Monday Morning Meeting - 24 August 2015

• DIAL-IN NUMBERS & PASSCODES:
• USA Toll Free Number: 866-901-8266
• USA Toll Number: +1-203-566-3863
• PARTICIPANT PASSCODE: 725760

Attendance

• Socorro: Jeff, Martin, Lindsey, Jim, Susan, Sanjay, Urvashi, Tak, Pam
• CV: Darrell, Ville, Andy, Akeem, Dave
• Garching: Julian, Dirk
• SCO:

News / Meetings / Visitors

Build, Release, Testing

• CleanChan test is failing
• All of tests are broken due to a checkin, will be reverted
  • tClean: CAS-7664, one blocker. Making a PB for non projection
    • Urbashi working on it
  • Bug w/ Mask
• SetJy and ephemeris objects
• Question about BDF flags, brought in Michel's changes. One problem, issue with ASDM.
• Andy:
  • Questions about google test / google Mock
  • VM Accepted tests are coming back
  • CASA Guides are coming up
• WidebandMosaic still failing; Sanjay to follow up

Architecture

• Technical Forums:

HPC

Development

Pipeline

• Development
  • Completed first version of standard ALMA interferometry imaging recipe
    • Includes combined (per band) spw continuum image, per band / per spw continuum image, per spw line cube
    • Includes automated continuum region finding algorithm from Todd and support for user defined continuum regions
• CASA 4.5
  • Tested ALMA pipeline support iof ephemeris table calibrator (planets and moons) and science target objects
  • Discovered CASA setjy ephemeris table bug
  • Confirmed that the ALMA pipeline can mosaic ephemeris table science targets correctly

AOB

Developer Reports

Monday Meeting

• Jeff -- reverse order!
• Sanjay Bhatnagar
• Sandra Castro
• Lindsey Davis
  • Small changes to the AQUA report generator
  • Ephemeris table support discussions with Dirk P., Todd, and others
  • Ephemeris table / setjy integration discussions with Dirk P., Todd, Crystal, Tak
  • Pipeline testing against CASA 5.5
  • CASA / Pipeline consultancy followup
• Pam Ford
  • Built release-4_5 branch
  • CAS-5730 - negative amplitudes not plotted for FLOAT_DATA - it was actually DATA, not a bug
  • CAS-7733 - shared axis/tarball seg fault - attempts to fix in the code were unsuccessful. Seg faults were in Qwt/Qt code so I looked into library paths again and noticed that libcasaqt.so in the tarball is dependent on /usr/lib64/casa/01/lib/libqwt.so.5.2.3 which is dependent on /usr/lib64/casa/01/lib/libQtCore.so.4.8.5, but at runtime in gdb the loaded libs ("info share") were in casa-test-4.6.1/lib not in /usr/lib64. When I set gdb to look for dynamic libs in the tarball's lib dir first, the plot completed. Sent email to B&T group.
  • CAS-7831 blocker RTT - incorrect virtual model - asked Kumar to look at the synthesis code and he found the issue
  • Fixed bug determining datacolumn in plotms when axes change
  • Continued work on improving test_plotms
• Kumar Golap
  • Fixed a couple of bugs related to virtual model data column
  • Changed clean behavior of default phasecenter with ms's that have field ids no in the ms
  • Went through the list of Jira tickets that Rafael assigned to me regarding VLASS and gave some (guess) time estimates and comments.
• Justo González Villalba
• Jim Jacobs
• Jeff Kern
  • 2 Days with Eduardo
  • CASA User Survey Response
  • Beginning Documentation Planning
• Susan Loveland
  • CARTA
    • Finished initial pass at contour work.
    • Miscellaneous bug fixes (selenium tests and code).
  • CAS-7329: Casacore Lattice Histograms gives a segfault when setting range outside of data range (tested)
• David Mehringer
  • Attended two hour CASA/Pipeline Consultancy "Workshop". Agreed with conclusions that dearth of well defined processes and disempowerment of developers to affect change are two major issues that the project has faced for quite some time.
  • Repaired test_boxit/boxit issue exposed by change of behavior in SpectralCoordinate::frequencyToVelocity(). Requested that consideration be given to add an extra optional parameter to this method to indicate that the caller would like an exception to be thrown in exceptional cases rather than having False returned. Such a modification would not break backward compatibility, and would provide more intuitive C++ behavior in exceptional situations IMO, and would mean callers wouldn't have to handle False returns and throw exceptions themselves, since one would normally just want to have an exception thown in the first place that they didn't have to explicitly handle. Request denied. IMO, this is an example of the much larger casacore maintenence and technical debt group of issues that exist that I find frustrating. My perception is that no one "in charge" of casacore these days seems to care about such issues, or even if they did, such issues couldn't be adequately addressed because the lack of governance and lack of cooperative decision making processes necessary for repairing and redesigning problematic parts of the casacore architecture.
  • (CAS-7770) msmd.sourcedirs returns 0,0 for ephemeris sources: More conversation about what is required.
  • (CAS-7622) msmd.*forscan(s) methods should default to "any" obsid: Completed implementation, RTT now.
  • (CAS-7523) MSMetaData::getNetSidedbands() should return actual values from MS: Made necessary casacore change, submitted pull request, still waiting for that to be merged before I can commit remaining gcwrap code. Merge delays such as this are one reason my productivity has been significantly and negatively impacted since the implementation of the new casacore management model.
  • Investigated and resolved intermittent OSX test_immoments failure. Most of the time was spent watching compile messages scroll slowly by.
  • Discovered that svn -r{DATE/TIME} does not work on the github implementation of the svn server. Added this to the list of things that continues to have significant negative impact on my productivity since the casacore unification effort began.
• George Moellenbrock
• Dirk Petry
  • ephemeris handling related work:
    • looking into CAS-7821 (Check that setjy code does not access the field table position columns directly)
      • the fact that also setjy is accessing the FIELD table and needs to be updated for the (relatively) new measures way of accessing, had fallen through the cracks
      • looking into this, the solution seems simple: make all retrievals of the phase center position use msmd.phasecenter()
        • waiting for Tak to apply this fix
      • furthermore, the ephemerides for fast moving objects should get higher time resolution, i.e. at ca. 0.1 days instead of 1 day.
    • resolved CAS-7270 (doc update for importasdm), and CAS-7789 (importasdm leaves ephemerides table open in cache)
  • CAS-7773: General BDF flagging work tracking ticket
    • merged last fixes from Michel into trunk and release branch
  • resolved (CAS-7327) SpectralCoordinate::frequencyToVelocity() returns nan's when rest frequency not set (rest freq = 0)
• Martin Pokorny
  • Further integration of code for updated visstat.
  • Looking at replacing LargeFiledesIO with FiledesIO in AsdmStMan.
  • [Still struggling with MPI-IO bug for VLA.]
• Urvashi Rao
  • Several minor edits to tclean for CAS-7664 (tests for ALMA use of tclean)
  • Worked a bit on VLASS tickets sent by Rafael + a bit on white book chapters.
• Darrell Schiebel
• Ville Suoranta
• Julian Taylor
• Takahiro Tsutsumi
  • Fixed a bug on cube tclean (the offset on the start velocity)
  • Discussion on setjy issue on the MS with attached ephemeris table (CAS-7821). Will check in the fix this week
  • User support/science time - some queries from a grad student at NAOJ on VLA S-band data reduction, looking into some issues on ALMA Cycle 2 data

Friday NAOJ Meeting

• Kanako Sugimoto
• Wataru Kawasaki
• Masaya Kuniyoshi
• Takeshi Nakazato
• Renaud Miel