Merge pull request #344 from alphatownsman/emsp

keep leading unicode spaces like emsp (full-width space)

src/mistune/markdown.py

@@ -53,7 +53,8 @@ def _iter_render(self, tokens, state):
             elif 'text' in tok:
                 text = tok.pop('text')
                 # process inline text
-                tok['children'] = self.inline(text.strip(), state.env)
+                # avoid striping emsp or other unicode spaces
+                tok['children'] = self.inline(text.strip(' \r\n\t\f'), state.env)
             yield tok
 
     def parse(self, s: str, state: Optional[BlockState]=None):

tests/test_misc.py

@@ -96,3 +96,10 @@ def test_ast_output(self):
             },
         ]
         self.assertEqual(result, expected)
+
+
+    def test_emsp(self):
+        md = mistune.create_markdown(escape=False, hard_wrap=True)
+        result = md('\u2003\u2003foo\nbar\n\n\u2003\u2003foobar')
+        expected = '<p>\u2003\u2003foo<br />\nbar</p>\n<p>\u2003\u2003foobar</p>'
+        self.assertEqual(result.strip(), expected)