msms workflow fails at step 6

So far, we know that the MS2 objects in steps 2-5 sometimes form badly and cause errors during aggregateSpectra call during the step six of the RMassBank workflow.

We need to identify which slots of the children objects cause this (ie, turn into NAs, or NULLs, or anything length==0).

I proposed a debugging procedure in which we pick two workspaces out of the res$wsflow table, one we know breaks the workflow and another that is successful. The workspaces should be taken at the stage before sp_done (so, after steps 1:4 have been completed). Then, we look for the differences in object slots in children.