validation work since the dec. collaboration meeting

-as announced, new bhabha test with release 21.3.0'
-used preskimmed bhabha files --> high speed test

test, part 1: objectivity

-crashed very soon: problem with code migration, needed to update some CalType objects, easy to fix
-so, after restarting everything went fine (nothing else expected from this test)
-output of the channel information works now, non validated channels will be printed out with absconstant and relconstant
-read out the constants and asked for a deletion of the constants and the temporal db content, but only the latter was deleted (will lead to some problems explained later)
-started offline computing, using the same files
-results from redoing:
-accumulated bhabhas in EmcSelectBhabha are the same in both types of processing --> inconsistencies vanished mysteriously
-accumulated bhabhas in the root application is not the same --> I suspect the dead channels haven't been counted here, in contradiction to EmcSelectBhabha
-calibration constants are the same, except:
-some (~50) channels differ because rounding issues (makes an 0.01% effect)
-and some (~40) differ because validation problems (weren't been cut out properly in the offline calib., but were by the online calibration)
-results derived ~1 hour ago, so more time needed to evaluate the differences
-but I expect no serious problems, it seems like a configuration issue (of the offline calibration!)
-conclusion for test, part 1: pretty much success! :-)

test, part 2:

-switched to usage of root based temporal/spatial db
-will be used for run 6 processing, so more important than part 1 :-)
-smothly accumulated 2e6 bhabhas, finished the calibration, stored the constants and crashed during the deletion of the temporal db content
-Andy fixed the bug and provided a new tag and we tried about 3 more times to finish and test the tag, but failed everytime with another (technical) problem
--> tag is not yet tested, but in release 22.0.1+
-nevertheless some results:
-as mentioned above, stored constants should be deleted, but weren't, because of ???
-so, the last run used the new calibration constants for computing the new constants --> the constants cannot be the same!
-some modification needed, but finally after first determining the relative change and then multiplying it twice to the old constant got the same results as in the objy test (neglecting rounding issues and validation issues)
-number of accumulated bhabhas is the same as well
-conclusion for test, part 2: not everything worked up to now, need to keep an eye on the functionality test of Andy's tag

other issues

-Concerning the 3Hz story: good news: we will get 3 Hz; bad news: confirmation came to late to make final changes in the current release --> we will calibrate after having 2e6 bhabhas (~13 days)
-plan was to accumulate 3e6 bhabhas and therefore calibrate after ~19 days
-not too bad because avoid interference between offline calibration and rolling calibration

-final conclusion: rolling bhabha calibration is basically ready for test phase in Run 6 :-)