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
In block 1089328 (#767 ) one of the problem Kagome team faced was related to crypto batch verification. Expected and implemented behaviour was the following:
Batch begins (ext_start_batch_verify)
Messages and signatures are queued
Not 100% about this part, but Internally ext_sr25519_verify postpones verification and just returns true even for invalid signatures if verification was invoked inside the batch (at least that way it was implemented in Kagome by this moment)
Batch ends and verification of all queued signatures begins. True is returned if all signatures are valid and false otherwise (ext_finish_batch_verify)
Surprisingly correct behaviour (at least for this block) was the following:
Batch begins (ext_start_batch_verify)
Signature verification is not postponed and it just goes to verify it (like here)
Batch ends and returns true even if inside batch there were invalid signatures (here)
So statement below is not true, at least for block 1089328
For now we will just keep things as it is done in smoldot, but later we need to investigate and return correct behaviour
The text was updated successfully, but these errors were encountered:
In block 1089328 (#767 ) one of the problem Kagome team faced was related to crypto batch verification. Expected and implemented behaviour was the following:
ext_start_batch_verify
)ext_finish_batch_verify
)Surprisingly correct behaviour (at least for this block) was the following:
ext_start_batch_verify
)So statement below is not true, at least for block 1089328
For now we will just keep things as it is done in smoldot, but later we need to investigate and return correct behaviour
The text was updated successfully, but these errors were encountered: