[netperf-dev] Are you parsing "classic" or default "omni human" netperf output with scripts?

[Rick Jones]

Folks -

A request has come my way to enhance the TCP_STREAM test to include some 
latency measurements.  Specifically the idea is to measure the time it 
takes to complete the connect() call on the data connection, and also 
measure the time it takes between the shutdown() and the read return of 
zero and use those to demonstrate the degree of bufferbloat during the test.

The request further asks that this be displayed automagically in the 
default netperf output.  The idea being to make it more immediately 
clear to people whether bufferbloat is a factor in their environments.

It is that latter bit which has me dealing with a bit of FUD on whether 
or not people are trying to parse the "classic" or default "omni human" 
netperf output of a TCP_STREAM test with a script.

Thus the question - are you parsing "classic" or default "omni human" 
output formats with a script?

happy benchmarking,

rick jones