On Sat, 2015-08-22 at 00:24 +0200, Volker Klasen wrote: > Hi Ian, > > you can use the patch I posted in the original report,
Uh, how on earth did I miss it! > there are two > places that need to be fixed. I'm using it since November. Thanks. Since loads is a classmethod I think the right fix for that issue is to s/self/cls/ in the body rather than s/cls/self/ in the declaration. I've updated the feature/bug763102 with the following extra patch: @@ -14,7 +14,7 @@ class SXPParser(object): def loads(cls, input): data = [] stack = [] - for match in self.tokenizer_re.finditer(input): + for match in cls.tokenizer_re.finditer(input): if match.group('open'): stack.append([]) elif match.group('close'): Ian.