bufsock.py 1.1

bufsock.py is a python module that makes it a little bit easier to work with sockets, and may also make your I/O faster if youre reading and/or writing lots of tiny packets.

Also, you may find that it makes your network applications more reliable, as the network is allowed to split apart your packets into multiple smaller packets, or aggregate two or more packets into one larger packet, if it decides it needs to do so for reliability or performance.

So you might do a s.recv expecting to get one line of input, but the network decided to batch together two remote s.sends into one s.recv, for example. But the network usually, but not always, will batch up your packets the way youd expect - so this can be a difficult problem to track down.

It supports the following methods:

read(length) brings in a specific number of bytes.
readto(char) reads up thru the next occurence of char
readtomax(char,length) reads up thru the next occurence of char, or length bytes, whichever is less
set_chunk_len(length) says "do reads and writes in increments of length". The chunk length defaults to 4096. You should make this larger on networks that employ jumbo frames to squeeze out more performance!
send(buf) writes the bytes in buf.
flush() is just like stdios (the C librarys) flush function. Call it when you want an output buffer flushed (written immediately).
shutdown(v) just like for a regular socket, except it flushes first.

Usage:

Code snippet:
sock = socket.socket(socket.AF_INET,socket.SOCK_STREAM)
sock.connect((hostname,port))
bs = bufsock.bufsock(sock)
version = bs.readto(n)
print version