chunkify with a zero-width regexp

lib/less/parser.js

@@ -120,15 +120,15 @@ less.parser = {
     //
     parse: function (str) {
         var tree;
-        input = str;
-        
-        inputLength = input.length;
-        chunks = input.split(/\n\n/g);
 
-        for (var k = 0; k < chunks.length; k++) {
-            if (k < chunks.length - 1) { chunks[k] += '\n' }
-            if (k) { chunks[k] = '\n' + chunks[k] }
-        }
+        input = str;
+        inputLength = input.length;
+
+        // Split the input into chunks,
+        // delmited by /\n\n/ (see rationale above)
+        // We use a lookahead, because we want to
+        // preserve the '\n's in the input.
+        chunks = input.split(/^(?=\n)/mg);
 
         // Start with the primary rule
         tree = new(node.Ruleset)([], $(this.parsers.primary, []));