Older automation harnesses get their results from binary-search.py
by reading a small JSON dump at the end of the output that has data
for each device, so set the default output mode to this ("device")
so that those tools continue to work
For newer harnesses this output is largely irrelevant and at worse
confusing (it potentially adds a bunch of useless output at the end
of the output) since they are directly loading the
binary-search.json output file and able to access all of the data
from a run from it. For these use cases we now allow this output to
be suppressed (using "none"). This mode is also set as an
"essential" parameter in multiplex.json so that the
Crucible/rickshaw stack always operates this way.
Older automation harnesses get their results from binary-search.py by reading a small JSON dump at the end of the output that has data for each device, so set the default output mode to this ("device") so that those tools continue to work
For newer harnesses this output is largely irrelevant and at worse confusing (it potentially adds a bunch of useless output at the end of the output) since they are directly loading the binary-search.json output file and able to access all of the data from a run from it. For these use cases we now allow this output to be suppressed (using "none"). This mode is also set as an "essential" parameter in multiplex.json so that the Crucible/rickshaw stack always operates this way.