Just noticed that after merging #95, the fuzzing code became invalid. This was not caught by CI because gofuzz hides the code behind a build flag, and we're not fuzzing on CI. We're still not fuzzing on CI (other than the corpus), but by using Go's native fuzzer we at least make sure that the fuzzing code still builds.
I tried to decode the existing corpus, but was all just random bytes (or at least it looked like random bytes).
Just noticed that after merging #95, the fuzzing code became invalid. This was not caught by CI because gofuzz hides the code behind a build flag, and we're not fuzzing on CI. We're still not fuzzing on CI (other than the corpus), but by using Go's native fuzzer we at least make sure that the fuzzing code still builds.
I tried to decode the existing corpus, but was all just random bytes (or at least it looked like random bytes).