addisoncrump
commented
2 weeks ago Oh, and second bonus: it sharply reduces the size requirements for our grammar-based inputs. For fields in the CFG which are represented by a higher complexity structure, we can simply store these as a queue of input segments ("to be dequeued when needed").
Context-free grammars are quite simple to enumerate: https://arxiv.org/pdf/2305.00522
This may be useful for implementing mutations as simple addition/subtraction.