Apply HTML 5 spec of not caring on casing for input (#2016)

* Sanitizer will still standardize on lowercase atm.
* Some optimize/minimize conditionals.
* Be specific on what/where to replace with open tags... not a huge deal atm but extra-cautious may be prudent.
* Restore some styling

Post #2014 #2015 and closes #1775

Auto-merge

libs/markdown.js

@@ -143,27 +143,10 @@ function sanitize(aHtml) {
 // Sanitize the output from the block level renderers
 blockRenderers.forEach(function (aType) {
   renderer[aType] = function () {
-    var matches = null;
-    var openTagName = null;
-    var closeTagName = null;
-
-    // Begin workaround for #1775
-    if (aType === 'html') {
-      matches = arguments[0].match(/^<([a-z]+)(?![^>]*\/>)[^>]*>$/);
-      if (matches) {
-        openTagName = matches[1];
-      }
-      matches = arguments[0].match(/^<\/([a-z]+)>$/);
-      if (matches && !!sanitize('<' + matches[1] + '></' + matches[1] + '>')) {
-        closeTagName = matches[1];
-      }
-    }
-
-    // Sanitize first to close any tags
+    // Render Markdown type first then sanitize HTML including any closing of tags
     var sanitized = sanitize(marked.Renderer.prototype[aType].apply(renderer, arguments));
 
-    // Autolink most usernames
-
+    // Autolink most usernames.
     var dom = new JSDOM('<div id="sandbox"></div>');
     var win = dom.window;
     var doc = win.document;
@@ -179,6 +162,8 @@ blockRenderers.forEach(function (aType) {
     var htmlContainer = null;
     var thisNode = null;
 
+    var matches = null;
+
     hookNode.innerHTML = sanitized;
 
     xpr = doc.evaluate(
@@ -217,14 +202,13 @@ blockRenderers.forEach(function (aType) {
       sanitized = hookNode.innerHTML;
     }
 
-    // End workaround for #1775
+    // Workaround for #1775
     if (aType === 'html') {
-      if (openTagName) {
-        sanitized = sanitized.replace(/<\/[a-z]+>/, '');
-      }
-
-      if (closeTagName) {
-        sanitized = '</' + closeTagName + '>';
+      matches = arguments[0].match(/^<(\/?)([a-z]+)(?![^>]*\/>)[^>]*>$/i);
+      if (matches && matches[2] && sanitize('<' + matches[2] + '></' + matches[2] + '>')) {
+        sanitized = matches[1]
+          ? '</' + matches[2].toLowerCase() + '>'
+          : sanitized.replace(new RegExp('<\/' + matches[2].toLowerCase() + '>$'), '');
       }
     }