Consider automatic trim and subset in server so reader codes can be simple

[jbfaden]

The python server should consider checking the timetags returned by the reader to see if trimming is needed. This would simply double-check that each timetag returned by the reader is after or equal to the request beginning, and before the request end. Similarly, the reader code could just return all the columns, and when the reader response is containing all columns, and not the subset of columns, then the server will also subset the response. This would allow "cat datafile.dat" to be valid reader.

[jbfaden]

It also occurred to me that this might be implemented with a native code, if Python is too slow. This would also be useful for other servers which read the output of a process.