Actually, I’m not so sure about using the cocycles because I’m still learning this. I opted for just returning them in a long vector and leaving cutting them up to the user.
And yeah, ripser calculates the persistent homology in \mathbb Z_p where p is the value of the modulus parameter, so extracting cycles should be possible.
Thanks about the slack tip, I think I’ll do that, even though I also get segfaults with clang on my (Linux) machine.