Stand-alone binary for the reconstruction from timeslice (using FairRun)

While the concept of executables seems to be partly new to the cbmroot repository, the addition should not be incompatible with the existing structure.
The only other kind of stand-alone applications I found are the ones based on FairMQ, which are located in a directory "MQ" at the top level.
Following this line of thought (framework as top directory name), this MR would have to add a directory "FairRun" or similar at top level, which I find rather confusing.
And when we add a variant of this application that does not rely on FairTasks/FairRun in the future ("cbmreco_plain"), we will need yet another top-level directory.

That's why I opted for a generic app directory at the top level as we also have it in the flesnet repository. What do you think, should this rather go somewhere else?

Also, the location of reco/tasks/CbmRecoConfigExample.json may not be canonical. Please regard this MR as an initial proposal and request changes as you see fit.

First of all: very nice! And very quick - would have taken me ages, not being familiar to boost::property_tree and program options.

To the file structure: Having in mind the (distant) goal that the repository should allow different builds (sim, ana, reco), I would suggest to put the app folder, now at the top level, to the reco directory (reco/app). Imagine we want similar executables for transport and detector response simulations, they would have quite different (additional) libraries to link to. But I must confess that the final folder structure to eventually arrive at is still not clear to me. The current top level folder MQ is not the result of some discussion that I remember. @f.uhlig, what is your opinion?