This issue came up while testing with someone else. I had left the a=
option in my DNS selector records (from an *early* version of the
draft), and their implementation choked because of the currently invalid
option.
I'm glad the consensus so far is for unrecognized options to be optional.
Tony Hansen
tony(_at_)att(_dot_)com
Mark Delany wrote:
If the spec doesn't say it, then it's an oversight. The intent has
always been to allow new tags in Selectors/policy and that existing
code should ignore unrecognized tags.
Arvel Hathcock wrote:
Yes we should. I was under the impression that unknown options would
simply be ignored. That's the only way to make upwardly compatible
change easily right?
Dave Crocker wrote:
I liked Arvel's response.
1. Additional options may be defined later.
2. A validator that does not recognize a particular option MUST ignore
it.
_______________________________________________
NOTE WELL: This list operates according to
http://dkim.org/ietf-list-rules.html