Merging of FDC cathode hits modified to merge like pulse height, not integrated charge

[nsjarvis]

The readout of the FDC cathodes produces pulse height, not integrated charge, so I modified the hit merging correspondingly, in a manner similar to that done for the CDC.

If there are both random and simulated hits on the same strip, the earlier hit is kept and the later one ignored.
If two hits arrive in the same time sample, q is summed.

The plots below show real data from run 70370 and simulated data with the old mcsmear and then the new mcsmear.
There is negligible difference between the two. However it can be seen that the simulated threshold is too high, as the first time peak is from the randoms. I'll look into fixing that next.

The new merging gives results only slightly differrent from the new merging because (for the run tested)
93% of new hits contained a random but no real hit on the same strip
4% of new hits were where a random hit arrived before a real hit and so replaced it
3% of new hits the real hit arrived first and was preserved
0.1% of new hits the real and random arrived in the same sample and q was summed.

Even though the new merging makes only a little difference to the output, I would like to use it because the method is correct.

Below, the new version is shown in red and the old in black. The only difference that I can see is that the two peaks in the time histogram are slightly higher in the new version (red).

Same histograms, log scale or zoomed in:

[gluex]

Test status for this pull request: SUCCESS

Summary: summary.txt
Logs: results/log

Build log: make_nsj_fdc_smear.log
Build report: report_nsj_fdc_smear.txt