This still needs quite a bit of cleanup and tests.
I think the header processing part could use some refactoring.
Processing incoming Autocrypt and Autocrypt-Gossip headers is
very similar.
Our internal data representation also does not destinguish gossip
and Autocrypt headers from the person themselves yet. Will need
to look into the update of the peerstate again.
We also lack tests for:
[ ] handling multiple gossip headers for the same recipient
[ ] ignoring gossip headers for email addresses that are not
part of the recipients
This still needs quite a bit of cleanup and tests.
I think the header processing part could use some refactoring. Processing incoming Autocrypt and Autocrypt-Gossip headers is very similar.
Our internal data representation also does not destinguish gossip and Autocrypt headers from the person themselves yet. Will need to look into the update of the peerstate again.
We also lack tests for: