From e26fe9078e8879006fdc4be1f34983d084ccf114 Mon Sep 17 00:00:00 2001
From: Alex Kocharin <alex@kocharin.ru>
Date: Wed, 13 May 2015 14:29:47 +0300
Subject: [PATCH] Process astral characters correctly

According to YAML 1.2 spec, #xD800-#xDFFF unicode range is considered
non-printables. But unicode range #x10000-#x10FFFF, represented
in UTF-16 by the same characters is not.

NOTE: Non-printable characters are still allowed in quoted scalars,
added a pending test for this case.

Fixes: https://github.com/nodeca/js-yaml/issues/191
---
 lib/js-yaml/loader.js       |  2 +-
 test/units/character-set.js | 25 +++++++++++++++++++++++++
 2 files changed, 26 insertions(+), 1 deletion(-)
 create mode 100644 test/units/character-set.js

diff --git a/lib/js-yaml/loader.js b/lib/js-yaml/loader.js
index af1b99b3..1012ff56 100644
--- a/lib/js-yaml/loader.js
+++ b/lib/js-yaml/loader.js
@@ -23,7 +23,7 @@ var CHOMPING_STRIP = 2;
 var CHOMPING_KEEP  = 3;
 
 
-var PATTERN_NON_PRINTABLE         = /[\x00-\x08\x0B\x0C\x0E-\x1F\x7F-\x84\x86-\x9F\uD800-\uDFFF\uFFFE\uFFFF]/;
+var PATTERN_NON_PRINTABLE         = /[\x00-\x08\x0B\x0C\x0E-\x1F\x7F-\x84\x86-\x9F\uFFFE\uFFFF]|[\uD800-\uDBFF](?![\uDC00-\uDFFF])|(?:[^\uD800-\uDBFF]|^)[\uDC00-\uDFFF]/;
 var PATTERN_NON_ASCII_LINE_BREAKS = /[\x85\u2028\u2029]/;
 var PATTERN_FLOW_INDICATORS       = /[,\[\]\{\}]/;
 var PATTERN_TAG_HANDLE            = /^(?:!|!!|![a-z\-]+!)$/i;
diff --git a/test/units/character-set.js b/test/units/character-set.js
new file mode 100644
index 00000000..86116ca9
--- /dev/null
+++ b/test/units/character-set.js
@@ -0,0 +1,25 @@
+'use strict';
+
+
+var assert = require('assert');
+var yaml = require('../../');
+
+
+test('Allow astral characters', function () {
+  assert.deepEqual(yaml.load('𝑘𝑒𝑦: 𝑣𝑎𝑙𝑢𝑒'), { '𝑘𝑒𝑦': '𝑣𝑎𝑙𝑢𝑒' });
+});
+
+test('Forbid non-printable characters', function () {
+  assert.throws(function () { yaml.load('\x01'); }, yaml.YAMLException);
+  assert.throws(function () { yaml.load('\x7f'); }, yaml.YAMLException);
+  assert.throws(function () { yaml.load('\x9f'); }, yaml.YAMLException);
+});
+
+test('Forbid lone surrogates', function () {
+  assert.throws(function () { yaml.load('\udc00\ud800'); }, yaml.YAMLException);
+});
+
+test.skip('Allow non-printable characters inside quoted scalars', function () {
+  assert.strictEqual(yaml.load('"\x7f\x9f\udc00\ud800"'), '\x7f\x9f\udc00\ud800');
+});
+