[Mnj] DOCAL bug in 31DEC13

[Eric Greisen]

There has been a bug since February in the functioning of DOCAL TRUE
when the CL table has a number of sources and gaps when antennas are
missing from the CL table followed by periods in which those antennas
are present.  My attempt to require source number matching along with
subarray and frequency ID made the long-standing bug important.  The
fixes I tried solved some issues but not enough.  The results of the
bug were visibility data that were calibrated wrongly - using values
from later in the table or something of that nature.  The fix has
required use of the NX table to define scans with the reading and use
of the CL table starting over at each scan boundary.  NOTE that CLCAL
which interpolates solutions to a CL table has no problem with these
boundaries and CL tables are required to have values at the beginning
and end of each scan as well as at intermediate points.

You should consider running a MNJ on your 31DEC13 installation if you
have installed or run a MNJ on 31DEC13 after February 19, 2013.

I made one (last I hope) fix this morning - so only MNJs run after this 
moment will be believed to have the corrected DOCAL.

Eric Greisen