Support chunked transfer encoding via a generator

With memory footprints and streaming connections in mind it would seem ideal to 
optionally have the request() method return a generator instead of the full 
response at once.

Maybe something like:

  (code, content) = http.request(uri, "GET", steaming=True)

  for chunk in content:
    print chunk

where streaming=True would indicate a generator is requested. Each chunk of a 
chunked HTTP transfer would map onto a yield in the generator.

Original issue reported on code.google.com by pieterennes on 21 Feb 2011 at 11:07

Merged into: #5

masperro / httplib2

Support chunked transfer encoding via a generator #132