Addressed some CR feedback.

src/compiler/emitter.ts

@@ -2211,7 +2211,7 @@ module ts {
 
             function getDoubleQuotedStringTextOfLiteral(node: LiteralExpression): string {
                 var result = escapeString(node.text);
-                result = replaceNonAsciiCharacters(result);
+                result = escapeNonAsciiCharacters(result);
 
                 return '"' + result + '"';
             }

src/compiler/scanner.ts

@@ -609,20 +609,26 @@ module ts {
             return +(text.substring(start, pos));
         }
 
+        /**
+         * Scans the given number of hexadecimal digits in the text,
+         * returning -1 if the given number is unavailable.
+         */
         function scanExactNumberOfHexDigits(count: number): number {
-            return scanHexDigits(/*minCount*/ count, /*maxCount*/ count);
+            return scanHexDigits(/*minCount*/ count, /*scanAsManyAsPossible*/ false);
         }
 
+        /**
+         * Scans as many hexadecimal digits as are available in the text,
+         * returning -1 if the given number of digits was unavailable.
+         */
         function scanMinimumNumberOfHexDigits(count: number): number {
-            return scanHexDigits(/*minCount*/ count, /*maxCount*/ undefined);
+            return scanHexDigits(/*minCount*/ count, /*scanAsManyAsPossible*/ true);
         }
 
-        function scanHexDigits(minCount: number, maxCount?: number): number {
-            var maxCountSpecified = maxCount !== undefined;
-
+        function scanHexDigits(minCount: number, scanAsManyAsPossible: boolean): number {
             var digits = 0;
             var value = 0;
-            while (!maxCountSpecified || digits < maxCount) {
+            while (digits < minCount || scanAsManyAsPossible) {
                 var ch = text.charCodeAt(pos);
                 if (ch >= CharacterCodes._0 && ch <= CharacterCodes._9) {
                     value = value * 16 + ch - CharacterCodes._0;
@@ -777,22 +783,19 @@ module ts {
                 case CharacterCodes.doubleQuote:
                     return "\"";
                 case CharacterCodes.u:
+                    // '\u{DDDDDDDD}'
                     if (pos < len && text.charCodeAt(pos) === CharacterCodes.openBrace) {
                         hasExtendedUnicodeEscape = true;
                         pos++;
                         return scanExtendedUnicodeEscape();
                     }
 
-                    // fall through
+                    // '\uDDDD'
+                    return scanHexadecimalEscape(/*numDigits*/ 4)
+
                 case CharacterCodes.x:
-                    var escapedValue = scanExactNumberOfHexDigits(ch === CharacterCodes.x ? 2 : 4);
-                    if (escapedValue >= 0) {
-                        return String.fromCharCode(escapedValue);
-                    }
-                    else {
-                        error(Diagnostics.Hexadecimal_digit_expected);
-                        return ""
-                    }
+                    // '\xDD'
+                    return scanHexadecimalEscape(/*numDigits*/ 2)
 
                 // when encountering a LineContinuation (i.e. a backslash and a line terminator sequence),
                 // the line terminator is interpreted to be "the empty code unit sequence".
@@ -810,6 +813,18 @@ module ts {
             }
         }
 
+        function scanHexadecimalEscape(numDigits: number): string {
+            var escapedValue = scanExactNumberOfHexDigits(numDigits);
+
+            if (escapedValue >= 0) {
+                return String.fromCharCode(escapedValue);
+            }
+            else {
+                error(Diagnostics.Hexadecimal_digit_expected);
+                return ""
+            }
+        }
+
         function scanExtendedUnicodeEscape(): string {
             var escapedValue = scanMinimumNumberOfHexDigits(1);
             var isInvalidExtendedEscape = false;

src/compiler/utilities.ts

@@ -1130,7 +1130,7 @@ module ts {
             newEndN = Math.max(newEnd2, newEnd2 + (newEnd1 - oldEnd2));
         }
 
-        return createTextChangeRange(createTextSpanFromBounds(oldStartN, oldEndN), /*newLength: */ newEndN - oldStartN);
+        return createTextChangeRange(createTextSpanFromBounds(oldStartN, oldEndN), /*newLength*/ newEndN - oldStartN);
     }
 
     // @internal
@@ -1213,8 +1213,12 @@ module ts {
         }
     }
 
-    var backslashOrDoubleQuote = /[\"\\]/g;
-    var escapedCharsRegExp = /[\u0000-\u001f\t\v\f\b\r\n\u2028\u2029\u0085]/g;
+    // This consists of the first 19 unprintable ASCII characters, canonical escapes, lineSeparator,
+    // paragraphSeparator, and nextLine. The latter three are just desirable to suppress new lines in
+    // the language service. These characters should be escaped when printing, and if any characters are added,
+    // the map below must be updated. Note that this regexp *does not* include the 'delete' character.
+    // There is no reason for this other than that JSON.stringify does not handle it either.
+    var escapedCharsRegExp = /[\\\"\u0000-\u001f\t\v\f\b\r\n\u2028\u2029\u0085]/g;
     var escapedCharsMap: Map<string> = {
         "\0": "\\0",
         "\t": "\\t",
@@ -1231,13 +1235,11 @@ module ts {
     };
 
     /**
-     * Based heavily on the abstract 'Quote'/ 'QuoteJSONString' operation from ECMA-262 (24.3.2.2),
-     * but augmented for a few select characters.
+     * Based heavily on the abstract 'Quote'/'QuoteJSONString' operation from ECMA-262 (24.3.2.2),
+     * but augmented for a few select characters (e.g. lineSeparator, paragraphSeparator, nextLine)
      * Note that this doesn't actually wrap the input in double quotes.
      */
     export function escapeString(s: string): string {
-        // Prioritize '"' and '\'
-        s = backslashOrDoubleQuote.test(s) ? s.replace(backslashOrDoubleQuote, getReplacement) : s;
         s = escapedCharsRegExp.test(s) ? s.replace(escapedCharsRegExp, getReplacement) : s;
 
         return s;
@@ -1254,7 +1256,7 @@ module ts {
     }
 
     var nonAsciiCharacters = /[^\u0000-\u007F]/g;
-    export function replaceNonAsciiCharacters(s: string): string {
+    export function escapeNonAsciiCharacters(s: string): string {
         // Replace non-ASCII characters with '\uNNNN' escapes if any exist.
         // Otherwise just return the original string.
         return nonAsciiCharacters.test(s) ?