Small change to allow get_algorithm_impls to correctly detect when lower
case algorithm strings are passed. I recently ran into a server that
sent 'md5' and so this function failed without this small change.

    def get_algorithm_impls(self, algorithm):
        # lambdas assume digest modules are imported at the top level
        if algorithm.lower() == 'md5':
            H = lambda x: hashlib.md5(x).hexdigest()
        elif algorithm.lower() == 'sha':
            H = lambda x: hashlib.sha1(x).hexdigest()
        ...

Function is about 2/3rds down in urllib2.

Note 1: hashlib checks upper and lower case versions of names.
So this seems like reasonable request.

Note 2: I would start hashlib.__get_builtin_constructor(name):
with 'name = name.lower()'
and replace first two 'name in' constructions with 'name =='
and shrink 2nd two lists.

Note 3. I would do same with get_algorithm_impls():
   algorithm = algorithm.lower()
Instead of possibly calling .lower twice.

Note 4: I consider 'name = lambda ...' inferior to 'def name...'
but I am not the one who will edit this.

Keyword says patch, but I do not see one.

The patch is inline. There's not much to it :-)

Agree with your suggestion to avoid calling lower() twice.

RFC2617 says that the authentication scheme should be case insensitive.
 Included is a patch which changes the string to uppercase before
comparison.

Looks like a sensible, simple fix to me :-)

Applied in r62713.