Improve multi-line comment parsing

* Make sure comments are skipped correctly with surrogates
* Fix regression in multiline comments with nlb

lib/parse.js

@@ -148,6 +148,21 @@ function get_full_char_code(str, pos) {
     return str.charCodeAt(pos);
 }
 
+function get_full_char_length(str) {
+    var surrogates = 0;
+
+    for (var i = 0; i < str.length; i++) {
+        if (str.charCodeAt(i) >= 0xd800 && str.charCodeAt(i) <= 0xdbff) {
+            if (str.charCodeAt(i + 1) >= 0xdc00 && str.charCodeAt(i + 1) <= 0xdfff) {
+                surrogates++;
+                i++;
+            }
+        }
+    }
+
+    return str.length - surrogates;
+}
+
 function from_char_code(code) {
     // Based on https://github.com/mathiasbynens/String.fromCodePoint/blob/master/fromcodepoint.js
     if (code > 0xFFFF) {
@@ -493,8 +508,9 @@ function tokenizer($TEXT, filename, html5_comments, shebang) {
         var i = find("*/", true);
         var text = S.text.substring(S.pos, i).replace(/\r\n|\r|\u2028|\u2029/g, '\n');
         // update stream position
-        forward(text.length /* doesn't count \r\n as 2 char while S.pos - i does */ + 2);
+        forward(get_full_char_length(text) /* text length doesn't count \r\n as 2 char while S.pos - i does */ + 2);
         S.comments_before.push(token("comment2", text, true));
+        S.newline_before = S.newline_before || text.indexOf("\n") >= 0;
         S.regex_allowed = regex_allowed;
         return next_token;
     });

test/compress/html_comments.js

@@ -69,3 +69,12 @@ html_comment_in_string_literal: {
     }
     expect_exact: 'function f(){return"\\x3c!--HTML--\\x3ecomment in\\x3c!--string literal--\\x3e"}';
 }
+
+html_comment_after_multiline_comment: {
+    input: {
+        var foo; /*
+*/-->   var bar;
+        var foobar;
+    }
+    expect_exact: "var foo;var foobar;"
+}

test/mocha/comment.js

@@ -31,7 +31,8 @@ describe("Comment", function() {
             "/*Some comment 2\r\n\r\n\r\n*/\r\n>\n\n\n\n\n\n",
             "/*Some comment 3\r\r\r*/\r>\n\n\n\n\n\n",
             "/*Some comment 4\u2028\u2028\u2028*/\u2028>\n\n\n\n\n\n",
-            "/*Some comment 5\u2029\u2029\u2029*/\u2029>\n\n\n\n\n\n"
+            "/*Some comment 5\u2029\u2029\u2029*/\u2029>\n\n\n\n\n\n",
+            "/*Some comment 6\udbff\udfff\udbff\udfff\n\n\n*/\n>\n\n\n\n\n"
         ];
 
         var fail = function(e) {
@@ -43,7 +44,7 @@ describe("Comment", function() {
 
         for (var i = 0; i < tests.length; i++) {
             assert.throws(function() {
-                uglify.parse(tests[i], {fromString: true})
+                uglify.parse(tests[i], {fromString: true});
             }, fail, tests[i]);
         }
     });