LoKiCore : add new class Polarization

LoKiCore : add new class Polarization with functions for calculation of various polarization quantities:

• frames
• direction cosines
• polarization variables

This MR for 2018-patches and master branches. If there are no conflicts and compilation errors, it could be propagates to 2017-patchesbranch, but it is not necessary.

changed the description

• [2018-03-23 00:08] Validation started with lhcb-2018-patches#154
• [2018-03-23 00:12] Validation started with lhcb-moore-test#3
• [2018-03-24 00:05] Validation started with lhcb-moore-test#4
• [2018-03-24 00:08] Validation started with lhcb-2018-patches#155
• [2018-03-25 00:05] Validation started with lhcb-moore-test#5
• [2018-03-25 00:07] Validation started with lhcb-2018-patches#157

added 1 commit

• 98d6960a - Polariztaion: use the trick with tagged_bool, thanks to Gerhard!

Compare with previous version

added 2 commits

• c1752605 - update LoKi:Tensors and Polarization
• 452f9969 - update LoKi::Tensors and Polarization

Compare with previous version

added 1 commit

• e7c038d4 - Polarization: add polarization vectors

Compare with previous version

@cattanem hi marco. Am I right that I am too late with this MR for LHCb v50r0 ?

Yes. But there will be other releases... In any case the master stack is not really intended for mainstream Run2 analysis

resolved all discussions

@ibelyaev Something is causing many of the lhcb-2018-patches tests crash in the dictionary system, I can't really tell whether it is due to this MR but I am making it WIP to test. Could you please have a look at the stderr of any of the failing tests in, e.g., https://lhcb-nightlies.cern.ch/logs/tests/nightly/lhcb-2018-patches/155/x86_64-slc6-gcc62-opt/LHCb/

marked as a Work In Progress

@cattanem I can't reproduce the error locally with LoKiGen, LoKiMC,LoKiNumbers, ...

What can I say? I restarted the nightly without this MR (https://lhcb-nightlies.cern.ch/nightly/lhcb-2018-patches/build/156/) and it's a lot happier than with it (https://lhcb-nightlies.cern.ch/nightly/lhcb-2018-patches/build/155/)

All I can suggest is that I put this back and try again, just in case it was a glitch... Will be build 157 tomorrow morning.

unmarked as a Work In Progress

I've managed to reproduce the failures (one needs to make a fresh build afte r make purge, otherwise the error is not reproducible. I think I've found a solution, will commit it soon

added 1 commit

• 573e4b24 - fix?

Compare with previous version

@cattanem I think I've "fixed" it. At least LoKi-tests are ok. The problem is a kind of unknown. I've replaced Gaudi::tagged_bool with plain bool, and the problem have been disappeared. Note just the presence of the header #include "GaudiKernel/TaggedBool.h" is enough to produce the problem, even without usage of this class @graven

Thanks @ibelyaev! @clemenci FYI.