chore(_ext_link_shadow): do not create memoryview of whole string

To same memory in rare cases that a lot of external_links are
created.

wikitextparser/_wikitext.py

@@ -1190,23 +1190,25 @@ def get_italics(self, recursive=True) -> List['Italic']:
         )
 
     @property
-    def _ext_link_shadow(self) -> memoryview:
+    def _ext_link_shadow(self) -> bytearray:
         """Replace the invalid chars of SPAN_PARSER_TYPES with b'_'.
 
         For comments, all characters are replaced, but for ('Template',
         'ParserFunction', 'Parameter') only invalid characters are replaced.
         """
         ss, se, _, _ = self._span_data
-        byte_array = bytearray(self._lststr[0], 'ascii', 'replace')
+        byte_array = bytearray(self._lststr[0][ss:se], 'ascii', 'replace')
         subspans = self._subspans
         for s, e, _, _ in subspans('Comment'):
-            byte_array[s:e] = (e - s) * b'_'
+            byte_array[s - ss : e - ss] = (e - s) * b'_'
         for s, e, _, _ in subspans('WikiLink'):
-            byte_array[s:e] = (e - s) * b' '
+            byte_array[s - ss : e - ss] = (e - s) * b' '
         for type_ in 'Template', 'ParserFunction', 'Parameter':
             for s, e, _, _ in subspans(type_):
-                byte_array[s:e] = INVALID_EXT_CHARS_SUB(b' ', byte_array[s:e])
-        return memoryview(byte_array)[ss:se]
+                byte_array[s - ss : e - ss] = INVALID_EXT_CHARS_SUB(
+                    b' ', byte_array[s:e]
+                )
+        return byte_array
 
     @property
     def external_links(self) -> List['ExternalLink']: