On 10/26/20 4:19 AM, Patrick Ben Koetter wrote:
There's only *one* SigningTable, but there are two KeyTables – one for rsa and the other one for ed25519. Maybe you are using an older version of dkimpy-milter. IIRC it had a related error in the man page.
oops, typo. yep, I've one ST & 2 KTs, one each for rsa & ed25519 using latest available via pip, v1.2.2. can try master branch.
Yes. I use it on mailop.org.
thx! ok. so it's local ... how are you generating your ed25519 data? for rsa, here, _either_ 'dknewkey' or 'openssl genrsa (etc)' works fine. for the ed25519, i get different fails -- in error logs -- with 'dknewkey' or 'openssl genpkey (etc)'. atm, with 'dknewkey' generated data, on attempted ed25519 signing I'm seeing: "sign_dkim: The seed must be exactly 32 bytes long" i know README says "in order to generate Ed25519 keys for dkimpy-milter, dkimpy specific tools must be used to be compatible " tho, i don't yet know _what_ the differences actually are ...
