction; whichever way the card goes, SOME of these 32 colours
	   are close to it. */
	box-shadow: 0 0 0 1px rgba(242, 242, 244, 0.45);
}

.sf-shop__grid li.wc-block-product .sf-shop__swatch.is-first {
	width: 16px;
	height: 16px;
	box-shadow: 0 0 0 1.6px rgba(242, 242, 244, 0.9);
}

/* The disc is a child rather than the span's own background, so a two-tone
   finish can be a gradient without the ring inheriting it. */
.sf-shop__grid li.wc-block-product .sf-shop__swatch i {
	display: block;
	width: 100%;
	height: 100%;
	border-radius: 50%;
	background: var(--sf-swatch, #6b6b70);
}

/* ---- the SALE badge ----
   WooCommerce's own element, switched on in templates/archive-product.html
   and restyled here to match the number beside it. It renders ONLY on a
   product carrying a sale price, so on most cards the corner holds the
   number alone.

   It is a child of the image wrapper, which is `position: relative` and
   starts at the card's top-left corner, so these offsets land in the same
   place the number does. The `left` is the number's own left plus its
   fixed width plus a 6px gap -- see the note on --sf-badge-w above.

   NOTHING ON THIS SITE IS ON SALE TODAY (checked 2026-10-22:
   wc_get_product_ids_on_sale() returns 0), so this rule has been verified
   against the element injected into the page by hand, not against a real
   discounted product. Look at it again the first time a sale price is set. */
.sf-shop__grid li.wc-block-product .wc-block-components-product-sale-badge {
	position: absolute;
	top: 24px;
	left: calc(24px + 40px + 8px);
	right: auto;
	bottom: auto;
	z-index: 3;
	display: inline-flex;
	align-items: center;
	justify-content: center;
	height: 29.8px;
	min-height: 0;
	margin: 0;
	padding: 0 12px;
	border: 0;
	border-radius: 4px;
	/* A PLATE WITH COLOURED WORDS, which is how the reference draws it -- its
	   sale badge is white with GREEN text, not a filled pill. Green is not a
	   SUPERFLY colour, so the brand magenta takes that role; the plate
	   matches the index badge beside it, dark rather than white, for the
	   reason in that rule. */
	background: rgba(8, 8, 9, 0.9);
	font-family: var(--sa-body);
	font-size: 12px;
	font-weight: 700;
	line-height: 13.8px;
	letter-spacing: normal;
	text-transform: uppercase;
	color: var(--sa-accent);
	pointer-events: none;
}

/* ---- the wishlist heart ----
   Injected INSIDE the price element by superfly_wishlist_on_card() in
   functions.php. ShopEngine draws the button itself; this only places it.

   On the price line, not pinned to a corner. An earlier version sat top
   right opposite the index number; it was moved here on the operator's
   instruction (2026-10-05) so the price and the heart read as one row. */
/* The price's own content, wrapped into one flex item by
   superfly_put_inside_price() so a variable product's RANGE
   ("RM574.00 - RM627.00", three spans) cannot split into three items and
   shove the heart out of the card. `min-width: 0` is what lets it actually
   shrink -- a flex item's default min-width is its min-content.

   It WRAPS rather than truncating. A two-up card at 375px leaves about 94px
   beside the heart, and a range needs ~131px on one line, so something has
   to give: an earlier version clipped it with an ellipsis and rendered
   "RM574.00 - RM627.0...". Never truncate a price. Wrapping lets the range
   take two lines inside its own item while the heart stays on the row beside
   it, which costs a few pixels of card height on mobile and nothing at all
   on desktop, where it fits on one line anyway. */
.sf-shop__grid li.wc-block-product .sf-shop__priceval {
	min-width: 0;
	flex: 0 1 auto;
}

.sf-shop__grid li.wc-block-product .sf-shop__wish {
	/* TOP RIGHT, OPPOSITE THE BADGES (operator, 2026-10-22).

	   The whole history, because this corner has now been argued twice: it
	   began top right, moved to the price line on 2026-10-05 so the price
	   and the heart read as one row, and came back here when the card was
	   rebuilt to the reference's anatomy -- the price row carries the words
	   ADD TO CART now and there is no room for a third item on it.

	   The old objection to this corner was that a fixed offset broke when a
	   title wrapped and moved the price. It does not apply up here: the top
	   of the card does not move, whatever the title does.

	   superfly_wishlist_on_image() in functions.php appends this after the
	   image block, so it is a child of the card `li`, which is
	   `position: relative` -- these offsets are against the card. */
	/* 24 from the top and the right -- the card's own padding, so the heart
	   lines up with the badge opposite it. */
	position: absolute;
	top: 24px;
	right: 24px;
	z-index: 3;
	display: inline-flex;
	align-items: center;
	flex: none;
	line-height: 1;
}

/* ---- the button ----
   ShopEngine ships it as a filled pink "badge" block. Only the look is
   overridden here — the element, its classes and its data-pid are still
   ShopEngine's, so the plugin's own list, its /wishlist/ page and its REST
   route are untouched.

   THE LOOK DID NOT CHANGE ON 2026-11-16. Everything in this block is the
   same bare outline heart it has drawn since 2026-10-22; what changed is
   that the control now works. See assets/js/wishlist-card.js for the four
   faults that were measured on the live archive, of which the first is the
   only one a customer could describe: a guest click ran ShopEngine's own
   alert("Please login first.") and nothing else.

   NO DISC, NO RING, NO PLATE, IN ANY STATE (operator, 2026-11-16: "i think
   A , but i dont want circle border"). Three round things were offered on
   the localhost proposal and all three were turned down — a glass plate
   under the glyph on hover, a faint magenta one behind the saved state, and
   a ring pulsing out of the glyph on press. That also puts the control back
   on this card's own earlier ruling, recorded above: "a bare outline heart
   in the corner, no background and no ring."

   WHAT CARRIES THE AFFORDANCE INSTEAD, now that nothing is painted behind
   the glyph: the glyph itself. Magenta and 12% larger on hover or focus.
   Colour and motion, no new shape. */
.sf-shop__grid li.wc-block-product .sf-shop__wish a,
.sf-shop__grid li.wc-block-product .sf-shop__wish .se-btn,
.sf-shop__grid li.wc-block-product .sf-shop__wish .badge {
	/* ShopEngine ships this button `position: absolute`, with no offsets set.
	   That is the trap: with all offsets auto an absolute box sits at its
	   STATIC position, so it looks perfectly in place while being out of
	   flow -- the wrapper span collapses to 0x0 and nothing stops the heart
	   being painted outside the card. Measured at 375px before this line:
	   wrapper 0x0 and the button 15-17px past the right edge of the card on
	   half the tiles. Forcing it back into flow is what makes the row
	   responsive, so do not drop this. */
	position: static;
	display: inline-flex;
	align-items: center;
	justify-content: center;

	/* 44px OF TARGET, 28px OF INK, AND NOTHING PAINTED ON EITHER.

	   It was a 28px box -- under the 44px a finger wants, with a 16px glyph
	   inside it and no plate, so a customer on a phone was aiming at a
	   hairline. The box is 44 now and the NEGATIVE MARGINS put the glyph
	   back where it has always been: the wrapper sits at top 24 / right 24,
	   so -8 on each pulls the larger box out until the glyph's centre lands
	   at 38 / 38 from the card's corner -- the pixel the 28px box centred on.
	   Measured live before and after; the control does not move.

	   CHANGE THE PAIR TOGETHER. A width without the matching margin shifts
	   the heart off the line it shares with the 01 badge opposite it. */
	width: 44px;
	height: 44px;
	margin: -8px -8px 0 0;
	padding: 0;

	border: 0;
	background: none;
	box-shadow: none;
	border-radius: 0;
	color: var(--sa-card-ink);
	font-size: 16px;
	line-height: 1;
	/* the icon-font fallback is a FONT, so it inherits letter-spacing, and
	   the 0.16em the card's type carries pushed it 2.6px off its own box --
	   measured, the 16px heart was rendering 18.6 wide and right of centre */
	letter-spacing: 0;
	cursor: pointer;
	/* colour only. Nothing else about this control transitions, because
	   nothing else about it changes. */
	transition: color .25s ease;
}

/* THE GLYPH IS AN INLINE SVG, NOT THE ICON FONT, and that is what lets the
   saved state be a FILL rather than a second glyph. An icon font paints a
   shape chosen by its codepoint; `fill: currentColor` on a path is one
   declaration. superfly_wishlist_button() in functions.php swaps it in
   server-side, so there is no flash of the old glyph on load and the
   control is correct with JavaScript off.

   IF THAT SWAP EVER FAILS the <i> ShopEngine rendered is still there and
   still inherits `color` from the rule above, so the card degrades to the
   outline heart it had before rather than to an empty corner. */
.sf-shop__grid li.wc-block-product .sf-shop__wish a .sfw-heart {
	width: 19px;
	height: 19px;
	flex: none;
	fill: none;
	stroke: currentColor;
	stroke-width: 1.8;
	stroke-linejoin: round;
	transition: transform .18s cubic-bezier(.34, 1.56, .64, 1), fill .18s ease;
}

/* hover and keyboard focus are the same state: there is one affordance and
   it should not depend on which input you arrived with */
.sf-shop__grid li.wc-block-product .sf-shop__wish a:hover,
.sf-shop__grid li.wc-block-product .sf-shop__wish a:focus-visible {
	color: var(--sa-accent);
}

.sf-shop__grid li.wc-block-product .sf-shop__wish a:hover .sfw-heart,
.sf-shop__grid li.wc-block-product .sf-shop__wish a:focus-visible .sfw-heart {
	/* 12%, not more: the glyph sits 24px from two card edges and anything
	   that reads as a jump up here pulls the eye off the photograph, which
	   is the card's subject. */
	transform: scale(1.12);
}

/* THE FOCUS RING IS THE ONE RING LEFT, and it is a rounded RECTANGLE, not a
   circle. It appears only for a keyboard user, for whom it is the sole
   indication of where they are on the page. The negative offset pulls it in
   off the 44px target so it hugs the glyph instead of floating in the
   card's corner. */
.sf-shop__grid li.wc-block-product .sf-shop__wish a:focus-visible {
	outline: 2px solid var(--sa-accent);
	outline-offset: -8px;
	border-radius: 6px;
}

/* SAVED: the outline fills. This is the one state a customer has to read at
   a glance down a row of cards, and it is why the glyph had to stop being
   an icon font. */
.sf-shop__grid li.wc-block-product .sf-shop__wish a.active {
	color: var(--sa-accent);
}

.sf-shop__grid li.wc-block-product .sf-shop__wish a.active .sfw-heart {
	fill: currentColor;
}

/* THE PRESS. A spring on the glyph, which is motion rather than a shape.
   Applied by wishlist-card.js, which removes the class 600ms later so the
   next press can restart it. */
.sf-shop__grid li.wc-block-product .sf-shop__wish a.is-sfw-pop .sfw-heart {
	animation: sfWishPop .42s cubic-bezier(.34, 1.56, .64, 1);
}

/* A MOUSE IS STILL OVER THE CONTROL AT THE MOMENT OF THE CLICK, so the
   plain keyframes would run 1 -> 1.32 -> 1 and then snap back up to the
   hover's 1.12 the instant they finished. This set starts and ends where
   the hover leaves it. A touch user never hovers and gets the plain one. */
.sf-shop__grid li.wc-block-product .sf-shop__wish a.is-sfw-pop:hover .sfw-heart,
.sf-shop__grid li.wc-block-product .sf-shop__wish a.is-sfw-pop:focus-visible .sfw-heart {
	animation-name: sfWishPopHover;
}

@keyframes sfWishPop {
	0%   { transform: scale(1); }
	38%  { transform: scale(1.32); }
	100% { transform: scale(1); }
}

@keyframes sfWishPopHover {
	0%   { transform: scale(1.12); }
	38%  { transform: scale(1.42); }
	100% { transform: scale(1.12); }
}

@media (prefers-reduced-motion: reduce) {
	.sf-shop__grid li.wc-block-product .sf-shop__wish a .sfw-heart {
		transition: fill .18s ease;
	}
	.sf-shop__grid li.wc-block-product .sf-shop__wish a:hover .sfw-heart,
	.sf-shop__grid li.wc-block-product .sf-shop__wish a:focus-visible .sfw-heart {
		transform: none;
	}
	.sf-shop__grid li.wc-block-product .sf-shop__wish a.is-sfw-pop .sfw-heart {
		animation: none;
	}
}


/* ---- the toast ----
   What alert("Please login first.") is replaced BY. Written by
   wishlist-card.js, one element reused for the life of the page.

   DELIBERATELY NOT ShopEngine's .parent-notification, which is appended
   fresh on every click and never removed -- assets/js/wishlist-toast.js
   exists solely to prune those corpses, and is still needed for the
   /wishlist/ page, which this control does not touch.

   Bottom right is ShopEngine's own `wishlist_position` default, so the two
   land in the same corner on the pages that still use theirs. */
.sf-wish-toast {
	position: fixed;
	right: 20px;
	bottom: 20px;
	z-index: 9999;
	display: flex;
	align-items: center;
	gap: 12px;
	max-width: min(360px, calc(100vw - 40px));
	padding: 14px 18px;
	border-radius: 4px;
	background: rgba(8, 8, 9, .96);
	/* the card's own emboss, so the toast is the same material as the grid
	   it is reporting on -- see --sa-lift / --sa-sink above */
	box-shadow: -3px -3px 8px var(--sa-lift),
	            4px 4px 18px rgba(0, 0, 0, .5);
	color: var(--sa-text);
	font-family: var(--sa-body);
	font-size: 13px;
	line-height: 1.45;
	transform: translateY(14px);
	opacity: 0;
	transition: opacity .22s ease, transform .22s ease;
	/* unreachable until it is shown, so the link inside cannot be tabbed
	   to while the toast is invisible */
	pointer-events: none;
}

.sf-wish-toast.is-on {
	opacity: 1;
	transform: none;
	pointer-events: auto;
}

.sf-wish-toast__bar {
	position: absolute;
	left: 0;
	top: 0;
	bottom: 0;
	width: 3px;
	background: var(--sa-accent);
	border-radius: 4px 0 0 4px;
}

.sf-wish-toast a {
	color: var(--sa-accent);
	text-decoration: underline;
}

@media (prefers-reduced-motion: reduce) {
	.sf-wish-toast { transition: opacity .22s ease; transform: none; }
}


/* ---- ShopEngine's own icon stack, hidden (operator: no eye, no compare)
   ----
   These came BACK when wp:woocommerce/product-button was added to the
   template on 2026-10-05. The button block puts the card into WooCommerce's
   classic loop context, and ShopEngine hooks that context to append its
   wishlist / quick-view / compare controls. Before the button existed the
   hooks never fired on this archive, which is why the card was clean
   without this rule.

   They are DIRECT CHILDREN of the <li>, not wrapped in anything, and
   absolutely positioned top-right -- so the child combinator is what makes
   this safe: the heart the card is supposed to have lives inside
   .sf-shop__wish within the price element, several levels down, and is not
   touched. Without the `>` this would hide that one too.

   The result is one heart per card, on the price line, and nothing else.
   No !important: at (0,3,1) these already outrank ShopEngine's own rules. */
.sf-shop__grid li.wc-block-product > a.shopengine_add_to_list_action,
.sf-shop__grid li.wc-block-product > a.shopengine-quickview-trigger,
.sf-shop__grid li.wc-block-product > a.shopengine_comparison_add_to_list_action,
.sf-shop__grid li.wc-block-product > a.shopengine-wishlist,
.sf-shop__grid li.wc-block-product > a.shopengine-comparison {
	display: none;
}


/* ---- the add-to-cart ICON on the price row ----
   Option A, chosen by the operator 2026-10-05 from
   revamp/card-button-options.html, replacing the 34px full-width bar that
   shipped earlier the same day. The bar was the heaviest element on a card
   whose subject is the photograph, and there was one on every tile.

   This costs the card NO extra height: the control moves onto the price row
   beside the heart, which is where the spare width already was.

   WHAT IT COSTS, measured and accepted: the pair of circles takes about
   66px off the price row, so on a narrow card a price RANGE
   ("RM574.00 - RM627.00") wraps to a second line. The row is built to wrap
   rather than clip -- see .sf-shop__priceval -- so nothing is ever lost, the
   card just grows a line on the two variable wheels.

   The markup is WooCommerce's own button block, captured and moved by
   superfly_capture_cart_button() in functions.php. It is not reimplemented,
   so the ajax add and the simple/variable distinction still work. */
.sf-shop__grid li.wc-block-product .sf-shop__acts {
	display: inline-flex;
	align-items: center;
	gap: 8px;
	flex: none;
}

.sf-shop__grid li.wc-block-product .sf-shop__cart,
.sf-shop__grid li.wc-block-product .sf-shop__cart .wp-block-button,
.sf-shop__grid li.wc-block-product .sf-shop__cart .wc-block-components-product-button {
	display: inline-flex;
	align-items: center;
	margin: 0;
	padding: 0;
	width: auto;
}

.sf-shop__grid li.wc-block-product .sf-shop__cart a,
.sf-shop__grid li.wc-block-product .sf-shop__cart button,
.sf-shop__grid li.wc-block-product .sf-shop__cart .wp-block-button__link {
	/* WORDS AND AN ICON, NOT A CIRCLE (operator, 2026-10-22).

	   It was a 28px disc beside the heart, chosen on 2026-10-05 because a
	   full-width bar was the heaviest thing on the card. The reference puts
	   the label itself on the price row, quietly -- small, unfilled, at the
	   far right -- which is lighter than a bar and says what it does.

	   THIS ALSO FIXES THE ONE REAL WEAKNESS OF THE CIRCLE. A variable
	   product renders "Select options", not "Add to cart", and an icon
	   cannot say that; the disc had to use a sliders glyph and hope. With
	   the label visible the control says SELECT OPTIONS in words, so the two
	   kinds of product are told apart by the thing that actually knows the
	   difference. The sliders icon below stays -- it reads as a mark beside
	   the words now rather than as the whole message.

	   `row-reverse` puts the ::before icon AFTER the text, which is the
	   order the reference uses and the order the words want.

	   NO PLATE. The disc needed a face to read as a control; words do not,
	   and a pill here would be the heaviest thing on the card again. The
	   heart keeps its circle -- it is still a 28px target in the corner. */
	position: static;
	display: inline-flex;
	flex-direction: row-reverse;
	align-items: center;
	justify-content: flex-end;
	width: auto;
	height: auto;
	min-height: 0;
	padding: 0;
	margin: 0;
	/* MEASURED: 12 / 400, rgb(48,49,51), 8px between the words and the cart
	   mark, no tracking. It was 10px tech at 0.14em -- that is this site's
	   control type, and it read as a button label beside a 16px price rather
	   than as the quiet line the reference sets. */
	/* !important ON FIVE PROPERTIES, AND IT IS NOT LAZINESS (2026-10-22).

	   WooCommerce renders this anchor with `wp-element-button` and
	   `has-small-font-size`, and the theme's global styles paint BOTH: a
	   filled magenta button at 14.4px. Measured on the live archive with
	   this rule already in place and no !important -- every card's control
	   was a magenta pill with white words, while the local preview showed
	   the quiet grey line it was written to be.

	   THE PREVIEW COULD NOT HAVE CAUGHT THIS. docs/shop-concepts loads this
	   stylesheet and nothing else, so a rule that loses to the THEME's own
	   sheet renders perfectly there and wrongly on the site. Anything on
	   this card that WordPress also styles has to be checked on the live
	   page, not in the lab.

	   Only the properties that actually lose are forced. gap, tracking,
	   transform and the icon are left alone -- nothing else contests them. */
	font-family: var(--sa-body);
	font-size: 12px !important;
	font-weight: 400 !important;
	letter-spacing: 0;
	line-height: 14px;
	text-transform: uppercase;
	white-space: nowrap;
	gap: 8px;
	border: 0;
	background: none !important;
	box-shadow: none;
	border-radius: 0 !important;
	padding: 0 !important;
	min-height: 0;
	color: var(--sa-card-ink2) !important;
	text-decoration: none;
	cursor: pointer;
	transition: box-shadow 0.25s ease, color 0.25s ease,
	            background-color 0.25s ease, transform 0.2s ease;
}

.sf-shop__grid li.wc-block-product .sf-shop__cart a:hover,
.sf-shop__grid li.wc-block-product .sf-shop__cart button:hover {
	/* the words carry the hover -- a filled magenta disc and a 1.1 scale
	   belonged to a 28px target and would be a magenta block here. Same
	   !important as the rest state, and for the same reason. */
	background: none !important;
	box-shadow: none;
	transform: none;
	color: var(--sa-accent) !important;
}

/* THE LABEL IS VISIBLE (2026-10-22). It was clipped to a screen-reader-only
   rectangle while the control was an icon disc. The clip and the
   `font-size: 0` that went with it are gone rather than commented out,
   because half a clip is worse than none: the `position: absolute
   !important` would have survived any attempt to show the words again and
   the label would have sat on top of the price. */
.sf-shop__grid li.wc-block-product .sf-shop__cart a > *,
.sf-shop__grid li.wc-block-product .sf-shop__cart button > * {
	position: static;
	width: auto;
	height: auto;
	margin: 0;
	padding: 0;
	overflow: visible;
	clip-path: none;
	white-space: nowrap;
	border: 0;
}

/* THE ICON, drawn with a mask so `currentColor` carries it -- that is what
   lets it go white on the magenta hover without a second image. */
.sf-shop__grid li.wc-block-product .sf-shop__cart a::before,
.sf-shop__grid li.wc-block-product .sf-shop__cart button::before {
	content: "";
	flex: none;
	width: 14px;
	height: 14px;
	background: currentColor;
	-webkit-mask: var(--sf-cart-icon) center / contain no-repeat;
	        mask: var(--sf-cart-icon) center / contain no-repeat;
	--sf-cart-icon: url('data:image/svg+xml;utf8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="black" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><circle cx="9" cy="20" r="1.5"/><circle cx="18" cy="20" r="1.5"/><path d="M2 3h3l2.6 12.2a1.6 1.6 0 0 0 1.6 1.3h8.6a1.6 1.6 0 0 0 1.6-1.3L21.5 7H6"/></svg>');
}

/* A VARIABLE PRODUCT GETS A DIFFERENT ICON — sliders, not a cart.
   This is the one real weakness of option A: the button block renders
   "Select options" for a variable product and an icon cannot say that. Two
   of the six wheels are variable, so a cart icon there would promise an add
   that does not happen — it goes to the product page to choose a colour.
   Sliders say "there is a choice to make" without words. */
.sf-shop__grid li.wc-block-product .sf-shop__cart a.product_type_variable::before,
.sf-shop__grid li.wc-block-product .sf-shop__cart a.product_type_grouped::before,
.sf-shop__grid li.wc-block-product .sf-shop__cart a.product_type_external::before {
	--sf-cart-icon: url('data:image/svg+xml;utf8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="black" stroke-width="2" stroke-linecap="round"><path d="M4 6h16M4 12h16M4 18h16"/><circle cx="9" cy="6" r="2.2" fill="white"/><circle cx="15" cy="12" r="2.2" fill="white"/><circle cx="8" cy="18" r="2.2" fill="white"/></svg>');
}

/* mid-add: WooCommerce sets this itself, so the card says something happened */
.sf-shop__grid li.wc-block-product .sf-shop__cart button.loading {
	opacity: 0.6;
	pointer-events: none;
}

/* Woo appends a "View cart" link after a successful add; the cart drawer
   and the header count already report that, so it is not wanted on the tile */
.sf-shop__grid li.wc-block-product .sf-shop__cart .added_to_cart {
	display: none;
}


/* ---- the scrim ----
   The words sit on the photograph now, so they need something to sit
   against. These product shots run from a near-black poster to a white
   studio backdrop, so the gradient has to be heavy enough for white text
   on white card 02 while still looking like shadow on dark card 01.

   It is transparent for the top half, so the product itself is barely
   touched, and reaches almost solid at the very bottom where the price is. */
.sf-shop__grid li.wc-block-product::after {
	/* OFF (2026-10-22). This gradient existed for one reason: the name and
	   price sat on the photograph and were unreadable over a white rim. They
	   sit on the card's own face now, so there is nothing to darken -- and
	   leaving it on would paint a gradient down a picture with no words on
	   it. The note below always said it could go once the text moved off the
	   picture; everything under here is kept intact so putting the overlay
	   card back is one line.

	   `content: none` removes the pseudo-element outright rather than
	   hiding a box that still computes. */
	content: none;
	/* THE EDGE VIGNETTE IS GONE (operator, 2026-10-08: "why the product
	   have shadow inside the image ? please remove it").

	   It was `box-shadow: inset 0 0 44px 16px rgba(0,0,0,.78)` on this
	   pseudo-element, added 2026-10-06 to give the neumorphic light a dark
	   surface to fall on where the photograph was a white studio shot. It
	   worked as a lighting trick and read as a dirty shadow painted inside
	   the picture, which is what was asked to be removed. The card is now
	   defined by its OUTER neumorphic shadow only, the same way contact.css
	   draws its panels.

	   If the white-backed shots ever need help again, do it on the CARD
	   (its face, its shadow), never inside the photo. And note the trap
	   that put it here in the first place: an inset box-shadow on the image
	   WRAPPER paints below its children, so the <img> covers it -- the
	   computed value looks right and the screen does not change.

	   The gradient below stays. That is the text scrim, not the vignette:
	   without it the name and price are unreadable over a white product. */
	position: absolute;
	/* matched to the photo, so the gradient darkens the PICTURE and not the
	   frame around it — at `inset: 0` it bled over the card's face and the
	   frame went muddy down the bottom edge */
	inset: var(--sa-inset);
	border-radius: 8px;
	z-index: 1;
	pointer-events: none;
	/* THE LAST STOP IS OPAQUE, AND IT IS THE PAGE'S OWN COLOUR
	   (operator, 2026-10-08: "why i stil see the border ? in product
	   cards").

	   It used to end at `rgba(10,10,12,0.96)`. Two things were wrong with
	   that and together they drew what looked exactly like a border:

	     * 0.96, not 1 -- so 4% of the PHOTOGRAPH still came through at the
	       very bottom of the card. These shots are mostly light grey
	       studio backdrops, so 4% of ~200 lifted the strip about eight
	       levels above the page.
	     * 10,10,12 against a page of 10,10,10 -- two more levels, blue.

	   The card is `overflow: hidden` with an 8px radius, so that lighter
	   strip was clipped to a rounded rectangle: a crisp, rounded, slightly
	   bright edge sitting exactly where a 1px border would be. Nothing in
	   the cascade declared one -- computed `border` read `0px none`
	   throughout, which is why it survived the first pass.

	   Now it reaches solid rgb(10,10,10) at 90% and holds it to the foot,
	   so the bottom of the card IS the page and there is no edge to see.
	   If the page background ever changes, change this with it. */
	background: linear-gradient(180deg,
		rgba(10, 10, 10, 0.55) 0%,
		rgba(10, 10, 10, 0) 26%,
		rgba(10, 10, 10, 0) 44%,
		rgba(10, 10, 10, 0.82) 72%,
		rgb(10, 10, 10) 90%,
		rgb(10, 10, 10) 100%);
}

/* the words ride above the scrim */
.sf-shop__grid li.wc-block-product .wp-block-post-terms,
.sf-shop__grid li.wc-block-product h2.wp-block-post-title,
.sf-shop__grid li.wc-block-product .wc-block-components-product-price {
	position: relative;
	z-index: 2;
}

/* ---- the photo ----
 * No panel of its own. It sits straight on the card face, and the card
 * is what frames it.
 *
 * TWO PANELS HAVE BEEN TRIED HERE AND BOTH WERE WRONG:
 *   #ffffff  drew a white halo around the (grey-backed) photographs
 *   #0f0f0f  drew a black border around every product on a dark page
 * The operator saw each of them. The card behind is enough.
 *
 * The photo is CONTAINed, never cropped: these shots carry the
 * RIDE.RACE.RESEARCH logo in a corner and cropping cuts it off.
 *
 * WooCommerce puts its own ratio modifier on this element
 * (--aspect-ratio-1-1), which is why the ratio is forced here. */
.sf-shop__grid li.wc-block-product .wc-block-components-product-image {
	/* IN FLOW AT THE TOP OF THE CARD (operator, 2026-10-22). It used to be
	   `position: absolute; inset: var(--sa-inset)`, pinned behind the words.
	   The words are under it now, so it is simply the first thing in the
	   column. --sa-inset is no longer read here; it is still used by the
	   scrim rule above, which is off. */
	position: relative;
	inset: auto;
	z-index: 0;
	/* INSET BY THE WORDS' OWN GUTTER (operator, 2026-10-22). The card has a
	   face now, so the photograph sits ON it rather than being it, and the
	   14px here is the same 14px the category line, the title and the price
	   already carry -- see those rules further down. Everything in the card
	   lines up on one left edge.

	   This is NOT the inset frame that was tried and rejected twice in 2026
	   (see the --sa-inset note). That one kept the card transparent, so the
	   frame was page showing through and the photograph still had to be its
	   own edge. Here the frame is a real surface. */
	/* 47.2px ABOVE IT, 12 BELOW. Both measured off the reference: the photo
	   box sits at y=71.2 in a card whose padding is 24, so 47.2 of clear air
	   holds the badge row above it, and the 12 below is the gap between the
	   photo and the text block. The card pads the sides, so there is no
	   horizontal margin here any more.

	   THE 12 BECAME 36 WHEN THE SWATCH ROW CAME OFF (2026-10-22). The
	   reference's 12 is the gap between the photograph and the TEXT BLOCK,
	   whose first child is the swatch holder; the name follows 24 later.
	   With the holder gone the name IS that first child, so the two gaps
	   meet and the honest value is 12 + 24. It is derived from the
	   reference's own numbers rather than picked -- do not round it to 32 or
	   40 without deciding to leave the measurement behind. */
	margin: 47.2px 0 36px;
	/* SQUARE, AND IT IS THE PHOTO BOX THAT SETS IT NOW (2026-10-22). The
	   card carried 4/5 while the picture was its backdrop; with the words
	   below it the card's height is its content and this box decides the
	   picture's shape. 1/1 matches the source shots exactly, so `cover`
	   below crops nothing -- the 35px-a-side crop the old 4/5 card cost is
	   gone with it. Every card in a row still lines up, because every
	   photo box is the same square.

	   4:3 NOW, NOT SQUARE (2026-10-22). Measured on the reference: 368.3 x
	   276.2 inside a 416.3 card, which is 4:3 exactly. Our source shots are
	   square, so `cover` would crop them top and bottom -- the image rule
	   below is switched to `contain` for that reason, which is also what the
	   reference does with its own padded 4:3 renditions. */
	aspect-ratio: 4 / 3;
	overflow: hidden;

	background: transparent;
	/* 12, matching the card's own corner -- measured off the reference, where
	   the photo box and the card share a radius. This said 8 while the line
	   above said 12, and the later one won; do not reintroduce the pair. */
	border-radius: 12px;

	/* NO FEATHER, NO SHADOW, NO BORDER -- the photograph is simply the card
	   (operator, 2026-10-08: "why the product still have the shadow ?
	   please remove").

	   A mask used to fade the outer 26px of the picture to nothing. On the
	   dark shots it was invisible; on the white studio shots -- and the
	   source JPEGs are pure 255,255,255 right to their edge, verified by
	   sampling the file rather than by eye -- it faded white straight into a
	   near-black page, which IS a dark halo round the product. Once the
	   card's box-shadow was gone it was the only soft edge left on screen,
	   and it read as the shadow that had just been removed.

	   THE WHOLE SEQUENCE, so nobody re-opens a closed question:
	     hairline border        removed
	     lighter card face      removed (it made any frame read as a panel)
	     inset frame            tried twice, worse both times
	     edge vignette          removed
	     neumorphic shadow      removed, after being strengthened and softened
	     feather mask           removed -- this
	   What is left is the photograph, full bleed, hard edged, with 8px
	   corners. The grid gap separates one product from the next and nothing
	   else does. That is the settled answer, reached by eliminating every
	   alternative in turn.

	   The one soft thing that REMAINS is the scrim gradient on ::after, and
	   it is not decoration: the name and price sit on the photograph and
	   over a white rim they are unreadable without it. Do not remove that
	   without moving the text off the picture first. */

}

.sf-shop__grid li.wc-block-product .wc-block-components-product-image a,
.sf-shop__grid li.wc-block-product .wc-block-components-product-image img {
	display: block;
	width: 100%;
	height: 100%;
}

.sf-shop__grid li.wc-block-product .wc-block-components-product-image img {
	/* COVER, NOT CONTAIN, now the card is taller than the photographs are.
	   The source shots are square; in a 4/5 card `contain` would letterbox
	   them and leave card colour above and below, which is the opposite of
	   filling the card.

	   WHAT THIS COSTS: cover scales the square up until it covers the card,
	   so about 35px is cropped from each SIDE. The products are centred, so
	   they survive it — but anything near the left or right edge of a
	   photograph is now outside the frame. Card 01's poster carries a QR
	   code in its bottom corner, which is the one to check. */
	/* CONTAIN, NOT COVER (2026-10-22). It was cover for a square box and a
	   square source, where the two agreed and nothing was cropped. The box
	   is 4:3 now and the photographs are still square, so cover would take
	   12.5% off the top and the bottom of every product. Contain fits the
	   whole square inside the 4:3 and leaves card colour either side, which
	   is what the reference's own c_pad renditions do. */
	/* !important, AND IT IS NOT LAZINESS. WooCommerce's product-image block
	   writes `style="object-fit:cover;aspect-ratio:1/1;"` on the <img>
	   ITSELF, and an inline declaration beats any selector -- only
	   !important can reach past it. Measured on the live archive after this
	   rule first went up without it: computed object-fit was `cover`, and
	   every square photograph was being cropped top and bottom to fill the
	   4:3 box.

	   The inline `aspect-ratio: 1/1` beside it needs no answer: the rule
	   above sets width and height to 100%, and an explicit pair of
	   dimensions makes aspect-ratio inert. */
	object-fit: contain !important;
	object-position: center;
	background: transparent;

	/* TRUE FULL BLEED (operator, 2026-10-05). shop-beast.css sets
	   `padding: 14px` on this image, which held the picture 14px inside a
	   box that was itself flush to the card — so the box reached the edge
	   and the picture did not, leaving a border of card colour around every
	   product. This overrides it; the selector here is (0,3,2) against that
	   file's (0,2,2), so it wins on specificity rather than load order.

	   Nothing gets cropped by this. The photographs are square and the box
	   is square, so the picture fills it exactly — the logo these shots
	   carry near their edge stays in frame. */
	padding: 0;
	transition: transform 0.5s cubic-bezier(0.22, 1, 0.36, 1);
}

.sf-shop__grid li.wc-block-product:hover .wc-block-components-product-image img {
	transform: scale(1.04);
}

/* everything under the tile sits on an 8px rhythm, not 24 */
.sf-shop__grid li.wc-block-product > * + * {
	margin-top: 0;
}

.sf-shop__grid li.wc-block-product .wc-block-components-product-rating {
	margin: 0;
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 500;
	line-height: 12px;
	letter-spacing: 0.1em;
	text-transform: uppercase;
	color: var(--sa-muted);
}

/* ---- the pink line above the name ----
   The product's CATEGORY, rendered by a wp:post-terms block added to the
   loop in templates/archive-product.html. It fills the slot the concept
   gave to the range name ("PROJECT A+"), which could not be taken from the
   title: only 12 of 68 titles begin "PROJECT x" and only 5 also carry a
   bracketed variant, so parsing would have produced a correct card for 5
   products and a broken one for 56. See the note in that template. */
/* THE CATEGORY LINE IS OFF (operator, 2026-10-22: the reference card has no
   such line, and "exactly 100% same" was the instruction). It is still in the
   template and the styling below is left whole, so bringing it back is one
   rule -- delete the display:none immediately under this.

   Worth knowing what it was for before removing it for good: it was chosen on
   2026-10-05 as the only range-like label that is real data on every product.
   The reference uses that slot for a one-line description instead, which is
   what the card now carries. */
.sf-shop__grid li.wc-block-product .wp-block-post-terms {
	display: none;
}

.sf-shop__grid li.wc-block-product .wp-block-post-terms {
	margin: 0;
	padding: 0;
	font-family: var(--sa-tech);
	font-size: 10.5px;
	font-weight: 700;
	line-height: 13px;
	letter-spacing: 0.2em;
	text-transform: uppercase;
	color: var(--sa-accent);
}

.sf-shop__grid li.wc-block-product .wp-block-post-terms a {
	color: inherit;
	text-decoration: none;
}

/* A product filed under more than one category would otherwise run to two
   lines and push every card in the row taller. Show the first, hide the
   rest and the separators between them. */
.sf-shop__grid li.wc-block-product .wp-block-post-terms a ~ a,
.sf-shop__grid li.wc-block-product .wp-block-post-terms .wp-block-post-terms__separator {
	display: none;
}

.sf-shop__grid li.wc-block-product h2.wp-block-post-title {
	/* MEASURED OFF THE REFERENCE (2026-10-22): 16px / 700, line-height 22,
	   letter-spacing -3%, uppercase, rgb(27,27,27). The card pads itself now
	   so there is no gutter here.

	   TWO DELIBERATE DEVIATIONS, both forced by our product names:

	     FAMILY. The reference's name is a bold grotesque. --sa-title is
	     Archivo BLACK, which is a weight heavier than that, so the body
	     family at 700 is the closer match -- it is Archivo, the same
	     typeface, at the reference's own weight.

	     TWO LINES, NOT ONE. The reference clamps to one line on desktop
	     because its names are short ("YETI 2"). Ours run to 55 characters
	     -- "PROJECT A+ SUPER SHOCKS ADJUSTABLE SUSPENSION FAT MAN" -- and
	     one line would cut most of that off. Two is its own md behaviour,
	     so the card still clamps; it just clamps one line later. */
	margin: 0;
	padding: 0;
	font-family: var(--sa-body);
	/* The reference sets 40/40 over two lines, but its product names are
	   short. Ours all begin "Superfly Project X 4 Spoke Sport Rims", which
	   at 40px ran FIVE lines and 200px. Sized to read rather than to
	   shout.

	   DOWN AGAIN to 12.5px for the five-across grid (2026-10-05). The
	   names are long — "PROJECT A+ SUPER SHOCKS ADJUSTABLE SUSPENSION
	   (FAT MAN)" is 55 characters — and the card is only about 172px
	   wide, so 20px ran six lines and made every row a different height. */
	font-size: 16px;
	font-weight: 700;
	line-height: 22px;
	letter-spacing: -0.03em;
	text-transform: uppercase;
	text-align: left;
}

/* THE CLAMP GOES ON THE LINK, NOT THE HEADING -- the same trap the phone
   block below documents, which this rule was written without and promptly
   fell into. The h2 is a flex item of the card, a flex item is BLOCKIFIED, so
   `display: -webkit-box` on it computes to `flow-root` and -webkit-line-clamp
   silently does nothing. Measured: the heading rendered at 22px with the clamp
   sitting in its computed style and the second line clipped mid-glyph. The
   anchor is not a flex item, so the legacy box display survives on it. */
.sf-shop__grid li.wc-block-product h2.wp-block-post-title a {
	display: -webkit-box;
	-webkit-line-clamp: 2;
	line-clamp: 2;
	-webkit-box-orient: vertical;
	overflow: hidden;
	color: var(--sa-card-ink);
	text-decoration: none;
	transition: color 0.2s ease;
}

.sf-shop__grid li.wc-block-product:hover h2.wp-block-post-title a {
	color: var(--sa-accent);
}

/* the product's short description, in the house italic */
.sf-shop__grid li.wc-block-product .wp-block-post-excerpt {
	/* 12px under the name, measured: the name's box ends at 433.4 and the
	   description starts at 445.4 on the reference card. */
	margin: 12px 0 0;
	padding: 0;
}

.sf-shop__grid li.wc-block-product .wp-block-post-excerpt__excerpt,
.sf-shop__grid li.wc-block-product .wp-block-post-excerpt p {
	margin: 0;
	font-family: var(--sa-body);
	/* UPRIGHT, NOT ITALIC (2026-10-22). The house italic was right when this
	   line was the only body copy on a card of technical type. The reference
	   sets it upright at 12/15.6 with -0.36px tracking, and that is what is
	   being matched. */
	font-style: normal;
	/* 18px was right beside a 413px photo. Beside a 172px one it is the
	   loudest thing on the card, so it is sized down and held to two
	   lines — cut short rather than allowed to set the row's height
	   (2026-10-05). Products with no short description are already
	   hidden by the :empty rule below. */
	font-size: 12px;
	line-height: 15.6px;
	letter-spacing: -0.36px;
	display: -webkit-box;
	-webkit-line-clamp: 2;
	-webkit-box-orient: vertical;
	overflow: hidden;
	color: var(--sa-card-ink2);
}

/* THE SHORT DESCRIPTION IS BACK ON (operator, 2026-10-22). It was switched
   off on 2026-10-05 because it mostly repeated the name above it and the grid
   read as a wall of text. The reference card carries exactly this line under
   its name, so it returns with the rest of that anatomy -- and the objection
   is real and still stands: if these cards look like they say everything
   twice, the fix is to rewrite the short descriptions, not to hide them
   again. */

/* THE NAME AND THE DESCRIPTION EACH HOLD A FIXED SLOT (operator,
   2026-10-22: "the gap i dont know what need to instruct you cause others
   product have the description, and certain product doesnt have
   description").

   THE PROBLEM. Grid rows stretch every card to the tallest in the row, and
   the price is bottom-anchored. So a card whose name runs to one line
   instead of two, or which has no description at all, does not come out
   SHORTER -- it comes out with the difference dumped as one empty hole above
   its price. On the suspensions archive that was up to ~100px on some cards
   and nothing on others, and the row read as broken rather than as varied.

   The reference never shows this because every one of its products has a
   short name and a one-line description, so its cards fill themselves.
   Ours do not: 103 of the 104 products have no short description at all, so
   this line is WordPress's fallback -- the first 18 words of the long
   description -- or, where there is no long description either, nothing.

   THE FIX IS GEOMETRY, NOT CONTENT. Both blocks reserve exactly what they
   use at their maximum: two lines of name (2 x 22 = 44) and two lines of
   description (2 x 15.6 = 31.2). Both already clamp to two lines, so
   neither can exceed its slot. With the two variable blocks fixed, every
   card in the catalogue is the same height, the auto margin on the price
   has nothing left to absorb, and the hole is gone -- what is left where a
   description is missing is one empty two-line slot, the same on every such
   card.

   THE REAL CURE IS STILL CONTENT. A written one-line description per
   product fills these slots and makes the card read like the reference's.
   Until then this keeps the grid even. */
.sf-shop__grid li.wc-block-product h2.wp-block-post-title {
	min-height: 44px;     /* 2 x the 22px line-height above */
}

.sf-shop__grid li.wc-block-product .wp-block-post-excerpt,
.sf-shop__grid li.wc-block-product .wp-block-post-excerpt:empty {
	display: block;
	min-height: 31.2px;   /* 2 x the 15.6px line-height above */
}

/* THE AUTO MARGIN HAS TO BE ON THE BLOCK WRAPPER, NOT THE PRICE (2026-10-22).

   `margin-top: auto` is set on .wc-block-components-product-price below, and
   it worked only because the card used to be `justify-content: flex-end` --
   everything was pushed to the foot regardless. With the card a normal
   top-aligned column it does nothing at all: the price element is not the
   card's flex child, the wp-block-woocommerce-product-price DIV around it is,
   and an auto margin on a grandchild cannot move its parent.

   Measured before this rule, at 1280px: the three single-price cards ended
   their price 54px above the card's bottom edge while the padding is 14, so
   40px of slack sat under them and the prices in one row did not line up.
   The range card had no slack and looked correct, which is exactly how this
   hides -- it only shows on the SHORT cards in a row. */
.sf-shop__grid li.wc-block-product .wp-block-woocommerce-product-price {
	/* Bottom-anchored, with a floor. The auto margin pins the price to the
	   foot of the card when there is slack; the padding is the MINIMUM gap
	   between the description and the price when there is not, so a product
	   with a two-line name and a two-line description still breathes. 24 is
	   the reference's own gap between those two blocks. */
	margin-top: auto;
	padding-top: 24px;
}

/* MUTED, not accent — see the note at the top of this file */
.sf-shop__grid li.wc-block-product .wc-block-components-product-price {
	display: block;
	margin: 8px 0 0;
	padding: 0;
	margin-top: 0;      /* the wrapper above carries the auto margin now */
	margin-left: 0;     /* the card pads itself -- no gutter here */
	margin-right: 0;
	padding-top: 0;
	/* the price and the heart share this line */
	display: flex;
	align-items: center;
	justify-content: space-between;
	gap: 10px;
	/* NO RULE ABOVE THE PRICE (operator, 2026-10-08: "why i stil see the
	   border ? in product cards"). This was
	   `border-top: 1px solid var(--sa-line)` -- a real border, drawn inside
	   the card across its full width, and the only actual border left on a
	   product card after the 2026-10-08 pass. The padding-top above keeps
	   the price off the title without it. */
	/* MEASURED: the amount is 16 / 700 in rgb(48,49,51), and the row's items
	   sit on their baselines rather than centred -- the CTA beside it is
	   12px, and centring put the words half a pixel above the price's
	   baseline. */
	align-items: flex-end;
	gap: 12px;
	font-family: var(--sa-body);
	font-size: 16px;
	font-weight: 700;
	line-height: 20px;
	letter-spacing: 0;
	text-transform: none;
	text-align: left;
	color: var(--sa-card-ink2);
}

/* the struck original on a sale card: the same size as the price it replaces,
   at half the ink, which is how the reference prints it */
.sf-shop__grid li.wc-block-product .wc-block-components-product-price del {
	color: var(--sa-faint);
	font-weight: 400;
}

.sf-shop__grid li.wc-block-product .wc-block-components-product-price ins {
	text-decoration: none;
	color: var(--sa-accent);
}

/* ---- the ShopEngine badges: REMOVED (operator, 2026-10-05) ----
 * This block pinned wishlist, quick-view and compare to the tile's top
 * right as 32px icons. All three reasons for it are gone:
 *   - quick-view and compare are hidden entirely (operator's decision)
 *   - the heart moved onto the PRICE ROW, inside .sf-shop__wish
 * and leaving it in was an active bug rather than dead weight. Its
 * `:hover` set `background: var(--sa-accent)` at (0,4,1); the newer rule
 * on .sf-shop__wish set `color: var(--sa-accent)` at (0,4,2). The newer
 * one won the colour, the old one still won the background, so hovering
 * the heart painted a MAGENTA ICON ON A MAGENTA CIRCLE and the heart
 * vanished. Reported, then reproduced by reading both computed values:
 * background rgb(236,30,121), icon rgb(236,30,121).
 *
 * The hover and saved states now live with the rest of the heart, up in
 * "the wishlist heart", and both set the icon explicitly to #fff. */


/* -------------------------------------------------------------
 * 6. NOTHING MATCHED
 * ------------------------------------------------------------- */
.sf-shop__grid .wp-block-woocommerce-product-collection-no-results {
	padding: 96px 0;
	text-align: center;
	font-family: var(--sa-title);
	font-size: 40px;
	/* 700, NOT 400. This was Archivo Black, which has ONE weight and is
	   black at 400; Inter at 400 in the same place reads as a caption.
	   category-grid.css records the same correction from the same cause
	   when section 2 changed face. */
	font-weight: 700;
	line-height: 40px;
	letter-spacing: 0.02em;
	text-transform: uppercase;
	color: rgba(242, 242, 244, 0.2);
}


/* -------------------------------------------------------------
 * 7. PAGINATION — "Grid plate" (option A, operator 2026-10-05,
 *    chosen from revamp/pagination-options.html)
 *
 * Replaces WordPress's default row of 10px grey numbers, which
 * was correct and completely anonymous — the one element on the
 * archive that could have belonged to any site, sitting as the
 * last thing a customer sees before the footer.
 *
 * Every number is a plate in the card's own neumorphic language:
 * same 6px radius, same light-above / shadow-below set. The
 * current page is FILLED magenta and pressed IN (inset shadow)
 * rather than raised, so "where I am" is a different physical
 * state, not just a different colour — that still reads for
 * someone who cannot separate the magenta from the grey.
 *
 * THE MARKUP IS WORDPRESS'S, UNCHANGED:
 *   nav.wp-block-query-pagination
 *     a.wp-block-query-pagination-previous      ("Previous Page")
 *     div.wp-block-query-pagination-numbers
 *       span.page-numbers.current / a.page-numbers / span.dots
 *     a.wp-block-query-pagination-next          ("Next Page")
 * Prev and next carry PLAIN TEXT and no arrow element — the
 * template does not enable them — so the chevrons below are drawn
 * with a mask, the same technique as the cart icon on the card.
 * ------------------------------------------------------------- */
.sf-shop__grid .wp-block-query-pagination {
	display: flex;
	align-items: center;
	justify-content: center;
	gap: 8px;
	padding: var(--sa-step) 0 0;
}

.sf-shop__grid .wp-block-query-pagination-numbers {
	display: flex;
	align-items: center;
	gap: 8px;
}

/* ---- the number plates ---- */
.sf-shop__grid .wp-block-query-pagination .page-numbers {
	display: inline-flex;
	align-items: center;
	justify-content: center;
	min-width: 38px;
	height: 38px;
	padding: 0 10px;
	border-radius: 6px;
	background: #16161a;
	box-shadow: 0 0 0 1px rgba(242, 242, 244, 0.14),
	            -2px -2px 6px rgba(255, 255, 255, 0.022),
	             3px 3px 8px rgba(0, 0, 0, 0.26);
	font-family: var(--sa-tech);
	font-size: 12px;
	font-weight: 600;
	letter-spacing: 0.06em;
	color: var(--sa-muted);
	text-decoration: none;
	transition: color 0.22s ease, box-shadow 0.22s ease;
}

.sf-shop__grid .wp-block-query-pagination a.page-numbers:hover {
	color: var(--sa-text);
	box-shadow: -3px -3px 8px rgba(255, 255, 255, 0.022),
	             4px 4px 10px rgba(0, 0, 0, 0.26);
}

/* the page you are on: filled, and pressed into the surface */
.sf-shop__grid .wp-block-query-pagination .page-numbers.current {
	color: #fff;
	background: var(--sa-accent);
	box-shadow: inset 2px 2px 6px rgba(0, 0, 0, 0.34),
	            0 4px 14px rgba(236, 30, 121, 0.34);
}

/* the ellipsis is not a destination, so it gets no plate */
.sf-shop__grid .wp-block-query-pagination .page-numbers.dots {
	min-width: 22px;
	padding: 0;
	background: none;
	box-shadow: none;
	color: var(--sa-faint);
}

/* ---- previous / next ---- */
.sf-shop__grid .wp-block-query-pagination-previous,
.sf-shop__grid .wp-block-query-pagination-next {
	display: inline-flex;
	align-items: center;
	gap: 8px;
	height: 38px;
	padding: 0 16px;
	border-radius: 6px;
	background: #16161a;
	box-shadow: 0 0 0 1px rgba(242, 242, 244, 0.14),
	            -2px -2px 6px rgba(255, 255, 255, 0.022),
	             3px 3px 8px rgba(0, 0, 0, 0.26);
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 700;
	letter-spacing: 0.14em;
	text-transform: uppercase;
	color: var(--sa-muted);
	text-decoration: none;
	transition: color 0.22s ease, box-shadow 0.22s ease;
}

.sf-shop__grid .wp-block-query-pagination-previous { margin-right: 4px; }
.sf-shop__grid .wp-block-query-pagination-next     { margin-left: 4px; }

.sf-shop__grid .wp-block-query-pagination-previous:hover,
.sf-shop__grid .wp-block-query-pagination-next:hover {
	color: var(--sa-accent);
	box-shadow: -3px -3px 8px rgba(255, 255, 255, 0.022),
	             4px 4px 10px rgba(0, 0, 0, 0.26);
}

/* the chevrons, drawn rather than marked up — see the note above */
.sf-shop__grid .wp-block-query-pagination-previous::before,
.sf-shop__grid .wp-block-query-pagination-next::after {
	content: "";
	flex: none;
	width: 14px;
	height: 14px;
	background: currentColor;
	-webkit-mask: var(--sf-chev) center / contain no-repeat;
	        mask: var(--sf-chev) center / contain no-repeat;
}

.sf-shop__grid .wp-block-query-pagination-previous::before {
	--sf-chev: url('data:image/svg+xml;utf8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="black" stroke-width="2.4" stroke-linecap="round" stroke-linejoin="round"><path d="M15 5l-7 7 7 7"/></svg>');
}

.sf-shop__grid .wp-block-query-pagination-next::after {
	--sf-chev: url('data:image/svg+xml;utf8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="black" stroke-width="2.4" stroke-linecap="round" stroke-linejoin="round"><path d="M9 5l7 7-7 7"/></svg>');
}


/* -------------------------------------------------------------
 * 8. NARROWER SCREENS
 *
 * The reference is a fixed 3 x 416 and stops working below that.
 * These breakpoints are the only additions to the measurement; the
 * desktop numbers above are untouched.
 * ------------------------------------------------------------- */
/* The columns step 3 - 2 (2026-10-22, was 4 - 3 - 2). They never reach
   one: a single cell per row is a list, not a catalogue, and two fit a
   phone comfortably.

   The step down to 3 that used to live at 1200px is gone with the change
   to three across -- it would have been a rule setting the value the grid
   already has. 980px still takes it to two. */
@media (max-width: 1200px) {
	.sf-shop__filters .wc-block-product-filters {
		grid-template-columns: repeat(2, minmax(0, 1fr));
		gap: 32px;
	}

	.sf-shop__grid li.wc-block-product h2.wp-block-post-title {
		font-size: 12.5px;
		line-height: 16.5px;
	}
}

/* THREE ACROSS FROM HERE DOWN (2026-11-18, with the desktop at four).
   1280 rather than the 1200 the block above uses, because the binding
   number is the grid's own track and not the window: at a 1280 client the
   gutter is 53.8, so the row is 1172 and the grid gets 804 of it once the
   320px rail and the 48px step are taken off. Four cards there measure
   187px; three measure 256.

   It must stay ABOVE the 980px block -- both match a phone, and the later
   one has to win. */
@media (max-width: 1280px) {
	.sf-shop__grid ul.wc-block-product-template {
		grid-template-columns: repeat(3, minmax(0, 1fr));
	}
}

@media (max-width: 980px) {
	.sf-shop__grid ul.wc-block-product-template {
		grid-template-columns: repeat(2, minmax(0, 1fr));
	}
}

@media (max-width: 760px) {
	.sf-shop {
		--sa-step: 20px;
	}

	.sf-shop__head .wp-block-term-description {
		font-size: 17px;
		line-height: 23px;
	}

	.sf-shop__grid ul.wc-block-product-template {
		grid-template-columns: repeat(2, minmax(0, 1fr));
	}

	.sf-shop__filters .wc-block-product-filters {
		grid-template-columns: minmax(0, 1fr);
		gap: 28px;
	}

	/* THE COUNT TAKES ITS OWN LINE HERE (2026-11-16). Three things on one
	   row is one too many at this width: measured at 390 the count wrapped
	   to two lines, the sorting truncated to "DEFA…" and the page carried
	   4px of horizontal overflow. It reads

	       SHOWING 1–16 OF 17 RESULTS
	       [ FILTERS + ]        SORT BY DEFAULT

	   with `order: -1` putting it first without moving it in the markup,
	   where it belongs after the Filters button for anyone tabbing. */
	.sf-shop__bar {
		flex-wrap: wrap;
		gap: 10px;
	}

	.sf-shop__bar .wc-block-product-results-count {
		flex: 1 0 100%;
		order: -1;
	}

	/* AND THE SORTING STOPS SHRINKING. This was `1 1 auto` so two boxed
	   controls could share a narrow row; as a line of type that only
	   bought a truncated value. */
	.sf-shop__bar .wc-block-catalog-sorting,
	.sf-shop__bar .sf-sort,
	.sf-shop__bar form.woocommerce-ordering {
		flex: 0 0 auto;
	}

	.sf-sort__btn {
		padding: 0;
		letter-spacing: 0.08em;
	}

	.sf-shop__filters-toggle {
		padding: 0 14px;
	}

	/* full width under the bar, so a long label is never clipped on a phone */
	.sf-sort__list {
		left: 0;
		right: 0;
		max-width: none;
	}

	.sf-sort__opt {
		white-space: normal;
	}

	/* ---- THE PHONE CARD (operator: "looks broken, not responsive at all")
	   ----
	   The card is `aspect-ratio: 4/5`, which is right at the desktop width
	   of 282 (giving 352 tall) and far too short at 162 (giving 202). The
	   words did not shrink with it: measured at 375px the category, a
	   SIX-line title and the price row came to 187px of a 202px card — 93%
	   — so the text covered essentially the whole photograph and ran over
	   the product's own printed branding. That is the "broken" look; it was
	   never a layout fault, it was a card too short for its contents.

	   Two changes, both aimed at giving the picture its half back:
	     - the title is clamped to two lines
	     - the card is taller in proportion (3/4 rather than 4/5)
	   Together they take the text block to roughly half the card, so the
	   product is visible above it. */
	/* THE CARD PADS ITSELF ON A PHONE TOO, just less (2026-10-22). 24 inside
	   a 162px card leaves 114 of photograph; 16 leaves 130. The badges and
	   the heart follow this value -- see the two rules under it.

	   ITS FIXED RATIO IS GONE WITH IT. `aspect-ratio: 3/4` was right while
	   the photograph WAS the card and its height could be dictated. The card
	   is padding, a 4:3 photo box, a swatch row, a name, a description and a
	   price now, which together do not fit a 3/4 box -- holding the ratio
	   just clipped the price off the bottom. */
	.sf-shop__grid li.wc-block-product {
		padding: 16px;
	}

	.sf-shop__grid li.wc-block-product::before {
		top: 16px;
		left: 16px;
		height: 24px;
		padding: 0 9px;
		font-size: 11px;
		min-width: 34px;
	}

	.sf-shop__grid li.wc-block-product .wc-block-components-product-sale-badge {
		top: 16px;
		left: calc(16px + 34px + 6px);
		height: 24px;
		padding: 0 9px;
		font-size: 11px;
	}

	.sf-shop__grid li.wc-block-product .sf-shop__wish {
		top: 16px;
		right: 16px;
	}

	/* the clear air above the photograph is the badge's height plus a little,
	   the same relationship the 47.2 has on desktop. The 24 under it is this
	   breakpoint's own 10 + 14 -- the gap to the text block plus the gap the
	   swatch row used to carry, added together now the row is gone. Same
	   arithmetic as the desktop rule. */
	.sf-shop__grid li.wc-block-product .wc-block-components-product-image {
		margin: 32px 0 24px;
	}

	.sf-shop__grid li.wc-block-product .sf-shop__swatches {
		height: 22px;
		gap: 9px;
		margin-bottom: 14px;
	}

	.sf-shop__grid li.wc-block-product .sf-shop__swatch { width: 10px; height: 10px; }
	.sf-shop__grid li.wc-block-product .sf-shop__swatch.is-first { width: 13px; height: 13px; }

	.sf-shop__grid li.wc-block-product h2.wp-block-post-title {
		min-height: 34px;     /* 2 x 17px at this breakpoint */
	}

	.sf-shop__grid li.wc-block-product .wp-block-post-excerpt,
	.sf-shop__grid li.wc-block-product .wp-block-post-excerpt:empty {
		margin-top: 8px;
		min-height: 28px;     /* 2 x 14px at this breakpoint */
	}

	.sf-shop__grid li.wc-block-product .wp-block-post-excerpt p {
		font-size: 11px;
		line-height: 14px;
	}

	.sf-shop__grid li.wc-block-product .wp-block-woocommerce-product-price {
		padding-top: 14px;
	}


	.sf-shop__grid li.wc-block-product h2.wp-block-post-title {
		font-size: 13px;
		line-height: 17px;
		letter-spacing: -0.03em;
	}

	/* THE CLAMP GOES ON THE LINK, NOT THE HEADING. The h2 is a flex item of
	   the card, and a flex item is BLOCKIFIED: `display: -webkit-box` on it
	   computes to `flow-root`, which silently disables -webkit-line-clamp.
	   Verified — the heading kept rendering four lines with `line-clamp: 2`
	   sitting right there in its computed style. The anchor inside it is not
	   a flex item, so the legacy box display survives on it. */
	.sf-shop__grid li.wc-block-product h2.wp-block-post-title a {
		display: -webkit-box;
		-webkit-line-clamp: 2;
		line-clamp: 2;
		-webkit-box-orient: vertical;
		overflow: hidden;
	}

	/* the full name is still on the product page; the card is a glance */
	.sf-shop__grid li.wc-block-product .wp-block-post-terms {
		font-size: 9px;
	}

	/* THE PRICE ROW ON A PHONE.
	   A 162px card leaves 134px of content width, and two 28px circles plus
	   their gaps took 72px of it — leaving 60px, in which
	   "RM574.00 - RM627.00" broke across THREE lines with the dash stranded
	   on its own. Shrinking the controls to 24px and the type to 11px gives
	   the price ~76px, which holds a range in two lines. The controls are
	   still a 24px target, above the 24px minimum for a secondary control
	   sitting next to a full-card link. */
	.sf-shop__grid li.wc-block-product .wc-block-components-product-price {
		font-size: 11px;
		gap: 6px;
	}

	.sf-shop__grid li.wc-block-product .sf-shop__acts {
		gap: 6px;
	}

	/* 44px OF TARGET HERE TOO, AND THIS IS THE BREAKPOINT IT IS FOR.
	   It was 24px square -- on the one device that has no pointer, no
	   hover and a finger instead of a cursor. A phone card is 165px
	   wide, so the control a customer is aiming at was a 16px hairline
	   inside a 24px box.

	   THE MARGIN IS -10, NOT THE -8 THE DESKTOP RULE USES, and the pair
	   has to be redone whenever either number moves. The wrapper sits at
	   top 16 / right 16 here rather than 24, and the glyph has always
	   been centred 28px in from the card's corner: a 44px box lands its
	   centre at 22, so it needs pulling out by 28 - 22 = 10. Taking the
	   desktop's -8 unchanged put the heart at 20/20 -- eight pixels off
	   the line it has sat on since this card was built, which is small
	   enough to look like nothing and big enough to see. */
	.sf-shop__grid li.wc-block-product .sf-shop__wish a,
	.sf-shop__grid li.wc-block-product .sf-shop__wish .se-btn,
	.sf-shop__grid li.wc-block-product .sf-shop__wish .badge {
		width: 44px;
		height: 44px;
		margin: -10px -10px 0 0;
	}

	/* 16, the size the icon font drew here before the SVG replaced it.
	   The desktop's 19 is pitched for a card twice this wide. */
	.sf-shop__grid li.wc-block-product .sf-shop__wish a .sfw-heart {
		width: 16px;
		height: 16px;
	}

	/* THE WORDS GO AWAY ON A PHONE AND THE DISC COMES BACK (2026-10-22).

	   The reference does the same thing: at 375px its ADD TO CART label is
	   replaced by a round icon button, because there is no room for a label
	   beside a price on a two-up card. The arithmetic here is the same one
	   the note above this media query already works through -- a card is
	   about 157px wide, the gutters take 28, and "ADD TO CART" at 10px with
	   0.14em tracking needs ~85px before the icon and its gap. That leaves
	   nothing for a price, let alone the two-line RANGE a variable product
	   prints.

	   So this is not a shrink, it is the icon control restored: the plate,
	   the circle and the screen-reader clip all come back, exactly as they
	   were before 2026-10-22. The icon-only weakness returns with them --
	   a variable product shows sliders rather than the words SELECT OPTIONS
	   -- which is the trade the width forces.

	   `.wp-block-button__link` is in this list on purpose. The full-size
	   rule carries it, which makes that selector (0,4,1) -- heavier than a
	   plain `.sf-shop__cart button` at (0,3,2) -- so without it these
	   declarations lose to the desktop ones, measured. */
	.sf-shop__grid li.wc-block-product .sf-shop__cart a,
	.sf-shop__grid li.wc-block-product .sf-shop__cart button,
	.sf-shop__grid li.wc-block-product .sf-shop__cart .wp-block-button__link {
		flex-direction: row;
		justify-content: center;
		gap: 0;
		width: 26px;
		height: 26px;
		/* FILLED, NOT A DARK PLATE (2026-10-22). It was #16161a with a
		   neumorphic pair, which was a raised disc on a near-black card and
		   is a black blot on a #EAEBED one. The reference fills this button
		   with its own accent on a phone, so this does the same in magenta
		   -- it is the only coloured thing on the card, which is right for
		   the one control on it. */
		/* !important to match the desktop rule above, which needs it to beat
		   the theme's wp-element-button styles -- without it here the disc
		   would lose its fill at this breakpoint instead */
		background: var(--sa-accent) !important;
		box-shadow: none;
		color: #fff !important;
		border-radius: 50% !important;
		padding: 0 !important;
	}

	.sf-shop__grid li.wc-block-product .sf-shop__cart a:hover,
	.sf-shop__grid li.wc-block-product .sf-shop__cart button:hover {
		background: #ffffff !important;
		color: var(--sa-accent) !important;
	}

	/* clipped, not hidden -- the button still announces "Add to cart" */
	.sf-shop__grid li.wc-block-product .sf-shop__cart a > *,
	.sf-shop__grid li.wc-block-product .sf-shop__cart button > * {
		position: absolute;
		width: 1px;
		height: 1px;
		margin: -1px;
		overflow: hidden;
		clip-path: inset(50%);
	}

	/* the variable-product anchor holds its text directly rather than in a
	   span, so it needs a zero line-box of its own */
	.sf-shop__grid li.wc-block-product .sf-shop__cart a {
		font-size: 0;
		line-height: 0;
	}

	.sf-shop__grid li.wc-block-product .sf-shop__cart a::before,
	.sf-shop__grid li.wc-block-product .sf-shop__cart button::before,
	.sf-shop__grid li.wc-block-product .sf-shop__cart .wp-block-button__link::before {
		width: 12px;
		height: 12px;
	}

	/* ---- PAGINATION ON A PHONE: ONE ROW (operator, 2026-10-05) ----
	   At 375px "PREVIOUS", five plates and "NEXT" came to more than the
	   screen, so NEXT wrapped onto a second line under the numbers.

	   The words are dropped here and the chevrons carry prev/next on their
	   own, which is what gets everything back onto one line. `font-size: 0`
	   rather than `display: none` on purpose: it hides the words from the
	   EYE but leaves "Previous Page" / "Next Page" in the accessibility
	   tree, so the links still announce themselves. The chevron is a
	   pseudo-element and sets its own size, so it is unaffected.

	   `flex-wrap: nowrap` is the guarantee — with the plates at 32px and the
	   gaps at 6, the whole row measures about 254px, so it still fits a
	   320px screen with the section's gutters. */
	.sf-shop__grid .wp-block-query-pagination {
		flex-wrap: nowrap;
		gap: 6px;
	}

	.sf-shop__grid .wp-block-query-pagination-numbers {
		gap: 6px;
	}

	.sf-shop__grid .wp-block-query-pagination .page-numbers {
		min-width: 32px;
		height: 32px;
		padding: 0 6px;
		font-size: 11.5px;
	}

	.sf-shop__grid .wp-block-query-pagination .page-numbers.dots {
		min-width: 16px;
	}

	.sf-shop__grid .wp-block-query-pagination-previous,
	.sf-shop__grid .wp-block-query-pagination-next {
		width: 32px;
		height: 32px;
		padding: 0;
		margin: 0;
		justify-content: center;
		gap: 0;
		font-size: 0;      /* hides the words, keeps them for a screen reader */
		letter-spacing: 0;
	}
}

@media (prefers-reduced-motion: reduce) {
	.sf-shop__grid li.wc-block-product .wc-block-components-product-image img,
	.sf-shop__filters-toggle {
		transition-duration: 0.001ms;
	}
}


/* THE NAVBAR'S OVERHANG STEPS DOWN WITH THE BAR (2026-11-13).
   navbar.css redefines `--sf-navbar-gutter` to 20px at this exact width, and
   `--sf-navbar-float-gap` reads it, so the pill stops hanging 24px past its
   own flow and hangs 20 instead. The masthead's clearance is derived from
   that number — see the note at section 1 — so it has to step here too, at
   THIS breakpoint and not at the 760px one above, which is ours and not the
   navbar's. Keep the two files in step: the literal is deliberate and the
   section 1 note says why. */
@media (max-width: 599px) {
	.sf-shop {
		/* Read by nothing since 2026-11-16 — the masthead margin it
		   corrected is gone and the section padding carries the navbar
		   clearance now. Kept because it is navbar.css own 599px gutter
		   and would be needed again the moment a margin comes back. */
		--sa-nav-overhang: 20px;
	}
}


/* =============================================================
 * 13. THE FILTER RAIL  (operator, 2026-11-16)
 * =============================================================
 * "i dont want the filter at modal ... in woocommerce should be the
 * title will on the left right ? you can open durian bikers website
 * also have filter in left side . but its too simple"
 *
 * Skin C of docs/home-concepts/shop-filter-proposal.html, chosen
 * from three, at a 320px rail. The references and the measurements
 * are in docs/research/2026-11-16-shop-filter-references.md.
 *
 * WHAT IT REPLACED: a full-screen role="dialog" aria-modal="true"
 * sheet that shop-archive.js built and MOVED these same blocks
 * into. The blocks did not change. Only the box did.
 *
 * WHAT IT IS NOT: Durian Bikers' rail, which is the thing it was
 * measured against -- 261px wide, position:static, and a category
 * widget 56 items and 2730px tall. By the time you reach the
 * products its filters are a screen and a half above you. This one
 * sticks, folds, and leads with the bike.
 * ------------------------------------------------------------- */

.sf-shop__body {
	display: grid;
	/* THE RAIL IS 320 AND 1280 IS A KNIFE EDGE. The grid's own track
	   floor is 260px, so at 1280 the three tracks land on 260.5 and
	   ONE PIXEL NARROWER drops the page to two columns. Measured:
	   1440 -> 3 x 311, 1280 -> 3 x 260.5, 1180 -> 2 x 353. Re-check
	   that row if the rail, this gap or that floor ever moves. */
	grid-template-columns: var(--sa-rail-w, 320px) 1fr;
	gap: var(--sa-step);
	align-items: start;
}

/* THE TWO COLUMNS START ON THE SAME LINE (operator, 2026-11-17, with the
   element highlighted in devtools: "look at this its not same row").

   WORDPRESS WAS PUSHING THE GRID COLUMN DOWN BY 19.2px. Its block layout
   emits

       :root :where(.is-layout-flow) > * { margin-block-start: 1.2rem }

   and .sf-shop__main is a wp-block-group, so it took 1.2rem = 19.2px that
   .sf-shop__side -- the first child, which the matching :first-child rule
   zeroes -- never got. The two are grid items of the same row, so nothing
   collapses it: the rail simply began 19.2px above the first card.

   IT WAS INVISIBLE BECAUSE THE HEADING CANCELLED IT. .sf-shop__side-title
   carried padding-top: 20px, so the WORD "FILTER" landed within a pixel of
   the card tops (measured 129.13 against 129.13 at a 1352px client) while
   the two BOXES were 19.2px apart. That is why this reads as correct until
   you put an inspector on it, and it is why both numbers had to go -- fixing
   one alone would have moved the label 20px off the cards in the other
   direction.

   THE SELECTOR NEEDS THE EXTRA CLASS. WordPress's rule is (0,1,0) -- :root
   plus a :where(), which contributes nothing -- and it is printed INLINE in
   <head>, after every enqueued stylesheet. A lone .sf-shop__main is also
   (0,1,0) and would lose on source order. navbar.css records the same trap
   with the block editor's flex layout. */
.sf-shop__body > .sf-shop__main {
	margin-block-start: 0;
}

/* IT IS __side AND NOT __rail, AND THAT IS NOT A TASTE DECISION.
   `.sf-shop__rail` IS ALREADY TAKEN — shop-beast.css carries a whole
   block of it (sticky top:96px, checkbox list, labels, chips, the range
   input, the counts, a 1024 breakpoint) left over from the 470px rail
   that was removed on 2026-10-05. Those rules matched nothing from that
   day until this one, and naming the new wrapper `__rail` brought every
   one of them back to life.

   It shipped that way for about a minute. Measured on the live page:
   TWO rules setting `top` on the same element, mine at 106 from
   shop-archive.css and shop-beast.css's at 96 — and shop-beast is
   enqueued AFTER this file, so the dead one won. The rail sat 10px high
   and half its controls were being styled from a file nobody had opened
   in two months.

   shop-beast.css WAS LEFT ALONE deliberately: it is not this change's
   file, those rules are inert again the moment nothing is called
   `__rail`, and widening a deploy to tidy someone else's dead code is
   how a one-file change becomes three. If it is ever cleaned out, that
   block can go with it — but leave this name alone either way.

   The lesson is the cheap check: grep the whole theme for a class name
   before using it, not just the file you are writing. */
.sf-shop__side {
	position: sticky;
	/* THE NAVBAR IS A FLOATING PILL AND IT PAINTS TO ABOUT 90, not 66.
	   <header> reserves 66px of flow and navbar.css then pushes the bar
	   down by its own gutter, so the bar paints well below the flow it
	   reserves. A rail stuck at 66 slides its first section under the
	   glass. 16 of air on top of that.

	   A LITERAL, not a var() read from navbar.css: a custom property
	   substitutes on the element where it is DECLARED, so a :root token
	   crossing stylesheets resolves by load order and fails silently.
	   This file and that one must change together -- the same decision
	   this file already records for --sa-gutter against --sfidx-gutter. */
	top: 106px;
	/* ONE EMBOSSED CONSOLE, which is what makes this skin C rather than
	   four separate panels. The face is the PAGE COLOUR and the shape is
	   the light alone -- see --sa-card above for why the two halves
	   cannot be separated, and why there is no border here. */
	background: var(--sa-bg);
	border-radius: 12px;
	box-shadow: -3px -3px 8px var(--sa-lift), 4px 4px 10px var(--sa-sink);

	/* NO PADDING ON TOP — THE HEADER CARRIES IT (2026-11-16). The
	   FILTER row is sticky inside this scroller, and a sticky child
	   pins to the SCROLLPORT, which is this element's PADDING box. With
	   padding-top here the header's resting position is 20px down and
	   `top: 0` lets it ride UP into that strip the moment you scroll —
	   so the air above it vanishes as soon as it sticks and comes back
	   when you scroll home. Moving the 20px onto the header itself
	   makes its resting position and its stuck position the same place,
	   so nothing moves at all. */
	padding: 0 20px 20px;

	/* A STICKY RAIL TALLER THAN THE VIEWPORT CANNOT BE REACHED, and this
	   is the cap that fixes it (2026-11-16).

	   `position: sticky` with only `top` set pins the rail's TOP at 106
	   and nothing ever brings its bottom up — so everything below the
	   fold is unreachable for as long as it is pinned. Measured on
	   /shop/ at 900px: the rail stood 1138, which left PRICE at y=1368
	   and reachable only at the very bottom of the page, after the rail
	   had given up and scrolled away with the document.

	   (It was 2647 before the `[hidden]` fix below — that is what made
	   this impossible to miss. The fix took it to 1138, which is better
	   and still does not fit. Both were needed.)

	   svh, not vh: the small viewport height is the one with a phone's
	   toolbars showing, so the cap holds at the worst moment — the same
	   reason home-opening.css uses it. The rule only applies above 1024
	   anyway, where the rail is sticky; below that it is static and this
	   would cap a column that is simply in the flow. */
	max-height: calc(100svh - 106px - 20px);
	overflow-y: auto;
	overscroll-behavior: contain;
	scrollbar-width: thin;
	scrollbar-color: var(--sa-line) transparent;
}

/* AND IT NEEDS data-lenis-prevent, WHICH shop-archive.js SETS.
   Lenis drives the real window scroll on every page here and calls
   preventDefault on every wheel event, so a panel that scrolls inside
   itself receives none: the page scrolls instead and the panel sits
   there looking frozen. This project has shipped that bug once already,
   on the dealer map's shop list. The attribute is what makes Lenis skip
   an event that starts inside the rail. */
.sf-shop__side::-webkit-scrollbar {
	width: 4px;
}

.sf-shop__side::-webkit-scrollbar-thumb {
	background: var(--sa-line);
	border-radius: 999px;
}

/* ---- the rail's own title ----
   The four section headings each name what they narrow; none of them
   said the column narrows anything. This is the <h2> they sit under.

   ONE VOICE (operator, 2026-11-16: "why have / ?"). It read
   "Filter / narrow it down" for an afternoon — a label and a sentence
   split by a slash — and nothing else on this site puts two voices in
   one heading. The word carries the row; the count and the Clear are
   information, set in the quieter inks, and both are added by
   shop-archive.js because both depend on the request. */
/* IT IS STATIC AND THE RAIL SCROLLS BEHIND IT (operator, 2026-11-16:
   "its a static , so when that sidebar is scroll its will scroll behind
   the static filter header").

   `position: sticky` inside .sf-shop__side's own scroller, NOT fixed
   and not a second element — so it is the same heading, in the same
   place, and it simply stops moving. It needs three things and all
   three are load-bearing:

     - top: 0, which is the scrollport's top edge. See the note on the
       rail's padding above for why that padding had to move here.
     - a BACKGROUND. Without one the filters scroll straight through the
       type. It is the rail's own --sa-bg, so the header is invisible
       until something passes under it.
     - z-index. A sticky element is positioned, but the filter sections
       below it are in flow and paint later; 5 clears them with room to
       spare and stays under the rail's own scrollbar.

   THE HAIRLINE IS STILL INSET BY THE RAIL'S GUTTER and does not need
   negative margins to cover the content. Nothing ever lays out in the
   left and right padding strips — children fill the content box — so
   the header only has to be as wide as the content it hides, which it
   already is. Vertically is the other story, which is what the padding
   move is for.

   IT IS OFF BELOW 1024, where the rail is static with overflow visible:
   with no scrollport of its own a sticky child resolves against the
   DOCUMENT and would pin FILTER to the top of the window, behind the
   floating navbar. See the 1023 block at the foot of this file.

   WHAT SCROLLS AWAY: the CLEAR link, which is a sibling below this
   heading rather than part of it. Pinning it too would mean the header
   changing height as filters come and go — it only exists while
   something is live — so it was left to scroll. */
.sf-shop__side-title {
	position: sticky;
	top: 0;
	z-index: 5;
	background: var(--sa-bg);

	display: flex;
	align-items: baseline;
	gap: 7px;
	margin: 0 0 14px;
	/* NO TOP PADDING SINCE 2026-11-17 -- see the note on .sf-shop__main above.
	   It was 20px, and with the stray 19.2px gone it would have pushed the
	   FILTER label that far BELOW the card tops instead of above them.

	   WHAT THE 20px WAS FOR: this heading is sticky at the top of the rail's
	   own scrollport, and the padding was the air between the scrollport edge
	   and the text once the filters scroll under it. The 14px BELOW the
	   hairline still separates the pinned header from the list, which is the
	   half that does the work; above it there is now nothing to separate the
	   text FROM, because the rail starts exactly where the cards start. */
	padding: 0 0 14px;
	border-bottom: 1px solid var(--sa-line);

	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 600;
	line-height: 1.2;
	letter-spacing: 0.16em;
	text-transform: uppercase;
	color: var(--sa-text);
}

/* THE COUNT takes the rest of the row, so it sits hard right whether or
   not Clear is there — the same margin-left:auto the filter counts use,
   which keeps every number in this rail on one edge. */
.sf-shop__side-count {
	margin-left: auto;
	font-size: 10px;
	letter-spacing: 0.12em;
	color: var(--sa-muted);
	font-variant-numeric: tabular-nums;
	white-space: nowrap;
}

/* CLEAR IS THE ONLY LINK IN THE RAIL, so it is the only thing in the
   accent — that is what marks it as something you can press rather than
   something you can read. It is absent until a filter is set, so it
   never offers to undo nothing.

   IT SITS BELOW THE HEADING'S LINE (operator, 2026-11-16), at the top
   of the active-filters area and on its own row above the chips, not
   beside the count. `width: fit-content` with `margin-left: auto` keeps
   the hit area the size of the word while putting it on the right —
   a plain `display: block; text-align: right` would make the whole row
   clickable, which is a wide target for a destructive action.

   THE SELECTOR IS THIS LONG BECAUSE WOOCOMMERCE OWNS EVERY <a> ON THE
   PAGE. woocommerce.css carries

       body.woocommerce a, body.woocommerce-page a { color: var(--wc-text) }

   which is 0,2,1 — two classes and an element — so a lone
   `.sf-shop__side-clear` at 0,1,0 loses and the link renders in the
   ink. It shipped that way for a minute and read as plain text.

   0,3,1 beats it outright rather than relying on load order, which is
   what an equal-specificity selector would have been betting on. It is
   also the only <a> this rail has: the chips and the rows are buttons
   and labels, so nothing else here needed it. */
.sf-shop .sf-shop__side a.sf-shop__side-clear {
	display: block;
	width: fit-content;
	margin: 0 0 12px auto;
	/* It used to inherit this from the heading it sat in; out here it has
	   to say so itself, or it is the one lowercase word in a rail of
	   tracked-out labels. */
	text-transform: uppercase;
	font-size: 10px;
	letter-spacing: 0.12em;
	color: var(--sa-accent);
	text-decoration: none;
	border-bottom: 1px solid transparent;
	white-space: nowrap;
	transition: border-color 0.2s;
}

.sf-shop .sf-shop__side a.sf-shop__side-clear:hover,
.sf-shop .sf-shop__side a.sf-shop__side-clear:focus-visible {
	border-bottom-color: var(--sa-accent);
}

/* the filter panel itself is now just a box inside the rail */
.sf-shop__side .sf-shop__filters {
	display: block;
	margin: 0;
	padding: 0;
	max-width: none;

	/* THE SECOND LINE UNDER THE HEADING WAS THIS (operator, 2026-11-16:
	   "at filter why have 2 line ? remove one").

	   `.sf-shop__filters` carries a hairline top and bottom from the
	   modal era, when it was a panel in its own right and needed edges.
	   Inside the rail it sits directly under the heading, whose own
	   border-bottom is 14px above — so the two drew as a double rule
	   with nothing between them. Measured: mine at y=464 and this one at
	   y=478.

	   The heading's line is the one that means something (it closes the
	   heading off), so this one goes. Scoped to the rail, because the
	   rules it overrides still apply wherever else that class is used. */
	border-top: 0;
	border-bottom: 0;
}

.sf-shop__side .wc-block-product-filters,
.sf-shop__side .wp-block-woocommerce-product-filters {
	display: block;
}

/* ---- the sections ---- */
.sf-shop__side .wp-block-woocommerce-product-filter-taxonomy,
.sf-shop__side .wp-block-woocommerce-product-filter-price {
	padding: 16px 0;
	margin: 0;
	counter-increment: sf-filter;
}

/* EVERY PAIR, NOT JUST THE ONES THAT EXISTED (operator, 2026-11-16:
   "under the filter price will have line , like others filter").

   This listed only `taxonomy + taxonomy` and `taxonomy + price`,
   because for as long as Price was LAST those were the only two
   adjacencies in the rail. Moving Price to the top made Bike follow a
   PRICE block, which no selector covered — so the one line under the
   price section quietly went missing while Model and Category kept
   theirs. Measured: Bike border-top 0px, the other two 1px.

   All four combinations of the two block types are listed now, so the
   rule survives any order. A THIRD filter type — status, rating,
   attribute — would need its own pairs adding; the alternative is a
   class of our own on each section from shop-archive.js, which was not
   worth a JS dependency for a hairline. */
.sf-shop__side .wp-block-woocommerce-product-filter-taxonomy + .wp-block-woocommerce-product-filter-taxonomy,
.sf-shop__side .wp-block-woocommerce-product-filter-taxonomy + .wp-block-woocommerce-product-filter-price,
.sf-shop__side .wp-block-woocommerce-product-filter-price + .wp-block-woocommerce-product-filter-taxonomy,
.sf-shop__side .wp-block-woocommerce-product-filter-price + .wp-block-woocommerce-product-filter-price {
	border-top: 1px solid var(--sa-line);
}

/* THE HEADING IS A BUTTON, built by shop-archive.js around the h3 the
   block already renders. The numbering is a counter, so a section added
   or removed in the template renumbers itself and cannot fall out of
   step -- the same reason the product index number is a counter. */
.sf-shop__side .sf-rail__head {
	all: unset;
	display: flex;
	width: 100%;
	align-items: baseline;
	gap: 10px;
	cursor: pointer;
}

.sf-shop__side .sf-rail__head::before {
	content: counter(sf-filter, decimal-leading-zero);
	font-family: var(--sa-tech);
	font-size: 10px;
	letter-spacing: 0.14em;
	color: var(--sa-muted);
}

.sf-shop__side .sf-rail__head h3,
.sf-shop__side .sf-rail__head .wp-block-group {
	flex: 1;
	margin: 0;
	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 600;
	letter-spacing: 0.16em;
	text-transform: uppercase;
	color: var(--sa-text);
}

.sf-shop__side .sf-rail__mark {
	font-family: var(--sa-tech);
	font-size: 13px;
	font-style: normal;
	color: var(--sa-muted);
}

.sf-shop__side .sf-rail__head[aria-expanded="true"] .sf-rail__mark {
	color: var(--sa-accent);
}

.sf-shop__side .sf-rail__body {
	padding: 12px 0 0 12px;
	/* THE MAGENTA SPINE. A section with a live filter carries it, which is
	   skin C's one signature and the only thing on the rail that says
	   "this one is doing something" without being read. */
	border-left: 1px solid var(--sa-line);
}

.sf-shop__side .is-sf-live > .sf-rail__body {
	border-left-color: var(--sa-accent);
}

.sf-shop__side .sf-rail__body[hidden] {
	display: none;
}

/* ---- chips (make, category) ---- */
.sf-shop__side .wc-block-product-filter-chips__items {
	display: flex;
	flex-wrap: wrap;
	gap: 6px;
}

.sf-shop__side .wc-block-product-filter-chips__item {
	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 600;
	letter-spacing: 0.1em;
	text-transform: uppercase;
	padding: 7px 11px;
	border: 1px solid var(--sa-line);
	border-radius: 999px;
	background: transparent;
	color: var(--sa-text);
	cursor: pointer;
	transition: border-color 0.2s, background-color 0.2s, color 0.2s;
}

.sf-shop__side .wc-block-product-filter-chips__item:hover {
	border-color: rgba(242, 242, 244, 0.4);
}

.sf-shop__side .wc-block-product-filter-chips__item[aria-checked="true"],
.sf-shop__side .wc-block-product-filter-chips__item.is-selected {
	background: var(--sa-accent);
	border-color: var(--sa-accent);
	color: #fff;
}

/* ---- the checkbox list (model) ----
   NO CHECKBOX. The row IS the control: the name on the left, the count
   hard right, and the name goes magenta when it is on. A 13px box beside
   a dozen model names is more furniture than information, and the count
   has to own the right edge or the list stops being scannable.

   The input is hidden with a clip, NOT display:none -- it is still the
   control the block's own Interactivity API listens to, and a
   display:none input is not focusable, so the list would stop working
   on a keyboard. */
.sf-shop__side .wc-block-product-filter-checkbox-list__items {
	display: flex;
	flex-direction: column;
	gap: 0;
}

/* THE WHITE BOX ROUND THE LIST WAS A <fieldset> (operator, 2026-11-16:
   "why have border white").

   WooCommerce wraps each filter's options in <fieldset><legend> — the
   right markup, and the legend is screen-reader-only — and a fieldset
   carries a 2px groove border from the UA stylesheet. Nothing had ever
   reset it here because the old sheet's lists were never inside a panel
   where it showed; in a bordered rail it reads as a stray frame.

   Reset on the ELEMENT, not on a WooCommerce class: the chips blocks
   use .wc-block-product-filter-chips__fieldset and the checkbox lists
   use a bare <fieldset>, so matching the tag covers both and covers
   whatever the next block type does. */
.sf-shop__side fieldset {
	border: 0;
	margin: 0;
	padding: 0;
	min-width: 0;
}

/* ---- "Show more", which is WooCommerce's own ----
   THE LIST IS NOT CAPPED BY US, AND AN EARLIER VERSION OF THIS FILE GOT
   THAT WRONG. /shop/ renders 64 model terms, so this first shipped with
   max-height: 288px, an internal scroll and a data-lenis-prevent set
   from script.

   All of that was redundant: the block already carries
   `displayLimit: 15` and renders its own Show more / Show less control,
   hiding the rest with data-wp-bind--hidden. Capping on top of it gave
   you fifteen rows in a scrollbox, and then Show more expanded 64 rows
   inside the same 288px scrollbox.

   The native control keeps the keyboard and the aria state correct and
   needs no Lenis handling at all, because nothing scrolls inside
   anything. Only the button is restyled.

   If the limit ever wants changing it is `displayLimit` on the block in
   templates/archive-product.html, not a height here. */
.sf-shop__side .wc-block-product-filter-checkbox-list__show-more,
.sf-shop__side .wc-block-product-filter-checkbox-list__show-less,
.sf-shop__side .wc-block-product-filter-chips__show-more,
.sf-shop__side .wc-block-product-filter-chips__show-less {
	margin-top: 10px;
	padding: 0;
	border: 0;
	background: none;
	cursor: pointer;
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 600;
	letter-spacing: 0.14em;
	text-transform: uppercase;
	color: var(--sa-accent);
	text-decoration: none;
}

.sf-shop__side .wc-block-product-filter-checkbox-list__show-more:hover,
.sf-shop__side .wc-block-product-filter-checkbox-list__show-less:hover,
.sf-shop__side .wc-block-product-filter-chips__show-more:hover,
.sf-shop__side .wc-block-product-filter-chips__show-less:hover {
	color: var(--sa-text);
}

.sf-shop__side .wc-block-product-filter-checkbox-list__item {
	display: flex;
	align-items: baseline;
	gap: 8px;
	padding: 6px 0;
	cursor: pointer;
	margin: 0;
}

/* ANY `display` DECLARATION BEATS THE `hidden` ATTRIBUTE, AND THAT RULE
   ABOVE SILENTLY BROKE "SHOW MORE" (found 2026-11-16, driving the rail
   rather than reading it).

   The block renders every term and hides the overflow with
   data-wp-bind--hidden, which sets the plain HTML `hidden` attribute.
   That attribute only works because the UA stylesheet says
   `[hidden] { display: none }` — the weakest rule there is. The flex
   above overrode it on every row, so on /shop/ all 64 model terms
   painted, "Show 49 more" sat there doing nothing, and the rail stood
   2647px on a 900px screen. Measured: 64 rendered, 49 carrying the
   attribute, all 64 painted at 31px.

   WHAT IT COST BEYOND THE LIST. The rail is `position: sticky`, and an
   element taller than the viewport cannot stay pinned — it scrolls away
   upward, so PRICE was only reachable at the very bottom of the page.
   The symptom looked like a sticky bug and the cause was one word in a
   display declaration.

   The lesson is general and this file should keep it: if you style a
   list the Interactivity API controls, restate `[hidden]` yourself.
   Checking the computed value of the ROW is what finds it — the
   attribute is present and correct the whole time. */
.sf-shop__side .wc-block-product-filter-checkbox-list__item[hidden] {
	display: none;
}

.sf-shop__side .wc-block-product-filter-checkbox-list__input-wrapper,
.sf-shop__side .wc-block-product-filter-checkbox-list__input {
	position: absolute;
	width: 1px;
	height: 1px;
	overflow: hidden;
	clip-path: inset(50%);
	white-space: nowrap;
}

.sf-shop__side .wc-block-product-filter-checkbox-list__label {
	display: flex;
	flex: 1;
	align-items: baseline;
	gap: 8px;
	cursor: pointer;
	font-family: var(--sa-body);
	font-size: 14px;
	color: var(--sa-text);
	margin: 0;
}

.sf-shop__side .wc-block-product-filter-checkbox-list__text {
	font-family: var(--sa-body);
	font-size: 14px;
	color: inherit;
	transition: color 0.2s;
}

.sf-shop__side .wc-block-product-filter-checkbox-list__item:hover .wc-block-product-filter-checkbox-list__text {
	color: var(--sa-accent);
}

.sf-shop__side .wc-block-product-filter-checkbox-list__item:has(:checked) .wc-block-product-filter-checkbox-list__text {
	color: var(--sa-accent);
}

/* the count, hard right */
.sf-shop__side .wc-block-product-filter-checkbox-list__count,
.sf-shop__side .wc-block-product-filter-chips__item-count,
.sf-shop__side .wc-block-product-filter-chips__count {
	margin-left: auto;
	font-family: var(--sa-tech);
	font-size: 11px;
	letter-spacing: 0.08em;
	color: var(--sa-muted);
	font-variant-numeric: tabular-nums;
}

/* ---- price ---- */
.sf-shop__side .wp-block-woocommerce-product-filter-price-slider {
	accent-color: var(--sa-accent);
}

/* THE SECOND PRICE BOX WAS BEING THROWN OUT OF THE RAIL (operator,
   2026-11-16: "this part is not responsive").

   WooCommerce lays the slider out as a grid with NAMED COLUMN LINES —
   the cells ask for `grid-column: left-input` and `right-input` — and
   at the rail's width the template computed

       grid-template-columns: 100.5px 100.5px 0px;   gap: 26px 28px

   so the right-hand cell landed in a THIRD column of zero width. Its
   input kept a 20px minimum, overflowed, and was clipped against the
   rail's edge: measured 277 of content in 257 of space, with the right
   box at width 0 sitting on x=334. On screen that reads as one price
   box where there should be two — "RM80" and nothing else.

   Two equal columns, and the cells flow into them. `grid-column: auto`
   is the load-bearing part: without it they would still be hunting for
   line names this template does not define. `min-width: 0` is the usual
   grid-overflow guard — a grid item defaults to `min-width: auto`,
   which refuses to shrink below its content.

   Scoped to the rail. Everywhere else the block keeps its own layout. */
/* FLEX, NOT GRID — AND THAT IS THE SECOND ATTEMPT (operator,
   2026-11-16: "the price still not responsive").

   The first fix overrode the grid template to two equal columns. It got
   both boxes visible, which was the worst of it, but the right-hand one
   still stopped 12px short of the panel edge while the Filter button
   below spanned the full width — so the block read as misaligned.

   THE TRACKS WOULD NOT BUDGE. WooCommerce's `auto 1fr auto` kept
   resolving — measured 116.5px 116.5px 0px, an empty third track eating
   a 12px gap — through a more specific override, through
   `repeat(2, minmax(0,1fr))`, through explicit `grid-column` on every
   child, and even through an INLINE style set on the element, with no
   `!important` anywhere to explain it. Four attempts, all measured, none
   of which moved the computed value.

   So the grid is sidestepped rather than argued with. Flex has no track
   model to inherit: the slider takes a full-width row, the two boxes
   share what is left. Measured after — left 77 w123, right ending on
   334, which is the content edge to the pixel, and the range the full
   257.

   The 2px of scrollWidth that remains is inside a box with visible
   overflow, so nothing clips and no scrollbar appears. */
.sf-shop__side .wc-block-product-filter-price-slider__content {
	display: flex;
	flex-wrap: wrap;
	gap: 14px 12px;
}

.sf-shop__side .wc-block-product-filter-price-slider__range {
	flex: 0 0 100%;
}

/* THE TWO TEXT INPUTS ARE CLIPPED, NOT REMOVED (operator, 2026-11-16).
   The range now reads as one line of type with the Filter button beside
   it — see .sf-rail__pricerow below — so the boxes are redundant. They
   stay in the DOM and stay focusable, because they are the controls the
   block binds to and the only way to type an exact figure on a
   keyboard; the readout is a mirror of them, never the source.

   This also retires the flex sizing above them: with both clipped, the
   slider is the only thing left in that row and takes the full width on
   its own. */
.sf-shop__side .wc-block-product-filter-price-slider__left,
.sf-shop__side .wc-block-product-filter-price-slider__right {
	position: absolute;
	width: 1px;
	height: 1px;
	overflow: hidden;
	clip-path: inset(50%);
	white-space: nowrap;
}

/* ---- the readout and the button, on one row ----
   Modelled on Durian Bikers' arrangement — slider, then the range in
   words with the button beside it — and deliberately not their design.

   WHAT IS OURS: no "Price:" prefix, because the section is already
   headed PRICE; the range in the band's own tracked tech face rather
   than body copy; and a magenta pill on the right of the row rather
   than a green one crowding the text. */
.sf-rail__pricerow {
	display: flex;
	align-items: center;
	gap: 10px;
	margin-top: 14px;
}

/* THE READOUT TAKES THE REST OF THE ROW (operator, 2026-11-16: "why
   its not full width beside the FILTER button").

   It was shrink-wrapped with `justify-content: space-between` holding
   the two apart, so the range sat in a short box at the left with 83px
   of nothing between it and the button — which reads as two things
   pushed to the ends rather than one row. `flex: 1` makes it the row,
   with the button as its end stop; the gap above is then the only
   space between them.

   The text stays left-aligned, which is where a range should start.
   `text-align: center` here would centre it in the new width if that
   is ever preferred — it is one line, and the only thing this change
   makes possible. */
.sf-rail__priceout {
	flex: 1 1 auto;
	min-width: 0;

	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 600;
	letter-spacing: 0.1em;
	color: var(--sa-text);
	font-variant-numeric: tabular-nums;
	white-space: nowrap;
}

.sf-shop__side .wc-block-product-filter-price-slider__right {
	text-align: right;
}

.sf-shop__side .wc-block-product-filter-price-slider__content input[type="text"] {
	width: 100%;
	min-width: 0;
	box-sizing: border-box;
}

/* THE APPLY BUTTON, built by shop-archive.js (2026-11-16). Price is the
   one filter that does not act on release — it waits for this, which is
   Durian Bikers' arrangement and was chosen deliberately: a slider that
   reloads the page on every nudge is unusable once the page reloads at
   all, and the whole rail reloads now.

   It is styled as the rail's own primary action rather than as a chip,
   because it is the only control here that DOES something rather than
   describing a choice. */
.sf-shop__side .sf-rail__apply {
	/* IT SITS IN THE ROW NOW, so it shrink-wraps to its label instead of
	   spanning the rail. It was a full-width block underneath the two
	   price boxes until 2026-11-16. */
	flex: 0 0 auto;
	margin: 0;
	padding: 9px 16px;
	border: 0;
	/* 12px, NOT A PILL (operator, 2026-11-16, pointing at it: "i want the
	   filter button is have border radius 12px"). It was 999px.

	   THE LABEL IS "Filter", WRITTEN BY shop-archive.js, NOT "Apply" -- see
	   that file around the apply.textContent line. So the pink pill the
	   operator photographed is THIS rule and not .sf-shop__filters-toggle,
	   which is the dark 6px plate up in the bar. Do not chase the toggle.

	   12px is the house radius: the navbar bar, the hero panel, the video
	   panel and section 2 cards all carry it, and the active-filter chips
	   beside this button deliberately keep their 999px -- a chip is a
	   removable token, this is a control. */
	border-radius: 12px;
	cursor: pointer;

	background: var(--sa-accent);
	color: #fff;
	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 600;
	letter-spacing: 0.14em;
	text-transform: uppercase;
	transition: filter 0.2s;
}

.sf-shop__side .sf-rail__apply:hover {
	filter: brightness(1.12);
}

/* ---- the active chips, and Clear ---- */
.sf-shop__side .wp-block-woocommerce-product-filter-active {
	padding: 0 0 14px;
	border-bottom: 1px solid var(--sa-line);
}

.sf-shop__side .wp-block-woocommerce-product-filter-active:empty {
	display: none;
}

.sf-shop__side .wc-block-product-filter-removable-chips__item {
	background: var(--sa-accent);
	border-color: var(--sa-accent);
	color: #fff;
	border-radius: 999px;
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 600;
	letter-spacing: 0.1em;
	text-transform: uppercase;
}

/* =============================================================
 * 13a. THE PHONE GETS A FOLD, NOT A SHEET
 * -------------------------------------------------------------
 * The whole point of this change was to be rid of the overlay, so
 * the phone does not get a smaller one. The Filters button in the
 * bar folds the rail open IN PLACE above the grid: no dialog, no
 * scroll lock, no focus trap, and the page behind it does not
 * move. The rail keeps its console face so it still reads as one
 * object when it opens.
 * ------------------------------------------------------------- */

/* ABOVE 1024 THE BUTTON IS GONE, because the rail is simply there.
   It stays in the markup rather than being removed from the template,
   so a visitor with no script keeps a control that works. */
@media (min-width: 1024px) {
	.sf-shop__bar .sf-shop__filters-toggle {
		display: none;
	}

	/* AND THE SORT GOES RIGHT (operator, 2026-11-16: "i want default
	   sorting is on the right not on the left").

	   The bar is `justify-content: space-between` with two children, which
	   put Filters left and sorting right. Hiding the button above 1024
	   leaves ONE child, and space-between parks a lone item at the start —
	   so the sort slid to the left edge the moment the rail replaced the
	   modal. Nothing was wrong with the sort; it was the button leaving.

	   flex-end rather than `margin-left: auto` on the control, because the
	   bar's own rule is what changed meaning and it is the bar that should
	   say so. Below 1024 space-between still applies and the pair sits
	   Filters-left, sorting-right as before.

	   IT IS `.sf-shop .sf-shop__bar`, AND THE EXTRA CLASS IS LOAD-BEARING.
	   WordPress prints the block's layout inline in <head>, AFTER the
	   enqueued stylesheets:

	       .wp-container-core-group-is-layout-788841da{
	           flex-wrap:nowrap; justify-content:space-between; }

	   That is 0,1,0 — level with a lone `.sf-shop__bar`, and later in the
	   cascade, so the single class loses and nothing moves. navbar.css
	   records the same trap with the same container hash.

	   DO NOT target that container class to win: the hash is shared with
	   the NAVBAR's own group (verified — the navbar bar carries
	   788841da too), so a rule on it would reach across the page. */
	/* COUNT LEFT, SORTING RIGHT. This was `flex-end` from 2026-11-16,
	   when hiding the Filters button left the sorting as the only child
	   and space-between parked it at the start. The count is back in the
	   row now, so there are two things to spread again. */
	.sf-shop .sf-shop__bar {
		justify-content: space-between;
	}
}

@media (max-width: 1023px) {
	.sf-shop__body {
		grid-template-columns: 1fr;
		gap: 24px;
	}

	/* AND THE HEADER STOPS STICKING WITH IT. With the rail static and
	   overflow visible there is no scrollport here, so a sticky child
	   resolves against the document and would pin FILTER to the top of
	   the window — behind the floating navbar, over the products. It
	   keeps its own padding-top, so the air above it is unchanged. */
	.sf-shop__side-title {
		position: static;
	}

	.sf-shop__side {
		position: static;
		top: auto;
		/* THE CAP COMES OFF WITH THE STICKINESS. It exists only because a
		   pinned rail cannot be scrolled; in the flow the page scrolls it
		   perfectly well, and leaving it on would turn the fold into a
		   full-height scroller inside a scrolling page — two scrollbars
		   for one list. */
		max-height: none;
		overflow: visible;
	}

	/* shut by default, and shop-archive.js is what opens it. WITH NO
	   SCRIPT the attribute is never written and the rail is simply
	   open -- the same failure direction the sheet had, and the right
	   one: filters you cannot hide beat filters you cannot reach. */
	.sf-shop__side[data-sf-collapsed="true"] {
		display: none;
	}
}


/* =============================================================
 * THE PRODUCT CARD — DESIGN B (operator, 2026-11-16)
 *
 * "i think this card product category is not perfect enough .. its
 * not aesthetic/professional or if user see this , user will not
 * interested even the interface looks like old school". Four designs
 * went up on localhost (docs/home-concepts/card-proposals.css, with
 * mirror-shop.js building the lab from the real archive) and B was
 * chosen, then narrowed twice: "no need to put this" (the fitment
 * chips) and "the price and button i want like this" (bold price,
 * the action as a mark, no label).
 *
 * WHAT WAS ACTUALLY WRONG, measured on the live wheels archive at
 * 1340px rather than described:
 *
 *   THE PRODUCT FILLED 26% OF THE CARD. The media box was 4/3 with
 *   `object-fit: contain` over a SQUARE source, so a 1024x1024
 *   photograph painted 175x175 inside a 234x175 box -- dead space
 *   down both sides, plus 47.2px of margin above it and 36px below.
 *   The card was 257x493 and the thing it sells had a quarter of it.
 *   That is most of why it read as empty, and it is why the first
 *   rule below is the media and not the type.
 *
 *   THE EXCERPT WAS LOUDER THAN THE NAME -- 16.8px against 16px, at
 *   full ink, and truncated mid-sentence. It is off.
 *
 * Measured after, same width: card 454 -> 422, photograph 26% -> 56%
 * of the card. On a 390px phone, 372 -> 305.
 *
 * THIS BLOCK SUPERSEDES RULES ABOVE IT rather than editing them,
 * which is a deliberate trade and not laziness: the rules it changes
 * are scattered over ~400 lines, each carrying the argument for a
 * decision that is still worth reading, and another session was
 * deploying to this box the same afternoon. Deleting it restores the
 * previous card exactly. CONSOLIDATING THE TWO IS WORTH DOING once
 * this design has settled.
 *
 * SELECTOR SHAPES ARE COPIED FROM THE RULES BEING SUPERSEDED, not
 * written fresh. The title is `h2.wp-block-post-title` (0,3,2) above,
 * so a plain `.wp-block-post-title` here would LOSE and the type
 * would silently stay as it was. Equal specificity plus source order
 * is what carries this block.
 * ============================================================= */

/* ---- 1. the card is a column, and the photograph is the top of it */
.sf-shop__grid li.wc-block-product {
	display: flex;
	flex-direction: column;
	/* was 24px; every child below carries its own 18px gutter so the
	   photograph can reach the card's edges */
	padding: 0;
	overflow: hidden;
}

.sf-shop__grid li.wc-block-product .wc-block-components-product-image {
	/* the whole of the 26% problem is these four lines */
	margin: 0;
	padding: 0;
	aspect-ratio: 1 / 1;
	height: auto;
	overflow: hidden;
}

/* THE SOURCES ARE 1024x1024, so a square box crops NOTHING -- `cover`
   here is pure gain over `contain`, not a trade. Re-check this if a
   product is ever uploaded at a different ratio: cover would then
   crop it, and the fix is the FILE, as category-grid.css records. */
.sf-shop__grid li.wc-block-product .wc-block-components-product-image a,
.sf-shop__grid li.wc-block-product .wc-block-components-product-image img {
	display: block;
	width: 100%;
	height: 100%;
	margin: 0;
	/* !important BECAUSE THE RULE IT SUPERSEDES HAS IT. The earlier
	   `object-fit: contain !important` sits on the identical selector,
	   so equal specificity plus source order is NOT enough -- an
	   important declaration beats a normal one whatever the order.
	   Measured live after the first deploy: still `contain`, with the
	   element's own inline `object-fit:cover` from WooCommerce losing
	   to it as well. shop-beast.css sets contain twice more. */
	object-fit: cover !important;
}

/* ---- 2. the excerpt is off ---------------------------------- */
/* Truncated marketing copy, and it was the loudest text on the card.
   Hidden rather than removed from the template: the block is in
   archive-product.html, and a CSS hide is one line to undo where a
   template edit is a deploy. It still costs the markup; take it out
   of the template if that ever matters. */
.sf-shop__grid li.wc-block-product .wp-block-post-excerpt,
.sf-shop__grid li.wc-block-product .wp-block-post-excerpt:empty {
	display: none;
}

/* ---- 3. the meta row: the index, then the category ----------- */
/* THE NUMBER IS THE SAME COUNTER THE CORNER BADGE USED. It is
   incremented on the <li> itself (see the card's own rule, far
   above), not on the ::before, so hiding that badge does not stop it
   counting and a descendant can still read it. That is what lets
   this row be pure CSS with no markup and no script. */
.sf-shop__grid li.wc-block-product::before {
	display: none;
}

/* the category label has been `display: none` since the card was
   rebuilt; it is the other half of the meta row now */
.sf-shop__grid li.wc-block-product .wp-block-post-terms {
	display: flex;
	align-items: center;
	gap: 8px;
	margin: 12px 18px 0;
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 600;
	letter-spacing: .16em;
	line-height: 1.4;
	text-transform: uppercase;
	color: var(--sa-muted);
}

.sf-shop__grid li.wc-block-product .wp-block-post-terms::before {
	content: counter( sf-item, decimal-leading-zero );
	color: var(--sa-accent);
	font-variant-numeric: tabular-nums;
}

.sf-shop__grid li.wc-block-product .wp-block-post-terms a {
	color: inherit;
	text-decoration: none;
}

/* THE SALE BADGE TAKES THE CORNER THE NUMBER VACATED. Its `left` was
   a hard offset clearing that badge -- 24 + the badge's 40 + 6 -- and
   with the badge gone it would have floated in the middle of the
   photograph's top edge. The pair moved together before and still
   does. */
.sf-shop__grid li.wc-block-product .wc-block-components-product-sale-badge {
	left: 24px;
}

/* ---- 4. the name -------------------------------------------- */
.sf-shop__grid li.wc-block-product h2.wp-block-post-title {
	margin: 8px 18px 12px;
	font-family: var(--sa-body);
	font-size: 13px;
	font-weight: 700;
	line-height: 1.42;
	letter-spacing: .02em;
	/* UPPERCASE STAYS, AND THE REASON IS THE DATA. Turning the
	   transform off to try sentence case showed what is actually
	   stored: product 889 is "Superfly Project 4 4 Spoke Sport Rims
	   For Yamaha Y15, LC135 5S" in title case while its neighbours are
	   full caps. With the transform off the grid shows BOTH, card by
	   card, which is worse than shouting consistently. Sentence-case
	   names are an edit to 104 product titles and the operator's
	   call. */
	text-transform: uppercase;
	/* two lines, always -- see the min-height rule above, which is the
	   other half of keeping a row's price line level */
	display: -webkit-box;
	-webkit-line-clamp: 2;
	-webkit-box-orient: vertical;
	overflow: hidden;
}

/* ---- 5. the price row --------------------------------------- */
/* THIS PUTS BACK A BORDER THE OPERATOR HAD REMOVED ON 2026-10-08
   ("why i stil see the border ? in product cards"), and that is worth
   knowing before anyone treats it as an oversight. It is deliberate:
   the hairline is part of design B, which was chosen from four with
   the line visible in every screenshot. What changed since October is
   the card around it -- it had a face then, so a rule across it read
   as a box being divided; the card is now the page colour with a
   full-bleed photograph, and the line is the only thing separating
   the name from the price. One line to drop if it is ever unwanted.

   THE HAIRLINE IS THIS BLOCK'S OWN BORDER, not an element. The
   proposal used a spacer div the script inserted; a border needs no
   markup, cannot fall out of order, and puts the gap above it and the
   gap below it in two properties that are read together.

   THE 24px PADDING THIS BLOCK CARRIES IS THE THEME'S, and it is the
   reason the gap looked wrong at first: the margin that appears to
   set it is on the INNER element, so a "12px" gap measured 36. */
.sf-shop__grid li.wc-block-product .wp-block-woocommerce-product-price {
	/* auto, so the price line stays level across a row whatever the
	   names do; 12px is then the minimum gap above the rule */
	margin: auto 18px 16px;
	padding-top: 12px;
	border-top: 1px solid rgba(255, 255, 255, .10);
}

.sf-shop__grid li.wc-block-product .wc-block-components-product-price {
	margin: 0;
	padding: 0;
	align-items: center;
}

.sf-shop__grid li.wc-block-product .sf-shop__priceval {
	font-family: var(--sa-body);
	font-size: 16px;
	font-weight: 700;
	letter-spacing: normal;
	color: var(--sa-text);
}

/* ---- 6. the action is a mark, not a label ------------------- */
/* `font-size: 0` collapses the text without removing it, so
   WooCommerce's own aria-label -- "Select options for ..." / "Add to
   cart: ..." -- survives for a screen reader. `display: none` would
   take that with it and leave an unlabelled icon.

   THE MARK IS A MASKED SVG, NOT AN ICON FONT. The theme's ::before
   carries `content: ""` with a mask-image data URI and
   `background-color: currentColor`, so FONT-SIZE DOES NOTHING here --
   sizing it needs width/height, and `color` is what tints it. A first
   cut set `font-size: 16px` and the mark stayed 14px.

   THIS REINSTATES SOMETHING 2026-10-22 DELIBERATELY ENDED: the
   control became the words SELECT OPTIONS because "an icon cannot
   say that" -- a variable product says Select options where a simple
   one says Add to cart. Checked before shipping: the mask is a
   DIFFERENT SVG per type, sliders against a trolley, verified across
   all 8 cards as exactly two distinct masks tracking
   product-type-variable / product-type-simple. So the two are still
   told apart visually -- but a first-time visitor will not know that
   sliders means "pick a colour first". Operator's call, made with
   the words visible in the alternative. */
/* THE CLASS IS IN THE SELECTOR ON PURPOSE. The rule being superseded
   is `.sf-shop__grid li.wc-block-product .sf-shop__cart a` with
   `font-size: 12px !important` -- identical specificity to a plain
   version of this, and TWO IMPORTANT DECLARATIONS AT EQUAL
   SPECIFICITY are decided by source order, which should have been
   enough. It was not: measured live after the first deploy, the
   label was still painting at 12px with this rule present, correct
   and last in the file. Adding `.wp-element-button` takes it to
   (0,4,2) so the question does not arise. WooCommerce puts that
   class on the anchor itself. */
.sf-shop__grid li.wc-block-product .sf-shop__cart a,
.sf-shop__grid li.wc-block-product .sf-shop__cart button,
.sf-shop__grid li.wc-block-product .sf-shop__cart a.wp-element-button,
.sf-shop__grid li.wc-block-product .sf-shop__cart button.wp-element-button {
	gap: 0 !important;
	font-size: 0 !important;
	letter-spacing: 0 !important;
}

.sf-shop__grid li.wc-block-product .sf-shop__cart a::before,
.sf-shop__grid li.wc-block-product .sf-shop__cart button::before {
	width: 16px !important;
	height: 16px !important;
	color: var(--sa-accent) !important;
}

/* ---- 7. the phone ------------------------------------------- */
/* The gutters come in with the card, which is 165px wide at 390 --
   18px each side there is a fifth of the measure. The breakpoint is
   the card's own 760px, not a new one. */
@media (max-width: 760px) {
	.sf-shop__grid li.wc-block-product .wp-block-post-terms {
		margin: 10px 14px 0;
		font-size: 9px;
	}

	.sf-shop__grid li.wc-block-product h2.wp-block-post-title {
		margin: 6px 14px 10px;
		font-size: 12px;
	}

	.sf-shop__grid li.wc-block-product .wp-block-woocommerce-product-price {
		margin: auto 14px 14px;
		padding-top: 10px;
	}

	.sf-shop__grid li.wc-block-product .sf-shop__priceval {
		font-size: 14px;
	}

	.sf-shop__grid li.wc-block-product .wc-block-components-product-sale-badge {
		left: 16px;
	}

	/* THE PHONE'S MAGENTA DISC GOES, AND NOT FOR TIDINESS. The
	   2026-10-22 rule above restores a 26px filled disc here, with
	   `color: #fff !important` for the glyph inside it -- and this
	   design tints that glyph with the accent, so the first deploy
	   painted magenta on magenta and every phone card showed a SOLID
	   PINK CIRCLE with no mark in it at all. Measured on the live page
	   at 390px; invisible on desktop, which is why it shipped.

	   Two ways out: put the glyph back to white and keep the disc, or
	   drop the disc so the phone reads like the desktop. The disc's
	   own argument was that a phone has no room for the LABEL beside a
	   price -- and the label is gone at every width now, so the disc
	   is solving a problem this design already solved. Dropping it
	   also makes the two look like one card.

	   THE TARGET GETS BIGGER, NOT SMALLER: 26px of disc becomes 44px
	   of tap area around a 16px mark, bought with padding and handed
	   straight back with an equal negative margin so the price row's
	   layout does not move. Same construction as the wishlist heart
	   two sections up.

	   `.wp-block-button__link` is in the list for the reason that rule
	   gives: without it these lose to the full-size selector at
	   (0,4,1). */
	.sf-shop__grid li.wc-block-product .sf-shop__cart a,
	.sf-shop__grid li.wc-block-product .sf-shop__cart button,
	.sf-shop__grid li.wc-block-product .sf-shop__cart .wp-block-button__link {
		width: auto !important;
		height: auto !important;
		min-width: 0 !important;
		min-height: 0 !important;
		padding: 14px !important;
		margin: -14px !important;
		background: none !important;
		border: 0 !important;
		border-radius: 0 !important;
		box-shadow: none !important;
	}
}

/* =============================================================
 * 14. THE HEADING GETS ITS AIR BACK, AND THE CARD GETS AN EDGE
 *     (operator, 2026-11-17: "why its doesnt have padding.. if you
 *     want add the padding please dont change the size of the box ..
 *     and please added the shadow , cause the card its not visible")
 * =============================================================
 * Two separate complaints, in one block because the first one is only
 * safe if it is done on BOTH columns at once.
 *
 * ---- 1. THE PADDING IS PAID FOR TWICE, OR THE ROW BREAKS ----
 *
 * The FILTER heading lost its 20px top padding an hour earlier, when the
 * two columns were brought onto one line: WordPress was pushing the grid
 * column down 19.2px and the padding was cancelling it, so both had to go
 * (see the note on .sf-shop__main). That left the heading flush against
 * the top of its own box, which is what "it doesnt have padding" is.
 *
 * PUTTING IT BACK ON THE HEADING ALONE WOULD UNDO THE FIX -- the label
 * would sit 20px BELOW the card tops instead of level with them. So the
 * SAME value goes on the grid column, and the two move down together:
 * boxes still level, text still level, and nothing changes size, which is
 * the other half of what was asked.
 *
 * SUPERSEDED BY SECTION 3 BELOW, the same day: pushing the grid column
 * down to meet the label is what took the rail's PANEL off the card's
 * panel, and the panels are what the eye lines up. The grid column takes
 * no top padding now; only the rail insets its own contents.
 *
 * ---- 2. THE CARD WAS THE PAGE COLOUR WITH A 2.4% LIGHT ON IT ----
 *
 * Measured live before this: card rgb(16,16,20) on a page of
 * rgb(16,16,20) -- the SAME colour -- with the shadow at
 * rgba(255,255,255,.024) and rgba(0,0,0,.26). Nothing was drawing an
 * edge, which is exactly "the card its not visible".
 *
 * THE FILE PREDICTED THIS. The note on --sa-lift above says, of the
 * 2026-11-16 move to the information pages' lighter shadow: "Our card is
 * the page colour exactly, so the light is doing all of it ... If that is
 * not enough, the fix is a face a few units off the page -- NOT a heavier
 * shadow, which is what was just asked to stop."
 *
 * So the face does most of the work here and the shadow only finishes it.
 * #0d0d10 is /contact-us/'s own panel colour on this same page colour --
 * the arrangement the operator approved on 2026-11-16 -- rather than a
 * value picked by eye.
 *
 * THE SHADOW IS A STEP, NOT A RETURN. .06 / .45 sits between the .024 /
 * .26 that was invisible and the .10 / .70 that was called "too much" on
 * 2026-11-16. IF IT IS TOO MUCH AGAIN, take the shadow back to
 * var(--sa-lift) / var(--sa-sink) and KEEP THE FACE: the card stays
 * visible, because the face is what makes it so.
 *
 * --sa-lift AND --sa-sink ARE DELIBERATELY NOT TOUCHED. They are shared
 * with the rail, the bar controls and the pagination, and with privacy,
 * contact and return-refund -- "change them together" as the note says.
 * This is the CARD only.
 *
 * THE HOVER HAS TO RESTATE THE SHADOW. The rule above holds the emboss
 * steady on hover by repeating it; left alone it would repeat the OLD
 * values and the card would flatten the moment the pointer arrived.
 * ============================================================= */

.sf-shop {
	/* THE RAIL'S INSET, 20 -> 28 (operator, 2026-11-17, with the rail
	   zoomed: "added more padding").

	   ONE VALUE FOR ALL FOUR SIDES, so the panel reads as evenly inset
	   rather than as a column with margins. The rail was padding: 0 20px
	   20px with its top carried by the sticky heading instead -- see
	   the .sf-shop__side-title rule below, which explains
	   why the top cannot simply be added here.

	   IT COSTS THE RAIL 16px OF CONTENT, 280 -> 264 inside the same 320px
	   track. Nothing in there is near its limit: the widest thing is the
	   price row (two boxes and the FILTER button) and the bike chips, which
	   wrap on their own. THE TRACK ITSELF DOES NOT MOVE, so the grid, the
	   cards and the row alignment are all untouched -- that is the point of
	   doing it with padding rather than with --sa-rail-w. */
	--sa-rail-pad: 28px;

	/* --sa-head-pad IS GONE. It was 28px, read by the heading AND by the
	   grid column, to push the first card down to meet the FILTER label.
	   That is reading (a) in the note below, and it is the one that moved
	   the rail's panel off the card's. The heading reads --sa-rail-pad
	   directly now, which is also what it always meant: the rail's top
	   inset. */

	--sa-card-face: #0d0d10;
	--sa-card-lift: rgba(255, 255, 255, 0.06);
	--sa-card-sink: rgba(0, 0, 0, 0.45);
}

/* ---- 3. THE TWO PANELS START ON ONE LINE, NOT THE TWO BITS OF TEXT
 *         (operator, 2026-11-17: "why the card for that filter card and
 *         the card for product not same row line ?") ----
 *
 * THE FIRST ANSWER TO THIS WAS THE WRONG ONE, and the difference is worth
 * stating because it is easy to make again. Both of these are defensible
 * readings of "same row":
 *
 *   a) the FILTER label level with the top of the first card -- align the
 *      two bits of CONTENT, which needs the rail's top inset repeated on
 *      the grid column so the card is pushed down to meet the label;
 *   b) the rail's PANEL level with the card's PANEL -- align the two
 *      BOXES, and let each one inset its own contents as it likes.
 *
 * (a) is what was built first, and it is wrong here for a reason that is
 * specific to this page: BOTH COLUMNS DRAW A PANEL. The rail has a face, a
 * 12px radius and an emboss; so does the card. Once two boxes are visible,
 * the eye lines up their EDGES and nothing else -- so pushing the cards
 * 28px down to meet the label just moved one panel off the other, which is
 * exactly what was reported. The rail insets its text by 28px because it
 * is a panel with padding; the card's photograph is full bleed because it
 * is a panel with a picture in it. Those two do not have to agree, and
 * making them agree is what broke the line.
 *
 * So the grid column takes NO top padding. It keeps margin-block-start: 0,
 * which is the separate WordPress fix above and still load-bearing.
 *
 * ---- 4. AND THE RAIL IS THE SAME PANEL AS THE CARD ----
 *
 * Not asked for in those words, but the operator called it "that filter
 * card", and it was failing the same test the product card had just been
 * fixed for: measured before this, the rail was rgb(16,16,20) on a page of
 * rgb(16,16,20) with the light at .024 -- the page colour with nothing
 * drawing an edge -- while the card beside it had just gone to #0d0d10 at
 * .06 / .45. Two panels side by side, one visible and one not.
 *
 * They read the CARD's three tokens now, so the pair cannot drift again.
 * The bar controls and the pagination keep the shared --sa-lift /
 * --sa-sink, which are also the information pages' set: this is the two
 * PANELS only.
 *
 * TO UNDO JUST THIS HALF, delete the .sf-shop__side rule's background and
 * box-shadow and leave its padding. The alignment above does not depend on
 * it. */

.sf-shop__side-title {
	padding-top: var(--sa-rail-pad);
	/* AND IT MUST CARRY THE RAIL'S FACE, NOT --sa-bg (2026-11-17).
	   This heading is sticky and has a background for one reason: the
	   filter rows scroll straight through the type without one. It was
	   written when the rail's face WAS --sa-bg, so it was invisible by
	   design -- and section 14 above then moved the rail to the card's
	   face and gave this padding but not a background, so it painted
	   #101014 over #0d0d10: a lighter band across the top of the rail,
	   with a sliver of card face showing down each side where the rail's
	   28px padding leaves it short. Reading the same token is what stops
	   the two drifting again the next time the face moves. */
	background: var(--sa-card-face);
}

.sf-shop__side {
	padding: 0 var(--sa-rail-pad) var(--sa-rail-pad);
	background: var(--sa-card-face);
	box-shadow: -4px -4px 10px var(--sa-card-lift),
	             5px 5px 14px var(--sa-card-sink);
}

.sf-shop__grid li.wc-block-product,
.sf-shop__grid li.wc-block-product:hover {
	background: var(--sa-card-face);
	box-shadow: -4px -4px 10px var(--sa-card-lift),
	             5px 5px 14px var(--sa-card-sink);
}

/* =============================================================
 * 15. THE BAR READS AT 12px, AND SORTING IS A CONTROL AGAIN
 *     (operator, 2026-11-17, with the row cropped: "i think the
 *     design is not perfect , and please increase a bit size card ,
 *     increase another 2 px")
 * =============================================================
 *
 * ---- 1. 10px -> 12px ----
 *
 * "SHOWING 1-16 OF 83 RESULTS" and "SORT BY DEFAULT" were the smallest
 * type on the page by a clear margin -- 10px against the card name's 13
 * and the rail labels' 11 -- which is most of why the row read as thin
 * and empty between two panels. The operator picked the text over the
 * bar's height when both were offered with their measurements.
 *
 * ---- 2. THE PLATE IS BACK, AND IT WAS DELIBERATELY REMOVED YESTERDAY ----
 *
 * READ THE NOTE ON .sf-sort__btn ABOVE BEFORE CHANGING THIS. It records
 * that on 2026-11-16 this control "went through a #16161a plate with a
 * ring and then the house card before the screenshot arrived; both put a
 * box inside a box, which is the one thing the reference does not do",
 * and the plate, ring, radius and emboss were all taken off so the
 * sorting was type sitting in the BAR's own card.
 *
 * THAT REASONING ASSUMED THE BAR READS AS A CARD, AND IT DOES NOT.
 * Measured on the live page the same day: the bar is rgb(16,16,20) on a
 * page of rgb(16,16,20) -- the identical colour -- with its light at
 * .024. It is the same fault the product card and the rail were both
 * fixed for hours earlier. So the "box inside a box" the plate was
 * removed to avoid never actually existed on screen: there is one box,
 * and until now it had nothing in it that looked pressable.
 *
 * So this is not a reversal of that decision so much as the other way of
 * settling it. THE OTHER WAY IS STILL OPEN AND IS ONE RULE: give the BAR
 * the card's face and emboss (--sa-card-face / -lift / -sink, as
 * .sf-shop__side now does) and delete this plate. Then the bar is a real
 * card and the sorting is type inside it, which is what 2026-11-16 was
 * reaching for. DO NOT DO BOTH -- a plate the same colour as the card it
 * sits on is invisible, and that is the trap in this corner of the page.
 *
 * THE PLATE IS THE FILTERS BUTTON'S, not the card's: #16161a with a 1px
 * ring at .14, which is what makes it legible against a bar that is the
 * page colour. The ring is load-bearing for the same reason the note on
 * .sf-shop__filters-toggle gives -- the neumorphic shadow alone is under
 * what a screen resolves at these levels. The radius is 12px rather than
 * that button's 6, because the page settled on 12 everywhere today.
 * ============================================================= */

.sf-shop__bar .wc-block-product-results-count,
.sf-shop__head .wc-block-product-results-count {
	font-size: 12px;
	line-height: 14px;
}

.sf-sort__btn {
	height: 38px;
	padding: 0 16px;
	border-radius: 12px;
	background-color: #16161a;
	box-shadow: 0 0 0 1px rgba(242, 242, 244, 0.14),
	            -2px -2px 6px rgba(255, 255, 255, 0.022),
	             3px 3px 8px rgba(0, 0, 0, 0.26);
	font-size: 12px;
	line-height: 14px;
}

.sf-sort__btn:hover,
.sf-sort__btn[aria-expanded="true"] {
	background-color: #1b1b20;
	box-shadow: 0 0 0 1px var(--sa-accent),
	            -3px -3px 8px rgba(255, 255, 255, 0.022),
	             4px 4px 10px rgba(0, 0, 0, 0.26);
}

/* A PHONE RUNS THE FILTERS BUTTON AND THIS ON ONE ROW, so the plate gives
   back the padding and the tracking that the 760px block took off while
   this was bare type. Measured at 375 after: the two controls and the gap
   fit the bar with room, and the count sits on its own line above them,
   which is that block's own arrangement. */
@media (max-width: 760px) {
	.sf-sort__btn {
		padding: 0 12px;
		letter-spacing: 0.08em;
	}
}

/* =============================================================
 * 16. SORTING IS THE REFERENCE'S BOX (operator, 2026-11-17, with an
 *     image: "i want like this")
 * =============================================================
 * A wide bordered box with the value hard LEFT and the chevron hard
 * RIGHT, reading "DEFAULT SORTING" -- WooCommerce's own words.
 *
 * FOUR THINGS CHANGED, AND THE FIRST TWO ARE A PAIR:
 *
 *   - the quiet "Sort by" ::before is OFF. It was there so the value
 *     could be shortened to "Default" / "Popularity"; the reference
 *     carries no prefix and shows the native string instead.
 *   - so shop-archive.js no longer strips "sort by" and "sorting" from
 *     the option text. PUT ONE BACK AND YOU MUST PUT BOTH BACK, or the
 *     control reads "Sort by Sort by popularity" -- which is the exact
 *     string the stripper was written to prevent, and it is recorded in
 *     that file at the point where it used to live.
 *   - justify-content is space-between, not flex-end. The two ends of
 *     the box are what make it read as a select rather than as a line of
 *     type with an arrow after it.
 *   - and it needs a MIN-WIDTH for that to show. Shrink-wrapped, "space
 *     between" has nothing to space: the box was 180px of content and
 *     the chevron sat against the word anyway.
 *
 * 230px AND 44px ARE MEASURED OFF THE REFERENCE, scaled from the image.
 * The radius stays at the page's 12px rather than the reference's ~10 --
 * every panel settled on 12 today and a 2px difference on one control is
 * not worth breaking that for.
 *
 * THE MIN-WIDTH IS DESKTOP ONLY. At 375 the bar runs Filters and this on
 * one row inside 295px of content; 230 plus the button plus the gap does
 * not fit, so below 760 it shrink-wraps as before and the ellipsis on
 * .sf-sort__label is what keeps a long value from pushing Filters off
 * the screen. That note is on the label rule above and still applies --
 * the values are LONGER now, since they are no longer shortened.
 * ============================================================= */

.sf-sort__btn::before {
	content: none;
}

.sf-sort__btn {
	justify-content: space-between;
	gap: 16px;
	height: 44px;
}

@media (min-width: 761px) {
	.sf-sort__btn {
		min-width: 230px;
	}
}

/* =============================================================
 * 17. SMALLER TYPE AND A SMALLER PLATE, SAME BAR (operator,
 *     2026-11-17: "reduce a bit the text and default sorting card
 *     and showing text .. i just want the bar only that added
 *     height")
 * =============================================================
 * The bar had grown to 72px because a 44px plate sat inside 14px of
 * padding. The HEIGHT was wanted; the things in it were not.
 *
 * SO THE PADDING ABSORBS EXACTLY WHAT THE PLATE GIVES UP.
 *
 *     before   14 + 44 + 14 = 72
 *     after    18 + 36 + 18 = 72
 *
 * The bar does not move by a pixel, the cards below it do not move,
 * and the row still starts level with the rail. THAT IS THE WHOLE
 * CONSTRAINT: if the plate is ever re-sized again, move the bar's
 * padding-block by half the difference in the other direction, or
 * the bar's height goes with it.
 *
 * 12px -> 11px on both the count and the sorting, which is "a bit"
 * -- it was 10 this morning and 12 an hour ago. 11 is where the
 * rail's own labels sit, so the three now agree.
 *
 * The plate also comes in: 230 -> 200 wide and 16 -> 14 of side
 * padding. The min-width is still desktop-only and the phone still
 * shrink-wraps; see section 16.
 * ============================================================= */

.sf-shop__bar {
	padding-block: 18px;
}

.sf-shop__bar .wc-block-product-results-count,
.sf-shop__head .wc-block-product-results-count {
	font-size: 11px;
	line-height: 13px;
}

.sf-sort__btn {
	height: 36px;
	padding: 0 14px;
	font-size: 11px;
	line-height: 13px;
}

/* AND THE FILTERS BUTTON COMES DOWN WITH IT. The two share a row on a
   phone -- it is hidden above 1024, where the rail is simply there -- and
   the bar centres them, so a 38px button beside a 36px plate sat ONE PIXEL
   out at the top. Measured at 375 before this: tops 1px apart, which is
   invisible until you look for it and then is all you can see.

   THEY ARE A PAIR. Re-size one and re-size the other. */
.sf-shop__filters-toggle {
	height: 36px;
}

@media (min-width: 761px) {
	.sf-sort__btn {
		min-width: 200px;
	}
}

/* =============================================================
 * 18. THE CATEGORY SECTION IS NAVIGATION (operator, 2026-11-17:
 *     "for categories , i want its same like durian bikers , its
 *     will navigate not filter")
 * =============================================================
 * A list of links to the category archives, where chips used to set
 * ?categories=<slug> on the current query. Built by
 * superfly_category_nav_shortcode(); that function carries the
 * argument for the swap and what it costs.
 *
 * IT IS DRAWN AS THE MODEL LIST, NOT AS THE CHIPS IT REPLACES --
 * name left, count right, one per row. Three reasons, and the first
 * is the point of the whole change: a row of links should not look
 * like a row of filter chips, or nobody can tell that pressing one
 * leaves the page. The second is that it is six names of uneven
 * length; the third is that the Model list directly above it is
 * already exactly this shape, so the rail reads as two lists rather
 * than as list-chips-list.
 *
 * THE COUNT IS NOT A LINK OF ITS OWN and sits inside the anchor, so
 * the whole row is the target -- the Model rows behave the same way.
 *
 * .is-selected IS THE CURRENT ARCHIVE, written by the shortcode. It
 * is also one of the two hooks shop-archive.js reads to light a
 * section's magenta spine, so the section marks itself live on a
 * category page with no extra wiring.
 * ============================================================= */

.sf-rail__nav-list {
	margin: 0;
	padding: 0;
	list-style: none;
}

.sf-rail__nav-link {
	display: flex;
	align-items: baseline;
	gap: 10px;
	padding: 7px 0;
	border: 0;
	background: none;
	color: var(--sa-text);
	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 500;
	line-height: 1.3;
	letter-spacing: 0.06em;
	text-transform: uppercase;
	text-decoration: none;
	transition: color 0.2s ease;
}

/* THE SELECTOR CARRIES .sf-shop__side BECAUSE WOOCOMMERCE OWNS EVERY
   <a> ON THIS PAGE. woocommerce.css sets

       body.woocommerce a, body.woocommerce-page a { color: var(--wc-text) }

   which is 0,2,1 -- two classes and an element -- so a lone
   .sf-rail__nav-link at 0,1,0 loses. The rail's Clear link records
   the same trap and is written the same way. */
.sf-shop__side .sf-rail__nav-link:hover,
.sf-shop__side .sf-rail__nav-link:focus-visible {
	color: var(--sa-accent);
}

.sf-shop__side .sf-rail__nav-link.is-selected {
	color: var(--sa-accent);
}

.sf-rail__nav-count {
	margin-left: auto;
	font-size: 10px;
	color: var(--sa-muted);
	font-variant-numeric: tabular-nums;
}


/* =============================================================
 * 13b. WHAT EACH SECTION IS DOING  (2026-11-17)
 * -------------------------------------------------------------
 * Ported from the localhost study in docs/home-concepts/brand-page,
 * built there across 2026-11-17 and chosen by the operator.
 *
 * Three additions, all drawn by shop-archive.js:
 *
 *   a badge on each section head   what it is filtering by, visible
 *                                  when the section is FOLDED, which
 *                                  is when the magenta spine is not
 *   a Clear inside each body       undo one section without losing
 *                                  the bike and the price with it
 *   a count on Filters             the phone folds the rail away, so
 *                                  the button is the only thing left
 *                                  that can say anything is set
 *
 * The fold itself and the URL state are NOT part of this: the fold
 * has been here since 2026-11-16 (section 13a) and WooCommerce's
 * filter blocks own the query string natively.
 * ============================================================= */

/* THE BADGE LIVES INSIDE THE HEAD, which is the whole point of it: a
   folded section hides its body, and the magenta spine is ON the body.
   Without this a section with three models set looks exactly like one
   with none for as long as it is shut. */
.sf-shop__side .sf-rail__badge {
	display: inline-flex;
	align-items: center;
	justify-content: center;
	min-width: 15px;
	height: 15px;
	padding: 0 4px;
	border-radius: 999px;
	background: var(--sa-accent);
	color: #fff;
	font-family: var(--sa-tech);
	font-size: 9px;
	font-weight: 700;
	letter-spacing: 0;
	font-variant-numeric: tabular-nums;
}

.sf-shop__side .sf-rail__badge[hidden] {
	display: none;
}

/* IT IS NOT IN THE HEAD, and that is forced rather than chosen: the
   head is itself a <button>, and a button inside a button is nested
   interactive content — browsers resolve it by dropping one of them
   and screen readers resolve it any way they like. So it sits at the
   top of the body instead, beside the controls it clears. */
.sf-shop__side .sf-rail__wipe {
	display: block;
	width: fit-content;
	margin: 0 0 10px auto;
	padding: 0;
	background: none;
	border: 0;
	cursor: pointer;
	font-family: var(--sa-tech);
	font-size: 9px;
	font-weight: 600;
	letter-spacing: 0.14em;
	text-transform: uppercase;
	color: var(--sa-faint);
	transition: color 0.18s ease;
}

.sf-shop__side .sf-rail__wipe:hover,
.sf-shop__side .sf-rail__wipe:focus-visible {
	color: var(--sa-accent);
}

.sf-shop__side .sf-rail__wipe[hidden] {
	display: none;
}

/* -------------------------------------------------------------
 * PRICE BANDS
 *
 * Two drags and a button press becomes one tap. A band IS the
 * slider — shop-archive.js moves both handles and lets one
 * navigation through — so the state stays a single price range
 * and the removable chip in the bar still reads it.
 *
 * THEY CARRY NO COUNTS, unlike the localhost version. Counting a
 * band needs every product's price in the current result set,
 * which the page does not have; it wants a PHP figure behind it
 * the way superfly_category_index_stats() already computes a
 * range per category. Left out of this deploy on purpose — it is
 * a functions.php change.
 * ------------------------------------------------------------- */

.sf-shop__side .sf-rail__bands {
	display: flex;
	flex-wrap: wrap;
	gap: 5px;
	margin: 0 0 12px;
}

.sf-shop__side .sf-rail__band {
	flex: 1 1 auto;
	padding: 5px 9px;
	background: none;
	border: 1px solid var(--sa-hair);
	border-radius: 999px;
	cursor: pointer;
	font-family: var(--sa-tech);
	font-size: 9px;
	font-weight: 600;
	letter-spacing: 0.08em;
	text-transform: uppercase;
	color: var(--sa-faint);
	white-space: nowrap;
	transition: border-color 0.16s ease, color 0.16s ease,
	            background-color 0.16s ease;
}

.sf-shop__side .sf-rail__band:hover,
.sf-shop__side .sf-rail__band:focus-visible {
	border-color: rgba(236, 30, 121, 0.5);
	color: var(--sa-text);
}

.sf-shop__side .sf-rail__band[aria-pressed="true"] {
	border-color: var(--sa-accent);
	background-color: rgba(236, 30, 121, 0.12);
	color: var(--sa-accent);
}

/* two per row on a phone, where four in a line would each be four
   characters wide and read as a single grey smear */
@media (max-width: 760px) {
	.sf-shop__side .sf-rail__band {
		flex: 1 1 calc(50% - 3px);
		text-align: center;
	}
}

/* -------------------------------------------------------------
 * THE COUNT ON THE FILTERS BUTTON
 *
 * Only ever seen on a phone, because the button is hidden above
 * 1024px where the rail is simply on the page — and a phone is
 * exactly where it is needed, since the fold takes the rail off
 * the screen and the button becomes the only thing left that can
 * say anything is set at all.
 * ------------------------------------------------------------- */

.sf-shop__filters-n {
	display: inline-flex;
	align-items: center;
	justify-content: center;
	min-width: 16px;
	height: 16px;
	padding: 0 4px;
	border-radius: 999px;
	background: var(--sa-accent);
	color: #fff;
	font-size: 9px;
	letter-spacing: 0;
	font-variant-numeric: tabular-nums;
}

.sf-shop__filters-n[hidden] {
	display: none;
}

/* the button goes magenta-on-pressed, and a magenta pill on a magenta
   label disappears — so the count inverts with it */
.sf-shop__filters-toggle[aria-expanded="true"] .sf-shop__filters-n {
	background: #fff;
	color: var(--sa-accent);
}


/* =============================================================
 * 13c. THE TYPE-AHEAD AND THE PRICE DISTRIBUTION  (2026-11-17)
 * -------------------------------------------------------------
 * The two halves of the localhost rail that did not come across
 * in the first deploy — operator: "why still not same like you
 * build at localhost , the price graph and search model in
 * filter model".
 *
 * Both are drawn by shop-archive.js. The bars are real figures:
 * every product's price for the current set, out of WooCommerce's
 * own Store API, so neither needed anything in functions.php.
 * ============================================================= */

/* -------------------------------------------------------------
 * THE TYPE-AHEAD
 *
 * It HIDES rows rather than rebuilding them, because these rows
 * are WooCommerce's — every one carries data-wp-each-child and a
 * data-wp-bind--hidden bound to their own store. So the match is
 * marked with a class and hidden here, and `hidden` itself is
 * left alone: writing it would fight the binding that owns it.
 * ------------------------------------------------------------- */

.sf-shop__side .sf-rail__q {
	margin: 0 0 10px;
}

.sf-shop__side .sf-rail__qin {
	width: 100%;
	padding: 8px 10px;
	background-color: rgba(242, 242, 244, 0.05);
	border: 1px solid var(--sa-line);
	border-radius: 8px;
	color: var(--sa-text);
	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 500;
	letter-spacing: 0.08em;
}

.sf-shop__side .sf-rail__qin::placeholder {
	color: var(--sa-faint);
	text-transform: uppercase;
	letter-spacing: 0.1em;
}

.sf-shop__side .sf-rail__qin:focus {
	outline: none;
	border-color: var(--sa-accent);
}

/* the platform's own clear cross is unstyleable and sits in the
   corner of a 294px field where it reads as a stray mark */
.sf-shop__side .sf-rail__qin::-webkit-search-cancel-button {
	-webkit-appearance: none;
}

/* TWO CLASSES, because this has to beat the block's own single-class
   rules on the same element rather than rely on source order. */
.sf-shop__side .sf-q-off {
	display: none;
}

.sf-shop__side .sf-rail__qnone {
	margin: 4px 0 0;
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 500;
	letter-spacing: 0.1em;
	text-transform: uppercase;
	color: var(--sa-faint);
}

/* -------------------------------------------------------------
 * THE PRICE DISTRIBUTION
 *
 * Linear buckets on the slider's own axis — a log scale reads
 * far better on a catalogue this skewed and is unusable, because
 * the slider underneath is linear and the bars would stop lining
 * up with the handles. A bar that is not under the handle
 * pointing at it is worse than no bar.
 * ------------------------------------------------------------- */

.sf-shop__side .sf-rail__hist {
	/* 8px each side, because that is as close to either end as a
	   16px handle can put its own centre. The accent fill on the
	   track is inset the same way, so the three agree. */
	margin: 0 8px 2px;
	height: 34px;
	display: flex;
	align-items: flex-end;
	gap: 1px;
}

.sf-shop__side .sf-rail__bar {
	flex: 1 1 0;
	min-width: 0;
	height: calc(4px + 30px * var(--h, 0));
	border-radius: 1px 1px 0 0;
	background-color: rgba(242, 242, 244, 0.17);
	transition: background-color 0.14s ease;
}

/* an empty bucket keeps a hairline — a gap in the row reads as a
   rendering fault, where a flat line reads as "none here" */
.sf-shop__side .sf-rail__bar[data-empty] {
	height: 2px;
	background-color: rgba(242, 242, 244, 0.09);
}

.sf-shop__side .sf-rail__bar.is-in {
	background-color: rgba(236, 30, 121, 0.55);
}

/* the chip's label and its count are separate nodes, and that is
   load-bearing: the label is written by applyBands and the count by
   paintBandCounts, so sharing one textContent makes the two erase
   each other on every pass of the observer */
.sf-shop__side .sf-rail__bandn {
	font-style: normal;
	margin-left: 5px;
	color: var(--sa-muted);
	font-variant-numeric: tabular-nums;
}

.sf-shop__side .sf-rail__band[aria-pressed="true"] .sf-rail__bandn {
	color: var(--sa-accent);
}

.sf-shop__side .sf-rail__band[disabled] {
	opacity: 0.38;
	cursor: default;
}

.sf-shop__side .sf-rail__band[disabled]:hover {
	border-color: var(--sa-hair);
	color: var(--sa-faint);
	background-color: transparent;
}

@media (max-width: 760px) {
	.sf-shop__side .sf-rail__hist {
		height: 28px;
	}

	.sf-shop__side .sf-rail__bar {
		height: calc(4px + 24px * var(--h, 0));
	}
}


/* =============================================================
 * 13d. ONE TYPE FOR THE WHOLE RAIL  (operator, 2026-11-17)
 * -------------------------------------------------------------
 * "is it the size font is same in filter" — it was not. Measured
 * before this block, the rail ran Chakra Petch 11px everywhere
 * except three places:
 *
 *   model rows      Inter 400 14px, mixed case, no tracking
 *   price bands     Chakra Petch 600 9px
 *   Show N more     Chakra Petch 600 8.75px
 *
 * THE MODEL LIST WAS THE ONE NOBODY HAD EVER STYLED. It is
 * WooCommerce's checkbox block coming straight through, which is
 * why it was the only Inter, the only 400 weight and the only
 * mixed case in the panel — 27% larger than the category list
 * directly below it, and with its own count set in Chakra Petch
 * beside a name set in Inter.
 *
 * The operator's own four values:
 *
 *   category rows   Chakra Petch 600 11px uppercase
 *   model rows      the same
 *   price bands     Chakra Petch 500 11px
 *   Show N more     Chakra Petch 500 11px
 * ============================================================= */

/* ---- the category list: 500 -> 600 ---- */
.sf-shop__side .sf-rail__nav-name {
	font-weight: 600;
}

/* ---- the model list, made to match it exactly ----
   Set on the LABEL so the checkbox and the row share one line
   height, and again on the text wrapper because WooCommerce gives
   that child its own 0.875em — without the second rule the name
   would come out at 9.6px while the row around it is 11. */
.sf-shop__side .wc-block-product-filter-checkbox-list__label {
	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 600;
	letter-spacing: 0.06em;
	text-transform: uppercase;
}

.sf-shop__side .wc-block-product-filter-checkbox-list__text-wrapper {
	font-size: inherit;
	font-weight: inherit;
	letter-spacing: inherit;
}

/* the count beside the name, matched to the category list's own —
   quieter and a step smaller, so the pair of lists read alike */
.sf-shop__side .wc-block-product-filter-checkbox-list__count {
	font-family: var(--sa-tech);
	font-size: 10px;
	font-weight: 500;
	letter-spacing: 0.06em;
}

/* ---- the price bands: 600/9 -> 500/11 ---- */
.sf-shop__side .sf-rail__band {
	font-size: 11px;
	font-weight: 500;
	letter-spacing: 0.06em;
}

/* AT 11px FOUR CHIPS NO LONGER FIT A ROW, and `flex: 1 1 auto` would
   leave a 3-1 rag. Two by two is the only even split a 294px rail
   has room for at this size — the same pair the phone already used,
   now at every width. */
.sf-shop__side .sf-rail__bands {
	gap: 6px;
}

.sf-shop__side .sf-rail__band {
	flex: 1 1 calc(50% - 3px);
	text-align: center;
}

/* ---- Show N more: 600/8.75 -> 500/11 ----
   8.75px was 0.875em of WooCommerce's own 10px, never a chosen
   size. It is the smallest text in the rail and it is a control. */
.sf-shop__side .wc-block-product-filter-checkbox-list__show-more-button {
	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 500;
	letter-spacing: 0.06em;
}

/* =============================================================
 * AND THE SECTION NUMBERS COUNTED WRONG
 * -------------------------------------------------------------
 * Price 01, Bike 02, Model 03, Category 03 — spotted by the
 * operator in a screenshot.
 *
 * The counter only incremented on WooCommerce's own filter
 * blocks, and CATEGORY stopped being one earlier today: it is
 * `.sf-rail__nav` now, a list of links to the archives rather
 * than a taxonomy filter. So it never incremented and reprinted
 * whatever Model had left behind.
 *
 * getSections() in shop-archive.js already names all three
 * selectors for exactly this reason. This is the stylesheet
 * catching up with it — keep the two lists in step.
 * ============================================================= */

.sf-shop__side .sf-rail__nav {
	counter-increment: sf-filter 1;
}


/* -------------------------------------------------------------
 * THE MODEL NAME LIVES ONE SPAN DEEPER THAN 13d STYLED
 * (operator, 2026-11-17: "why the model like Y15ZR looks like
 * not same with YAMAHA , BODYKITS fonts")
 *
 * Because it was still Inter 600 14px. 13d set the row on
 * `…__label` and reset `…__text-wrapper` to inherit, and neither
 * is the element that holds the words:
 *
 *   label                 <- 13d set 11px Chakra Petch here
 *     input-wrapper
 *     text-wrapper        <- 13d set font-size: inherit here
 *       text              <- THE WORDS ARE HERE, and this file
 *                            already carried an explicit
 *                            `font-family: var(--sa-body);
 *                             font-size: 14px` on it
 *       count             <- already matched
 *
 * INHERITANCE CANNOT BEAT A DIRECT RULE, however specific the
 * ancestor's selector is, so the name was never going to pick up
 * the row's type while that rule existed.
 *
 * AND THE EARLIER CHECK MISSED IT because it measured the two
 * elements the rule had been written for and found them both
 * correct. MEASURE THE PARENT OF THE TEXT NODE — walk to the text
 * and read what owns it, rather than reading the element you
 * aimed at.
 *
 * 0,3,0, so it does not rely on source order against the 0,2,0
 * rule above it. Colour is deliberately left alone: the hover and
 * :checked rules paint this same element accent and still do.
 * ------------------------------------------------------------- */

.sf-shop__side .wc-block-product-filter-checkbox-list__label
	.wc-block-product-filter-checkbox-list__text {
	font-family: var(--sa-tech);
	font-size: 11px;
	font-weight: 600;
	letter-spacing: 0.06em;
	text-transform: uppercase;
}
