use sanitize-html's textFilter callback to only apply highlights to textNodes when highlighting HTML. fixes https://github.com/vector-im/vector-web/issues/294

src/HtmlUtils.js

@@ -17,6 +17,7 @@ limitations under the License.
 'use strict';
 
 var React = require('react');
+var ReactDOMServer = require('react-dom/server')
 var sanitizeHtml = require('sanitize-html');
 var highlight = require('highlight.js');
 
@@ -57,24 +58,17 @@ class Highlighter {
         this._key = 0;
     }
 
-    applyHighlights(safeSnippet, highlights) {
+    applyHighlights(safeSnippet, safeHighlights) {
         var lastOffset = 0;
         var offset;
         var nodes = [];
 
-        // XXX: when highlighting HTML, synapse performs the search on the plaintext body,
+        var safeHighlight = safeHighlights[0];
-        // but we're attempting to apply the highlights here to the HTML body.  This is
-        // never going to end well - we really should be hooking into the sanitzer HTML
-        // parser to only attempt to highlight text nodes to avoid corrupting tags.  
-        // If and when this happens, we'll probably have to split this method in two between
-        // HTML and plain-text highlighting.
-
-        var safeHighlight = this.html ? sanitizeHtml(highlights[0], sanitizeHtmlParams) : highlights[0];
         while ((offset = safeSnippet.toLowerCase().indexOf(safeHighlight.toLowerCase(), lastOffset)) >= 0) {
             // handle preamble
             if (offset > lastOffset) {
                 var subSnippet = safeSnippet.substring(lastOffset, offset);
-                nodes = nodes.concat(this._applySubHighlights(subSnippet, highlights));
+                nodes = nodes.concat(this._applySubHighlights(subSnippet, safeHighlights));
             }
 
             // do highlight
@@ -86,15 +80,15 @@ class Highlighter {
         // handle postamble
         if (lastOffset != safeSnippet.length) {
             var subSnippet = safeSnippet.substring(lastOffset, undefined);
-            nodes = nodes.concat(this._applySubHighlights(subSnippet, highlights));
+            nodes = nodes.concat(this._applySubHighlights(subSnippet, safeHighlights));
         }
         return nodes;
     }
 
-    _applySubHighlights(safeSnippet, highlights) {
+    _applySubHighlights(safeSnippet, safeHighlights) {
-        if (highlights[1]) {
+        if (safeHighlights[1]) {
             // recurse into this range to check for the next set of highlight matches
-            return this.applyHighlights(safeSnippet, highlights.slice(1));
+            return this.applyHighlights(safeSnippet, safeHighlights.slice(1));
         }
         else {
             // no more highlights to be found, just return the unhighlighted string
@@ -132,7 +126,7 @@ module.exports = {
      *
      * content: 'content' of the MatrixEvent
      *
-     * highlights: optional list of words to highlight
+     * highlights: optional list of words to highlight, ordered by longest word first
      *
      * opts.onHighlightClick: optional callback function to be called when a
      *     highlighted word is clicked
@@ -142,28 +136,38 @@ module.exports = {
 
         var isHtml = (content.format === "org.matrix.custom.html");
 
-        var safeBody;
+        var safeBody, body;
         if (isHtml) {
+            // XXX: We sanitize the HTML whilst also highlighting its text nodes, to avoid accidentally trying
+            // to highlight HTML tags themselves.  However, this does mean that we don't highlight textnodes which
+            // are interrupted by HTML tags (not that we did before) - e.g. foo<span/>bar won't get highlighted
+            // by an attempt to search for 'foobar'.  Then again, the search query probably wouldn't work either
+            if (highlights && highlights.length > 0) {
+                var highlighter = new Highlighter(isHtml, "mx_EventTile_searchHighlight", opts.onHighlightClick);
+                // XXX: hacky bodge to temporarily apply a textFilter to the sanitizeHtmlParams structure.
+                sanitizeHtmlParams.textFilter = function(safeText) {
+                    var html = highlighter.applyHighlights(safeText, highlights).map(function(span) {
+                        // XXX: rather clunky conversion from the react nodes returned by applyHighlights
+                        // (which need to be nodes for the non-html highlighting case), to convert them
+                        // back into raw HTML given that's what sanitize-html works in terms of.
+                        return ReactDOMServer.renderToString(span);
+                    }).join('');
+                    return html;
+                };
+            }
             safeBody = sanitizeHtml(content.formatted_body, sanitizeHtmlParams);
+            delete sanitizeHtmlParams.textFilter;
+            return <span className="markdown-body" dangerouslySetInnerHTML={{ __html: safeBody }} />;
         } else {
             safeBody = content.body;
-        }
+            if (highlights && highlights.length > 0) {
-
+                var highlighter = new Highlighter(isHtml, "mx_EventTile_searchHighlight", opts.onHighlightClick);
-        var body;
+                return highlighter.applyHighlights(safeBody, highlights);
-        if (highlights && highlights.length > 0) {
-            var highlighter = new Highlighter(isHtml, "mx_EventTile_searchHighlight", opts.onHighlightClick);
-            body = highlighter.applyHighlights(safeBody, highlights);
-        }
-        else {
-            if (isHtml) {
-                body = <span className="markdown-body" dangerouslySetInnerHTML={{ __html: safeBody }} />;
             }
             else {
-                body = safeBody;
+                return safeBody;
             }
         }
-
-        return body;
     },
 
     highlightDom: function(element) {