Closed michaelosthege closed 6 years ago
The trailing .
is not required.
In a DNS message, a domain name is never encoded as a "string" but as a sequence of length-prefixed labels with a ending length-prefix of zero. For example, both a.b
and a.b.
are encoded as "01 61 01 62 00".
Also, a domain name can be compressed, if the leading label(s) have already been encoded; see Section 4.1.4 of RFC1035.
See WireWriter.WriteDomainName for the details.
@michaelosthege I'm happy to add features you need for your project. Please raise issues.
@richardschneider thank you for this clarification. We will change to net-mdns
in the near future.
I raised issue #29 because this is of importance for our context (communication protocol for laboratory devices). We have to deal with corporate networking, where an often-seen configuration is that the firewall prevents opening UDP:5353
on some network interfaces for security reasons.. As I described in my updated issue description, this causes some trouble with the way how the UdpClient
is created. I will probably fork net-mdns
and see if I can add optional network interface params
.
@michaelosthege thanks for the update.
As you may guess, I'm a solo developer and don't deal with "corporate firewalls." However, I would expect a firewall to simply not forward the datagram on a restricted interface/port; which should not cause an issue with creating a UdpClient
.
Can you open a new issue that shows the actual exception (and stack trace) you are getting?
As always, I'm happy to get a PR.
Hello @richardschneider ,
Together with @gaviriar (and others) we intend to use
Makaretu.Mdns
as a dependency for a protocol that relies on mDNS. After reading RFC6762 our understanding is that for mDNS every service should announce itself with alocal
domain, however we are not certain if this should be.local
or the fully-qualified.local.
.In ServiceProfile.cs#L137 it looks like the
FullyQualifiedName
does not include a trailing.
, which makes it (by definition, as I understand it) not fully-qualified, but just relative.Would you mind having a look at this, to determine if this is a bug?
kind regards, Michael