upgrading >v4 header cells to unicode

adding test

IPython/nbformat/v4/convert.py

@@ -91,7 +91,7 b' def upgrade_cell(cell):'
     elif cell.cell_type == 'heading':
         cell.cell_type = 'markdown'
         level = cell.pop('level', 1)
-        cell.source = '{hashes} {single_line}'.format(
+        cell.source = u'{hashes} {single_line}'.format(
             hashes='#' * level,
             single_line = ' '.join(cell.get('source', '').splitlines()),
         )

IPython/nbformat/v4/tests/test_convert.py

@@ -1,3 +1,4 b''
+# -*- coding: utf-8 -*-
 import copy
 
 import nose.tools as nt
@@ -33,6 +34,10 b' def test_upgrade_heading():'
             v3h(source='foo\nbar\nmulti-line\n', level=4),
             v4m(source='#### foo bar multi-line'),
         ),
+        (
+            v3h(source='unicode–conversion', level=4),
+            v4m(source=u'#### unicode–conversion'),
+        ),
     ]:
         upgraded = convert.upgrade_cell(v3cell)
         nt.assert_equal(upgraded, expected)