Giving ngramread a text file where lines are terminiated with space causes models built with this to have a missing symbol. (BER: fixed -- deals with leading and trailing spaces)
Several commands seg fault if given missing files (BER: fixed to be similar to openfst in this regard)
the ngramapply example in the quick tour segfaults if --keep_isymbols is not given (BER: fixed, and will also update quick tour)
the ngramapply example says there are 2 sentences but there is only one (BER: fixed)
Giving ngramread a blank line causes models built with missing symbol (at least in the version I'm using).
Giving ngramprint --ARPA an FST with no symbol tables segfaults
Shrinking does not terminate (after 10 hours) on a 1 gb input
Checking for normalized model is perhaps too exact (allow small delta)