Comment 13 for bug 1894619

Yeah I'm also not entirely sure.
But I'm assuming that the maintainers insight into all the interactions is better than mine and therefore tend to follow.

About the "why the difference DHCPv6 vs SLAAC" I think that is fine because the code now allows to behave different (reasonable). And the defaults as configured by NM or anything else could configure lifetimes for DHCPv6+SLAAC then no behavior change would happen.

Thereby (with a lot of un-sureness left) I'd intend to drop this change now (somewhat early/mid in 21.10) to identify any worse fallout better now than later.

I've closed my Debian bug that I spawned back in the day.
I will prepare an upload of dnsmasq, then we will see which tests or behavior break on it nowadays.