/**
 * SITE FOOTER — the things parts/footer.html cannot express in block markup.
 *
 * Design source: Figma 14474:2190 "Footer - Desktop" (measured from
 * .claude/runs/extracts/component--FooterDesktop2.js), plus the tablet
 * (768x923) and mobile (390x1842) footers inside TabletViews / MobileViews.
 * Every number below is a literal from those style objects, not an estimate.
 *
 * HOW THIS LOADS: functions.php already registers it —
 *   $framework_skin( 'core/template-part', 'framework-footer', 'assets/css/footer.css' )
 * — i.e. wp_enqueue_block_style() with `path` data, so the inliner embeds it
 * in <head> (verified: the served homepage carries
 * <style id="framework-separator-inline-css"> from the identical mechanism,
 * and now framework-footer-inline-css beside it). It therefore loads for ANY
 * template part on the page, header included, which is why every rule below
 * is scoped under `.site-footer`.
 *
 * NOT LOADED IN THE SITE-EDITOR CANVAS. wp_enqueue_block_style() only hooks
 * render_block / wp_enqueue_scripts; it does not add the handle to the block
 * type's style_handles, so the editor never sees it (same gap as
 * separator.css / image.css / spacer.css). Design values that MUST survive in
 * the canvas — heading weight/size, the lead line, column widths — are block
 * attributes in parts/footer.html instead of rules here. See the report's
 * open questions for the one-line add_editor_style() that would close it.
 *
 * SPECIFICITY NOTE: nothing here uses !important except where core ships its
 * own !important (column flex-basis, columns flex-wrap) — each such case is
 * commented at the rule.
 */

/* -------------------------------------------------------------------------
 * 0. THE `has-link-color` COLLISION  (measured, not theorised)
 *
 * theme.json's palette contains a slug named `link` (#002056). WordPress
 * emits a text-colour class for every palette slug, so global styles ship
 *
 *   .has-link-color { color: var(--wp--preset--color--link) !important; }
 *
 * — and `has-link-color` is ALSO the marker class core stamps on any block
 * that sets style.elements.link.color, which the footer band does (its links
 * ride the on-dark cream). The two meanings collide: the band inherited
 * #002056 to every descendant that did not declare its own colour. Measured
 * in the browser: getComputedStyle(.site-footer).color === rgb(0, 32, 86) on
 * a rgb(19, 34, 48) band = 1.05:1. It was invisible before only because the
 * previous placeholder pinned an inline colour on every single paragraph.
 *
 * The part no longer sets elements.link (see §5), so the class is gone from
 * today's markup and this rule is DEFENSIVE: it fires the moment anyone picks
 * a link colour for this band in the site editor, which would otherwise paint
 * the whole footer navy again. !important is not a preference here — the
 * colliding rule carries one, so only a more specific !important answers it.
 * (0,2,0) over (0,1,0); inert while the class is absent.
 *
 * THE REAL FIX IS IN theme.json (not this agent's file): rename the palette
 * slug `link` to something that cannot collide with a core element-support
 * marker (`link`, `text`, `background`, `border` are all reserved in
 * practice) and sweep the var() references. Every block on the site that
 * sets elements.link.color is carrying this landmine, not just the footer.
 * ---------------------------------------------------------------------- */

.site-footer.has-link-color {
	color: var(--wp--preset--color--white) !important;
}

/*
 * CANVAS BELT (2026-09-02). Editing the footer PART, the editor stamps
 * `wp-elements-N` on the band instead of `has-link-color`, so the disarm
 * above never matches there and the nav links computed ink-on-ink
 * (measured in the site editor). Editor-only by construction —
 * `.editor-styles-wrapper` never exists on the front — and `!important`
 * for the same reason the disarm itself carries it: the editor's own
 * elements stylesheet is what it has to beat.
 */
.editor-styles-wrapper .site-footer a,
.editor-styles-wrapper .site-footer .wp-block-navigation-item__content {
	color: var(--wp--preset--color--white) !important;
}

/* -------------------------------------------------------------------------
 * 1. LOGO
 *
 * core/site-logo's `width` attribute emits NO CSS — it only filters the
 * <img> width/height attributes (CLAUDE.md §Layout gotchas), so the 307 has
 * to come from a rule here. Note `site-footer__logo` sits on the SAME
 * element as `wp-block-site-logo`: a descendant selector reaches the <img>,
 * never the block wrapper.
 * Design lockup: 307 x 46 at every breakpoint (desktop, tablet AND mobile
 * all draw the same 271px wordmark + 29px shield at x=278).
 *
 * WHY THE 307 IS ON THE <img> AND NOT ON THE WRAPPER  (2026-08-23, client:
 * "el logo del footer debe ir a la izquierda. ahora esta centrado.")
 *
 * The wrapper is a constrained-layout child of the footer band, and core
 * ships auto inline margins for those. MEASURED with
 * .claude/runs/qa-round/probe-logo.js on /resources/ at 1440 — two identical
 * rules match, from global-styles-inline-css and core-block-supports-inline-css:
 *
 *   .is-layout-constrained > :where(:not(.alignleft):not(.alignright):not(.alignfull)),
 *   .wp-container-core-group-is-layout-5bc57418 > :where(:not(.alignleft)…)
 *   { max-width:1320px; margin-left:auto!important; margin-right:auto!important }
 *
 * Real specificity (0,1,0) — :where() contributes nothing — but !important.
 * Giving the WRAPPER `width:307px` therefore handed those auto margins a
 * 307px box to centre: measured margin-left/right 506.5px = (1320-307)/2,
 * logo box x=567 in a band whose content box starts at x=60.
 *
 * The comp draws it FLUSH LEFT at every artboard (measured, not estimated —
 * .claude/runs/qa-round/comp-footer-logo.js + map-resources/probe-region.js):
 *   Resources.dc.html  band x0 y3222 1440x646, pad "50px 60px 30px" -> lockup x=60,  307x46
 *   Tablet Views.dc.html  footer band 768x892.7, pad "50px 24px 30px" -> lockup x=24,  307x46
 *   Mobile Views.dc.html  footer band 390x1797.7, pad "50px 16px 30px" -> lockup x=16, 307x46
 * i.e. always ON the band's content-box left edge, level with the column row
 * and the legal row.
 *
 * So the wrapper is left ALONE and core's constrained layout is allowed to do
 * its job: it gets the same 1320px content box every other footer child gets
 * (x=60 at 1440, x=300 at 1920 where the 1320 is centred inside the 1800
 * padding box, x=16 at 390), and the lockup lands on its left edge because
 * the wrapper is text-align:start with core's `.wp-block-site-logo a
 * {display:inline-block}` inside. Overriding the auto margins with a more
 * specific !important was the obvious fix and is the WRONG one: it pins the
 * logo to the band's 60px padding, which at 1920 is 240px left of every other
 * footer element.
 *
 * Only `width` is declared: core's own block stylesheet already ships
 * `.wp-block-site-logo img{height:auto;max-width:100%}`, which supplies the
 * aspect ratio and the shrink cap for viewports narrower than 307 + 2 gutters.
 * ---------------------------------------------------------------------- */

.site-footer .site-footer__logo img {
	width: 307px;
}

/* -------------------------------------------------------------------------
 * 2. COLUMN ROW GEOMETRY
 *
 * The design row is 306 / gap 145 / 193 / gap 32 / 306 / gap 32 / 306, which
 * sums to EXACTLY the 1320px content box. core/columns has ONE blockGap, so
 * the part sets it to the 32px that serves three of the four positions and
 * the extra 113px of the first gap is added here as a margin (32 + 113 = 145).
 *
 * Why not nested flex groups: core/column's `width` attribute is the only
 * width that SERIALISES into the saved markup (`style="flex-basis:306px"`),
 * which is what keeps the widths right in the site editor too. A nested
 * core/group would have needed its width from this file, invisible in the
 * canvas. One margin rule buys back the whole design structure.
 *
 * No flex-wrap work is needed at desktop: core ships
 * `@media (min-width:782px){.wp-block-columns{flex-wrap:nowrap!important}}`,
 * and with the tracks summing to exactly the container that nowrap is what
 * prevents the sub-pixel wrap CLAUDE.md warns about.
 *
 * WHY THE TRACKS ARE PROPORTIONAL AND NOT THE DESIGN'S LITERAL PIXELS.
 * The container is min(1320, viewport - 2*gutter), and the gutter clamp only
 * reaches its 60px maximum at 1440 — so between 1024 (where the tablet arm
 * takes over) and 1440 the row has LESS than 1320 to work with. Left as fixed
 * pixels, flexbox shrinks the four tracks proportionally while the 145px gap
 * stays rigid: measured at 1024 that gave 204 / 145 / 128 / 204 / 204, i.e.
 * a first gap wider than half the nav column. Expressing every track AND the
 * extra first-gap margin as a share of the free space keeps the whole row in
 * the design's proportions at every width, and is EXACT at 1440 because
 * 1224 = 1320 - 3*32 divides by 306 four times over:
 *
 *   306/1224 = 0.25        193/1224 = 0.1576797    113/1224 = 0.0923203
 *   3*0.25 + 0.1576797 + 0.0923203 = 1.0000000
 *
 * !important is forced: core/column serialises its width as an INLINE
 * `style="flex-basis:306px"`, and nothing but !important outranks inline.
 * Keeping the attribute anyway is deliberate — the site editor does not load
 * this file, and 306px is what the design wants the canvas to show.
 * ---------------------------------------------------------------------- */

.site-footer .site-footer__col--mission,
.site-footer .site-footer__col--contact,
.site-footer .site-footer__col--newsletter {
	flex-basis: calc((100% - 96px) * 0.25) !important;
}

.site-footer .site-footer__col--nav {
	flex-basis: calc((100% - 96px) * 0.1576797) !important;
}

/* 32px of blockGap + this = the design's single 145px gap. */
.site-footer .site-footer__col--mission {
	margin-right: calc((100% - 96px) * 0.0923203);
}

/* -------------------------------------------------------------------------
 * 3. MISSION COPY + SOCIAL CHIPS
 *
 * Chip: 38 x 38 cream disc, 20px navy glyph, 10px apart. No core preset
 * produces 38px — core's sizes are 16/24/36/48px glyphs at `padding: .25em`
 * — so the geometry is font-size 20px on the list (core sizes the SVG at
 * 1em) plus 9px of anchor padding: 20 + 2*9 = 38.
 *
 * Colours are painted here rather than through the block's
 * iconColor/iconBackgroundColor attributes on purpose: core/social-links is a
 * STATIC block, so those attributes only reach the front end if the saved
 * <ul> markup is hand-forged to match what save() would write, and a mismatch
 * invalidates the block in the editor. The chip is one uniform skin for all
 * five services (not per-item editorial colour), so it belongs with the
 * geometry it cannot be separated from anyway.
 *
 * Contrast: #002056 on #FBF9F8 = 14.9:1; the cream disc on #132230 = 15.4:1.
 * ---------------------------------------------------------------------- */

.site-footer .site-footer__social {
	font-size: 20px;
}

/* (0,3,0) — beats core's per-service brand fills at (0,1,0)
   (`:where(.wp-block-social-links:not(.is-style-logos-only)) .wp-social-link`). */
.site-footer .site-footer__social .wp-block-social-link {
	background-color: var(--wp--preset--color--white);
	color: var(--wp--preset--color--link);
}

/* (0,4,0) — beats core's `:root :where(… .wp-social-link a){padding:.25em}`. */
.site-footer .site-footer__social .wp-block-social-link .wp-block-social-link-anchor {
	padding: 9px;
}

/*
 * HOVER. Core lifts social chips (`transform: scale(1.1)`); CLAUDE.md rules
 * hover a colour swap ONLY — no lift, no shadow. (0,4,0) beats core's (0,2,0).
 *
 * The comp this file was built from drew no hover state, so an earlier pass
 * only cancelled core's lift and left the chip inert. The client supplied the
 * missing state on 2026-08-20: the cream disc becomes GOLD-LIGHT and the glyph
 * does not move or change colour. That is the same on-dark hover colour the
 * button system uses (`is-style-inverse` → gold-light), so the footer is not
 * inventing a one-off here.
 *
 * Measured contrast, so the swap is not a downgrade:
 *   glyph #002056 on gold-light #EDCE91 = 10.3:1  (rest was 14.9:1 on cream)
 *   gold-light disc on the ink band #132230 = 10.7:1
 * Both clear AA for non-text UI (3:1) and for text (4.5:1) with room to spare.
 *
 * Not gated behind `@media (hover: hover)`, deliberately and consistently with
 * assets/css/buttons.css: a tap latches :hover either way, and gating only some
 * of the theme's hovers would leave the site half-applied on touch. A latched
 * gold disc is cosmetic; a half-applied system is a bug.
 */
.site-footer .site-footer__social .wp-block-social-link:hover {
	transform: none;
	background-color: var( --wp--preset--color--gold-light );
}

/*
 * FORWARD GUARD. functions.php's framework_social_icon() swaps core's glyph
 * for assets/icons/social/<service>.svg when that file exists, and those
 * design lockups draw their OWN plate — a plate under a plate is a circle
 * inside a circle, which is exactly what the function's docblock says this
 * file has to prevent. The folder does not exist yet (core's glyphs render
 * today), so the plate above is live and this turns it off the moment the
 * assets land. :has() takes the specificity of its argument, so these are
 * (0,4,0)/(0,5,0) and win over the two rules above.
 */
.site-footer .site-footer__social .wp-block-social-link:has(.saconnects-social-icon) {
	background-color: transparent;
}

.site-footer .site-footer__social .wp-block-social-link:has(.saconnects-social-icon) .wp-block-social-link-anchor {
	padding: 0;
}

/* -------------------------------------------------------------------------
 * 4. NAVIGATION COLUMN
 *
 * The design row is padding-free — 16/600/1.4 ink with the 15px blockGap
 * carrying the whole rhythm.
 *
 * Core does ship `.wp-block-navigation-item > .wp-block-navigation-item__content
 * {display:flex;flex-grow:1;padding:.5em 1em}`, but a CONTROL TEST (delete the
 * rule below through the CSSOM, re-measure, put it back) says it does not
 * reach a VERTICAL nav: padding stayed 0px and the anchor stayed 45.1 x 22.4
 * either way. So this is a GUARANTEE, not an override — it holds the design's
 * zero if the orientation is ever flipped or core's selector widens. Do not
 * read it as evidence that core is padding these rows today.
 * ---------------------------------------------------------------------- */

.site-footer .site-footer__nav .wp-block-navigation-item > .wp-block-navigation-item__content {
	padding: 0;
}

/*
 * WCAG 2.2 SC 2.5.8. With the padding gone the ink is 16 x 1.4 = 22.4px tall,
 * under the 24px minimum. ONE absolutely-positioned pad restores the target
 * without moving a pixel of ink (the carousel-dot technique). It cannot reach
 * a neighbour: row centres are 22.4 + 15 = 37.4px apart and each pad extends
 * 12px from its centre.
 */
.site-footer .site-footer__nav .wp-block-navigation-item__content {
	position: relative;
}

.site-footer .site-footer__nav .wp-block-navigation-item__content::before {
	content: "";
	position: absolute;
	top: 50%;
	right: 0;
	left: 0;
	height: 24px;
	transform: translateY(-50%);
}

/*
 * The external-link mark on "Donate" and "CRD".
 *
 * The design glyph is a STROKED 12x12 (path M1.5 10.5 L10.5 1.5 / M3.5 1.5
 * L10.5 1.5 L10.5 8.5, stroke-width 1.8, round caps). The theme owns no such
 * asset — assets/icons/arrow-12-diagonal.svg is a FILLED mark from a
 * different Design System node and would ship a visibly heavier arrow — and
 * core/navigation-link has no icon support, so `is-external` was dead markup
 * (zero consumers anywhere in the theme before this file).
 *
 * Drawn as a mask so the fill is currentColor and the glyph tracks the link
 * colour in every state. Decorative by construction: a pseudo-element is
 * invisible to assistive tech, and the new tab is already announced by
 * core's target/rel.
 */
.site-footer .site-footer__nav .is-external .wp-block-navigation-item__content::after {
	content: "";
	/*
	 * inline-block, NOT the default inline. In a VERTICAL footer nav core's
	 * `.wp-block-navigation-item > …__content{display:flex}` does not apply and
	 * the anchor stays `display:block`, so an inline pseudo-element ignores
	 * width/height entirely: measured, the glyph contributed 0px to the anchor
	 * (Donate = 52.9px label + 5px margin + 0) and painted nothing at all,
	 * while getComputedStyle still reported "12px". inline-block sizes in a
	 * block context and blockifies harmlessly if the anchor ever is a flex
	 * container. flex:none keeps it from shrinking in that case.
	 */
	display: inline-block;
	vertical-align: middle;
	flex: none;
	width: 12px;
	height: 12px;
	margin-left: 5px;
	background-color: currentColor;
	/* LONGHANDS, not the `mask` shorthand. The shorthand form
	   `url() center / contain no-repeat` is valid CSS, but the Nu validator
	   rejects it ("Too many values") — its mask grammar predates the
	   position/size syntax. Longhands render identically and validate. */
	-webkit-mask-image: url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 12 12' fill='none' stroke='black' stroke-width='1.8' stroke-linecap='round' stroke-linejoin='round'%3E%3Cpath d='M1.5 10.5L10.5 1.5M3.5 1.5L10.5 1.5L10.5 8.5'/%3E%3C/svg%3E");
	mask-image: url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 12 12' fill='none' stroke='black' stroke-width='1.8' stroke-linecap='round' stroke-linejoin='round'%3E%3Cpath d='M1.5 10.5L10.5 1.5M3.5 1.5L10.5 1.5L10.5 8.5'/%3E%3C/svg%3E");
	-webkit-mask-position: center;
	mask-position: center;
	-webkit-mask-size: contain;
	mask-size: contain;
	-webkit-mask-repeat: no-repeat;
	mask-repeat: no-repeat;
}

/* -------------------------------------------------------------------------
 * 5. TEXT LINKS INSIDE FOOTER PARAGRAPHS
 *
 * The design draws the phone number, the email and the legal row as PLAIN
 * TEXT. The part links them (better UX, and the recorded convention), which
 * makes them links whose colour is IDENTICAL to the copy around them —
 * no cue at all, not even the colour cue 1.4.1 already rejects. The theme's
 * paragraph-link grammar (block-conventions.md §Consolidation rulings) is a
 * from-font underline, so that is what they get. Nav links are exempt: a
 * <nav> list conveys interactivity by grouping, and the design shows none.
 *
 * padding-block: 3px is ASRI sign-off D4, carried over from the
 * `.footer-legal p a, .footer-address a` rule in navigation.css that this
 * part no longer matches (see the report). Free: vertical padding on an
 * INLINE box paints and hit-tests but adds nothing to line-box height.
 */

/*
 * LINK COLOUR — the design paints every footer link #FBF9F8 SOLID (15.4:1 on
 * the band; the whole footer has no low-alpha text value anywhere).
 *
 * This lived on the band as `style.elements.link.color.text` until it was
 * measured. The value `var:custom|on-dark|link` emits NO RULE AT ALL:
 * WordPress's `var:` shorthand resolves PRESETS (`var:preset|color|white`)
 * and has no `var:custom|…` resolver, so the attribute produced the classes
 * and not one line of CSS — verified in the browser, zero `.wp-elements-*`
 * rules in any served stylesheet — and every footer link fell through to
 * global styles' `a:where(:not(.wp-element-button)){color:…--link}` =
 * #002056 on #132230 = 1.05:1. navigation.css's own footer rules were
 * masking it on the nav column only.
 *
 * Painting it here instead of restoring a working attribute value also drops
 * the `has-link-color` class core stamps beside elements.link, which is what
 * keeps §0's collision out of the SITE-EDITOR CANVAS (global styles load
 * there; this file does not).
 */
.site-footer a:where(:not(.wp-element-button)) {
	color: var(--wp--preset--color--white);
}

/*
 * UNDERLINE IS A HOVER STATE, NOT A RESTING ONE — client, 2026-08-26:
 * "segun el diseño en el footer los links no tienen underline. aplica el
 * underline al hover."
 *
 * `padding-block: 3px` is unrelated and stays (ASRI sign-off D4, the +3px
 * target height on inline footer links — see §5's note).
 *
 * WHAT THIS COSTS, recorded so the next reader does not "fix" it back:
 * CLAUDE.md's contrast rule asks for a PERMANENT underline on links in
 * running copy whose colour matches the text around them, because colour
 * alone then carries the whole affordance (WCAG 1.4.1). Two footer links are
 * exactly that case — "…New York 10994 | 1-800-SAL-ARMY" and "Nonprofit
 * Website Design by Elevation" — both white on white beside plain white copy.
 * The design is the client's call and it is the newer instruction, so it
 * wins; the exposure is theirs to weigh, and it is flagged in the QA notes,
 * not silently absorbed. Every OTHER footer link is either alone in its
 * paragraph or inside the nav list, where grouping is the affordance.
 */
.site-footer p a {
	padding-block: 3px;
	text-decoration: none;
}

/*
 * The hover the client asked for, on the two link families that carry TEXT:
 * paragraph links and the navigation columns. The site logo and the social
 * row are anchors too, but they wrap an image and an SVG — an underline on
 * either draws nothing and only muddies the rule.
 *
 * `:is()` takes the weight of its heaviest argument, so this lands at
 * (0,3,1) — above core's own `text-decoration: none` on nav item content.
 * NOT mirrored on `:focus-visible`: per CLAUDE.md a focus state is never a
 * hover clone; the keyboard affordance is the global focus ring.
 */
.site-footer :is(p, .wp-block-navigation-item) a:hover {
	text-decoration: underline;
	text-decoration-skip-ink: none;
	text-underline-position: from-font;
}

/* -------------------------------------------------------------------------
 * 6. "STAY CONNECTED" MARKER — THE CREST
 *
 * This shipped as a 9px accent CIRCLE, and the open question recorded here
 * (report Q6) was whether a plain dot belonged on a site whose every other
 * section marker is the Salvation Army crest. The client answered it on
 * 2026-11-03: "podemos ajustar el bullet del footer y cambiarlo por el
 * shield?" — so it is the crest, and the footer stops being the one place
 * that marks a heading with a shape found nowhere else.
 *
 * ONE DRAWING. The path is assets/icons/shield-22.svg's, byte-identical to
 * the one framework_eyebrow_crest() inlines on fourteen blocks. It cannot
 * drift into a second crest, because there is no second crest to maintain.
 *
 * THE ARTWORK, NOT THE SLOT. That file is a 22x30 icon slot holding
 * 17.334 x 19.336 of artwork at left 2.332 / top 5; the padding exists to
 * reproduce the eyebrow's COLUMN spacing, which this heading already has in
 * its own 10px margin. Taking the artwork alone (the viewBox the asset's own
 * comment names) paints the crest at exactly the size it appears everywhere
 * else — measured on /sld/: a 22x30 box, 10px from a 16px/22.4px label, the
 * same label metrics this heading has — while fitting inside the 22.4px line
 * box, so the footer row does not grow. A full 30px slot would have pushed it.
 *
 * A DATA URI, not url('../icons/shield-22.svg'). This stylesheet is enqueued
 * with `path` data and is INLINED into <head>, and a relative url() resolves
 * against the document once that happens, not against the file — it would
 * 404 in production and nowhere else (CLAUDE.md §Performance playbook). Same
 * reason, same longhand form, as the external-link arrow above.
 *
 * Inline-block rather than a flex row on purpose: the heading holds a text
 * run, and CLAUDE.md rules out display:flex on containers whose children can
 * come from the editor (flex blockifies loose text).
 * ---------------------------------------------------------------------- */

.site-footer .site-footer__heading--marker::before {
	content: "";
	display: inline-block;
	width: 17.334px;
	height: 19.336px;
	margin-right: 10px;
	vertical-align: middle;
	background-color: var(--wp--preset--color--accent);
	/* LONGHANDS, not the `mask` shorthand — see the arrow above for why. */
	-webkit-mask-image: url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 17.334 19.336' fill='none'%3E%3Cpath fill='black' fill-rule='nonzero' d='M 8.667 19.336 C 9.204 18.814 10.416 17.926 11.578 17.586 C 12.907 17.196 16.348 16.594 17.151 13.177 C 17.781 10.494 16.589 7.478 16.452 6.128 C 16.331 4.934 16.702 2.227 16.762 1.813 C 16.841 1.265 16.306 1.054 15.95 1.186 C 15.36 1.403 14.092 1.793 12.641 1.819 C 11.082 1.847 9.158 0.737 8.667 0 C 8.174 0.739 6.249 1.847 4.692 1.819 C 3.242 1.793 1.974 1.401 1.384 1.186 C 1.028 1.056 0.49 1.265 0.572 1.813 C 0.632 2.225 1.003 4.932 0.882 6.128 C 0.744 7.48 -0.448 10.494 0.183 13.177 C 0.986 16.594 4.427 17.196 5.756 17.586 C 6.915 17.926 8.129 18.814 8.667 19.336 Z'/%3E%3C/svg%3E");
	mask-image: url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 17.334 19.336' fill='none'%3E%3Cpath fill='black' fill-rule='nonzero' d='M 8.667 19.336 C 9.204 18.814 10.416 17.926 11.578 17.586 C 12.907 17.196 16.348 16.594 17.151 13.177 C 17.781 10.494 16.589 7.478 16.452 6.128 C 16.331 4.934 16.702 2.227 16.762 1.813 C 16.841 1.265 16.306 1.054 15.95 1.186 C 15.36 1.403 14.092 1.793 12.641 1.819 C 11.082 1.847 9.158 0.737 8.667 0 C 8.174 0.739 6.249 1.847 4.692 1.819 C 3.242 1.793 1.974 1.401 1.384 1.186 C 1.028 1.056 0.49 1.265 0.572 1.813 C 0.632 2.225 1.003 4.932 0.882 6.128 C 0.744 7.48 -0.448 10.494 0.183 13.177 C 0.986 16.594 4.427 17.196 5.756 17.586 C 6.915 17.926 8.129 18.814 8.667 19.336 Z'/%3E%3C/svg%3E");
	-webkit-mask-position: center;
	mask-position: center;
	-webkit-mask-size: contain;
	mask-size: contain;
	-webkit-mask-repeat: no-repeat;
	mask-repeat: no-repeat;
}

/* -------------------------------------------------------------------------
 * 7. NEWSLETTER SURFACE CONTRACT
 *
 * The form itself is NOT built here — it needs a saconnects/newsletter-form
 * block (block-conventions.md §Forms) plus a Gravity Form, neither of which
 * is in this agent's file set. See the report.
 *
 * These are the custom properties that block's skin
 * (_shared/_gravity-forms.scss `engine()` + `on-dark`) reads, pre-declared on
 * the column so the form inherits the measured design surface the day it
 * lands. The field ring is deliberately NOT the mockup's
 * rgba(152,151,153,.6): over #132230 that composites to rgb(99,104,111) =
 * 2.81:1 and fails WCAG 1.4.11's 3:1 floor for a control boundary, and the
 * ring is the ONLY thing delimiting the input. --wp--custom--on-dark--input-bounds
 * (rgba(251,249,248,.55)) composites to rgb(147,152,158) = 5.43:1.
 * ---------------------------------------------------------------------- */

.site-footer .site-footer__col--newsletter {
	--saconnects-gf-input-bg: transparent;
	--saconnects-gf-input-ink: var(--wp--preset--color--white);
	--saconnects-gf-input-bounds: var(--wp--custom--on-dark--input-bounds);
	--saconnects-gf-body: var(--wp--preset--color--white);
	--saconnects-gf-label: var(--wp--preset--color--white);
	--saconnects-gf-btn-bg: var(--wp--preset--color--white);
	--saconnects-gf-btn-ink: var(--wp--preset--color--contrast);
	--saconnects-gf-btn-bg-hover: var(--wp--preset--color--gold-light);
}

/* -------------------------------------------------------------------------
 * 8. LEGAL ROW
 *
 * Desktop: one space-between row, left group gapped 29px, right group 26px
 * (the design really does use two different gaps). Both groups get their gap
 * from block attributes; only the cross-axis alignment lives here.
 * ---------------------------------------------------------------------- */

.site-footer .site-footer__legal {
	align-items: center;
}

/*
 * The rule above the legal row: 1px, rgba(152,151,153,.3), full 1320 content
 * box. The colour and the 20px of air either side come from the part (block
 * attribute + the wrapping group's blockGap); the THICKNESS has to come from
 * here because core ships two competing declarations —
 * `.wp-block-separator{border:none;border-top:2px solid}` in the block's
 * style.css and `border:none;border-bottom:2px solid` in its theme.css — and
 * which one wins depends on the order those two stylesheets happen to be
 * enqueued. Measured before this rule: a 2px box where the design draws 1px.
 * (0,3,0) settles it either way.
 *
 * Deliberately NOT brightened. Composited over #132230 the design's hairline
 * is rgb(59,69,80) = 1.75:1 — that is a whisper on purpose, it is decorative,
 * and WCAG 1.4.11 does not reach it (it delimits nothing and conveys nothing).
 * The previous placeholder used the cream on-dark border at 4.85:1, roughly
 * three times as bright as the design.
 */
.site-footer .wp-block-separator.has-background {
	border: 0;
	height: 1px;
}

/* =========================================================================
 * 9. TABLET  (<= 1023px)
 *
 * Measured from the 768x923 footer at TabletViews offset 61561 (an UNNAMED
 * frame — the bundle carries no layer names; it was found structurally by
 * its rgb(19,34,48) container and the mission string).
 *
 * The mission + social band goes full width above a three-column row of
 * 162 / 24 / 224 / 24 / 286 = 720 = the 768 content box. Those tracks are
 * expressed as fractions of (100% - 48px) so the row keeps the design's
 * proportions across the whole 782-1023 band, not just at 768:
 *   162/672 = 24.107%   224/672 = 33.333%   286/672 = 42.560%
 *
 * The container gap drops 80 -> 40 and the band gap is 70.
 * ====================================================================== */

@media (max-width: 1023px) {

	/* Doubled class = (0,2,0), over core's per-container blockGap rule at
	   (0,1,0) (`.wp-container-…-is-layout-… > * + *`). */
	.site-footer.site-footer > * + * {
		margin-block-start: 40px;
	}

	/* Two competing core !importants have to be answered here: `flex-wrap:
	   nowrap!important` above 782px, and `flex-wrap:wrap!important` below it.
	   (0,3,0) clears both. */
	.site-footer .wp-block-columns.site-footer__columns {
		flex-wrap: wrap !important;
		row-gap: 70px;
		column-gap: 24px;
	}

	/*
	 * (0,4,0) — and it has to be. Core stacks columns with
	 * `@media (max-width:781px){.wp-block-columns:not(.is-not-stacked-on-mobile)
	 * > .wp-block-column{flex-basis:100%!important}}`, which is (0,3,0) AND
	 * !important, so the 601-781 half of this arm would otherwise lose and a
	 * 768 tablet — the width the tablet artboard is drawn at — would render
	 * the mobile stack. Measured before this fix: at 768 all four columns came
	 * back 705px wide. The !important is for the inline flex-basis (see §2).
	 */
	.site-footer .site-footer__columns > .wp-block-column.site-footer__col--mission {
		flex-basis: 100% !important;
		margin-right: 0;
	}

	.site-footer .site-footer__columns > .wp-block-column.site-footer__col--nav {
		flex-basis: calc((100% - 48px) * 0.2410714) !important;
	}

	.site-footer .site-footer__columns > .wp-block-column.site-footer__col--contact {
		flex-basis: calc((100% - 48px) * 0.3333333) !important;
	}

	.site-footer .site-footer__columns > .wp-block-column.site-footer__col--newsletter {
		flex-basis: calc((100% - 48px) * 0.4255953) !important;
	}

	/* Mission -> social drops 70 -> 40. (0,2,0) over the column's own
	   blockGap rule at (0,1,0). */
	.site-footer .site-footer__col--mission > * + * {
		margin-block-start: 40px;
	}

	/* A measure cap, not a column width: the tablet mission span is 596 wide
	   inside a 720 content box. */
	.site-footer .site-footer__mission {
		max-width: 596px;
	}

	/* The legal row restructures — it does not merely wrap. Two centred rows,
	   15px apart; each row keeps its own 29px / 26px gap. */
	.site-footer .site-footer__legal {
		flex-direction: column;
		gap: 15px;
	}
}

/* =========================================================================
 * 10. MOBILE  (<= 600px)
 *
 * Measured from the 390x1842 footer at MobileViews offset 60440. One column,
 * 358px wide (390 - 2*16, which is exactly what --wp--custom--gutter clamps
 * to at 390 — nothing to restate). Container gap 40, inter-column gap 50,
 * mission -> social 40, and the legal block fully stacks and centres.
 *
 * 600 is the switch, NOT core's 781. The design's tablet artboard is drawn
 * at 768, which is BELOW 781 — stacking there would have thrown the whole
 * tablet layout away on the exact width it was measured at (verified: at 768
 * every column rendered 705px wide before this was corrected). 600 is the
 * "single column, gutter 16, mobile type scale" tier block-conventions.md
 * §Responsive already defines, and it sits between the two artboards the
 * design actually supplies (768 and 390). The wide arm of this pair is the
 * ungated block above — never a min-width.
 * ====================================================================== */

@media (max-width: 600px) {

	/* (0,3,0) to match the tablet arm's own row-gap rule — at (0,2,0) this
	   silently lost and mobile kept the tablet's 70px (measured at 375). */
	.site-footer .wp-block-columns.site-footer__columns {
		row-gap: 50px;
	}

	/* Restates the stack at the tablet arm's own (0,4,0) — those rules also
	   match down here, and would otherwise hold the three-column row. Later in
	   the file, so the tie breaks this way. */
	.site-footer .site-footer__columns > .wp-block-column.site-footer__col {
		flex-basis: 100% !important;
	}

	.site-footer .site-footer__legal {
		gap: 40px;
	}

	.site-footer .site-footer__legal-group {
		flex-direction: column;
		align-items: center;
		gap: 10px;
	}
}
