Anoncreds revoke and publish-revocations endorsement #2782
Conversation
I think this happened when I did the original merge of the |
Signed-off-by: jamshale <[email protected]>
jamshale
force-pushed
the
anoncreds-endorsement-revoke
branch
from
336416b to
6f6ad0d
Compare
|
Overall looks good, is it ready to go or are you still working on it? |
|
My manual testing and integration testing has been good, so I think it's ready. The last integration test run failed but I'm pretty sure it was a random unrelated fail. |
|
…dation#2782) Signed-off-by: jamshale <[email protected]> Co-authored-by: Ian Costanzo <[email protected]> Signed-off-by: Golda Velez <[email protected]>
Opening for visibility. I'm still going to manually test more but the Integration tests are passing for endorsement and publishing revocations.
One area of confusion for me is what happens with the local wallet verse the ledger. Write now an endorsement transaction is created when the
update_revocation_listis called with changes. Then the local wallet is updated. At the same time the endorsement manager receives the114event and notifies the anoncredsRevocationListFinishedevent handler. It checks if the wallet contains a record for the list viarev_reg_def_idand if it doesn't it creates the record. If it does then it assumes the record has been updated already and does nothing.I was trying some other stuff like deleting the record and creating a new one with a
waitstate but couldn't decide if this was necessary, or something we wanted to do.I deleted the
/anoncreds/revokeand/anoncreds/publish-revocationendpoints and instead just have the/revocation/revokeand/revocation/publish-revocation. I thought this was appropriate because the revocation endpoint are all loaded for anoncreds specifically and the rest of the anoncreds endpoints are to do with creating objects. I changed the params for requesting a transaction manually to use the body options and updated the integration tests.