"Pascal Thubert (pthubert)" <[email protected]> writes:
> I agree... I think that the idea was to make sure that the buffer you
> prepare is big enough for the uncompressed form.

That is a useful property.

> When Jonathan and I wrote a proposed rework we changed that as you suggest:
> https://tools.ietf.org/html/draft-thubert-6lo-forwarding-fragments-02 
[...]
> The rework addressed the problems discussed in the attached slides,
> and which are still fully present today.
> The work needs support to move on. Last I checked interest was really limited.

As a general rule, people will show interest if they're being bitten by
the problem in practice.  If there isn't much interest, that would mean
that people aren't running into these problems in practice.  It's not
clear to me why that would be.

> Note also this text in RFC 6282: [...]

OK, I can see how that makes the packet reconstructable, as complicated
as that system is.  I do wonder why it was chosen, though, as it seems
both more complicated and less natural than separating compression and
fragmentation into two layers.

Dale

_______________________________________________
6lo mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lo

Reply via email to