Pull request #1004: CAS-14567
Merge in CASA/casa6 from CAS-14567 to master
* commit '60f42dd7b7b7bef2e2251f7c4980a098d3c27e2f':
For CAS-14567, fixed one final(?) indexing error in SolvableVisJones::listCal for the case of field_id=-1; Linux seems to be stricter about this than MacOS
For CAS-14567, re-enable test_gainCurveVLBA now that ba123a.gc in casatestdata repo updated for new listcal tests
For CAS-14567, added test_gaincurveListing to test_task_listcal.py to test VLBA gaincurve listcal ouput and other refinements to listcal tests; refined file naming semantics in gaincurve-related tests in test_task_gencal.py so that the testing intent is clearer. NB: tests involving casatestdata/table/ba123a.gc (VLBA gaincurve) await an update to this table on the DR.
For CAS-14567, in SolvableVisJones::listCal, repaired/refined formatting of field names in the listing, including retrieval of field and antenna names from the caltable rather from the MS (which seemed to be broken for non-trivial field cases); to test_task_listcal.py, added a test of the fringefit caltable listing header, and updated other tests for consistency with format changes.
For CAS-13619, in VisEquation.h, make nVisTotal variabla an Int64, to avoid overflow when accumulating the number is atomic visibilities corrected by applycal.
For CAS-14567, in SolvableVisCal.cc added "uHz" units for dispersive delay in listcal listing for fringefit caltables; also minor general formatting refinements.
For CAS-14567, in EJones.cc, EPowerCurve::specify(), reshape gain coeffiencient Array to have a degenerate channel axis, so that resulting caltable is generally conformant (FPARAM col has shape [npar,nchan,nrow]). In SolvableVisCal::listCal() and ::writeHeader(), various improvements to listcal formatting to better support rendering EPowerCurve listings, and added proper formatting support for fringefit caltables, including some scaling/unit juggling so the printed values are sensible. Also cleared up some (otherwise inconsequential) compiler warnings in SolvableVisCal.cc.
For CAS-14567, fixed one final(?) indexing error in SolvableVisJones::listCal for the case of field_id=-1; Linux seems to be stricter about this than MacOS
Akeem Wells
Automatically assigned