Somehow there's still one test failed after ce7cff8.

after some digging, I'm unable to locate the cause. @chancharles92 you are probably more familiar with this gadget, do you have any clue where did i forget to update?

update: fixed in 44b7c42 ✅

reminder for myself after I address comments for this PR:

• update changelog, and bump version as I forgot to do in a previous PR

Should not we add an if statement to raise an error here if the plonk instance does not support lookups?

because anyway it's not used to construct the hasher.

yes, but previously, we were using it. in this PR, we removed it.

Same field also exists in the RescueTranscript, we should consider removing that as well.

that would require refactoring of the recursion circuit, since we are not using it atm, maybe we leave it as it is for now.

cc @mrain

Should not we add an if statement to raise an error here if the plonk instance does not support lookups?

we have already does the protection from the caller side: in snark.rs

            let plookup_evals = if circuits[i].support_lookup() {
                let evals = prover.compute_plookup_evaluations(
                    prove_keys[i],
                    &challenges,
                    &online_oracles[i],
                )?;
                transcript.append_plookup_evaluations::<E>(&evals)?;
                Some(evals)
            } else {
                None
            };

we only invoke it if .supports_lookup() evaluates to true.
cc @philippecamacho

maintainer check: @mrain are you happy with jf-plonk-only bump in 4c421e7?

(if so, i'll create a tag named jf-plonk-v0.5.0 later)