/*
 * PE-CW-01 -03c — the `.reveal` CSS workaround that used to live at the top of this file is
 * GONE, and this note is what replaced it.
 *
 * -03 could not ship client JavaScript, so it made `.reveal`'s revealed state the static
 * default in CSS (RETURN-03.md decision D-03-1) — 47 elements on this route carry `.reveal`
 * and lxf-page-home.css starts them at `opacity:0;transform:translateY(24px)`, so without the
 * IntersectionObserver the page rendered almost entirely invisible.
 *
 * -03c carries the real observer as admitted island `lxf-page-home`, so the workaround is
 * deleted: THE LIVE BEHAVIOUR WINS. The revealed state is now reached the same way on both
 * sides, by the same code, and `.reveal`/`.reveal.visible` are governed only by the vendored
 * lxf-page-home.css exactly as on the live page.
 *
 * The cost is honest and must not be read around: the candidate is now subject to the same
 * settle flakiness as the live page. RETURN-03b recorded that the candidate settled 47/47
 * every time *because* this rule made it static, and that the live page sometimes settles
 * 41/47 or 45/47 within the 30 s wait. CONTRACT.md § 6.2 as hardened in -03d already rules on
 * that: a viewport whose settle precondition fails on either side is INVALID with no ratio,
 * not a FAIL. Expect more INVALID pixel runs; that is the gate working, not the port
 * regressing.
 */

/*
 * PE-CW-01 -03b — PLACEHOLDER, not final behaviour. Operator decision D-03-5b.
 *
 * WHY IT EXISTS
 * On the live page, lxf-page-home.js fetches /wp-json/lxf/v1/reviews and appends three
 * `.review-card` children to #reviews-grid. That fetch is not ported, so the container
 * renders 0px tall here while the live grid occupies real layout:
 *
 *   viewport   live #reviews-grid   candidate   columns
 *   390x844    1050px              0px         1fr        (@media max-width:767px)
 *   768x1024    692px              0px         1fr 1fr    (@media max-width:900px)
 *   1440x900    334px              0px         repeat(3,1fr)
 *
 * Measured from the mask geometry recorded on BOTH sides in
 * wedges/PE-CW-01/parity/report-05-pixel.json, not estimated. (Note the ordering: the
 * NARROWER the viewport, the TALLER the grid, because the cards stack. The coordinator's
 * brief quoted these three numbers against the opposite viewports; the report is the
 * authority and is what is used here.)
 *
 * #reviews-grid is a CONTRACT.md § 3 mask selector, so its CONTENT is correctly excluded
 * from both screenshots. A mask cannot exclude its SIZE: with the height missing, every
 * pixel below the grid is offset and the whole lower half of the page reads as differing.
 * RETURN-05.md R2 measured that this one empty container is the entire residual height
 * delta once the fonts are restored.
 *
 * The breakpoints are the live grid's own, taken from lxf-page-home.css
 * (`.reviews-grid` at `max-width:900px` -> 2 columns, at `max-width:767px` -> 1 column),
 * so the reserved height changes where the real layout changes rather than at invented
 * widths.
 *
 * WHY THIS IS A PLACEHOLDER AND MUST NOT BE READ AS A FIX
 * The live height is a function of FETCHED CONTENT: three reviews of variable text length,
 * capped at 280 characters. These three numbers are one observation of it, not an
 * invariant, and a longer review set would move them. Reserving space for content that
 * never arrives also leaves a visible empty band for a real visitor, which is a worse
 * outcome than either real alternative. It is here so that -05 can function as a gate at
 * all, and it is scoped to `:empty` so it evaporates the moment anything populates the
 * grid.
 *
 * D-03-5b — the operator decides: admit the reviews fetch as an island (the grid then
 * sizes itself and this rule is deleted), or ship a static server-rendered review set (same
 * outcome), or accept a reserved placeholder and specify its height as a contract value
 * rather than a measurement.
 *
 * -03c STATUS: still here, now demoted to the NO-JAVASCRIPT FALLBACK it should always have
 * been. Admitted island `lxf-page-home` fetches the review feed and appends real cards, and
 * island `lxf-reviews-feed` is the static same-origin snapshot it resolves to off the
 * WordPress origin, so `:empty` no longer matches in a JS-enabled browser on either side and
 * these three min-heights are inert there. They are kept, unchanged, because a visitor with
 * JavaScript off still gets an empty container, and because deleting them would remove the
 * only thing holding the page's lower half in place if an island ever fails to load.
 */
#reviews-grid:empty {
  min-height: 334px;
}

@media (max-width: 900px) {
  #reviews-grid:empty {
    min-height: 692px;
  }
}

@media (max-width: 767px) {
  #reviews-grid:empty {
    min-height: 1050px;
  }
}
