Allow for more accepting states

[gaperez64]

@michaelcadilhac I've found a benchmark for which the assertion below does not hold. See 07.tlsf in the realizable benchmarks in tests. Why did we not use at least int?

acacia-bonsai/src/boolean_states/forward_saturation.hh

Line 32 in 9e75f86

assert (nb_accepting_states < UCHAR_MAX);

[michaelcadilhac]

I'd call this premature optimization. This allows for a slightly better data structure for the loop that follows, in the code. I'm "refreshing" all integer types with fixed width integer types right now, and this will solve that.

[gaperez64]

Is it solved now @michaelcadilhac ?

[gaperez64]

I still see the assertion in the code.