You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Possibly relates to; #834. Or atleast it's in the same code area to be reworked.
Problem
Currently the implementation of build_rev_states_json in prover_internal.rs forces that the same cred used on multiple referents must present the exact same timestamp in it's NRP. E.g. if "referent1" is being presented for interval to: 10 and "referent2" for to: 20, then build_rev_states_json will force both to the same timestamp for presentation (10 or 20, depending on ordering).
Bigger Problem?
This problem is possibly bigger than the main problem, but heavily related. In build_rev_states_json it looks like any credential that is revocable, is forced into creating a NRP, even if the proof request did not request it... Testing this against ACA-py 0.8.2 results in a verification failure, with the following error: "VALUE_ERROR::Timestamp on sub-proof #0 is superfluous vs. requested attribute {referent-here}". We are creating a NRP that is not necessary, and ACA-py sees this as superfluous.
Expected Behaviour
referents that do not need to present a NRP (no referent interval nor global interval) should not be forced to present one
Happy to take this one on if we agree it is a problem; i think prover_internal may just need a fresh lookover
gmulhearn-anonyome
changed the title
Presenting same credential with different revocation timestamps (or lack thereof)
Cannot present same credential with different revocation timestamps (or lack thereof)
Aug 4, 2023
Possibly relates to; #834. Or atleast it's in the same code area to be reworked.
Problem
Currently the implementation of
build_rev_states_json
inprover_internal.rs
forces that the same cred used on multiple referents must present the exact same timestamp in it's NRP. E.g. if "referent1" is being presented for intervalto: 10
and "referent2" forto: 20
, then build_rev_states_json will force both to the same timestamp for presentation (10 or 20, depending on ordering).Bigger Problem?
This problem is possibly bigger than the main problem, but heavily related. In
build_rev_states_json
it looks like any credential that is revocable, is forced into creating a NRP, even if the proof request did not request it... Testing this against ACA-py 0.8.2 results in a verification failure, with the following error:"VALUE_ERROR::Timestamp on sub-proof #0 is superfluous vs. requested attribute {referent-here}"
. We are creating a NRP that is not necessary, and ACA-py sees this as superfluous.Expected Behaviour
The text was updated successfully, but these errors were encountered: