Speeding up text processing in Python (is hard)

IIRC a interesting case in this area was the Eric Raymond attempt to migrate GCC repository from Subversion to Git with a Python-based tool he created, called Reposurgeon. After several attempts, they gave up and ended migrating the whole tool to another language: http://esr.ibiblio.org/?p=8161

In this case, the problem was less about speed and more about excessive memory load. I wonder, does the Faster CPython project also try to adress Python memory usage in addition to speed?

By the way, unfortunately it appears ESR stopped updating his blog. I enjoyed his technical posts.