Oddbean new post about | logout
▲ ▼
Vitor Pamplona | 22 days ago (raw) | root | parent | reply | flag +1 
 As long as both encode in the exact same way so that one can decide the other, it should be fine.
▲ ▼
Nuh 🔻 | 22 days ago (raw) | root | parent | reply | flag +1 
 and we went with DNS packet encoding, we don't invent wheels that are already good enough and proven and simple.

if someone wants to use a different encoding, I honestly think they are the ones deliberately breaking compatibility. (of course the information in records can still require more interoperability and conventions).
▲ ▼
Vitor Pamplona | 22 days ago (raw) | root | parent | reply | flag +1 
 By DNS packet encoding, do you mean RFC 1101? Or something else?
▲ ▼
Nuh 🔻 | 22 days ago (raw) | root | parent | reply | flag +3 
 At least 1035 ... but most implementations support more, we use this library so we support these rfcs https://github.com/balliegojr/simple-dns/blob/main/simple-dns/README.md#dns-packet-parserbuilder

Pubky specifically cares about SVCB and HTTPs  records in in RFC 9460 

but the thing is, unsupported dns records are still parsed by anything that supports 1035... just not usefully.