logParser: cross-section checks
From Frank (!167 (merged)):
also with 10 events one can encounter negative cross sections, and for some samples (e.g. interference) cross sections will always be negative. So maybe we should just remove that check from logParser (or make it a WARNING).
Overriding the test would be easy by:
- making use of the already available
-t
flag - moving it to the generator-specific tests and e.g. not applying it to Sherpa
- applying it only if
CountHepMC
(number of generated events) is higher than a given threshold
It's not clear to me in which cases we expect whole samples to have negative cross-sections.
If it's a O(1%) case we can just live with [skip logParser]
for such cases?