Remove track::v2 from the configuration and fix PrCheatedSciFiTracking

Remove Track::v2 from the configuration

Depends on Rec!2470 (merged) and Rec!2481 (merged)

Looking at all the diffs (https://lhcb-nightlies.web.cern.ch//compare?a=nightly/lhcb-master-ref/1470&b=nightly/lhcb-master-mr/2597), there are a few things:

• The hlt1_reco_muonmatching_with_mcchecking still uses v2::Track in some places. This will be removed in the next step (not this MR).
• The PrCheatedSciFiTracking was broken before, as it did not add any tracks to the output. I fixed this and converted the output to PrSeedingTracks. This now gives reasonable numbers in the hlt2_reco_baseline_with_cheatedseeding test, but maybe @lohenry can confirm.
• The Calo tests in the fast sequence show a bit an odd behaviour, in hlt2_fastest_reco_with_mcchecking, hlt2_lead_lead_light_reco_pr_kf, hlt2_lead_lead_light_reco_pr_kf_with_mcchecking, hlt2_light_reco_pr_kf_with_mcchecking and hlt2_light_reco_with_mcchecking, however, it actually seems to improve the performance. Given that this MR does not change the tracks itself and it only appears for the fast sequence, my suspicion is that unfitted tracks are used, and the dummy value for the covariance changed, which leads to this behaviour. See !900 (comment 4672289) for more details.