Hello,

the urlparse() function accepts a parameter default_scheme, to be used
if the address given does not contain one, but I cannot make
use of it, because I would expect these two returning
identical values:

>>> from urlparse import urlparse
>>> urlparse("www","http")
('http', '', 'www', '', '', '')
>>> urlparse("http://www","http")
('http', 'www', '', '', '', '')

This has been reported about six years ago but apparently
the behaviour hasn't changed.  I cannot imagine that this
really is the intended behaviour.

Regards,

    pk

and this is the url to the old report:

http://mail.python.org/pipermail/python-list/2002-August/157171.html

In http://bugs.python.org/issue754016 there's already a discussion about
this.
The RFC that urlparse is following (rfc 1808) requires the net_loc
component to start with // even if the scheme component is missing,
which is why urlparse("www","http") puts the 'www' in to the path
component instead of net_loc.

It seems that this is indeed the intended behavior, and the patch for
issue 754016 adds a docfix clarifying this.

Thanks (pk) and Anthony!