MuonClusterSegmentFinder - Introduce rough pointing cut in phi

Hi everybody,

08:38:16 MuonSegmentMaker.MuonClusterSegmentFinderTool          7     6 SUCCESS Fit segment from ([x,y,z]=(-3509.48,526.458,-7593.58) [theta/eta/phi]=(154.952,-1.50446,171.469)  pointing to [x,y,z]=(-0.293402,-0.284357,-0.912719) [theta/eta/phi]=(155.884,-1.54358,-135.897). Contained measurements in candidate:
08:38:16 pseudo measurement                              r     0 z     0 phi 0.000 theta 0.000 lPos  0.00 error 100.00
08:38:16 MM MML eta -2 phi  5 chlay 1 lay 3 channel 536  r  3546 z -7585 phi 2.992 theta 2.704 lPos -449.88 error  1.65
08:38:16 MM MML eta -2 phi  5 chlay 2 lay 1 channel 616  r  3577 z -7681 phi 3.002 theta 2.706 lPos -413.93 error  1.65
08:38:16 MM MML eta -2 phi  5 chlay 2 lay 2 channel 677  r  3577 z -7698 phi 3.003 theta 2.707 lPos -386.31 error  1.66
08:38:16 MM MML eta -2 phi  5 chlay 2 lay 3 channel 661  r  3582 z -7714 phi 3.005 theta 2.707 lPos -394.36 error  1.65
08:38:16 MM MML eta -2 phi  5 chlay 2 lay 4 channel 673  r  3587 z -7731 phi 3.006 theta 2.707 lPos -388.72 error  1.65
08:38:16 
08:38:16    Caught FPE 3 (floating point divide by zero) at address 0x7f4fc878f8c1)
08:38:16    backtrace() returned 27 addresses

this MR introduces a rough cut on the delta phi between the muon crossing point in the NSW and its direction. This prevents bogous segment fits that triggered an FPE as reported in ATLASRECTS-7366