Preserve en/em/non-breaking/hair space etc. while minifying (#849)

kangax · Sep 11, 2017 · 7683fbc · 7683fbc
1 parent 5241da9
commit 7683fbc
Show file tree

Hide file tree

Showing 2 changed files with 50 additions and 14 deletions.
diff --git a/src/htmlminifier.js b/src/htmlminifier.js
@@ -8,20 +8,16 @@ var TokenChain = require('./tokenchain');
 var UglifyJS = require('uglify-js');
 var utils = require('./utils');
 
-var trimWhitespace = String.prototype.trim ? function(str) {
+function trimWhitespace(str) {
   if (typeof str !== 'string') {
     return str;
   }
-  return str.trim();
-} : function(str) {
-  if (typeof str !== 'string') {
-    return str;
-  }
-  return str.replace(/^\s+/, '').replace(/\s+$/, '');
-};
+  return str.replace(/^[ \n\r\t\f]+/, '').replace(/[ \n\r\t\f]+$/, '');
+}
 
 function collapseWhitespaceAll(str) {
-  return str && str.replace(/\s+/g, function(spaces) {
+  // Non-breaking space is specifically handled inside the replacer function here:
+  return str && str.replace(/[ \n\r\t\f\xA0]+/g, function(spaces) {
     return spaces === '\t' ? '\t' : spaces.replace(/(^|\xA0+)[^\xA0]+/g, '$1 ');
   });
 }
@@ -30,17 +26,18 @@ function collapseWhitespace(str, options, trimLeft, trimRight, collapseAll) {
   var lineBreakBefore = '', lineBreakAfter = '';
 
   if (options.preserveLineBreaks) {
-    str = str.replace(/^\s*?[\n\r]\s*/, function() {
+    str = str.replace(/^[ \n\r\t\f]*?[\n\r][ \n\r\t\f]*/, function() {
       lineBreakBefore = '\n';
       return '';
-    }).replace(/\s*?[\n\r]\s*$/, function() {
+    }).replace(/[ \n\r\t\f]*?[\n\r][ \n\r\t\f]*$/, function() {
       lineBreakAfter = '\n';
       return '';
     });
   }
 
   if (trimLeft) {
-    str = str.replace(/^\s+/, function(spaces) {
+    // Non-breaking space is specifically handled inside the replacer function here:
+    str = str.replace(/^[ \n\r\t\f\xA0]+/, function(spaces) {
       var conservative = !lineBreakBefore && options.conservativeCollapse;
       if (conservative && spaces === '\t') {
         return '\t';
@@ -50,7 +47,8 @@ function collapseWhitespace(str, options, trimLeft, trimRight, collapseAll) {
   }
 
   if (trimRight) {
-    str = str.replace(/\s+$/, function(spaces) {
+    // Non-breaking space is specifically handled inside the replacer function here:
+    str = str.replace(/[ \n\r\t\f\xA0]+$/, function(spaces) {
       var conservative = !lineBreakAfter && options.conservativeCollapse;
       if (conservative && spaces === '\t') {
         return '\t';
@@ -1253,7 +1251,7 @@ function minify(value, options, partialMarkup) {
         return collapseWhitespace(chunk, {
           preserveLineBreaks: options.preserveLineBreaks,
           conservativeCollapse: !options.trimCustomFragments
-        }, /^\s/.test(chunk), /\s$/.test(chunk));
+        }, /^[ \n\r\t\f]/.test(chunk), /[ \n\r\t\f]$/.test(chunk));
       }
       return chunk;
     });

diff --git a/tests/minifier.js b/tests/minifier.js
@@ -347,6 +347,44 @@ QUnit.test('space normalization around text', function(assert) {
   assert.equal(minify(input, { collapseWhitespace: true }), output);
 });
 
+QUnit.test('types of whitespace that should always be preserved', function(assert) {
+  // Hair space:
+  var input = '<div>\u200afo\u200ao\u200a</div>';
+  assert.equal(minify(input, { collapseWhitespace: true }), input);
+
+  // Hair space passed as HTML entity:
+  var inputWithEntities = '<div>&#8202;fo&#8202;o&#8202;</div>';
+  assert.equal(minify(inputWithEntities, { collapseWhitespace: true }), inputWithEntities);
+
+  // Hair space passed as HTML entity, in decodeEntities:true mode:
+  assert.equal(minify(inputWithEntities, { collapseWhitespace: true, decodeEntities: true }), input);
+
+
+  // Non-breaking space:
+  input = '<div>\xa0fo\xa0o\xa0</div>';
+  assert.equal(minify(input, { collapseWhitespace: true }), input);
+
+  // Non-breaking space passed as HTML entity:
+  inputWithEntities = '<div>&nbsp;fo&nbsp;o&nbsp;</div>';
+  assert.equal(minify(inputWithEntities, { collapseWhitespace: true }), inputWithEntities);
+
+  // Non-breaking space passed as HTML entity, in decodeEntities:true mode:
+  assert.equal(minify(inputWithEntities, { collapseWhitespace: true, decodeEntities: true }), input);
+
+  // Do not remove hair space when preserving line breaks between tags:
+  input = '<p></p>\u200a\n<p></p>\n';
+  assert.equal(minify(input, { collapseWhitespace: true, preserveLineBreaks: true }), input);
+
+  // Preserve hair space in attributes:
+  input = '<p class="foo\u200abar"></p>';
+  assert.equal(minify(input, { collapseWhitespace: true }), input);
+
+  // Preserve hair space in class names when deduplicating and reordering:
+  input = '<a class="0 1\u200a3 2 3"></a>';
+  assert.equal(minify(input, { sortClassName: false }), input);
+  assert.equal(minify(input, { sortClassName: true }), input);
+});
+
 QUnit.test('doctype normalization', function(assert) {
   var input;