Simon Kelley <simon@thekelleys.org.uk> writes:

> Great, thanks. I've done that and confirmed what you are seeing. At this point I
> need to trace through the whole validation process in parallel on working and
> broken platforms, to see where the two diverge.
> (I've had to do this before. It was not fun, but it did produce results.) I've
> installed gdb and can run dnsmasq under gdb, but the binary is very thoroughly
> stripped, so I can't get any useful information out. Best way to get a
> non-stripped dnsmasq binary on there?

I've rebuilt the packages non-stripped; same links as before (warning,
they're quite big now) :)

-Toke