Add reconstructed time to CAF particle record

[PetrilloAtWork]

Pandora is sometimes able to tag a particle with a time (cathode-crossing tracks are the case in ICARUS and presumably SBND).
The information of that time should be stored in CAF data.
In fact, it may be conceivable that several time estimations are available (from matching with light, with CRT, and from TPC reconstruction, for example) and one might even want to accommodate all of them and let the analyser choose which one they believe most.

Anyway, for this one, a default value of "non-assigned" is needed, and then I recommend the time to be stored in the standard CAF time unit and with the same time reference as the reconstructed optical flashes (beam gate time? this point is topic of a different discussion).

[wesketchum]

Is this PR fixing this issue? #308

[gputnam]

@PetrilloAtWork can confirm -- but I believe it goes some of the way.

It does add in the cathode crossing time for slices that have one.

[wesketchum]

@PetrilloAtWork should this be left open? going to move to a different release at the least ... but would be good to close if we can.

[PetrilloAtWork]

I think PR #308 addressed it for the time being.
We'll want the time from PMT and CRT when they become available, but that's not for the current release.
Up to you if you prefer to close this one or move its target.