feat: implement line-break.txt v13 (#15)

scripts/generate_line_break_tests.ts

@@ -4,10 +4,9 @@ import {BREAK_MANDATORY, BREAK_NOT_ALLOWED, BREAK_ALLOWED} from '../src/LineBrea
 
 const data = readFileSync(resolve(__dirname, '../tests/LineBreakTest.txt')).toString();
 const tests: string[] = [];
-data
-    .split('\n')
-    .filter(s => s.length > 0)
-    .forEach(s => {
+data.split('\n')
+    .filter((s) => s.length > 0)
+    .forEach((s) => {
         let [input, comment] = s.split('#');
         input = input.trim();
 
@@ -16,7 +15,7 @@ data
             const inputs = input.split(/\s+/g);
             const codePoints: string[] = [];
             const breaks: string[] = [];
-            inputs.forEach(input => {
+            inputs.forEach((input) => {
                 if ([BREAK_ALLOWED, BREAK_MANDATORY, BREAK_NOT_ALLOWED].indexOf(input) !== -1) {
                     breaks.push(input);
                 } else {

scripts/generate_line_break_trie.ts

@@ -12,7 +12,7 @@ let rangeType: number | null = null;
 
 rawData
     .split('\n')
-    .map(s => {
+    .map((s) => {
         const index = s.indexOf('#');
         const first = (index === -1 ? s : s.substring(0, index)).trim();
         return index === -1

src/LineBreak.ts

@@ -56,6 +56,8 @@ const RI = 41; //  Keep pairs together. For pairs; break before and after other
 const SA = 42; //  Provide a line break opportunity contingent on additional, language-specific context analysis
 const XX = 43; //  Have as yet unknown line breaking behavior or unassigned code positions
 
+const ea_OP = [0x2329, 0xff08];
+
 export const classes: {[key: string]: number} = {
     BK,
     CR,
@@ -304,7 +306,7 @@ const _lineBreakAtIndex = (
     }
 
     // LB8a Do not break between a zero width joiner and an ideograph, emoji base or emoji modifier.
-    if (UnicodeTrie.get(codePoints[currentIndex]) === ZWJ && (next === ID || next === EB || next === EM)) {
+    if (UnicodeTrie.get(codePoints[currentIndex]) === ZWJ) {
         return BREAK_NOT_ALLOWED;
     }
 
@@ -378,8 +380,8 @@ const _lineBreakAtIndex = (
         return BREAK_NOT_ALLOWED;
     }
 
-    // LB22 Do not break between two ellipses, or between letters, numbers or exclamations and ellipsis.
-    if (next === IN && ALPHABETICS.concat(IN, EX, NU, ID, EB, EM).indexOf(current) !== -1) {
+    // LB22 Do not break before ellipsis.
+    if (next === IN) {
         return BREAK_NOT_ALLOWED;
     }
 
@@ -476,7 +478,9 @@ const _lineBreakAtIndex = (
 
     // LB30 Do not break between letters, numbers, or ordinary symbols and opening or closing parentheses.
     if (
-        (ALPHABETICS.concat(NU).indexOf(current) !== -1 && next === OP) ||
+        (ALPHABETICS.concat(NU).indexOf(current) !== -1 &&
+            next === OP &&
+            ea_OP.indexOf(codePoints[afterIndex]) === -1) ||
         (ALPHABETICS.concat(NU).indexOf(next) !== -1 && current === CP)
     ) {
         return BREAK_NOT_ALLOWED;
@@ -532,17 +536,14 @@ interface IOptions {
     wordBreak?: WORD_BREAK;
 }
 
-const cssFormattedClasses = (
-    codePoints: number[],
-    options?: IOptions
-): [number[], number[], boolean[] | undefined] => {
+const cssFormattedClasses = (codePoints: number[], options?: IOptions): [number[], number[], boolean[] | undefined] => {
     if (!options) {
         options = {lineBreak: 'normal', wordBreak: 'normal'};
     }
     let [indicies, classTypes, isLetterNumber] = codePointsToCharacterClasses(codePoints, options.lineBreak);
 
     if (options.wordBreak === 'break-all') {
-        classTypes = classTypes.map(type => ([NU, AL, SA].indexOf(type) !== -1 ? ID : type));
+        classTypes = classTypes.map((type) => ([NU, AL, SA].indexOf(type) !== -1 ? ID : type));
     }
 
     const forbiddenBreakpoints =

src/Trie.ts

@@ -74,7 +74,7 @@ const slice16 = (view: number[] | Uint16Array, start: number, end?: number) => {
         return view.slice(start, end);
     }
 
-    return new Uint16Array(Array.prototype.slice.call(view, start, end))
+    return new Uint16Array(Array.prototype.slice.call(view, start, end));
 };
 
 const slice32 = (view: number[] | Uint32Array, start: number, end?: number) => {

src/TrieBuilder.ts

@@ -14,7 +14,7 @@ import {
     UTRIE2_INDEX_2_MASK,
     UTRIE2_SHIFT_1_2,
     Trie,
-    int
+    int,
 } from './Trie';
 
 import {encode} from 'base64-arraybuffer';
@@ -288,10 +288,10 @@ export class TrieBuilder {
      */
     setRange(start: int, end: int, value: int, overwrite: boolean = false): TrieBuilder {
         /*
-     * repeat value in [start..end]
-     * mark index values for repeat-data blocks by setting bit 31 of the index values
-     * fill around existing values if any, if(overwrite)
-     */
+         * repeat value in [start..end]
+         * mark index values for repeat-data blocks by setting bit 31 of the index values
+         * fill around existing values if any, if(overwrite)
+         */
         let block, rest, repeatBlock;
         if (start > 0x10ffff || start < 0 || end > 0x10ffff || end < 0 || start > end) {
             throw new Error('Invalid code point range.');
@@ -351,32 +351,32 @@ export class TrieBuilder {
                 /* already allocated */
                 if (overwrite && block >= UNEWTRIE2_DATA_0800_OFFSET) {
                     /*
-             * We overwrite all values, and it's not a
-             * protected (ASCII-linear or 2-byte UTF-8) block:
-             * replace with the repeatBlock.
-             */
+                     * We overwrite all values, and it's not a
+                     * protected (ASCII-linear or 2-byte UTF-8) block:
+                     * replace with the repeatBlock.
+                     */
                     setRepeatBlock = true;
                 } else {
                     /* !overwrite, or protected block: just write the values into this block */
                     this.fillBlock(block, 0, UTRIE2_DATA_BLOCK_LENGTH, value, this.initialValue, overwrite);
                 }
             } else if (this.data[block] !== value && (overwrite || block === this.dataNullOffset)) {
                 /*
-             * Set the repeatBlock instead of the null block or previous repeat block:
-             *
-             * If !isWritableBlock() then all entries in the block have the same value
-             * because it's the null block or a range block (the repeatBlock from a previous
-             * call to utrie2_setRange32()).
-             * No other blocks are used multiple times before compacting.
-             *
-             * The null block is the only non-writable block with the initialValue because
-             * of the repeatBlock initialization above. (If value==initialValue, then
-             * the repeatBlock will be the null data block.)
-             *
-             * We set our repeatBlock if the desired value differs from the block's value,
-             * and if we overwrite any data or if the data is all initial values
-             * (which is the same as the block being the null block, see above).
-             */
+                 * Set the repeatBlock instead of the null block or previous repeat block:
+                 *
+                 * If !isWritableBlock() then all entries in the block have the same value
+                 * because it's the null block or a range block (the repeatBlock from a previous
+                 * call to utrie2_setRange32()).
+                 * No other blocks are used multiple times before compacting.
+                 *
+                 * The null block is the only non-writable block with the initialValue because
+                 * of the repeatBlock initialization above. (If value==initialValue, then
+                 * the repeatBlock will be the null data block.)
+                 *
+                 * We set our repeatBlock if the desired value differs from the block's value,
+                 * and if we overwrite any data or if the data is all initial values
+                 * (which is the same as the block being the null block, see above).
+                 */
                 setRepeatBlock = true;
             }
             if (setRepeatBlock) {
@@ -484,9 +484,9 @@ export class TrieBuilder {
             }
 
             /*
-     * write the index-2 array values for supplementary code points,
-     * shifted right by UTRIE2_INDEX_SHIFT, after adding dataMove
-     */
+             * write the index-2 array values for supplementary code points,
+             * shifted right by UTRIE2_INDEX_SHIFT, after adding dataMove
+             */
             for (i = 0; i < this.index2Length - index2Offset; i++) {
                 index[destIdx++] = (dataMove + this.index2[index2Offset + i]) >> UTRIE2_INDEX_SHIFT;
             }
@@ -615,17 +615,17 @@ export class TrieBuilder {
             this.map[i] = start;
         }
         /*
-     * Start with a block length of 64 for 2-byte UTF-8,
-     * then switch to UTRIE2_DATA_BLOCK_LENGTH.
-     */
+         * Start with a block length of 64 for 2-byte UTF-8,
+         * then switch to UTRIE2_DATA_BLOCK_LENGTH.
+         */
         blockLength = 64;
         blockCount = blockLength >> UTRIE2_SHIFT_2;
         for (start = newStart; start < this.dataLength; ) {
             /*
-         * start: index of first entry of current block
-         * newStart: index where the current block is to be moved
-         *           (right after current end of already-compacted data)
-         */
+             * start: index of first entry of current block
+             * newStart: index where the current block is to be moved
+             *           (right after current end of already-compacted data)
+             */
             if (start === UNEWTRIE2_DATA_0800_OFFSET) {
                 blockLength = UTRIE2_DATA_BLOCK_LENGTH;
                 blockCount = 1;

src/Util.ts

@@ -39,7 +39,7 @@ export const fromCodePoint = (...codePoints: number[]): string => {
             codeUnits.push(codePoint);
         } else {
             codePoint -= 0x10000;
-            codeUnits.push((codePoint >> 10) + 0xd800, codePoint % 0x400 + 0xdc00);
+            codeUnits.push((codePoint >> 10) + 0xd800, (codePoint % 0x400) + 0xdc00);
         }
         if (index + 1 === length || codeUnits.length > 0x4000) {
             result += String.fromCharCode(...codeUnits);