Editing Dealing with binary data (section)

====Incremental parsing====

If you have to deal with a protocol which isn't length prefixed, or otherwise
chunkable, from the network then you are faced with the problem of knowing when
you have enough data to parse something semantically useful. You could run a
strict <tt>Get</tt> over what you have and catch the truncation result, but
that means that you're parsing the data multiple times etc.

Instead, you can use an incremental parser. There's an incremental version of
the <tt>Get</tt> monad in <tt>Data.Binary.Strict.IncrementalGet</tt> (you'll
need the <tt>binary-strict</tt> package).

You use it as normal, but rather than returning an <tt>Either</tt> value, you
get a [http://hackage.haskell.org/package/binary-strict/docs/Data-Binary-Strict-IncrementalGet.html#t%3AResult Result]. You need to go follow that link and look at the documentation for <tt>Result</tt>.

It reflects the three outcomes of parsing possibly truncated data. Either the
data is invalid as is, or it's complete, or it's truncated. In the truncated
case you are given a function (called a continuation), to which you can pass
more data, when you get it, and continue the parse. The continuation, again,
returns a <tt>Result</tt> depending on the result of parsing the additional
data as well.