[evlatests] Modcomp free pointing

[Ken Sowinski]

There was a five hour pointing run last night, but no data was
produced even though the antennas appeared to have been controlled
properly.

The executor log shows that the reason is that at the switchover
to Modcompless observing the script modcorr, which tells the
executor to let the Modcomps control the executor and produce
data, was run instead of execorr, which tells the executor to
control the correlator and idcaf to produce data.

The switch back to Modcomp control at the end of the pointing
run appears to have succeeded as there is reasonable looking
data in the archive starting around 1830 LST which was to have
been the end of the pointing run.