urllib.urlopen / open_http should put the
http-error-code somewhere in .info() 

the "fancy" error handling may be ok. But the errcode &
errmsg should not get lost and always be accessible ex
post.
In current style bare urllib.urlopen is hard to use.
(web pages are often down). I always need to fumble a
extra urllib.URLopener even for the simplest task.

E.g.:

f=urllib.urlopen('http://www.python.org/')
httpcode=f.info().get_errcode()