Distinguish `tez` and `mutez` literals at CST level
Currently lexer transforms tez literals like 10tez
to mutez literals like 10000000mutez
This leads to ligo print pretty
transforming all tez
to mutez
as well.
As in https://tezos-dev.slack.com/archives/GLGGHDDMH/p1695645238844719?thread_ts=1695629705.617119&cid=GLGGHDDMH, its not an expected behaviour. We could distinguish tez
and mutez
values in CST, nad do the transformation probably at the unification level.