/**
 * Freesentation — the Korean face.
 *
 * The type system is Supreme, which has no Hangul. Rather than a second scale, the same
 * three roles are mapped onto Freesentation's Regular/Medium/Bold and appended to each
 * token's family list: a Latin glyph still comes from Supreme and only Hangul falls
 * through, so one rule set serves both languages and nothing has to be sized twice.
 *
 * 92%, measured: at 100px Supreme's cap is 70 tall and Freesentation's Hangul inks 84 —
 * Hangul fills the em box where a Latin cap does not, so set at the same px it reads
 * noticeably larger. 92% brings it to 77 against that 70 cap, which is the small overhang
 * Korean normally carries. Line boxes are unaffected: size-adjust scales the glyphs, and
 * every line-height here is unitless off the element's own size.
 *
 * Subset to KS X 1001 — the 2350 syllables of written modern Korean — not to the strings
 * this build happens to print. A syllable the file does not carry does not go missing; it
 * falls through to the system Korean face, which has no size-adjust, so it comes back 9%
 * LARGER than the syllables around it. That is what "different font sizes in the same
 * sentence" was: 엉 and 덩 were outside a 538-syllable subset cut from the copy of the day.
 * Sized to the language rather than to the copy, new wording cannot reopen it.
 * 2462 glyphs, ~137kB a weight, down from 2.7MB.
 */
@font-face {
  font-family: 'Freesentation';
  src: url('../fonts/freesentation/Freesentation-Regular.woff2?v=1094') format('woff2');
  font-weight: 400; font-style: normal; font-display: swap;
  size-adjust: 92%;
}
@font-face {
  font-family: 'Freesentation-Medium';
  src: url('../fonts/freesentation/Freesentation-Medium.woff2?v=1094') format('woff2');
  font-weight: 500; font-style: normal; font-display: swap;
  size-adjust: 92%;
}
@font-face {
  font-family: 'Freesentation-Bold';
  src: url('../fonts/freesentation/Freesentation-Bold.woff2?v=1094') format('woff2');
  font-weight: 700; font-style: normal; font-display: swap;
  size-adjust: 92%;
}

/* Korean breaks between 어절, not inside them. Without this a phone-width column splits a
   word across two lines, which is the one thing that makes Korean UI copy hard to read.
   Only under lang=ko, so English keeps its own wrapping. */
html[lang="ko"] { word-break: keep-all; overflow-wrap: break-word; }