Fix parsing AMTRELAY no gateway relay

[mozzieongit]

As mentioned in #256 and nsd#444, the AMTRELAY RFC requires the relay to be . in the presentation format:

If the relay type field is 0, the relay field MUST be ".".

Fixes: #256

I noticed that dig also omits the . when printing the AMTRELAY RR with gateway type 0.

[wcawijngaards]

From reading RFC 8777, it looks like it certainly says that the relay field has to be '.' when the type is 0 no gateway. The fix looks fine for code, I wanted to improve the error printout to explain what is causing error if the field is wrong or omitted from input.

[mozzieongit]

I ended up forgetting to check, but did it now:
When omitting the field, it only prints Missing relay in AMTRELAY. Only when adding a bogus relay, it prints the nice message explaining the requirement of ..

Unfortunately, adding the nice description to the missing relay error would require extending the take_contiguous and maybe_take_contiguous functions to take an optional string to concatenate to the existing error log (as they call the SYNTAX_ERROR macro stating the missing field, and not the parsing function, see types.h parser.h parser.h)

[wcawijngaards]

Okay that is maybe not really worth it. The extra printout in the error case where it is now is nice, but extending the functions may not be best for parse performance by complicating them. So I guess we keep it as it is.