personal schedule star/unstar count handling is incorrect on multi-session talks

[particle]

starring and unstarring the second of two back-to-back sessions of http://www.yapcna.org/yn2013/talk/4734 modifies the starred count for the first session of the talk by appending a new string value to the star count, rather than replacing the numeric starred value . it does not affect the starred count for the second session of the talk at all.

e.g. the count for the first session ('43') becomes '4344' after the first star/unstar, while the count for the second session remains '43'. after a second star/unstar, the first session count becomes '434443', and the second session count remains '43'.