Hi all, I just noticed that I sent my last three messages today with “Autocrypt” [1] headers – I'm working on support for it in Balsa. As it it /really/ untested yet, I hope it doesn't break anything. If it does, please tell me… Hopefully, I can provide a first patch implementing the most simple version of “Autocrypt Level 1” soon. I plan to include * scanning the Autocrypt headers (but /not/ Autocrypt-Gossip) from incoming messages, * maintaining the peer state, * injecting Autocrypt headers (but /not/ Autocrypt-Gossip) in outgoing messages if requested and * implement the recommendation for encryption as required by the standard. I do *not* plan to add gossip (which is a little questionable IMHO) and key management which is better done by specialised applications (gpa, seahorse, etc.). Of course, the feature can be disabled separately for every identity. As the peer management is a little complex, I now use a sqlite3 database for that purpose. It simplifies the code, but is a little against the “lightweight” philosophy of Balsa, as it adds an extra dependency. And there are some other open questions, e.g. how spam could be detected (the standard says they should be ignored, which makes sense). A solution might be defining a header indicating spam. Then it would be nice to have an optional special “Junk” folder, and move new spam messages /automatically/ to it (like Thunderbird). As only some basic stuff is usable in my test environment at the moment, I could easily modify my plan if you think it's dumb… As always, any comment is really welcome! Cheers, Albrecht. [1] <https://autocrypt.org/>
Attachment:
pgpoQvd7gbsoS.pgp
Description: PGP signature