/**
 * core/heading rules, each scoped to the unit that owns it.
 *
 * This file is shared by construction: every unit that places a core/heading lands here,
 * and that is nearly every unit. Nothing in it may be written unscoped -- a rule with no
 * unit hook class repaints every heading on the site, which is a theme decision and
 * belongs to wp-theme-scaffold.
 */

/*
 * .wm-section-content: the blue double-bar before the FIRST item's heading.
 *
 * The contract's `chrome.heading-marker` row. It is a pseudo-element in the source
 * (`.image-position-left .section-heading:before`, scss/modules/_paragraph.scss:279-300),
 * and it has to be one here, for a reason that has nothing to do with which block could
 * hold the glyph:
 *
 *   - it is `position: absolute` at `left: -70px` inside the `h2`, which is its containing
 *     block, so it sits OUTSIDE the heading's box and over the media column. No block
 *     attribute places a block outside its parent's box.
 *   - `core/icon` would register and paint it BLACK:
 *     `WP_Icons_Registry::sanitize_icon_content()` keeps only fill/fill-rule/d/transform on
 *     a `<path>` and deletes `<defs>` and `<style>` outright, and this glyph's #013385
 *     lives in a `.cls-1` class inside the `<defs><style>` that is deleted
 *     (reference/sources/wordpress-core.md; silent-failures.md row 40).
 *
 * ITEM ONE ONLY. The source selects it with `.image-position-left`, and item two carries
 * `image-position-right`, so it has no marker -- which is the asymmetry between the two
 * headings. `wm-section-content__item--media-left` is this build's spelling of that class.
 *
 * THE HEIGHT IS THE HEADING'S, not a fixed number: `height: 100%` over a heading measured
 * 48px tall at 1440x900 and 31px at the three narrower viewports gives the contract's
 * 50x48 / 50x31 / 50x31 / 50x31 with one declaration and no breakpoint.
 *
 * NO `background-size`, DELIBERATELY. The asset is 50x57 intrinsic in a 50x48 box, so the
 * source clips the bottom 9px at 1440x900 and the bottom 26px at the other three.
 * `background-size: contain` would scale the bars instead of clipping them, which is a
 * different picture. `background-repeat: repeat` is the source's computed value and is also
 * CSS's initial value; it is written because the row records it, and no tile repeats --
 * one 57px-tall tile overflows the 48px box on its own.
 *
 * AT 390 IT RENDERS OFF-VIEWPORT AND THAT IS FAITHFUL. The heading's x is 20, so the
 * marker's box is x=-50..0. The source measures `overflow: visible` on html, body, main and
 * the unit root alike and nothing scrolls to it -- leftward overflow is not scrollable in
 * LTR -- so the rule carries no breakpoint and matches at all four viewports.
 */
/*
 * .wm-inner-section-content JOINS BOTH RULES BELOW rather than getting copies of them
 * (2026-09-11), and its contract says in as many words that it reaches the marker "through the
 * SAME rule rather than a parallel one": the compiled selector
 * `.paragraph__inner_section_content, .paragraph__section_content_item` is a two-member list
 * opening at scss/modules/_paragraph.scss:149, and this unit is the FIRST member.
 *
 * THE PROBE THAT FOUND IT IS ON THE CONTRACT. `.paragraph__inner_section_content.image-position-left
 * .section-heading` matched 1 on this unit's reference page at 1440x900 and 390x844; the looser
 * `.image-position-left .section-heading` matched 3, the other two being
 * `.paragraph__text_image_carousel` headings that get NO marker.
 *
 * EVERY DECLARATION IS IDENTICAL AND ONLY THE HOST BOX DIFFERS, which no declaration names: the
 * `height: 100%` rides a heading measured 48px tall at 1440x900 and 31px at the three narrower
 * viewports, giving the contract's 50x48 / 50x31 / 50x31 / 50x31 with one rule and no breakpoint.
 * The asset is the same `m-dc7d758a` / header-graphic.svg, already in assets/images/, and its own
 * #013385 is this theme's `secondary` -- but it is not written as a token here because the colour
 * lives inside the SVG file, not in this rule.
 *
 * NO `image-position-right` VARIANT SELECTOR, unlike `.wm-section-content` above. This unit has
 * one captured instance and it carries `image-position-left`, which the contract records as
 * measured off the stored markup's root class list. The contract also records that an
 * `image-position-right` sibling exists and is NOT captured, and that "which core spelling carries
 * it ... is not decided here" -- so there is no second class to scope against yet and the marker
 * hangs off the unit root.
 *
 * AT 390 IT RENDERS OFF-VIEWPORT AND THAT IS FAITHFUL, for the reason the note above gives: this
 * unit's heading measures x=10 at 390x844, so the marker's box is x=-60..-10, and the contract
 * records `overflow: visible` probed on html, body, main, the unit root and its content column
 * alike with nothing scrolling to it. No breakpoint.
 */
.wm-section-content .wm-section-content__item-heading,
.wm-inner-section-content .wm-inner-section-content__heading {
	position: relative;
}

.wm-section-content .wm-section-content__item--media-left .wm-section-content__item-heading::before,
.wm-inner-section-content .wm-inner-section-content__heading::before {
	content: "";
	position: absolute;
	top: 0;
	left: -70px;
	display: inline-block;
	width: 50px;
	height: 100%;
	background-image: url("../../images/header-graphic.svg");
	background-repeat: repeat;
}

/*
 * .wm-testimonials: the testimonial heading's LINE HEIGHT is 1.4, and theme.json's `h2` is 1.1.
 *
 * The contract asserts `level: 2` and `fontSize: xx-large` and nothing about leading, but its
 * `why` records the heading's measured BOX at all four viewports -- 555x123 at 44px, 467x78 at
 * 28px, 708x39 at 28px, 350x78 at 28px -- and every one of those divides out to 1.393-1.398:
 *
 *   1440  123 / 2 lines / 44px = 1.3977
 *   1024   78 / 2 lines / 28px = 1.3929
 *    768   39 / 1 line  / 28px = 1.3929
 *    390   78 / 2 lines / 28px = 1.3929
 *
 * THE 768 ROW IS THE ONE THAT SETTLES IT, because it needs no guess about where the string
 * wraps: a 708px-wide box holding a 39px line at a 28px font is one line, and 39/28 is 1.393
 * whatever the other three do. The rest agree with it.
 *
 * `styles.elements.h2.lineHeight` is 1.1, which renders the same heading 97px tall at 1440 and
 * 62px at 1024 -- and because the quote body and the CTA sit below it in a flow column, that
 * 26px shortfall moves EVERY box under it by the same amount. One declaration closes all of it.
 *
 * SCOPED TO THIS UNIT, AND THE CORPUS SAYS THAT IS EXACTLY WHERE IT BELONGS. theme.json's own
 * `styles.elements.heading.lineHeight` is already 1.4; `h2` alone is specialised to 1.1, which
 * was a token decision taken on an earlier unit. An earlier version of this comment said the
 * rule should be DELETED if that 1.1 turned out to be wrong for the site. It has since been
 * measured and it is not wrong: a sweep of every stored reference payload puts the h2 leading
 * at 1.1 on nine units and 1.4 on six, and this band is one of the six. So the token stands,
 * the outlier is scoped, and the note that invited a token change is retired -- the sweep is
 * the check that closes it, and re-running it is what a future doubt costs.
 *
 * THE SAME BAND'S OTHER THREE TYPOGRAPHY CORRECTIONS ARE BLOCK ATTRIBUTES, NOT RULES HERE --
 * the quote's weight and leading, and the CTA's leading, written on the blocks themselves in
 * patterns/testimonials.php items 5 and 6. An attribute is the higher rung (CLAUDE.md, Build
 * preference order) and core/paragraph has no class to scope a rule to at all: its
 * `supports.className` is false, so a rule would have to select a bare `p` inside the unit and
 * would catch every future paragraph in it. This heading rule is the one survivor, from the
 * pass before those three; unifying it with them would change no computed value, so it is
 * raised in the hand-back rather than churned here.
 *
 * NOT A BREAKPOINT RULE. The ratio is the same at all four viewports; the font size steps and
 * the leading follows it, which is what a unitless line-height does by construction.
 */
.wm-testimonials .wp-block-heading {
	line-height: 1.4;
}

/*
 * .wm-image-link-cards: the card label is 28px AT EVERY VIEWPORT, and a block attribute cannot
 * say so on this theme.
 *
 * The contract's `core/heading` node measures the label's anchor "at 28px" and its wrapper at
 * 360x84 -- 84 / 2 lines / 28px = 1.5, which is the `lineHeight` written on the block itself in
 * patterns/image-link-cards.php. The SIZE could not go with it, and that is a measurement
 * rather than a preference:
 *
 * `settings.typography.fluid` is true on this theme, so core's `wp_render_typography_support()`
 * rewrites a block's own `style.typography.fontSize` into a clamp AT RENDER. Written as
 * `"fontSize":"28px"`, the pattern file on disk says `font-size:28px` and
 * `curl http://127.0.0.1:8000/?page_id=34` says
 * `font-size:clamp(17.905px, 1.119rem + ((1vw - 3.2px) * 1.147), 28px)` -- measured 2026-09-11.
 * The two disagree and nothing reports it, because 1440 lands on the clamp's own maximum and
 * looks correct; the label computed 25.98px at 1024, 23.04px at 768 and 18.71px at 390, which
 * took its 84px box to 78, 69 and 56 and the band's measured 1923 to 1905.
 *
 * The contract's own note is the other half of this: it says "No `fontSize`" because `xx-large`
 * is fluid {28px, 44px} and 28 is its MINIMUM, so the slug renders 44px at 1440. Both routes an
 * attribute offers -- a preset slug and a literal -- are fluid on this theme, and this rule is
 * the one spelling that is not. theme.json's `styles.elements.h3` is `x-large` (26px), which a
 * 0,2,0 selector beats without `!important`; there is no inline font-size left to beat, which
 * is exactly why the attribute had to be dropped rather than supplemented.
 *
 * NOT A TOKEN CHANGE. 28px is off this theme's fontSizes ladder entirely, and bending a preset
 * to suit one unit would move every unit that reads it (wp-theme-scaffold's call, not a
 * build's). Scoped to this unit's own label class for the same reason.
 */
.wm-image-link-cards .wm-image-link-cards__label {
	font-size: 28px;
}

/*
 * .wm-link-list: THE HEADING'S ANCHOR -- RED, UN-UNDERLINED, 32px, WITH A RED CHEVRON AFTER IT.
 *
 * Four measured differences, none of which a block attribute can carry, because the thing they
 * describe is an `<a>` inside `core/heading`'s rich text rather than a block. The contract's
 * heading node records the markup ("THE LINK IS INSIDE THE HEADING, not beside it") and asserts
 * nothing about its treatment; the unit contract carries no `chrome` key at all, so the four
 * facts below reached this build through a verification pass and are named as a discovery
 * defect in its hand-back.
 *
 * 1. COLOUR IS `primary`, rgb(219, 36, 11), at all four viewports. It rendered BLACK before
 *    this rule, and the cause is a theme token rather than an omission: theme.json's
 *    `styles.elements.link.color.text` is `currentColor`, so every link on this site takes its
 *    parent's colour, and an `h3` has none of its own. `:root :where(a:where(:not(.wp-element-button)))`
 *    is 0,1,0; this selector is 0,2,1 and needs no `!important`.
 *
 *    IT IS WRITTEN ON THE ANCHOR AND NOT AS `textColor` ON THE HEADING, which would have been
 *    the higher rung. The anchor is the heading's whole content, so the two render identically
 *    -- but the source's `h3` is not red, only its link is, and the comparator measures `color`
 *    on both nodes. A `textColor` attribute would close one difference and open another.
 *
 * 2. NO UNDERLINE. `textDecoration` is not a property this toolkit measures ANYWHERE -- it is
 *    in no payload, no count and no check -- so no comparison this build has run or could run
 *    would have reported it, and four passes did not. theme.json sets
 *    `elements.link[":hover"].textDecoration: none` and nothing at rest, so the rest state was
 *    the user-agent's underline.
 *
 * 3. 32px OVER A 44.8px LINE, AND THE `h3` ITSELF STAYS AT 26px. This is the correction that
 *    refuted the earlier build's own note, which had reasoned that the missing size belonged on
 *    the heading and put it at 32px there. It does not: the source's `h3` measures 26px, which
 *    is exactly what this theme's `styles.elements.h3` gives it with no attribute written, and
 *    it is the ANCHOR the source enlarges. No `line-height` is written -- theme.json's 1.4 is
 *    unitless, so it inherits as a ratio and 32 x 1.4 = 44.8 follows on its own, which is also
 *    what makes the heading's box the 45px the contract measures.
 *
 *    BANDED ABOVE 1123.98px, AND ONLY THE 1440 VALUE IS MEASURED. Below that edge this theme's
 *    own media query steps `--wp--preset--font-size--x-large` from 1.625rem to 1.25rem, so the
 *    heading and its anchor render at 20px, and the four root heights close on that: the
 *    contract's 403 at 1024 is 30 + 30 band padding + 30 item margin + 60 item padding + 188
 *    icon column + (28 + 8 + 29) content, exactly. A 32px anchor at 1024 would add 9px to
 *    every narrow viewport and take 403 to 412 and 430 to 441. So the source steps this size
 *    too, this theme's own edge is where it is written, and the narrow value is whatever
 *    `x-large` is -- which is the spelling that needs no second measurement to stay true.
 *
 * 4. THE CHEVRON, which is a pseudo-element and therefore in no payload, no count and no check
 *    either. Its only trace in the measurements is the anchor's own `padding-right`: 22px at
 *    1440 and 18px at the other three, against 0 before this rule.
 *
 *    THE GLYPH AND ITS COLOUR ARE THE ONES THE OTHER UNITS ALREADY DRAW. Same icomoon
 *    codepoint (U+E900, `$icon-arrow`), reproduced from its own path as a MASK with the colour
 *    supplied by `background-color` so it stays on the `primary` token -- the decision and its
 *    reasoning are in assets/css/blocks/button.css, where `section-content`, `centered-text`
 *    and `cta-band` share it, and the font is deliberately not shipped. The trap that file
 *    names applies here unchanged: a rule naming `font-family: icomoon` with no `@font-face`
 *    renders U+E900 from a fallback font, where it is a private-use codepoint with no glyph,
 *    and nothing reports the blank.
 *
 *    IT IS IN THIS FILE AND NOT IN button.css, for two reasons. The host is a heading's anchor,
 *    not a `core/button`, so `core/button`'s stylesheet would not even load on a page that had
 *    only this unit on it. And button.css was locked to another package for the pass that wrote
 *    this. The duplication is one `mask-image` URL and is named in the hand-back: the promotion
 *    button.css's own comments have been asking for -- one unscoped chevron rule owned by
 *    wp-theme-scaffold -- is now wanted by five units across two stylesheets.
 *
 *    CORRECTED 2026-09-16: THE GLYPH IS A FLAT 14x14 AT EVERY VIEWPORT, NOT 18x18 ABOVE
 *    1123.98px, AND IT MUST SIT IN FLOW, NOT `position: absolute`. The contract's own `chrome`
 *    entry did not exist when the paragraph above was written (it says so: "the unit contract
 *    carries no `chrome` key at all"). It exists now and measures the pseudo directly at
 *    1440x900: `box: {w: 14, h: 14}`, `display: inline-block`, `position: relative` -- no
 *    banding, and NOT `position: absolute`. That is not decoration: an absolutely positioned
 *    pseudo is taken out of flow and contributes nothing to its host anchor's own rendered
 *    width, while a `position: relative` / `display: inline-block` one still occupies its
 *    14px in the line before any visual offset is applied. A fresh measurement caught the
 *    consequence of building it the other way: the anchor rendered exactly 14px narrower than
 *    the source at EVERY viewport -- 424px against 438px at 1440, 269px against 283px at
 *    1024/768/390 -- the same 14px gap under two different font-size bands (32px and 20px),
 *    which is the signature of a missing flat contribution, not a text-rendering or
 *    font-metrics difference (those would scale with the font size, not hold flat across a
 *    32-to-20px change). Backed out algebraically at both bands, using each viewport's own
 *    built anchor width and this rule's own unchanged `padding-right`: 1440 is
 *    424 - 22 = 402px of text, and 402 + 14 (the chevron, in flow) + 22 = 438, the measured
 *    source width, exactly; narrow is 269 - 18 = 251px of text, and 251 + 14 + 18 = 283,
 *    again exact. Both bands close on the same +14px with no other number moved, which is
 *    why the fix is confined to how the chevron participates in layout and its own dimensions,
 *    and touches neither `padding-right` nor the anchor's font-size banding in point 3 above,
 *    both of which stay measured-correct as they were.
 *
 *    `vertical-align: middle` REPLACES `top: 50%` / `transform: translateY(-50%)`, which were
 *    only meaningful for an absolutely positioned box centred against its `position: relative`
 *    containing block. An in-flow inline-block box is centred against the line with
 *    `vertical-align` instead; there is no measurement of the source's exact vertical mid-line
 *    (its own centring comes from icomoon glyph metrics this build does not reproduce), so
 *    `middle` is a visual choice, not a transcription, and is confirmed by eye against the
 *    screenshot rather than by a number. Named in the hand-back as the one part of this rule
 *    no measurement backs -- the same caveat this paragraph carried before the correction, now
 *    attached to the vertical alignment rather than to the glyph's size, which the chrome entry
 *    settled.
 */
.wm-link-list .wm-link-list__item-heading a {
	position: relative;
	padding-right: 18px;
	color: var(--wp--preset--color--primary);
	text-decoration: none;
}

.wm-link-list .wm-link-list__item-heading a::after {
	content: "";
	display: inline-block;
	vertical-align: middle;
	width: 14px;
	height: 14px;
	background-color: var(--wp--preset--color--primary);
	mask-image: url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 1024 1024'%3E%3Cpath d='M328.692 1024l-139.062-139.062 379.259-372.938-379.259-372.938 139.062-139.062 505.679 512z'/%3E%3C/svg%3E");
	mask-position: center;
	mask-repeat: no-repeat;
	mask-size: 14px 14px;
}

@media (width > 1123.98px) {
	.wm-link-list .wm-link-list__item-heading a {
		padding-right: 22px;
		font-size: 32px;
	}
}

/*
 * .wm-section-content: THE BAND HEADING'S OWN BOX IS 30px NARROWER THAN THE ROW IT SITS IN,
 * and no block attribute on this theme can write that.
 *
 * WHAT THE CONTRACT MEASURES, quoted from its band-heading node: `div.section-heading-wrapper.row`
 * measures 1170x68 at x=135 and "the `h2.section-heading.field__item` inside it 1140x48 at
 * x=150". The 30px is the inner level of Foundation's grid gutter -- `padding: 0 15px` on the
 * `div.columns.small-12` that WRAPS the `h2` -- so in the source it is the wrapper's padding and
 * the heading's own border box is the 1140.
 *
 * THE ATTRIBUTE THE CONTRACT ASSERTS REPRODUCES THE CONTENT BOX AND NOT THAT BOX. `padding:
 * {left: 15px, right: 15px}` on the heading itself leaves the heading's own box at the 1170 the
 * constrained parent gives it and insets only its text; the node's own `why` checks the 1140 as a
 * CONTENT width and marks it ✓. Measured on the built instance before this rule, at the four
 * captured viewports: 1170x48 at x=135, 994x31 at x=15, 738x31 at x=15, 370x92 at x=10 -- against
 * the contract's 1140/x=150, 964/x=30, 708/x=30, 350/x=20. Eight rows of box difference, w and x
 * at every viewport, from an attribute that paints the text in exactly the right place.
 *
 * AND MARGIN CANNOT SPELL IT EITHER, which is why this is a rule rather than a corrected
 * attribute. A constrained layout writes `margin-left: auto !important; margin-right: auto
 * !important` on every non-aligned child (wp-includes/block-supports/layout.php:583-584, 592-594
 * on this install), and an `!important` author declaration beats a block's inline style -- so
 * `style.spacing.margin.left` on this heading is inert. The 15px padding attribute and its
 * `@mobile` 10px companion are removed from the block in the same pass; leaving them would inset
 * the text a second time.
 *
 * TWO DECLARATIONS, RIGHT AT ALL FOUR VIEWPORTS, because the gutter is subtracted from whatever
 * the row is and the 1170 cap only binds at the top end. `width: calc(100% - 30px)` over the root
 * group's content box -- 1410 at 1440, 994 at 1024, 738 at 768 -- gives 1380, 964 and 708, and
 * `max-width: 1140px` takes the first of those down to the contract's 1140. Core's own
 * `max-width` for the constrained child sits inside a `:where()` at zero specificity, so this
 * 0,2,0 selector wins it without `!important`; the centring is core's own `auto !important`
 * margins, which put the narrower box at x=150, x=30, x=30 and x=20 -- the measured values.
 *
 * THE `@mobile` BAND IS THE GUTTER'S OWN, not a second mechanism: Foundation's
 * `$grid-column-gutter` is 20px below `medium` and 30px from it up, halved per side, and 639.98px
 * is this theme's `settings.viewport.mobile` to the hundredth. 370 - 20 = 350 at x=20 ✓.
 *
 * IT DOES NOT CLOSE THE HEIGHT AT 390, and that is a different defect with a different owner.
 * The built heading is 92px tall there against the contract's 31 because the placeholder string
 * is 42 characters and the measured instance's is short enough to set one line in 350px at 28px.
 * The band heading's text is content, the projection carries none, and no rule here can invent
 * it. Raised in the build hand-back.
 */
.wm-section-content .wm-section-content__band-heading {
	width: calc(100% - 30px);
	max-width: 1140px;
}

@media (max-width: 639.98px) {
	.wm-section-content .wm-section-content__band-heading {
		width: calc(100% - 20px);
	}
}

/*
 * `tall-image-nav-row` -- THE SUBHEADING'S 8px OF SPACE, WHICH EXISTS AT THREE VIEWPORTS AND NOT
 * AT THE FOURTH.
 *
 * The contract's subheading node measures `h3.section-subheading.text-center` at 1170x109 at
 * y=2587 at 1440x900 with the card row starting at y=2696 -- 2587 + 109 = 2696, so there is NO
 * space under it there -- and says of the other three: "At 1024x768, 768x1024 and 390x844 it
 * measures 20px w400 with an 8px bottom margin; at 1440 there is no bottom margin."
 *
 * IT CANNOT BE A BLOCK ATTRIBUTE. A `style.spacing.margin.bottom` is one value at every viewport,
 * and `style["@tablet"]` would band at this theme's 1123.98px edge in the right place but writes a
 * non-important declaration that still has to beat the band's own `blockGap: "0"`, which the
 * contract asserts on the root group. So one media query, scoped to this unit's own hook class.
 *
 * SPECIFICITY, the same reading assets/css/blocks/group.css records for `.wm-raw-html`:
 * `blockGap: "0"` compiles to `.wp-container-core-group-is-layout-<hash> > * { margin-block-start:
 * 0; margin-block-end: 0 }` at 0,1,0, and the two classes below are 0,2,0.
 *
 * THE EDGE IS THE THEME'S OWN `settings.viewport.tablet` (1123.98px), which is also where
 * theme.json's `styles.css` steps `x-large` from 26px to 20px -- the same step the contract
 * measures on this element. So the 8px arrives with the smaller type rather than at an edge
 * nothing supports.
 *
 * 2026-09-23: CHANGED TO 10px, AN OPERATOR DEVIATION -- unlike most of this session's other
 * spacing fixes, the 8px above WAS a real, confirmed measurement (three viewports, not a gap
 * in the record), not an undocumented literal. Standardized anyway: this exact number recurred
 * as an ungrounded literal elsewhere on the same page (section-content.php's h3->p gap, fixed
 * earlier the same day) and the operator's decision was to prioritize one consistent scale
 * over reproducing a source value that happens to sit off it, the same call already made for
 * the band-padding and vertical-rhythm fixes above.
 */
@media (width <= 1123.98px) {

	.wm-tall-image-nav-row .wm-tall-image-nav-row__subheading {
		margin-block-end: var( --wp--preset--spacing--20 );
	}
}

/*
 * .wm-product-teaser-card: THE SUBHEADER'S 18px, FLAT AT EVERY VIEWPORT AND OFF THIS THEME'S
 * LADDER -- the same trap `.wm-image-link-cards__label` above records and the same fix.
 *
 * The contract measures `.teaser-subheader` at 18px/w400 on node-308 ("Ceramic Media", the
 * product this field is actually populated on), which conflicts with a REAL shared token:
 * `styles.elements.h4` gives every h4 on this site `fontSize: large` (21-24px fluid), so an
 * unstyled h4 here would render 21-24px where the source renders a flat 18. `settings.typography.
 * fluid` is `true` on this theme, so a literal `style.typography.fontSize` on the block would be
 * rewritten into a `clamp()` at render rather than surviving as 18px -- the same measured trap
 * `.wm-image-link-cards__label`'s own note above documents. Scoped to this unit's own subheader
 * class rather than the shared `h4` element, which is exactly what a token change would touch.
 *
 * fontWeight and lineHeight are NOT overridden: the measured 400/1.4 already match
 * `styles.elements.h4` exactly, so nothing here would change them.
 */
.wm-product-teaser-card .wm-product-teaser-card__subheader {
	font-size: 18px;
}

/*
 * 16px BELOW 640px -- 2026-09-25 (Direction A, item 5): "both lines 16px" for the teaser copy in its
 * mobile grid cell (assets/css/blocks/group.css, "PRODUCT CARD DIRECTION A -- MOBILE"). The excerpt's own
 * mobile override is the same rule in paragraph.css.
 *
 * TEXT-ALIGN NEEDED A THIRD CLASS, 2026-09-25 (ONE FIX PASS, checkpoint item 2) -- THE "NO `!important`
 * NEEDED" REASONING ABOVE WAS WRONG, MEASURED: at 390 on the instance page this subheader still computed
 * `text-align: center`, not `left`. The comment's own arithmetic (0,2,0) missed that WordPress core's own
 * alignment utility is not a plain one-class rule either: `wp-includes/css/dist/block-library/common.css`
 * ships `:root .has-text-align-center { text-align: center; }` -- `:root` is a pseudo-CLASS, so that
 * selector is ALSO (0,2,0), a tie this selector does not win by specificity, and `common.css` is enqueued
 * after this theme's block stylesheets on this install (confirmed by reading the served page's own
 * `<head>`), so the tie went to core. Adding `.has-text-align-center` to this selector -- the exact class
 * the block emits for its own `textAlign: center` attribute (patterns/product-teaser-card.php) -- reaches
 * (0,3,0), which beats (0,2,0) regardless of source order. VERIFIED AFTER: computes `text-align: left` at
 * 390 on both the instance page and /products.
 */
@media (max-width: 639.98px) {
	.wm-product-teaser-card .wm-product-teaser-card__subheader {
		font-size: 16px;
	}

	.wm-product-teaser-card .wm-product-teaser-card__subheader.has-text-align-center {
		text-align: left;
	}
}

/*
 * Rich-text flow inside the product body (`.wm-product-detail .wp-block-post-content`), 2026-09-21.
 * The source's untyped body headings and paragraphs (no component class) keep Foundation's own defaults, measured on
 * /products/alumina-zirconia/norzon-nv-1620/ at 1440: h2 44px / line-height 61.6px (1.4) with margin-bottom 8px;
 * p margin-bottom 16px. The theme's `h2` element style is 1.1 (measured from the classed `section-heading` h2, which
 * is 48.4px) and the build's blockGap is 0, so the untyped heading rendered 48px tall and touching its paragraph.
 * Scoped to the product body because that is the only place this theme renders unclassed imported headings.
 */
/*
 * `>` (DIRECT CHILD), NOT A PLAIN DESCENDANT COMBINATOR -- fixed 2026-09-25 (site-wide
 * vertical-spacing correction pass). A genuinely unclassed imported body heading (measured
 * on /products/alumina-zirconia/norzon-nv-1620/: `<h2 class="wp-block-heading">Applications
 * </h2>`, no other class) IS a direct child of `.wp-block-post-content` -- raw imported HTML
 * with nothing wrapping it. The plain descendant form below used to also match any heading
 * nested INSIDE another block that happens to live in the same post's content, which on a
 * product page includes `washington-mills/related-products`' own "Products" heading (it is
 * placed in the product's post_content, so it is a grandchild of `.wp-block-post-content`,
 * inside `.wm-related-products`) -- a component heading this rule was never meant to reach,
 * measured giving it an unwanted 20px margin-bottom on top of the band's own 20px
 * post-template padding-top, doubling its gap to 40px against the approved 20
 * (spacing-standard.md §2 "Proximity": the gap below a heading is ONE mechanism, never two).
 */
.wm-product-detail .wp-block-post-content > :is(h2, h3, h4, h5, h6) {
	line-height: 1.4;
	/* Was 8px (the source's own measured value): superseded 2026-09-25, spacing-standard.md §2
	   "Proximity" -- a heading's own gap to its content is spacing|30 (20px) everywhere on the
	   site, not a per-unit re-measurement. This is the one place it has to be a stylesheet rule
	   rather than a block attribute: wp:post-content renders raw imported HTML with no
	   attributes of its own. */
	margin-block: 0 var(--wp--preset--spacing--30);
}

/* The h1 title owns no bottom margin of its own -- the gap to its content
   (washington-mills/page-hero-content) is `.wm-product-detail__text > h1 + *` in group.css,
   spacing|30 (20px). Was 8px (spacing-standard.md §2, superseded 2026-09-25). */
.wm-product-detail__text .wp-block-heading {
	margin-bottom: 0;
}

/*
 * Footer "Quick Links": a FIXED 30px at every width (computed on the live source, 1440 / 1024 / 390). It was written as a
 * `fontSize: 30px` block attribute, which core's fluid typography turns into clamp(18.959px, ..., 30px) -- 27.8px at
 * 1024 and 19.8px at 390. 30px is off this theme's preset ladder, so it is the one CSS declaration.
 */
.wm-footer .wm-footer__heading {
	font-size: 30px;
}

/*
 * .wm-page-hero__title--section (2026-09-21): THE SOURCE'S `.section-heading` TITLE SPEC ON A HERO BAND. Read off the live
 * source with a computed-style walk at 1440x900 and 390x844: /news `h1.section-heading` and /thank-you-0 `h2.section-heading` both
 * compute 57px/62.7px 700 (1.1; 32px/35.2 at 390) with margin 0 0 20px, where the band's default `node-title` h1 is 1.0 at the wide
 * step and carries a 30px margin. Used on block-photo-left (news) and block-image-band (vibratory-finishing).
 * The cover-borne selector is 0,4,1 because cover.css's `.wm-page-hero.wp-block-cover .wm-page-hero__col h1` (0,3,1) sets the
 * 30px margin, and the leading selector is 0,2,1 because the wide-step `.wm-page-hero .wp-block-heading` (0,2,0) sets 1.
 */
.wm-page-hero h1.wm-page-hero__title--section {
	line-height: 1.1;
	margin-bottom: var(--wp--preset--spacing--30);
}

.wm-page-hero.wp-block-cover .wm-page-hero__col h1.wm-page-hero__title--section {
	margin-bottom: var(--wp--preset--spacing--30);
}

/*
 * .wm-page-hero__title--marker: THE BLUE DOUBLE-BAR BEFORE THE TITLE, the same `header-graphic.svg` marker and the same box as
 * the section-content one above (50px wide, `left: -70px`, the title's own height), measured on the source's /news title at
 * 1440x900. It is `display: none` on the source's /thank-you-0 h2, so it is opt-in by a second class. It sits over the photo
 * column, outside the title's box, which no block attribute can place; a pseudo-element it stays.
 */
.wm-page-hero__title--marker {
	position: relative;
}

.wm-page-hero__title--marker::before {
	content: "";
	position: absolute;
	top: 0;
	left: -70px;
	display: inline-block;
	width: 50px;
	height: 100%;
	background-image: url("../../images/header-graphic.svg");
	background-repeat: repeat;
}
