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
I was looking at your conformance work and the limitation you noted about the three SDKs sharing the same author.
I wrote a separate Python verifier using the published spec, schema, ADRs and vectors at d52962fa, without reading the SDK implementations.
It agreed with 73 of 76 tested vectors. The three disagreements appear to come from the same issue: receipt IDs and action IDs that don't match the schema's UUID patterns.
Positive cases
In v020_vectors.json, terminalChain and parametersDisclosureReceipt contain IDs such as urn:receipt:v020-terminal-1 and act_v020_1.
The schema requires UUID shaped identifiers, and §7.8 requires schema validation before signature verification. So these receipts fail before their signatures or chains are checked, even though the corpus marks them valid.
The receipt_signature_preservation_legacy_0_2_0 case hits the same problem on its signature verification expectation.
Disabling only the two ID pattern checks as a diagnostic makes these positive cases verify successfully, including their signatures and expected hashes.
Negative cases
All 11 cases in malformed_vectors.json also contain action IDs that fail the UUID pattern.
A schema first verifier therefore rejects all 11 at the action ID pattern before reaching the specific check each case is intended to exercise.
With the ID pattern checks disabled, nine reach their intended failure checks, including three schema checks. The remaining two, truncated_proof_value and missing_previous_receipt_hash_mid_chain, fail schema validation before reaching the signature or linkage checks they name.
This means an implementation could report 11/11 rejected without demonstrating that it detects the underlying signature or linkage failures.
Question
Is schema validation intended to be mandatory for these older pinned vectors?
If so, the positive and negative fixtures need correction, with their signatures, hashes and frozen version status handled explicitly. If the old IDs are intentionally accepted, §7.8 or the version specific schema requirements need to say so.
The verifier and runner were written in one job, so this is not a second independent reproduction. It uses rfc8785 for canonicalization and cryptography for Ed25519. It doesn't cover every part of the protocol.
I can publish the verifier, per case results and documented interpretation choices, pinned to d52962fa, so you can reproduce the finding.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
I was looking at your conformance work and the limitation you noted about the three SDKs sharing the same author.
I wrote a separate Python verifier using the published spec, schema, ADRs and vectors at
d52962fa, without reading the SDK implementations.It agreed with 73 of 76 tested vectors. The three disagreements appear to come from the same issue: receipt IDs and action IDs that don't match the schema's UUID patterns.
Positive cases
In
v020_vectors.json,terminalChainandparametersDisclosureReceiptcontain IDs such asurn:receipt:v020-terminal-1andact_v020_1.The schema requires UUID shaped identifiers, and §7.8 requires schema validation before signature verification. So these receipts fail before their signatures or chains are checked, even though the corpus marks them valid.
The
receipt_signature_preservation_legacy_0_2_0case hits the same problem on its signature verification expectation.Disabling only the two ID pattern checks as a diagnostic makes these positive cases verify successfully, including their signatures and expected hashes.
Negative cases
All 11 cases in
malformed_vectors.jsonalso contain action IDs that fail the UUID pattern.A schema first verifier therefore rejects all 11 at the action ID pattern before reaching the specific check each case is intended to exercise.
With the ID pattern checks disabled, nine reach their intended failure checks, including three schema checks. The remaining two,
truncated_proof_valueandmissing_previous_receipt_hash_mid_chain, fail schema validation before reaching the signature or linkage checks they name.This means an implementation could report 11/11 rejected without demonstrating that it detects the underlying signature or linkage failures.
Question
Is schema validation intended to be mandatory for these older pinned vectors?
If so, the positive and negative fixtures need correction, with their signatures, hashes and frozen version status handled explicitly. If the old IDs are intentionally accepted, §7.8 or the version specific schema requirements need to say so.
The verifier and runner were written in one job, so this is not a second independent reproduction. It uses
rfc8785for canonicalization andcryptographyfor Ed25519. It doesn't cover every part of the protocol.I can publish the verifier, per case results and documented interpretation choices, pinned to
d52962fa, so you can reproduce the finding.All reactions