Use Uglify's ascii_only to avoid converting UTF-8 escapes.

Thanks @curiousdannii!

grunt.js

@@ -112,7 +112,11 @@ module.exports = function( grunt ) {
 			],
 			tasks: "dev"
 		},
-		uglify: {}
+		uglify: {
+			codegen: {
+				ascii_only: true
+			}
+		}
 	});
 
 	// Default grunt.

src/core.js

@@ -605,8 +605,7 @@ jQuery.extend({
 	},
 
 	// Use native String.trim function wherever possible
-	// Uglify converts escapes to UTF-8; this circumvents (#12690)
-	trim: core_trim && !core_trim.call( String.fromCharCode( 0xFEFF, 0xA0 ) ) ?
+	trim: core_trim && !core_trim.call("\xFEFF\xA0") ?
 		function( text ) {
 			return text == null ?
 				"" :