Skip to content

The RSKj-parity fixtures cannot be regenerated: the harness is not in the PR #378

Open

Description

@SDL-A1

From the external review of #374.

Six testdata files state they were produced by running the RSKj 9.0.4 jar
through a harness. That harness is not in the PR.

The consequence is precise and worth stating carefully: tests built on these
files show that Rustock agrees with the contributor's file. They do not show
that Rustock agrees with RSKj. The fixture is the oracle, and we cannot
re-derive it.

All eight fixture files are used by some test, so none is dead weight — the
coverage is real. What is missing is its provenance.

Why this matters more here than usual

This is the same class of problem as the committed .der TLS fixtures (#368):
an opaque artefact that cannot be regenerated. The difference is that those were
1.9 KB of self-signed certificates whose properties could be read back with
openssl. These encode consensus answers, and nothing on this host can
reproduce them.

What would settle it

Either the harness lands in the repo (ideally under docs/ with the jar version
pinned), or each fixture gains a header recording the exact RSKj commit, the
command that produced it, and a checksum of the jar.

Reported by the external reviewer. The claim that all eight files are used by a
test has not been re-verified here.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions