Allow higher unicode characters in XMB files.

grit/tool/xmb.py

Index: grit/tool/xmb.py
diff --git a/grit/tool/xmb.py b/grit/tool/xmb.py
index aaefeecad4b54b554402fb21049d89f22ea30072..0e7950ccde1e237eec79b5e115ff9d7e95ce26f1 100644
--- a/grit/tool/xmb.py
+++ b/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:

[newt (away), 2015/11/10 17:22:52]

All of Chrome's grd files pass this stricter error checking (as you'd hope!).
It's possible that some other project contains illegal characters but I think
they'd rather know about that just continue to get translations with missing
characters.

+    raise Exception('String contains characters disallowed in XML: %s' %
+                    repr(result))
+  return result.encode('utf-8')
 
 
 def _WriteAttribute(file, name, value):