Jakob Heitz <jakob.he...@ericsson.com> wrote: > > Current usages of BGP probably will not have anything matching > a marker in the byte stream other than the marker.
I can't think of any way for it to get there that isn't an error... > But future additions could well do that. Sounds like a very bad idea to me. > There is a draft to increase the maximum length, > so you can't rely on that being <= 4096. Indeed: it will probably grow past 4096 sometime. > What about a notification message or a possible future > operational message that replays the errant update. Definitely a bad idea! Replay of an errant update can't help, and would likely crash its receiver. Don't do that! > That will look just like an update to the resync code. I'm not saying resync is necessarily a good thing -- but I believe a marker (16 bytes all ones) in the middle of _any_ message deserves an error of some kind. Lacking a TLV structure for messages, we should not let go of a marker which separates messages. -- John Leslie <j...@jlc.net> _______________________________________________ GROW mailing list GROW@ietf.org https://www.ietf.org/mailman/listinfo/grow