Fix incorrect linking of URLs without trailing punctuation in HTML bibliographies

chrome/content/zotero/xpcom/csl.js

@@ -407,13 +407,11 @@ Zotero.CSL.prototype.formatBibliography = function(itemSet, format) {
 		if(format == "HTML") {
 			var coins = Zotero.OpenURL.createContextObject(item.zoteroItem, "1.0");
 			
-			// Wrap URLs in <a href=""> links, in a very unsophisticated manner
+			// Wrap URLs and DOIs in HTML links
 			//
-			// This should be done earlier when the data is still in variables
+			// This should be handled when the values are still in variables
-			//
+			// (and presumably will be in the new engine)
-			// Ignore URLs preceded by '>', since these are likely already links
+			string = Zotero.Utilities.prototype.autoLink(string);
-			string = string.replace(/([^>])(https?:\/\/[^\s]+)([\."'>:\]\)\s])/g, '$1<a href="$2">$2</a>$3');
-			string = string.replace(/(doi:[ ]*)(10\.[^\s]+[0-9a-zA-Z])/g, '$1<a href="http://dx.doi.org/$2">$2</a>');
 			
 			var span = (coins ? ' <span class="Z3988" title="'+coins.replace("&", "&amp;", "g")+'">&nbsp;</span>' : '');
 			

chrome/content/zotero/xpcom/utilities.js

@@ -203,6 +203,27 @@ Zotero.Utilities.prototype.unescapeHTML = function(/**String*/ str) {
 	return nsISUHTML.unescape(str);
 }
 
+
+/**
+ * Wrap URLs and DOIs in <a href=""> links in plain text
+ *
+ * Ignore URLs preceded by '>', just in case there are already links
+ */
+Zotero.Utilities.prototype.autoLink = function (str) {
+	// "http://www.google.com."
+	// "http://www.google.com. "
+	// "<http://www.google.com>" (and other characters, with or without a space after)
+	str = str.replace(/([^>])(https?:\/\/[^\s]+)([\."'>:\]\)](\s|$))/g, '$1<a href="$2">$2</a>$3');
+	// "http://www.google.com"
+	// "http://www.google.com "
+	str = str.replace(/([^">])(https?:\/\/[^\s]+)(\s|$)/g, '$1<a href="$2">$2</a>$3');
+	
+	// DOI
+	str = str.replace(/(doi:[ ]*)(10\.[^\s]+[0-9a-zA-Z])/g, '$1<a href="http://dx.doi.org/$2">$2</a>');
+	return str;
+}
+
+
 /**
  * Parses a text string for HTML/XUL markup and returns an array of parts. Currently only finds
  * HTML links (&lt;a&gt; tags)