Reminder to work on adding a (or multiple?) valgrind run to the CI, to catch memory issues sooner. Potentially, we could think about some sanitizers as well?