Allow higher unicode characters in XMB files.

The XMB tool has a regex of invalid XML characters, which erroneously
contained all unicode characters in the supplementary planes (U+10000 to
U+10FFFF). The tool would silently replace these characters with spaces
when generating XMB files, which caused problems recently when an emoji
character was added to a grd file. The translation console supports
these characters, so GRIT should too.

The XMB tool now supports these characters, and now raises an exception
if an invalid character is used (instead of silently replacing it with a
space).

BUG=498288

Review URL: https://codereview.chromium.org/1440313003

Cr-Commit-Position: refs/heads/master@{#359621}

2 files changed, 16 insertions, 5 deletions

diff --git a/tools/grit/grit/tool/xmb.py b/tools/grit/grit/tool/xmb.py
index aaefeec..0e7950c 100755
--- a/tools/grit/grit/tool/xmb.py
+++ b/tools/grit/grit/tool/xmb.py
@@ -28,8 +28,10 @@ _XML_QUOTE_ESCAPES = {
     u"'":  u''',
     u'"':  u'"',
 }
+# See http://www.w3.org/TR/xml/#charsets
 _XML_BAD_CHAR_REGEX = lazy_re.compile(u'[^\u0009\u000A\u000D'
-                                      u'\u0020-\uD7FF\uE000-\uFFFD]')
+                                      u'\u0020-\uD7FF\uE000-\uFFFD'
+                                      u'\U00010000-\U0010FFFF]')
 
 
 def _XmlEscape(s):
@@ -40,7 +42,11 @@ def _XmlEscape(s):
   if not type(s) == unicode:
     s = unicode(s)
   result = saxutils.escape(s, _XML_QUOTE_ESCAPES)
-  return _XML_BAD_CHAR_REGEX.sub(u'', result).encode('utf-8')
+  illegal_chars = _XML_BAD_CHAR_REGEX.search(result)
+  if illegal_chars:
+    raise Exception('String contains characters disallowed in XML: %s' %
+                    repr(result))
+  return result.encode('utf-8')
 
 
 def _WriteAttribute(file, name, value):
diff --git a/tools/grit/grit/tool/xmb_unittest.py b/tools/grit/grit/tool/xmb_unittest.py
index 10f81d7..df8e84b 100755
--- a/tools/grit/grit/tool/xmb_unittest.py
+++ b/tools/grit/grit/tool/xmb_unittest.py
@@ -37,18 +37,23 @@ class XmbUnittest(unittest.TestCase):
             <message name="IDS_BONGOBINGO">
               Yibbee
             </message>
+            <message name="IDS_UNICODE">
+              Ol\xe1, \u4eca\u65e5\u306f! \U0001F60A
+            </message>
           </messages>
           <structures>
             <structure type="dialog" name="IDD_SPACYBOX" encoding="utf-16" file="grit/testdata/klonk.rc" />
           </structures>
         </release>
-      </grit>'''), '.')
+      </grit>'''.encode('utf-8')), '.')
     self.xmb_file = StringIO.StringIO()
 
   def testNormalOutput(self):
     xmb.OutputXmb().Process(self.res_tree, self.xmb_file)
-    output = self.xmb_file.getvalue()
-    self.failUnless(output.count('Joi') and output.count('Yibbee'))
+    output = self.xmb_file.getvalue().decode('utf-8')
+    self.failUnless(output.count('Joi'))
+    self.failUnless(output.count('Yibbee'))
+    self.failUnless(output.count(u'Ol\xe1, \u4eca\u65e5\u306f! \U0001F60A'))
 
   def testLimitList(self):
     limit_file = StringIO.StringIO(