* Search method now returns a list instead of an iterator. This allows us to catch maxrows internally and automatically
make subsequent reqeusts with offsets rather than forcing the client to catch and adapt. As the size of the reply is often
small, getting even ~32k listings of data was >300k in memory, this results should not affect the memory footprint.
* Added the `auto_offset` parameter for Session.search method. Defaults to True for making subsequent search requests
if the RETS server truncated the number of requested listings.