Merge pull request #9 from 'markcarver/patch-2'

Author: rayd

lib/parse.js

@@ -4,13 +4,21 @@ var parseTag = require('./parse-tag');
 // re-used obj for quick lookups of components
 var empty = Object.create ? Object.create(null) : {};
 // common logic for pushing a child node onto a list
-function pushTextNode(list, html, start) {
+function pushTextNode(list, html, level, start, ignoreWhitespace) {
     // calculate correct end of the content slice in case there's
     // no tag after the text node.
     var end = html.indexOf('<', start);
     var content = html.slice(start, end === -1 ? undefined : end);
-    // if a node is nothing but whitespace, no need to add it.
-    if (!/^\s*$/.test(content)) {
+    // if a node is nothing but whitespace, collapse it as the spec states:
+    // https://www.w3.org/TR/html4/struct/text.html#h-9.1
+    if (/^\s*$/.test(content)) {
+        content = ' ';
+    }
+    // don't add whitespace-only text nodes if they would be trailing text nodes
+    // or if they would be leading whitespace-only text nodes:
+    //  * end > -1 indicates this is not a trailing text node
+    //  * leading node is when level is -1 and list has length 0
+    if ((!ignoreWhitespace && end > -1 && level + list.length >= 0) || content !== ' ') {
         list.push({
             type: 'text',
             content: content
@@ -53,7 +61,7 @@ module.exports = function parse(html, options) {
             }
 
             if (!current.voidElement && !inComponent && nextChar && nextChar !== '<') {
-                pushTextNode(current.children, html, start);
+                pushTextNode(current.children, html, level, start, options.ignoreWhitespace);
             }
 
             byTag[current.tagName] = current;
@@ -81,10 +89,15 @@ module.exports = function parse(html, options) {
                 // if we're at the root, push a base text node. otherwise add as
                 // a child to the current node.
                 parent = level === -1 ? result : arr[level].children;
-                pushTextNode(parent, html, start);
+                pushTextNode(parent, html, level, start, options.ignoreWhitespace);
             }
         }
     });
 
+    // If the "html" passed isn't actually html, add it as a text node.
+    if (!result.length && html.length) {
+        pushTextNode(result, html, 0, 0, options.ignoreWhitespace);
+    }
+
     return result;
 };

test/parse.js

@@ -338,6 +338,10 @@ test('parse', function (t) {
             { type: 'text', content: 'something' }
         ]
     },{
+        type: 'text',
+        content: ' '
+    },
+    {
         type: 'tag',
         name: 'a',
         attrs: {},
@@ -347,8 +351,66 @@ test('parse', function (t) {
         type: 'text', content: 'else '
     }], 'should handle text nodes in the middle of tags at the top-level');
 
-    html = '<div>Hi</div>\n\n <span>There</span> \t <iframe>\n\t</iframe>';
+    html = '\n<div>  <span>Hi!</span> </div>\n';
+    parsed = HTML.parse(html);
+    t.deepEqual(parsed, [{
+        type: 'tag',
+        name: 'div',
+        attrs: {},
+        voidElement: false,
+        children: [
+            { type: 'text', content: ' '},
+            {
+                type: 'tag',
+                name: 'span',
+                attrs: {},
+                voidElement: false,
+                children: [
+                    { type: 'text', content: 'Hi!' }
+                ]
+            },
+            { type: 'text', content: ' '}
+        ]
+    }], 'should trim whitespace from the beginning and end of the parsed nodes');
+
+    html = '<div>Hi</div>\n\n <span>There</span> \t <div>\n\t</div>';
     parsed = HTML.parse(html);
+    t.deepEqual(parsed, [{
+        type: 'tag',
+        name: 'div',
+        attrs: {},
+        voidElement: false,
+        children: [
+            { type: 'text', content: 'Hi' }
+        ]
+    },{
+        type: 'text',
+        content: ' '
+    },
+    {
+        type: 'tag',
+        name: 'span',
+        attrs: {},
+        voidElement: false,
+        children: [
+            { type: 'text', content: 'There' }
+        ]
+    },{
+        type: 'text',
+        content: ' '
+    },{
+        type: 'tag',
+        name: 'div',
+        attrs: {},
+        voidElement: false,
+        children: [
+            { type: 'text', content: ' ' }
+        ]
+    }], 'should collapse whitespace');
+    // See https://www.w3.org/TR/html4/struct/text.html#h-9.1
+
+    html = '<div>Hi</div>\n\n <span>There</span> \t <iframe>\n\t</iframe>';
+    parsed = HTML.parse(html, { ignoreWhitespace: true });
     t.deepEqual(parsed, [{
         type: 'tag',
         name: 'div',
@@ -371,7 +433,7 @@ test('parse', function (t) {
         attrs: {},
         voidElement: false,
         children: []
-    }], 'should remove text nodes that are nothing but whitespace');
+    }], 'should remove text nodes that are nothing but whitespace if the ignoreWhitespace option is passed');
 
     html = '<!--\n\t<style type="text/css">\n\t\t.header {\n\t\t\tfont-size: 14px;\n\t\t}\n\t</style>\n\n-->\n<div>Hi</div>';
     parsed = HTML.parse(html);
@@ -410,6 +472,13 @@ test('parse', function (t) {
         voidElement: false,
         children: []
     }], 'should ignore nested HTML comments');
+
+    html = 'Hello, world!';
+    parsed = HTML.parse(html);
+    t.deepEqual(parsed, [{
+        type: 'text',
+        content: 'Hello, world!'
+    }], 'should handle just text');
     t.end();
 });