Currently we print the Field elements of the witness into a single file. There are two inconveniences with this.
1) Field elements don't serialize well for use outside of libsnark (by default they have a binary representation and need to be further transformed before using e.g. in DIZK).
2) Solvers require to count and split primary and auxiliary input. Having a single concatenated file is not great for that (also because different solvers might assume different ordering).
This change makes it so that we print the decimal representation of the field elements into two file (one for primary, one for auxiliary) separated by newline instead of space.
Currently we print the Field elements of the witness into a single file. There are two inconveniences with this. 1) Field elements don't serialize well for use outside of libsnark (by default they have a binary representation and need to be further transformed before using e.g. in DIZK). 2) Solvers require to count and split primary and auxiliary input. Having a single concatenated file is not great for that (also because different solvers might assume different ordering).
This change makes it so that we print the decimal representation of the field elements into two file (one for primary, one for auxiliary) separated by newline instead of space.