Welcome to Bjorn Emonts! - Charlottesville - CASA Liason
Next two weeks - CASA Lead in-person interviews - one or two in CV, one in Socorro
CASA 5.0 inching it's way out the door - release looks solid, Juergen will deploy upon his return (if not sooner)
5.0 Has been released
5.1 First Official release notes are available. Workflow for 5.1. is on course
Pipeline does not get officially released with 5.0
CAS-10120 - Unit test for buildmytasks
Unscheduled until we can figure out a method to test CASA executables outside of CASA (e.g. tablebrowser, buildmytasks, casaviewer, casalogger, etc.)
CAS-10334 - lightweight version of mpi4casa
New branch to build and test it on, due to current workflow (though it's a minor fix)
Currently testing it on a branch now to add to master later
An issue on the OSX run, but not significant, since we don't do MPI on OSX, yet
CASAguide updates (on the ALMA side)
ALMA guides will not be updated for 5.0
Expected to be updated for 5.1
Anybody know the VLA casguide status?
CAS-10367 - return asdmSummary binary to the CASA distribution
This needs to be done before the unit test for asdmsummary can be added (CAS-9883)
Suggested to use feature or bugfix when naming a prefix for branches because those workflows are clearly defined. The others are in undefined states. Need to define a state for engineering tasks suggested to add task prefix
Final last minute testing on 5.0 on OSX. Nothing serious to stop release. Effort on pipeline testing
Statwt testing on 2nd tester.
Low on resources, looking for testers not on pipeline testing
Branch type: if one cannot use the Custom type, why does it exist in the first place? In my ticket CAS-10334 to create a new test for the system, which for me is a type "task" not feature or bugfix , I created a branch type custom and of course I don't expect Bamboo to run any tests on it. Why should I now change the branch type?
ALMA pipeline testing
as reported in CAS-10078, the convergence problem in tclean that was blocking any parallel tests of the pipeline magically disappeared. The first build that was successfull was in 17 June, casa-prerelease-5.1.0-18.el6. Federico found a suspect, which could be this commit. Can someone from tclean confirm it?
In the meantime, several datasets that previously failed are now going through the full pipeline without any errors. We will resume our batch of tests and monitor those tests in the ESO Jenkins for any possible failures in the near future.
Only 1 development ticket left : running time windows. All others under validation by B.Mason (Jen - perhaps we need someone from VLA too?)
Two addressed as 'wont fix' : CAS-7746 (special case for single channel data) and CAS-10018 (combine across spw)
Polarization Calibration : 2 dev items ongoing and nearly completed (i.e. on track, even though GM leaves for 2 weeks)
Imager ( 9 out of the 12 'Imager_BugFixes_5.1' tickets are being addressed. 0 of 2 development items have begun. )
Aftermath of backed out changes : Master OK for pipeline, Fix for unrelated bug by MP - ready for pull to master ?
Parallel Continuum : CAS-9787 : Fixed, and under validation. ( One more left : CAS-10337; SB is looking at it )
CAS-10318 : Divergence upon restart : Under investigation by UR.
Autoboxing tests : New algorithmic issues coming up. Tak working with Amanda.
KG's list (when he returns) : Speed issues with FIELD table access, re-fix of CAS-10280/10348, mosaic slowdown investigation.
Official ALMA cycle 5 testing started. Tests mostly worked. Some trouble in SB, changes to OT. Looking through data. Trouble getting Tsys into data (issue with 12m data).
CASA HPC hopes to resolve items that are blocking pipeline testing
When are the 5.2 stakeholder requirements expected to come in ? Main topics of focus ?
CASA internal stakeholders: Group has been labelling JIRA tickets to track requirements
Will Look for more detailed answers (perhaps from Juergen or Morgan) May need to draft requirements into a document and send around
Next stakeholders meeting: Anand to reply on date
Workflow for JIRA tickets internal to the team ? Not everything needs to go via stakeholder validation, or perhaps better control on what types of queries/issues actually become JIRA tickets that are tracked through all the workflow stages ?
Non-features or bugfixes that still need testing. Can move straight to ready to validate. If it doesn't required user testing group, it should be labeled a engineering task.
Should questions from users become JIRA tickets? Suggest make ticket, answer question, close ticket.
Bjorn may be new resource for deciding helpdesk tickets to JIRA workflow
Kana needs write access to CASA Guides site. To follow up with Josh Malone)
Why sometimes bamboo ignores or forgets to build the package branches (for the first time: CAS-5727, or after new commits: CAS-9089). Random glitches, or some bug that we could try to circumvent?
Perhaps Issue Matching tickets that are features but branches are bugfixes. Ville to take a look at issues with CAS-9089.
Deferred CAS-10142 to 5.2 as the pipeline doesn't need it in 5.1.
While creating a test for flagdata using a virtual MODEL column (CAS-4888), I realised that flagdata should fail when datacolumn="RESIDUAL" is requested and the MS doesn't have a MODEL or virtual MODEL column. Right now it doesn't fail because the VI2 creates a unit point source model for the calibrations. George created a ticket to take the flagdata case into account in the VI2 and I will in the meantime make flagdata fail in the absence of such column(s). CAS-10383.
Working on Plone documentation - CAS-9950: include the flagging developer notes into Plone.
plotms overlays (atmospheric transmission, sky temperature) ready for validation (CAS-9053)
investigating flag extensions and resultant changes in amplitude, probably due to vector averaging (CAS-9879)
Discussions on proposed ASDM changes. At least 2 expected for the next ALMA COMMON release. Resulting new shared/generated code should move to CASA as soon as possible (hopefully for 5.1).
asdmSummary issues. Should be fixed, not yet verified.
Various GBT issues. Some in preparation for their planned release later this summer.
Federico M Pouzols
CAS-9089 "Automatic Generation of Antenna Position Corrections for ALMA" - ready for validation. Calculation results look sensible in principle.
CAS-10283: "wt extrapolation explodes to >1e34 values" - seems fixed after the changes in mstransform regridding, but waiting for validation of CAS-10221 in CASA 5.1/master to re-validate.
CAS-10078: "ALMA pipeline triggers non-convergence of parallel-cube imaging" - testing datasets from CAS-9519 after the "magic" fix.
Some debugging for CAS-9787, CAS-10318 and failing acceptance tests (now waiting on MP's fix).
Various discussions : statwt ( with Dave and stakeholders Emmanuel and Claire ), HPC issues ( with James Robnett and Kumar ), responded to 'input required' on a couple of tickets, CARTA+InteractiveImager (with Erik/Jamie).
More summer student work. One left yesterday => some casa code work can begin now.
Looked into parallel cube tclean issue (CAS-10078)