I wrote:
Additionally, if there is a CONTENT-LENGTH specified for non-binary data, the read_binary method should be used for it also, because it is much more efficient than readlines... less scanning of the data, and fewer outer iterations.  This goes well with the technique of leaving that data in binary until read from the file.

I further elucidate:
Sadly, while the browser (Firefox) seems to calculate an overall CONTENT-LENGTH for the HTTP headers, it does not seem to calculate CONTENT-LENGTH for individual parts, not even file parts where it would be extremely helpful.