Add ParticleMassMonitors to bremsstrahlung test

Add ParticleMassMonitorss for the Jpsi and B candidates in test_bremsstrahlung.py and check their results in Hlt2Conf.test_hlt2_bremsstrahlung.

This is my naive attempt at addressing this discussion. The new consumer introduced in Phys!939 (merged) adds counters to the log of the type:

Hlt2DiEBremJpsiMonitor                 INFO Number of counters : 2
 |    Counter                                      |     #     |    sum     | mean/eff^* | rms/err^*  |     min     |     max     |
 | "Candidate mass"                                |        33 |   99755.79 |     3022.9 |     413.13 |      2245.9 |      3947.6 |
 | "Candidates in 1 sigma"                         |        15 |
Hlt2NoBremJpsiMonitor                  INFO Number of counters : 2
 |    Counter                                      |     #     |    sum     | mean/eff^* | rms/err^*  |     min     |     max     |
 | "Candidate mass"                                |        20 |   54113.91 |     2705.7 |     295.90 |      2282.0 |      3341.2 |
 | "Candidates in 1 sigma"                         |         4 |
Hlt2SingleBremJpsiMonitor              INFO Number of counters : 2
 |    Counter                                      |     #     |    sum     | mean/eff^* | rms/err^*  |     min     |     max     |
 | "Candidate mass"                                |        39 |   117891.4 |     3022.9 |     400.07 |      2245.9 |      3947.6 |
 | "Candidates in 1 sigma"                         |        17 |

where the "Candidates in 1 sigma" counter is used in Hlt2Conf.test_hlt2_bremsstrahlung to check that the resolution improves when bremsstrahlung corrections are used. It also fills histograms of the type: that should provide more information in case of test failure.

Needs Phys!939 (merged)