postext 1.8.6 → 1.9.0

package/dist/measure/cjkClasses.js

@@ -0,0 +1,273 @@
+export const CJK_LINE_BREAK_LEVELS = ['none', 'basic', 'gb', 'strict'];
+/** The level lines of CJK text break at when a measurement names none:
+ *  the document's `cjk.lineBreak`, which the build sets before it lays the
+ *  document out (as it sets the hyphenation language), so captions, table
+ *  cells, notes and boxes break like the body. */
+let documentLineBreak = 'gb';
+/** Set the level {@link getCjkLineBreak} returns (the build does, from the
+ *  resolved `cjk.lineBreak`). */
+export function setCjkLineBreak(level) {
+    documentLineBreak = level;
+}
+/** The document's line-break level for CJK text (see
+ *  {@link setCjkLineBreak}); `gb` until a build sets another. */
+export function getCjkLineBreak() {
+    return documentLineBreak;
+}
+/** Whether a line may break between two graphemes that touch: `a` (class
+ *  `aCls`, a CJK grapheme when `aCjk`) and `b` after it. Only next to a CJK
+ *  grapheme — two Western graphemes never part here — and never before the
+ *  ideographic space; at `level`, no prohibited mark opens or closes a
+ *  line. */
+export function cjkBreakAllowed(aCls, aCjk, bCls, bCjk, level) {
+    if (aCls === 'zwsp' || bCls === 'zwsp')
+        return true;
+    if (!aCjk && !bCjk)
+        return false;
+    if (bCls === 'ideoSpace')
+        return false;
+    return !isLineStartProhibited(bCls, level) && !isLineEndProhibited(aCls, level);
+}
+/** Whether a grapheme of `cls` may not open a line at `level`. */
+export function isLineStartProhibited(cls, level) {
+    switch (level) {
+        case 'none':
+            return false;
+        case 'strict':
+            if (cls === 'dash' || cls === 'ellipsis')
+                return true;
+        // falls through
+        case 'gb':
+            if (cls === 'solidus')
+                return true;
+        // falls through
+        case 'basic':
+            return cls === 'pause' || cls === 'stop' || cls === 'closing' || cls === 'connector'
+                || cls === 'interpunct' || cls === 'iteration' || cls === 'postfix';
+    }
+}
+/** Whether a grapheme of `cls` may not close a line at `level`. */
+export function isLineEndProhibited(cls, level) {
+    switch (level) {
+        case 'none':
+            return false;
+        case 'strict':
+        case 'gb':
+            if (cls === 'solidus')
+                return true;
+        // falls through
+        case 'basic':
+            return cls === 'opening' || cls === 'prefix';
+    }
+}
+const OPENING = new Set('（〔［｛【〖《〈「『〘〚〝“‘([{«‹︵︷︹︻︽︿﹁﹃﹇﹙﹛﹝︗｟｢⦅');
+const CLOSING = new Set('）〕］｝】〗》〉」』〙〛〞〟”’)]}»›︶︸︺︼︾﹀﹂﹄﹈﹚﹜﹞︘｠｣⦆');
+const PAUSE = new Set('、，；：,;:﹐﹑﹔﹕︐︑︓︔､');
+const STOP = new Set('。．！？.!?‼⁇⁈⁉﹒﹖﹗︒︕︖｡');
+const INTERPUNCT = new Set('·‧・･');
+const ITERATION = new Set('々〻ゝゞヽヾーｰ〱〲〳〴〵');
+const DASH = new Set('—―⸺⸻︱︲﹘');
+const ELLIPSIS = new Set('…‥⋯︙︰');
+const CONNECTOR = new Set('–～〜~〰゠');
+const SOLIDUS = new Set('/／');
+const PREFIX = new Set('¥$€£₩￥＄￡￦±−＋﹩₽₹');
+const POSTFIX = new Set('%‰‱°℃℉′″％￠﹪');
+/** The squared Latin abbreviations of units in the CJK compatibility
+ *  block: ㍱ hPa to ㍺ IU, ㎀ pA to ㏟ A∕m, ㏿ gal. They follow a number as
+ *  % does, in any script, and are no CJK text. */
+export function isUnitSquare(cp) {
+    return (cp >= 0x3371 && cp <= 0x337A) || (cp >= 0x3380 && cp <= 0x33DF) || cp === 0x33FF;
+}
+/** Fullwidth digits ０–９. */
+export function isFullwidthDigit(cp) {
+    return cp >= 0xFF10 && cp <= 0xFF19;
+}
+/** Fullwidth digits and Latin letters (０–９, Ａ–Ｚ, ａ–ｚ): set one em
+ *  wide like ideographs, but a number or a word a line never breaks
+ *  inside. */
+export function isFullwidthAlnum(cp) {
+    return (cp >= 0xFF10 && cp <= 0xFF19) || (cp >= 0xFF21 && cp <= 0xFF3A) || (cp >= 0xFF41 && cp <= 0xFF5A);
+}
+/** Code points of the CJK scripts and their symbol blocks, classed
+ *  `ideograph` unless a punctuation set above names them. */
+function isCjkCodePoint(cp) {
+    if (cp >= 0x3371 && cp <= 0x33FF && isUnitSquare(cp))
+        return false;
+    return (cp >= 0x1100 && cp <= 0x11FF) // Hangul Jamo
+        || (cp >= 0x2E80 && cp <= 0x2FDF) // CJK and Kangxi radicals
+        || (cp >= 0x2FF0 && cp <= 0x2FFF) // ideographic description characters
+        || (cp >= 0x3000 && cp <= 0x303F) // CJK symbols and punctuation
+        || (cp >= 0x3040 && cp <= 0x30FF) // kana
+        || (cp >= 0x3100 && cp <= 0x312F) // bopomofo
+        || (cp >= 0x3130 && cp <= 0x318F) // Hangul compatibility Jamo
+        || (cp >= 0x3190 && cp <= 0x31FF) // kanbun, bopomofo extended, CJK strokes, katakana extension
+        || (cp >= 0x3200 && cp <= 0x33FF) // enclosed CJK, CJK compatibility
+        || (cp >= 0x3400 && cp <= 0x4DBF) // Extension A
+        || (cp >= 0x4E00 && cp <= 0x9FFF) // unified ideographs
+        || (cp >= 0xA960 && cp <= 0xA97F) // Hangul Jamo extended A
+        || (cp >= 0xAC00 && cp <= 0xD7FF) // Hangul syllables, Jamo extended B
+        || (cp >= 0xF900 && cp <= 0xFAFF) // compatibility ideographs
+        || (cp >= 0xFE10 && cp <= 0xFE1F) // vertical forms
+        || (cp >= 0xFE30 && cp <= 0xFE6F) // CJK compatibility forms, small form variants
+        || (cp >= 0xFF00 && cp <= 0xFFEF) // halfwidth and fullwidth forms
+        || (cp >= 0x16FE0 && cp <= 0x16FFF) // ideographic symbols (small 儿)
+        || (cp >= 0x1B000 && cp <= 0x1B16F) // kana supplement and extended
+        || (cp >= 0x1F200 && cp <= 0x1F2FF) // enclosed ideographic supplement
+        || (cp >= 0x20000 && cp <= 0x3FFFF); // Extensions B and later
+}
+const PICTOGRAPHIC_RE = /\p{Extended_Pictographic}/u;
+/** The class of a grapheme, read from its first code point. */
+export function cjkClassOf(grapheme) {
+    const ch = grapheme.length === 1 ? grapheme : String.fromCodePoint(grapheme.codePointAt(0));
+    if (ch === '\u3000')
+        return 'ideoSpace';
+    if (ch === '\u200B')
+        return 'zwsp';
+    if (OPENING.has(ch))
+        return 'opening';
+    if (CLOSING.has(ch))
+        return 'closing';
+    if (PAUSE.has(ch))
+        return 'pause';
+    if (STOP.has(ch))
+        return 'stop';
+    if (INTERPUNCT.has(ch))
+        return 'interpunct';
+    if (ITERATION.has(ch))
+        return 'iteration';
+    if (DASH.has(ch))
+        return 'dash';
+    if (ELLIPSIS.has(ch))
+        return 'ellipsis';
+    if (CONNECTOR.has(ch))
+        return 'connector';
+    if (SOLIDUS.has(ch))
+        return 'solidus';
+    if (PREFIX.has(ch))
+        return 'prefix';
+    if (POSTFIX.has(ch))
+        return 'postfix';
+    const cp = ch.codePointAt(0);
+    if (isUnitSquare(cp))
+        return 'postfix';
+    if (isCjkCodePoint(cp))
+        return 'ideograph';
+    if (cp >= 0x2190 && PICTOGRAPHIC_RE.test(ch))
+        return 'ideograph';
+    return 'western';
+}
+/** Whether a grapheme is set as a Chinese character in a CJK paragraph: it
+ *  opens a unit of its own instead of joining a Western run. The ASCII
+ *  marks (`,` `.` `(` `%`…) and the signs of numbers stay in the Western
+ *  run they touch; the marks Chinese shares with Latin text (— … · “ ” ‘ ’
+ *  – ~) are Chinese marks, except the apostrophe or the interpunct of a
+ *  Latin word, which the unit builder keeps in it. */
+export function isCjkGrapheme(grapheme) {
+    const cp = grapheme.codePointAt(0);
+    if (cp < 0x80)
+        return false;
+    if (isCjkCodePoint(cp))
+        return true;
+    switch (grapheme[0]) {
+        case '\u2014':
+        case '\u2015':
+        case '\u2E3A':
+        case '\u2E3B': // — ― ⸺ ⸻
+        case '\u2026':
+        case '\u2025':
+        case '\u22EF': // … ‥ ⋯
+        case '\u00B7':
+        case '\u2027': // · ‧
+        case '\u201C':
+        case '\u201D':
+        case '\u2018':
+        case '\u2019': // “ ” ‘ ’
+        case '\u2013': // –
+        case '\u200B':
+            return true;
+    }
+    return cp >= 0x2190 && PICTOGRAPHIC_RE.test(grapheme);
+}
+/** Whether a grapheme is a letter or digit of Western text (the neighbours
+ *  of an apostrophe or an interpunct kept inside a word). */
+export function isWesternWordChar(grapheme) {
+    if (grapheme === undefined)
+        return false;
+    const cp = grapheme.codePointAt(0);
+    if (isCjkCodePoint(cp))
+        return false;
+    return /[\p{L}\p{N}]/u.test(grapheme);
+}
+/** Whether an apostrophe or an interpunct sits inside a Latin word
+ *  ("don’t", "l·l"): between two letters or digits of Western text. It
+ *  belongs to the word then — the composer keeps it in the word's run, and
+ *  vertical text sets it sideways with the word — not to the Chinese text
+ *  around it. */
+export function isWordInnerMark(grapheme, prev, next) {
+    return (grapheme === '\u2019' || grapheme === '\u00B7') && isWesternWordChar(prev) && isWesternWordChar(next);
+}
+/** Whether a code point is a mark a browser sets half width when it meets
+ *  another (CSS `text-spacing-trim`, OpenType `chws`/`halt`): an opening or
+ *  closing bracket or quote, a pause or stop mark, an interpunct, the
+ *  ideographic space. The marks of the CJK blocks always; the marks Latin
+ *  text shares with Chinese (“ ” ‘ ’ · ‧ « »…) only when `shared` is set
+ *  (the text is Chinese). ASCII never. */
+export function isTrimmableMark(cp, shared) {
+    if (cp < 0xAB)
+        return false;
+    if (cp < 0x2E80 && !shared)
+        return false;
+    switch (cjkClassOf(String.fromCodePoint(cp))) {
+        case 'opening':
+        case 'closing':
+        case 'pause':
+        case 'stop':
+        case 'interpunct':
+        case 'ideoSpace':
+            return true;
+        default:
+            return false;
+    }
+}
+/**
+ * Where to cut `text` so that no piece holds two CJK marks side by side:
+ * the offsets (UTF-16) between each such pair. Chrome (CSS
+ * `text-spacing-trim: normal`, and fonts with `chws`) sets the first of
+ * `）》`, `”“`, `》。` or `。《` half width when a run is measured or painted
+ * whole — `本）》录` in Noto Serif SC is 3.5 em as one run and 4 em
+ * character by character — while the CJK composer measures every
+ * character alone and adjusts the marks itself (`cjk.punctuationWidth`).
+ * Measuring and painting such text piece by piece keeps the browser out of
+ * it. `shared` counts the marks Latin text shares with Chinese (“ ” ‘ ’ ·):
+ * set it for Chinese text only, so a Latin `’”` is never cut. Empty for
+ * text with no such pair.
+ */
+export function cjkMarkCuts(text, shared) {
+    let cuts;
+    let prev = false;
+    for (let i = 0; i < text.length; i++) {
+        const cp = text.codePointAt(i);
+        const mark = isTrimmableMark(cp, shared);
+        if (mark && prev)
+            (cuts ??= []).push(i);
+        prev = mark;
+        if (cp > 0xFFFF)
+            i++;
+    }
+    return cuts ?? [];
+}
+/** `text` cut at {@link cjkMarkCuts}: one piece when there is no cut. */
+export function cjkMarkPieces(text, shared) {
+    const cuts = cjkMarkCuts(text, shared);
+    if (cuts.length === 0)
+        return [text];
+    const out = [];
+    let from = 0;
+    for (const at of cuts) {
+        out.push(text.slice(from, at));
+        from = at;
+    }
+    out.push(text.slice(from));
+    return out;
+}
+//# sourceMappingURL=cjkClasses.js.map

package/dist/measure/cjkClasses.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"cjkClasses.js","sourceRoot":"","sources":["../../src/measure/cjkClasses.ts"],"names":[],"mappings":"AA0BA,MAAM,CAAC,MAAM,qBAAqB,GAAiC,CAAC,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,QAAQ,CAAC,CAAC;AA4CrG;;;kDAGkD;AAClD,IAAI,iBAAiB,GAAsB,IAAI,CAAC;AAEhD;iCACiC;AACjC,MAAM,UAAU,eAAe,CAAC,KAAwB;IACtD,iBAAiB,GAAG,KAAK,CAAC;AAC5B,CAAC;AAED;iEACiE;AACjE,MAAM,UAAU,eAAe;IAC7B,OAAO,iBAAiB,CAAC;AAC3B,CAAC;AAED;;;;YAIY;AACZ,MAAM,UAAU,eAAe,CAAC,IAAc,EAAE,IAAa,EAAE,IAAc,EAAE,IAAa,EAAE,KAAwB;IACpH,IAAI,IAAI,KAAK,MAAM,IAAI,IAAI,KAAK,MAAM;QAAE,OAAO,IAAI,CAAC;IACpD,IAAI,CAAC,IAAI,IAAI,CAAC,IAAI;QAAE,OAAO,KAAK,CAAC;IACjC,IAAI,IAAI,KAAK,WAAW;QAAE,OAAO,KAAK,CAAC;IACvC,OAAO,CAAC,qBAAqB,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,mBAAmB,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;AAClF,CAAC;AAED,kEAAkE;AAClE,MAAM,UAAU,qBAAqB,CAAC,GAAa,EAAE,KAAwB;IAC3E,QAAQ,KAAK,EAAE,CAAC;QACd,KAAK,MAAM;YACT,OAAO,KAAK,CAAC;QACf,KAAK,QAAQ;YACX,IAAI,GAAG,KAAK,MAAM,IAAI,GAAG,KAAK,UAAU;gBAAE,OAAO,IAAI,CAAC;QACxD,gBAAgB;QAChB,KAAK,IAAI;YACP,IAAI,GAAG,KAAK,SAAS;gBAAE,OAAO,IAAI,CAAC;QACrC,gBAAgB;QAChB,KAAK,OAAO;YACV,OAAO,GAAG,KAAK,OAAO,IAAI,GAAG,KAAK,MAAM,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,WAAW;mBAC/E,GAAG,KAAK,YAAY,IAAI,GAAG,KAAK,WAAW,IAAI,GAAG,KAAK,SAAS,CAAC;IAC1E,CAAC;AACH,CAAC;AAED,mEAAmE;AACnE,MAAM,UAAU,mBAAmB,CAAC,GAAa,EAAE,KAAwB;IACzE,QAAQ,KAAK,EAAE,CAAC;QACd,KAAK,MAAM;YACT,OAAO,KAAK,CAAC;QACf,KAAK,QAAQ,CAAC;QACd,KAAK,IAAI;YACP,IAAI,GAAG,KAAK,SAAS;gBAAE,OAAO,IAAI,CAAC;QACrC,gBAAgB;QAChB,KAAK,OAAO;YACV,OAAO,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,QAAQ,CAAC;IACjD,CAAC;AACH,CAAC;AAED,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,sCAAsC,CAAC,CAAC;AAChE,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,uCAAuC,CAAC,CAAC;AACjE,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,kBAAkB,CAAC,CAAC;AAC1C,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,oBAAoB,CAAC,CAAC;AAC3C,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,MAAM,CAAC,CAAC;AACnC,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,eAAe,CAAC,CAAC;AAC3C,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,SAAS,CAAC,CAAC;AAChC,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,CAAC;AAClC,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,CAAC;AACpC,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC;AAC9B,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,iBAAiB,CAAC,CAAC;AAC1C,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,aAAa,CAAC,CAAC;AAEvC;;kDAEkD;AAClD,MAAM,UAAU,YAAY,CAAC,EAAU;IACrC,OAAO,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,IAAI,EAAE,KAAK,MAAM,CAAC;AAC3F,CAAC;AAED,4BAA4B;AAC5B,MAAM,UAAU,gBAAgB,CAAC,EAAU;IACzC,OAAO,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC;AACtC,CAAC;AAED;;cAEc;AACd,MAAM,UAAU,gBAAgB,CAAC,EAAU;IACzC,OAAO,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC;AAC5G,CAAC;AAED;6DAC6D;AAC7D,SAAS,cAAc,CAAC,EAAU;IAChC,IAAI,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,IAAI,YAAY,CAAC,EAAE,CAAC;QAAE,OAAO,KAAK,CAAC;IACnE,OAAO,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,cAAc;WAC/C,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,0BAA0B;WACzD,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,qCAAqC;WACpE,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,8BAA8B;WAC7D,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,OAAO;WACtC,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,WAAW;WAC1C,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,4BAA4B;WAC3D,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,6DAA6D;WAC5F,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,kCAAkC;WACjE,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,cAAc;WAC7C,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,qBAAqB;WACpD,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,yBAAyB;WACxD,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,oCAAoC;WACnE,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,2BAA2B;WAC1D,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,iBAAiB;WAChD,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,+CAA+C;WAC9E,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC,CAAC,gCAAgC;WAC/D,CAAC,EAAE,IAAI,OAAO,IAAI,EAAE,IAAI,OAAO,CAAC,CAAC,gCAAgC;WACjE,CAAC,EAAE,IAAI,OAAO,IAAI,EAAE,IAAI,OAAO,CAAC,CAAC,+BAA+B;WAChE,CAAC,EAAE,IAAI,OAAO,IAAI,EAAE,IAAI,OAAO,CAAC,CAAC,kCAAkC;WACnE,CAAC,EAAE,IAAI,OAAO,IAAI,EAAE,IAAI,OAAO,CAAC,CAAC,CAAC,yBAAyB;AAClE,CAAC;AAED,MAAM,eAAe,GAAG,4BAA4B,CAAC;AAErD,+DAA+D;AAC/D,MAAM,UAAU,UAAU,CAAC,QAAgB;IACzC,MAAM,EAAE,GAAG,QAAQ,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,aAAa,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC,CAAE,CAAC,CAAC;IAC7F,IAAI,EAAE,KAAK,QAAQ;QAAE,OAAO,WAAW,CAAC;IACxC,IAAI,EAAE,KAAK,QAAQ;QAAE,OAAO,MAAM,CAAC;IACnC,IAAI,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,SAAS,CAAC;IACtC,IAAI,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,SAAS,CAAC;IACtC,IAAI,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,OAAO,CAAC;IAClC,IAAI,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,MAAM,CAAC;IAChC,IAAI,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,YAAY,CAAC;IAC5C,IAAI,SAAS,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,WAAW,CAAC;IAC1C,IAAI,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,MAAM,CAAC;IAChC,IAAI,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,UAAU,CAAC;IACxC,IAAI,SAAS,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,WAAW,CAAC;IAC1C,IAAI,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,SAAS,CAAC;IACtC,IAAI,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,QAAQ,CAAC;IACpC,IAAI,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,SAAS,CAAC;IACtC,MAAM,EAAE,GAAG,EAAE,CAAC,WAAW,CAAC,CAAC,CAAE,CAAC;IAC9B,IAAI,YAAY,CAAC,EAAE,CAAC;QAAE,OAAO,SAAS,CAAC;IACvC,IAAI,cAAc,CAAC,EAAE,CAAC;QAAE,OAAO,WAAW,CAAC;IAC3C,IAAI,EAAE,IAAI,MAAM,IAAI,eAAe,CAAC,IAAI,CAAC,EAAE,CAAC;QAAE,OAAO,WAAW,CAAC;IACjE,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;;;;sDAKsD;AACtD,MAAM,UAAU,aAAa,CAAC,QAAgB;IAC5C,MAAM,EAAE,GAAG,QAAQ,CAAC,WAAW,CAAC,CAAC,CAAE,CAAC;IACpC,IAAI,EAAE,GAAG,IAAI;QAAE,OAAO,KAAK,CAAC;IAC5B,IAAI,cAAc,CAAC,EAAE,CAAC;QAAE,OAAO,IAAI,CAAC;IACpC,QAAQ,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC;QACpB,KAAK,QAAQ,CAAC;QAAC,KAAK,QAAQ,CAAC;QAAC,KAAK,QAAQ,CAAC;QAAC,KAAK,QAAQ,CAAC,CAAC,UAAU;QACtE,KAAK,QAAQ,CAAC;QAAC,KAAK,QAAQ,CAAC;QAAC,KAAK,QAAQ,CAAC,CAAC,QAAQ;QACrD,KAAK,QAAQ,CAAC;QAAC,KAAK,QAAQ,CAAC,CAAC,MAAM;QACpC,KAAK,QAAQ,CAAC;QAAC,KAAK,QAAQ,CAAC;QAAC,KAAK,QAAQ,CAAC;QAAC,KAAK,QAAQ,CAAC,CAAC,UAAU;QACtE,KAAK,QAAQ,CAAC,CAAC,IAAI;QACnB,KAAK,QAAQ;YACX,OAAO,IAAI,CAAC;IAChB,CAAC;IACD,OAAO,EAAE,IAAI,MAAM,IAAI,eAAe,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;AACxD,CAAC;AAED;6DAC6D;AAC7D,MAAM,UAAU,iBAAiB,CAAC,QAA4B;IAC5D,IAAI,QAAQ,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IACzC,MAAM,EAAE,GAAG,QAAQ,CAAC,WAAW,CAAC,CAAC,CAAE,CAAC;IACpC,IAAI,cAAc,CAAC,EAAE,CAAC;QAAE,OAAO,KAAK,CAAC;IACrC,OAAO,eAAe,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;AACxC,CAAC;AAED;;;;iBAIiB;AACjB,MAAM,UAAU,eAAe,CAAC,QAAgB,EAAE,IAAwB,EAAE,IAAwB;IAClG,OAAO,CAAC,QAAQ,KAAK,QAAQ,IAAI,QAAQ,KAAK,QAAQ,CAAC,IAAI,iBAAiB,CAAC,IAAI,CAAC,IAAI,iBAAiB,CAAC,IAAI,CAAC,CAAC;AAChH,CAAC;AAED;;;;;0CAK0C;AAC1C,MAAM,UAAU,eAAe,CAAC,EAAU,EAAE,MAAe;IACzD,IAAI,EAAE,GAAG,IAAI;QAAE,OAAO,KAAK,CAAC;IAC5B,IAAI,EAAE,GAAG,MAAM,IAAI,CAAC,MAAM;QAAE,OAAO,KAAK,CAAC;IACzC,QAAQ,UAAU,CAAC,MAAM,CAAC,aAAa,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC;QAC7C,KAAK,SAAS,CAAC;QAAC,KAAK,SAAS,CAAC;QAAC,KAAK,OAAO,CAAC;QAAC,KAAK,MAAM,CAAC;QAAC,KAAK,YAAY,CAAC;QAAC,KAAK,WAAW;YAC5F,OAAO,IAAI,CAAC;QACd;YACE,OAAO,KAAK,CAAC;IACjB,CAAC;AACH,CAAC;AAED;;;;;;;;;;;;GAYG;AACH,MAAM,UAAU,WAAW,CAAC,IAAY,EAAE,MAAe;IACvD,IAAI,IAA0B,CAAC;IAC/B,IAAI,IAAI,GAAG,KAAK,CAAC;IACjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,MAAM,EAAE,GAAG,IAAI,CAAC,WAAW,CAAC,CAAC,CAAE,CAAC;QAChC,MAAM,IAAI,GAAG,eAAe,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC;QACzC,IAAI,IAAI,IAAI,IAAI;YAAE,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACxC,IAAI,GAAG,IAAI,CAAC;QACZ,IAAI,EAAE,GAAG,MAAM;YAAE,CAAC,EAAE,CAAC;IACvB,CAAC;IACD,OAAO,IAAI,IAAI,EAAE,CAAC;AACpB,CAAC;AAED,yEAAyE;AACzE,MAAM,UAAU,aAAa,CAAC,IAAY,EAAE,MAAe;IACzD,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;IACvC,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,CAAC,IAAI,CAAC,CAAC;IACrC,MAAM,GAAG,GAAa,EAAE,CAAC;IACzB,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,KAAK,MAAM,EAAE,IAAI,IAAI,EAAE,CAAC;QACtB,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC;QAC/B,IAAI,GAAG,EAAE,CAAC;IACZ,CAAC;IACD,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;IAC3B,OAAO,GAAG,CAAC;AACb,CAAC"}

package/dist/measure/cjkCompose.d.ts

@@ -0,0 +1,86 @@
+/**
+ * The CJK line composer: Chinese (and Japanese, Korean) paragraphs, on both
+ * the plain and the formatted path, broken between characters under the
+ * clreq prohibitions of the document's `cjk.lineBreak` and, when justified,
+ * spread between characters.
+ *
+ * Linear in the paragraph: every unit (one grapheme, a Western run a line
+ * never breaks inside, a 2-em dash or ellipsis, an atomic box) is measured
+ * once — the width cache holds single characters, not prefixes — and the
+ * breaker walks the units once with a running width. A line that would
+ * start with a prohibited mark gives up characters to the next line
+ * (push-out), back to the last break the rules allow.
+ *
+ * Justification (clreq §6.2.2.4): a justified line that is not its
+ * paragraph's last is stretched to the measure — word spaces first, up to
+ * ½ em each, then every gap between characters equally, never inside a
+ * Western run (a word, a number with its signs), a 2-em dash or ellipsis,
+ * and never next to a connector or a solidus. The spread is carried as
+ * segment `tracking` (px after each grapheme, in the segment's width); a
+ * Western run or a dash pair that a gap follows gives its last grapheme a
+ * segment of its own. A line that needs more than the cap (½ em, or
+ * `bodyText.maxJustifyTracking` when set) is set with the cap and flagged
+ * `cjkLoose` and `ragged`; a line with no CJK character on it (the head of
+ * a long web address) is set `ragged` without the flag.
+ *
+ * Segments: a Western run keeps one of its own, and single characters
+ * share one only when they have one style, link and spacing and advance
+ * alike, so a caret spread evenly over a segment lands on its characters
+ * and a Markdown link covers only its own.
+ */
+import type { InlineSpan } from '../parse';
+import type { VDTSegmentMarks } from '../vdt';
+import { type MeasuredBlock, type MeasureBlockOptions } from './types';
+import { type CjkLineBreakLevel } from './cjkClasses';
+/**
+ * Whether a paragraph is set by the CJK composer: it holds more CJK letters
+ * than word spaces. A word space is a run of spaces between two characters
+ * that are not CJK. A space that touches a CJK character or mark is the
+ * one web text types at each boundary (`2026 年 9 月 28 日`, `安装 Node.js
+ * 和 Git`), which the composer replaces with the Han–Latin space.
+ * A Latin paragraph that quotes a Chinese title or name ("the novel 紅樓夢
+ * was …") has more word spaces, stays with Knuth–Plass and breaks next to
+ * the characters it quotes; a Chinese paragraph with Latin words in it
+ * ("用 iPhone 拍照") is composed, with its typed spaces or without them.
+ */
+export declare function isCjkParagraph(text: string): boolean;
+/** Whether the CJK composer sets a paragraph of this text: it holds CJK
+ *  text (not only unit squares or marks Latin shares with Chinese) and more
+ *  CJK letters than word spaces (see {@link isCjkParagraph}). Both
+ *  measuring paths ask this, so they agree. */
+export declare function composesAsCjk(text: string): boolean;
+/** The marks a span sets on its characters (#193), the emphasis dots'
+ *  defaults filled in: a filled dot (an open circle), under the text, or
+ *  right of it in vertical text. */
+export declare function spanMarks(span: InlineSpan, vertical: boolean): VDTSegmentMarks | undefined;
+/**
+ * Set a Chinese, Japanese or Korean paragraph (see the module comment):
+ * `options.cjkLineBreak` (else the document's level) decides where lines
+ * may break; `textAlign: 'justify'` spreads every line but the last to the
+ * measure. First-line and hanging indents, the block's tracking
+ * (`letterSpacingPx`), other measures for later lines (`restWidths`) and a
+ * positive `looseness` (one line more, for column balancing: the lines are
+ * broken a little short of the measure and spread to it) are honoured;
+ * Knuth–Plass options are not used.
+ */
+export declare function composeCjkParagraph(spans: InlineSpan[], normalFont: string, boldFont: string, italicFont: string, boldItalicFont: string, maxWidthPx: number, lineHeightPx: number, options: MeasureBlockOptions | undefined): MeasuredBlock;
+/** Where a word that holds CJK characters may break, and its widths: see
+ *  {@link cjkWordBreaks}. */
+export interface CjkWordBreaks {
+    /** Offsets into the word a line may break at, ascending. */
+    breaks: number[];
+    /** The word's advance: the sum of its units'. */
+    width: number;
+    /** The advance of the word's first `idx` characters. */
+    widthBefore: (idx: number) => number;
+}
+/**
+ * The breaks of one word (no breaking space inside) that holds CJK
+ * characters, in a paragraph set by the word-by-word breaker (a Latin
+ * paragraph quoting CJK): next to its CJK characters, under the document's
+ * line-break rules. Each unit is measured once and the widths before a
+ * break come from their running sums, so a long run of ideographs costs
+ * linear time.
+ */
+export declare function cjkWordBreaks(word: string, font: string, smallCaps: boolean | undefined, letterSpacingPx: number, level?: CjkLineBreakLevel): CjkWordBreaks;
+//# sourceMappingURL=cjkCompose.d.ts.map

package/dist/measure/cjkCompose.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cjkCompose.d.ts","sourceRoot":"","sources":["../../src/measure/cjkCompose.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6BG;AAEH,OAAO,KAAK,EAAc,UAAU,EAAiB,MAAM,UAAU,CAAC;AACtE,OAAO,KAAK,EAA6C,eAAe,EAAc,MAAM,QAAQ,CAAC;AAErG,OAAO,EAAe,KAAK,aAAa,EAAE,KAAK,mBAAmB,EAAE,MAAM,SAAS,CAAC;AAkBpF,OAAO,EAWL,KAAK,iBAAiB,EACvB,MAAM,cAAc,CAAC;AA4CtB;;;;;;;;;;GAUG;AACH,wBAAgB,cAAc,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAsBpD;AAED;;;+CAG+C;AAC/C,wBAAgB,aAAa,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAEnD;AA6GD;;oCAEoC;AACpC,wBAAgB,SAAS,CAAC,IAAI,EAAE,UAAU,EAAE,QAAQ,EAAE,OAAO,GAAG,eAAe,GAAG,SAAS,CAc1F;AA2rDD;;;;;;;;;GASG;AACH,wBAAgB,mBAAmB,CACjC,KAAK,EAAE,UAAU,EAAE,EACnB,UAAU,EAAE,MAAM,EAClB,QAAQ,EAAE,MAAM,EAChB,UAAU,EAAE,MAAM,EAClB,cAAc,EAAE,MAAM,EACtB,UAAU,EAAE,MAAM,EAClB,YAAY,EAAE,MAAM,EACpB,OAAO,EAAE,mBAAmB,GAAG,SAAS,GACvC,aAAa,CA4Ff;AAaD;6BAC6B;AAC7B,MAAM,WAAW,aAAa;IAC5B,4DAA4D;IAC5D,MAAM,EAAE,MAAM,EAAE,CAAC;IACjB,iDAAiD;IACjD,KAAK,EAAE,MAAM,CAAC;IACd,wDAAwD;IACxD,WAAW,EAAE,CAAC,GAAG,EAAE,MAAM,KAAK,MAAM,CAAC;CACtC;AAED;;;;;;;GAOG;AACH,wBAAgB,aAAa,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,GAAG,SAAS,EAAE,eAAe,EAAE,MAAM,EAAE,KAAK,GAAE,iBAAqC,GAAG,aAAa,CAoC9K"}