Monday Morning Meeting Nov 10, 2014
- DIAL-IN NUMBERS & PASSCODES:
- USA Toll Free Number: 866-901-8266
- USA Toll Number: +1-203-566-3863
- PARTICIPANT PASSCODE: 725760
Attendance
- Socorro: Jeff, Rob, George, Susan, Tak, Kumar, Jim, Lindsey
- CV: Mark, Dave, Andy, Darrell
- Garching: Sandra, Justo
- SCO:
News / Meetings / Visitors
- Kickoff of Alma Study in Europe
- Magix is a fitting code
- Artist
- CASA Developer Mailing List
- Pre-releases announced here
Build, Release, Testing
4.3 Release
- BDF Flagging:
- importASDM: Failing on EL6 talk to Michel
- WeightSpectrum infrastructure is in place, modify to take flags into account.
Architecture
- Technical Forums:
- Jim working on preferences
Development
Pipeline
* Testing with 4.3
* What we can do for November release
AOB
Developer Reports
Monday Meeting
- Sanjay Bhatnagar
- Sandra Castro
- Followed up on split2 and cvel2 users testing. Removed excessive stacktrace messages from mstransform/cvel2.
- Fixed a small bug in the way the temporary MMS directory is created. Fixed in release and trunk branches.
- Started to use pipeline data sets for MMS testing.
- Still running tests on cycle 2 script and MMSs. We now have a small cluster of 2 computers, 8 and 12 cores, with a SDD disk shared between them for the mpi tests on this sript.
- Lindsey Davis
- Pam Ford
- Set up Mac workstation and installed CASA trunk and 4.3 for Plotms tickets
- CAS-7044 RTT
- working on CAS-7042 (not a blocker)
- Out on leave Monday 11/10/14
- Kumar Golap
- Working on cleaning out bugs.
- Profiling on Multi-MS partitioned in time (did not help).
- Justo González Villalba
- Finalized testing casacore of support for SORTED_TABLEs bigger than 2GB
- Implemented propagation of Weight Spectrum in mstransform to be used by channel average (all transformations in general)
- Follow up science testing, preparation of a testing cluster in Garching, minor unit test fixes
- Jim Jacobs
- Fixed cache issue with old VI/VB
- Looking into ROSubtable
- Jeff Kern
- Two days in Utah working on visualization
- Lots of meetings
- Susan Loveland
- Plotms: CAS-7050, CAS-7074
- Viewer: CAS-6782
- CARTA
- Fixed two errors from previous check-in: server version not working & segfault for invalid image path
- Work on region selection.
- David Mehringer
- Continued stats framework implementation.
- Was pinged re (CAS-5836) Create new task imrebin. Apparently only a warning is now wanted. I had thought more was involved. Requested exact verbiage of requested warning and exactly the circumstances under which requested warning is requested to occur. Pointed out that a single message in one place in the code is not feasible at this point because of architecture issues. Asked requestor his preference on how this should be done.
- Responded re: release notes and noted several ia features that seem to have been left out.
- Responded re: known issues re: immath. I'm not aware of any problems like this on Linux and have no such tickets to support this immath "bug" on Liunx. If there is such a Linux defect, someone needs to open a ticket to me or it will never get fixed. Most of these "issues" have been resolved in the past once users have RTM. I do have a ticket which has an example of this issue on OSX, but until I can actually build and debug on OSX that will remain uninvestigated.
- George Moellenbrock
- Finished my PEP
- Resurrected holography data parsing python class (for S. African SKA collaboration) I started many months ago
- Advised several EVLA workshop participants (that stayed for another week of data reduction) on polarization calibration
- Initial testing of Justo's SIGMA_SPECTRUM work
- Jansky Symposium
- Dirk Petry
- Mark pointing out 4.3 Undertest Item wvrgcal issue (CAS-6982)
- Urvashi Rao
- Darrell Schiebel
- EL6 tests are now running
- Automating OSX tests
- Rob Selina
- Julian Taylor
- Takahiro Tsutsumi
- Makemask modification (fixed copy mode behavior for the output with an internal mask) - CAS-6912
- Fluxscale issue (CAS-7064) - turned out to be the user error (he ran gaincal on the same field with append=T which caused the caltable to contain duplicated entries per timestamp) and so it was not a bug but added a check for such case to issue more user-friendly error message
- Setjy: started to look at CAS-7066 ("Butler-JPL-Horizons 2012 standard shows strange behaviour")
- Ger van Diepen
- Created a MultiFile class to combine multiple files in a single file in order to reduce the nr of files used by the table system. Changed Tables such that the storage managers can optionally use MultiFile (controlled by a Table constructor argument or aipsrc variable). All casacore test programs run well when using MultiFile. I need to do some performance testing though. It would also be interesting to do performance tests on Lustre; could somebody at NRAO do that?
- Some time ago I made a python script to count the number of code lines (ignoring lines with alphanumerics, e.g. a line containing only a brace). The casacore statistics are:
Type Files Lines Code Comment Blank Header Other
casacore library
C++ 1033 395068 237233 60.0% 29044 7.4% 37813 26867 64111
C++Hdr 1071 260810 72268 27.7% 111087 42.6% 36519 27463 13473
C 21 15718 7542 48.0% 4075 25.9% 1863 0 2238
Fortran 145 51798 30728 59.3% 12613 24.4% 1915 877 5665
Flex 13 2380 1372 57.6% 457 19.2% 279 0 272
Bison 13 4950 2883 58.2% 609 12.3% 415 0 1043
Python 2 233 150 64.4% 40 17.2% 31 2 10
Perl 4 1458 854 58.6% 177 12.1% 137 55 235
sh 75 25643 17153 66.9% 1432 5.6% 1989 4035 1034
csh 2 522 303 58.0% 47 9.0% 42 119 11
tcsh 11 1632 1059 64.9% 45 2.8% 81 357 90
CMake 39 4565 3426 75.0% 311 6.8% 333 323 172
Config 3 138 128 92.8% 5 3.6% 5 0 0
doxygen 14 682 568 83.3% 0 0.0% 27 0 87
xml 2 378585 378581 100.0% 0 0.0% 4 0 0
html 3 79 65 82.3% 0 0.0% 14 0 0
binary 23 0 0 0.0% 0 0.0% 0 0 0
ignore 160 0 0 0.0% 0 0.0% 0 0 0
other 128 41462 0 0.0% 0 0.0% 2962 0 38500
casacore testcode
C++ 517 150231 92180 61.4% 14741 9.8% 11521 13076 18713
C++Hdr 7 1004 293 29.2% 314 31.3% 177 182 38
test-run 57 1318 1039 78.8% 0 0.0% 224 0 55
test-in 14 138 135 97.8% 0 0.0% 3 0 0
test-out 164 35869 32856 91.6% 0 0.0% 1874 0 1139
Python 1 163 135 82.8% 4 2.5% 23 1 0
CMake 34 863 771 89.3% 7 0.8% 46 0 39
binary 10 0 0 0.0% 0 0.0% 0 0 0
ignore 50 0 0 0.0% 0 0.0% 0 0 0
other 14 1461 0 0.0% 0 0.0% 119 0 1342
Type Files Lines Code Comment Blank Header Other
Friday NAOJ Meeting
- Kanako Sugimoto
- removed old SD tasks from trunk
- updating the mirroring script of CASA at NAOJ
- circulated base list of release notes to call for comments and corrections (Bunyo will collect release notes and known issue list for single dish)
- Wataru Kawasaki
- updated SD unit codes for testing flag data handling
- fixed a sd.rasterutil function to return TaQL query to select a raster scan. committed to the trunk, and merged it to the 4.3 branch
- fixed a bug in sd.scantable.stat() that it fails for spectra with all channels flagged. committed to the trunk also merged it to the 4.3 branch.
- started working on CAS-6985. this seems to be an issue of NRO filler.
- Masaya Kuniyoshi
- Takeshi Nakazato
- worked on importasdm issues CAS-7053, CAS-7054, and CAS-7075
- CAS-6986 velocity and rest frequency issue on NRO data: will discuss with NRO testers offline
- test support for 4.3 release: CAS-6589, CAS-6596, and CAS-6121
- code review: unit tests for sdbaseline, sdcoadd, sdlist, sdsave, sdstat, sdscale, sdimaging, sdmath, and sdcal
- Shinnosuke Kawakami
- building GC-casacore + casa-code completed 100% by modifying some source codes.
--
JeffKern - 2014-11-06