implementation of CellId.to_token and from_token slightly different from the official one

[wenjinshao]

Specifically, the official (?) implementation strips all trailing '0's from the token generated from a CellId, and always right-pads '0's to make the input token a 16-character hex representation before constructing a CellId with it.

[jongillham]

Thanks for the report. I'll have a go at fixing it.

[wenjinshao]

Thanks, sounds great!