The original intent was to allow the pledge to signal “I can act on this parameter but not on that value that you gave me”, cherry-picking specific keys from the key set or just copying the whole thing if nothing works. The previous text optimized for that by mandating the value or its CBOR-encoded subset to be included all the time.
I rewrote the paragraph now in order to allow the pledge to send a null for the case where the JRC should not bother to retry configuring that specific parameter again. The commit is at: https://bitbucket.org/6tisch/draft-ietf-6tisch-minimal-security/commits/60e2a8a2acbae5cca624932e88c1f650a95d9ca6?at=minimal-security-10 <https://bitbucket.org/6tisch/draft-ietf-6tisch-minimal-security/commits/60e2a8a2acbae5cca624932e88c1f650a95d9ca6?at=minimal-security-10> Let me know if it reads better. Mališa > On 5 Apr 2019, at 16:32, Christian Amsüss <christ...@amsuess.com> wrote: > > Only a small question remains: When a pledge sends [0 /* unsupported */, > 42 /* some fancy extension */, X] as an Unsupported Configuration, why > should it send the full value that was sent with the join response and > not just a syntactic placeholder (null)?
_______________________________________________ 6tisch mailing list 6tisch@ietf.org https://www.ietf.org/mailman/listinfo/6tisch
