scanner: More XML unicode fixes

Coerce input temporarily inside xmlwriter into Unicode to do
string manipulation, and then only convert it to "UTF8str" when
writing into the data buffer.

1 files changed, 13 insertions, 9 deletions

diff --git a/giscanner/xmlwriter.py b/giscanner/xmlwriter.py
index 84c24c0f..76880de0 100755
--- a/giscanner/xmlwriter.py
+++ b/giscanner/xmlwriter.py
@@ -95,10 +95,10 @@ class XMLWriter(object):
             tag_name, attributes, self._indent,
             self._indent_char,
             len(tag_name) + 2)
-        self.write_line('<%s%s>' % (tag_name, attrs))
+        self.write_line(u'<%s%s>' % (tag_name, attrs))
 
     def _close_tag(self, tag_name):
-        self.write_line('</%s>' % (tag_name, ))
+        self.write_line(u'</%s>' % (tag_name, ))
 
     # Public API
 
@@ -113,17 +113,19 @@ class XMLWriter(object):
     def get_xml(self):
         return self._data.getvalue()
 
-    def write_line(self, line='', indent=True, do_escape=False):
+    def write_line(self, line=u'', indent=True, do_escape=False):
+        if isinstance(line, str):
+            line = line.decode('utf-8')
+        assert isinstance(line, unicode)
         if do_escape:
-            line = escape(str(line))
-
+            line = escape(str(line)).decode('utf-8')
         if indent:
             self._data.write('%s%s%s' % (
                     self._indent_char * self._indent,
                     line.encode('utf-8'),
                     self._newline_char))
         else:
-            self._data.write('%s%s' % (line, self._newline_char))
+            self._data.write('%s%s' % (line.encode('utf-8'), self._newline_char))
 
     def write_comment(self, text):
         self.write_line('<!-- %s -->' % (text, ))
@@ -131,11 +133,13 @@ class XMLWriter(object):
     def write_tag(self, tag_name, attributes, data=None):
         if attributes is None:
             attributes = []
-        prefix = '<%s' % (tag_name, )
+        prefix = u'<%s' % (tag_name, )
         if data is not None:
-            suffix = '>%s</%s>' % (escape(data), tag_name)
+            if isinstance(data, str):
+                data = data.decode('UTF-8')
+            suffix = u'>%s</%s>' % (escape(data), tag_name)
         else:
-            suffix = '/>'
+            suffix = u'/>'
         attrs = collect_attributes(
             tag_name, attributes,
             self._indent,