Don't allow escaped surrogated identifiers + introduce ascii_identifiers

Don't use 2 characters for surrogates in identifiers because there is
support for the \u{} syntax when escaped identifiers were introduced.

Also catch eof errors while reading identifier names

Introduce ascii_identifiers:

By setting ascii_identifiers to undefined (default value),
ascii_identifiers will print identifiers using the same setting as
ascii_only within the limits of the ecmascript 6 grammar.

ascii_identifiers accept true and false, allowing identifiers to be
printed under different settings than strings with the ascii_only setting.

test/mocha/eof.js

@@ -0,0 +1,36 @@
+var assert = require("assert");
+var uglify = require("../../");
+
+describe("EOF", function() {
+    it("Should test code for at least throwing syntax error when incomplete", function() {
+        var error = function(e) {
+            return e instanceof uglify.JS_Parse_Error &&
+                /^SyntaxError: /.test(e.message);
+        }
+        var parse = function(test) {
+            return function() {
+                uglify.parse(test);
+            }
+        }
+        // Chops off 1 char at a time until limit or start of string is reached
+        // The passed code must still be valid when unchopped
+        var test_eol = function(input, chopLimit) {
+            if (chopLimit === undefined) {
+                chopLimit = input.length - 1;
+            }
+
+            assert.doesNotThrow(parse(input), "Expected valid code for \n" + input);
+
+            for (var i = input.length - 1; chopLimit > 0; chopLimit--, i--) {
+                var code = input.substr(0, i);
+                assert.throws(parse(code), error, code);
+            }
+        }
+
+        test_eol("var \\u1234", 7); // Incomplete identifier
+        test_eol("'Incomplete string'");
+        test_eol("/Unterminated regex/");
+        test_eol("` Unterminated template string`");
+        test_eol("/* Unfinishing multiline comment */");
+    });
+});

test/mocha/unicode.js

@@ -52,7 +52,11 @@ describe("Unicode", function() {
         var tests = [
             'var \\u{0} = "foo";',
             'var \\u{10ffff} = "bar";',
-            'var \\u000a = "what\'s up";'
+            'var \\u000a = "what\'s up";',
+             // Valid ID_Start, but using up 2 escaped characters and not fitting in IdentifierStart
+            'var \\ud800\\udc00 = "Hello";',
+            'var \\udbff\\udfff = "Unicode";', // Same as previous test
+            'var \\ud800\udc01 = "Weird unicode";', // Same as above, but mixed escaped with unicode chars
         ];
 
         var exec = function(test) {