Pull request #689: CAS-14038
Merge in CASA/casa6 from CAS-14038 to master
* commit '9a9320d45cac6c6529ab56530d5c2430675f61bc': (50 commits)
For CAS-14038, improve concatspws xml doc info, and remove some diagnostic console messages CAS-13266: Explicitly pass MS name instead of reconstructing it CAS-13266: Handle Multi-MS properly when reconstructing MS name
For CAS-14038, fixed errors in conflict resolutions for merges from CAS-12765 and CAS-13361
For CAS-12765, fixed calibrater_cmpt.cc merge-from-master conflict.
For CAS-13266: In MSMetaInfoForCal.cc, restore full path to formation of MS name, because it is needed elsewhere, and without it, causes MMS operations to fail.
The range for test_task_listcal test_small_gcal_sel_ant was adjusted. The results before this branch were very close to the minimum threshold already, and changes on the branch cause the file path to be ommited. This causes the filesize to fall below the treshold and fail. Since the changes to the written file name appear intentional then changing the lower bound seems appropriate.
For CAS-13266: Moved logging of missing calibration from caltable read / interpolation prep stages to interpolation calculation stage to avoid generating spurious log messages (e.g., cal missing for unselected data)
Added a new row to the observation table in the test_selectVisObs. Missing this would cause a seg-fault
For CAS-13266, in Calibrater.cc and MSMetaInfoForCal.cc, ensure that the MSMetaInfoForCal object used in constructing VisCals is constructed from the _selected_ MS, so that CTPatchedInterp considers only relevant data when evaluating calibration availability. Also temporarily disabled a broken test in test_tool_calibrater.py.
For CAS-13266, in CTPatchedInterp, improved byObs/byScan semantics to refer to timeSegs generically, to avoid confusion. Also added a logger warning for when perbos and perscan are both specified (perscan pertains)
removed CASA5 related code and imports
re-added the new class to the return
re-formatted the newly added test case
Added new testcase to fringefit for combining pols
Changed the correlation combination parameter to a string corrcomb, set as "none" by default
Reconciled casatasks/src/private/task_fringefit.py and casa5/gcwrap/python/scripts/task_fringefit.py
Provisional version of DFT wideband
An updated version of the wideband fringefitter, with a DFT used to stack SPWs
Added a polarization combination feature to the fringefit task
...
Pull request #688: First version of the new regression test script based on the CASAguides on
Merge in CASA/casa6 from CAS-14076 to master
* commit '51eb3fbe73aeebe723e9ef3bb0799d287a125c6e':
Added new regression script to component_to_test_map for running automatically in Bamboo.
First version of the new regression test script based on the CASAguides on simulation using component lists. This test script leaves four png files in disk, which are to be visualized, if needed, when running the tests on release branches.
Pull request #687: CAS-13191
Merge in CASA/casa6 from CAS-13191 to master
* commit 'e74fb9a6daf174dc71f117c76c0add8e0cc8222f': (31 commits)
Fixed a problem when cfcache='' and running with awp gridder in serial mode
remove a race condition for making cfcache. Also fixed a problem in parallel where awproject does not transport the telecopename across properly...for now in those cases assuming it is the default EVLA
Fix a cfcache create problem for cube imaging and having multiprocessing
removed some debugging print outs
Making sure that cfs are even numbered pixels
More odd number origin error correction
removed some print outs
enable wterm_awproject test
one more place where oversampling has to be dealt with
another place where center can be wrong when odd number of pixels are chosen for CF
Fixed the w-term problem of awp w.r.t a pixel off
Try to avoid loading the weight image a lot
counters were not initialized in some cases
Put some timers on AWProjectFT
Brought fixed from CAS-13518 here
recover 13212
some changes from CAS-13581 CAS-13212 : Enabled awproject in task_sdintimaging and added 3 verification tests toexercise it and make comparisons with tclean for the equivalent intonly runs for cube and mtmfs(with cube major cycles).
Remove print statement and trigger build.
some fixes to get unittest to pass
...
