cmplog: only add tokens that were found to fit?

create from thin air if no good seed after a cycle and dict large enough?
(static u32 no_of_struct_inputs;) 

splicing -> check if whitespace/token is needed

whitespace/token check only AFTER mutation

analyse welche einen DICT haben, und welche davon rein ascii

corpus analyse:
	+ libxml
	- hardbuzz
	- sqlite
	- libpcap
min len, max len, % wenn 95/98/99/100 ascii

funktion und env für menge an mutationen

env für menge an per mutation run

only add inital dictionary, not furher finds, e.g. cmplog