Resolve peaks with identical m/z in one scan in read_ms_data

[ridderl]

Should keep only the highest intensity in this case (test with Justin's "problem.mzXML", noise filter = 1000)