Merge pull request #8 from cto-af/string-width

String width

docs/classes/spacebreaker.SpaceBreaker.html

@@ -23,7 +23,7 @@ <h4>Hierarchy</h4>
 <ul class="tsd-hierarchy">
 <li><span class="target">SpaceBreaker</span></li></ul></li></ul></section><aside class="tsd-sources">
 <ul>
-<li>Defined in <a href="https://github.com/cto-af/linewrap/blob/68c9344/lib/spacebreaker.js#L138">lib/spacebreaker.js:138</a></li></ul></aside>
+<li>Defined in <a href="https://github.com/cto-af/linewrap/blob/b63e3d6/lib/spacebreaker.js#L138">lib/spacebreaker.js:138</a></li></ul></aside>
 <section class="tsd-panel-group tsd-index-group">
 <section class="tsd-panel tsd-index-panel">
 <details class="tsd-index-content tsd-index-accordion" open><summary class="tsd-accordion-summary tsd-index-summary">
@@ -82,7 +82,7 @@ <h4 class="tsd-returns-title">Returns <a href="spacebreaker.SpaceBreaker.html" c
 <div class="tsd-comment tsd-typography"></div><aside class="tsd-sources">
 <p>Overrides Rules.constructor</p>
 <ul>
-<li>Defined in <a href="https://github.com/cto-af/linewrap/blob/68c9344/lib/spacebreaker.js#L143">lib/spacebreaker.js:143</a></li></ul></aside></li></ul></section></section>
+<li>Defined in <a href="https://github.com/cto-af/linewrap/blob/b63e3d6/lib/spacebreaker.js#L143">lib/spacebreaker.js:143</a></li></ul></aside></li></ul></section></section>
 <section class="tsd-panel-group tsd-member-group">
 <h2>Properties</h2>
 <section class="tsd-panel tsd-member tsd-is-private tsd-is-inherited tsd-is-external"><a id="_private" class="tsd-anchor"></a>

docs/index.html

@@ -15,6 +15,7 @@ <h2>@cto.af/linewrap</h2></div>
 <div class="tsd-panel tsd-typography"><a id="md:ctoaflinewrap" class="tsd-anchor"></a><h1><a href="#md:ctoaflinewrap">@cto.af/linewrap</a></h1><p>Wrap lines using Unicode UAX #14 line breaking rules.</p>
 <a id="md:installation" class="tsd-anchor"></a><h2><a href="#md:installation">Installation</a></h2><pre><code class="language-sh"><span class="hl-0">npm</span><span class="hl-1"> </span><span class="hl-2">install</span><span class="hl-1"> </span><span class="hl-2">@cto.af/linewrap</span>
 </code><button>Copy</button></pre>
+<a id="md:cli" class="tsd-anchor"></a><h2><a href="#md:cli">CLI</a></h2><p>A command line interface is available: <a href="https://github.com/cto-af/linewrap-cli">@cto.af/linewrap-cli</a></p>
 <a id="md:api" class="tsd-anchor"></a><h2><a href="#md:api">API</a></h2><pre><code class="language-js"><span class="hl-3">import</span><span class="hl-1"> {</span><span class="hl-4">LineWrap</span><span class="hl-1">} </span><span class="hl-3">from</span><span class="hl-1"> </span><span class="hl-2">&#39;@cto.af/linewrap&#39;</span><br/><span class="hl-5">const</span><span class="hl-1"> </span><span class="hl-6">w</span><span class="hl-1"> = </span><span class="hl-5">new</span><span class="hl-1"> </span><span class="hl-0">LineWrap</span><span class="hl-1">()</span><br/><span class="hl-4">w</span><span class="hl-1">.</span><span class="hl-0">wrap</span><span class="hl-1">(</span><span class="hl-2">&#39;Lorem ipsum dolor sit amet...&#39;</span><span class="hl-1">)  </span><span class="hl-7">// A string, wrapped to your console length</span><br/><span class="hl-3">for</span><span class="hl-1"> (</span><span class="hl-5">const</span><span class="hl-1"> </span><span class="hl-6">line</span><span class="hl-1"> </span><span class="hl-5">of</span><span class="hl-1"> </span><span class="hl-4">w</span><span class="hl-1">.</span><span class="hl-0">lines</span><span class="hl-1">(</span><span class="hl-2">&#39;Lorem ipsum dolor sit amet...&#39;</span><span class="hl-1">)) {</span><br/><span class="hl-1">  </span><span class="hl-7">// `line` does not have a newline at the end</span><br/><span class="hl-1">}</span>
 </code><button>Copy</button></pre>
 <p>Full <a href="https://cto-af.github.io/linewrap/">API docs</a> are available.</p>
@@ -145,6 +146,7 @@ <h3><svg width="20" height="20" viewBox="0 0 24 24" fill="none"><use href="#icon
 <li>
 <ul>
 <li><a href="#md:installation"><span>Installation</span></a></li>
+<li><a href="#md:cli"><span>CLI</span></a></li>
 <li><a href="#md:api"><span>API</span></a></li>
 <li>
 <ul>

docs/modules/index.html

@@ -16,7 +16,7 @@
 <li><a href="index.html">index</a></li></ul>
 <h1>Module index</h1></div><aside class="tsd-sources">
 <ul>
-<li>Defined in <a href="https://github.com/cto-af/linewrap/blob/68c9344/lib/index.js#L1">lib/index.js:1</a></li></ul></aside>
+<li>Defined in <a href="https://github.com/cto-af/linewrap/blob/b63e3d6/lib/index.js#L1">lib/index.js:1</a></li></ul></aside>
 <section class="tsd-panel-group tsd-index-group">
 <section class="tsd-panel tsd-index-panel">
 <h3 class="tsd-index-heading uppercase">Index</h3>

docs/modules/spacebreaker.html

@@ -16,7 +16,7 @@
 <li><a href="spacebreaker.html">spacebreaker</a></li></ul>
 <h1>Module spacebreaker</h1></div><aside class="tsd-sources">
 <ul>
-<li>Defined in <a href="https://github.com/cto-af/linewrap/blob/68c9344/lib/spacebreaker.js#L1">lib/spacebreaker.js:1</a></li></ul></aside>
+<li>Defined in <a href="https://github.com/cto-af/linewrap/blob/b63e3d6/lib/spacebreaker.js#L1">lib/spacebreaker.js:1</a></li></ul></aside>
 <section class="tsd-panel-group tsd-index-group">
 <section class="tsd-panel tsd-index-panel">
 <h3 class="tsd-index-heading uppercase">Index</h3>

lib/index.js

@@ -1,5 +1,6 @@
 import {Break} from '@cto.af/linebreak'
 import {SpaceBreaker} from './spacebreaker.js'
+import {StringWidth} from '@cto.af/string-width'
 
 const DEFAULT_LOCALE = new Intl.Segmenter().resolvedOptions().locale
 const DEFAULT_IS_NEWLINE = /[^\S\r\n\v\f\x85\u2028\u2029]*[\r\n\v\f\x85\u2028\u2029]+\s*/gu
@@ -40,7 +41,7 @@ export class LineWrap {
    */
   static OVERFLOW_ANYWHERE = Symbol('overflow-clip')
 
-  #graphemes
+  #stringWidth
   #rules
   #opts
 
@@ -57,19 +58,19 @@ export class LineWrap {
   #indent
 
   /**
-   * Length of indent in graphemes
+   * Length of indent in display cells
    * @type {number}
    */
   #indentWidth
 
   /**
-   * Width of #opts.newlineReplacement, in graphemes
+   * Width of #opts.newlineReplacement, in display cells
    * @type {number}
    */
   #replacementWidth
 
   /**
-   * Width of #opts.ellipsis or #opts.hyphen, in graphemes
+   * Width of #opts.ellipsis or #opts.hyphen, in display cells
    * @type {number}
    */
   #enderWidth
@@ -88,7 +89,7 @@ export class LineWrap {
    *   Example 7 of UAX #14.
    * @prop {number} [firstCol=NaN] If indentFirst is false, how many columns
    *   was the first line already indented?  If NaN, use the indent width,
-   *   in graphemes.  If indentFirst is true, this is ignored.
+   *   in display cells.  If indentFirst is true, this is ignored.
    * @prop {string} [hyphen='-'] String to use when long word is
    *   split to next line with LineWrap.OVERFLOW_ANYWHERE.
    * @prop {number|string} [indent=''] If a string, indent every line with
@@ -101,6 +102,7 @@ export class LineWrap {
    * @prop {boolean} [indentFirst=true] Indent the first line?  If not, treat
    *   the first line as if it was already indented, giving a short first
    *   line.
+   * @prop {boolean} [isCJK] If specified, override CJK detection by locale.
    * @prop {RegExp} [isNewline=DEFAULT_IS_NEWLINE] Regular expression that
    *   finds newlines for replacement with `newlineReplacement`.  Ensure you
    *   do not create a regular expression denial of service
@@ -139,6 +141,7 @@ export class LineWrap {
     indentChar: ' ',
     indentEmpty: false,
     indentFirst: true,
+    isCJK: false,
     isNewline: DEFAULT_IS_NEWLINE,
     locale: DEFAULT_LOCALE,
     newline: '\n',
@@ -163,9 +166,21 @@ export class LineWrap {
       example7: this.#opts.example7,
       verbose: this.#opts.verbose,
     })
-    this.#graphemes = new Intl.Segmenter(this.#opts.locale, {
-      granularity: 'grapheme',
-    })
+
+    /**
+     * Don't use this.#opts because it's polluted by the default
+     * @type {{locale?: string, isCJK?: boolean}}
+     */
+    const swOpts = {}
+    if (typeof opts.locale === 'string') {
+      swOpts.locale = opts.locale
+    }
+    if (typeof opts.isCJK === 'boolean') {
+      swOpts.isCJK = opts.isCJK
+    }
+    this.#stringWidth = new StringWidth(swOpts)
+    this.#opts.locale = this.#stringWidth.locale
+    this.#opts.isCJK = this.#stringWidth.isCJK
 
     if (typeof this.#opts.indent === 'number') {
       this.#indent = ''.padEnd(
@@ -175,7 +190,7 @@ export class LineWrap {
       this.#indentWidth = this.#opts.indent
     } else {
       this.#indent = this.#opts.indent
-      this.#indentWidth = this.#graphemeCount(this.#indent)
+      this.#indentWidth = this.#stringWidth.width(this.#indent)
     }
     this.#firstIndent = (this.#opts.indentFirst || isNaN(this.#opts.firstCol)) ?
       this.#indentWidth :
@@ -185,16 +200,18 @@ export class LineWrap {
     if (this.#workingWidth <= 0) {
       throw new Error(`No space to wrap, incompatible width and indent: ${this.#workingWidth}`)
     }
-    this.#replacementWidth = this.#graphemeCount(this.#opts.newlineReplacement)
+    this.#replacementWidth = this.#stringWidth.width(
+      this.#opts.newlineReplacement
+    )
     switch (this.#opts.overflow) {
       case LineWrap.OVERFLOW_VISIBLE:
         this.#enderWidth = 0
         break
       case LineWrap.OVERFLOW_CLIP:
-        this.#enderWidth = this.#graphemeCount(this.#opts.ellipsis)
+        this.#enderWidth = this.#stringWidth.width(this.#opts.ellipsis)
         break
       case LineWrap.OVERFLOW_ANYWHERE:
-        this.#enderWidth = this.#graphemeCount(this.#opts.hyphen)
+        this.#enderWidth = this.#stringWidth.width(this.#opts.hyphen)
         break
       default:
         throw new Error(`Invalid overflow style: "${String(this.#opts.overflow)}"`)
@@ -205,18 +222,17 @@ export class LineWrap {
   }
 
   /**
-   * How many graphemes are in this string?
-   *
-   * @param {string} str
-   * @returns {number}
+   * Did we determin that we are in a CJK context?  Useful for testing.
    */
-  #graphemeCount(str) {
-    // TODO: count widths better, including ZW and ea=F
-    let ret = 0
-    for (const _ of this.#graphemes.segment(str)) {
-      ret++
-    }
-    return ret
+  get isCJK() {
+    return this.#opts.isCJK
+  }
+
+  /**
+   * The calculated locale.  Useful for testing.
+   */
+  get locale() {
+    return this.#opts.locale
   }
 
   /**
@@ -228,7 +244,7 @@ export class LineWrap {
    */
   *#fragments(brk, first) {
     const seg = /** @type {string} */ (brk.string)
-    const graphemes = this.#graphemeCount(seg)
+    const graphemes = this.#stringWidth.width(seg)
     const width = (first && !this.#opts.indentFirst) ?
       this.#opts.width - this.#firstIndent :
       this.#workingWidth
@@ -263,31 +279,34 @@ export class LineWrap {
       case LineWrap.OVERFLOW_CLIP: {
         // Clip it, and end with an ellipsis
         const b = new Break(-1, false)
-        b.string = [...this.#graphemes.segment(seg)]
-          .slice(0, width - this.#enderWidth)
-          .map(s => s.segment)
-          .join('') +
-          this.#opts.ellipsis
-        b.props = {...brk.props, graphemes: width}
+        const breaks = this.#stringWidth.break(
+          seg,
+          width - this.#enderWidth
+        )
+        // Note that breaks[0].cells might be less than width - enderWidth
+        // assert(breaks.length > 0)
+        // assert(!breaks[0].last)
+        b.string = breaks[0].string + this.#opts.ellipsis
+        b.props = {...brk.props, graphemes: breaks[0].cells + this.#enderWidth}
         yield b
         break
       }
       case LineWrap.OVERFLOW_ANYWHERE: {
-        const g = [...this.#graphemes.segment(seg)]
-        const page = width - this.#enderWidth
-        // Might be more that one line long.
-        for (let offset = 0; offset < g.length; offset += page) {
+        const breaks = this.#stringWidth.break(
+          seg,
+          width - this.#enderWidth
+        )
+        for (const {string, cells, last} of breaks) {
           const b = new Break(-1, false)
-          const pg = g.slice(offset, offset + page)
-          b.string = pg.map(s => s.segment).join('')
-          b.props = {...brk.props, graphemes: pg.length}
-          if (offset + page < g.length) {
-            b.string += this.#opts.hyphen
-            b.props.graphemes += this.#enderWidth
+          if (last) {
+            b.string = string
+            b.props = {...brk.props, graphemes: cells}
+          } else {
+            b.string = string + this.#opts.hyphen
+            b.props = {...brk.props, graphemes: cells + this.#enderWidth}
           }
           yield b
         }
-        break
       }
     }
   }