raise UnicodeError, "label too long"

["Martin v. Löwis"]

Flavio schrieb:
> something like this, for instance:
> http://.wikipedia.org/wiki/Copper%28II%29_hydroxide
> 
> but even url with any non-ascii characters such as this
> 
> http://.wikipedia.org/wiki/Ammonia
> 
> also fail when passed to urlopen :
> File "/usr/lib/python2.4/encodings/idna.py", line 72, in ToASCII
>     raise UnicodeError, "label too long"
> UnicodeError: label too long
> 
> very strange, because I tried other unicode urls  from the python
> console like this

It's the host name that starts with a dot that makes it fails:

py> u".wikipedia.org".encode("idna")
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "encodings/idna.py", line 163, in encode
  File "encodings/idna.py", line 72, in ToASCII
UnicodeError: label too long
py> u"wikipedia.org".encode("idna")
'wikipedia.org'

The exception is certainly misleading; I'll have to find out
whether there is a bug beyond that (i.e. whether host names
with empty labels should be accepted).

Regards,
martin