Re: dnscurve updates

On Sun, Jul 5, 2009 at 6:30 PM, Matthew Dempsky<[email protected]> wrote:
> On Sun, Jul 5, 2009 at 9:19 AM, Andy
> Bradford<[email protected]> wrote:
>> I for one  look forward to running  DNSCurve as soon as I  find a decent
>> implementation.
>
> Cache, forwarder, or both?  Have you looked at any of the existing
> prototype-ish implementations?  Feedback/constructive criticism is a
> big motivator for me to work on them. :)

Same here.  I've got your second patch merged into zinq-djbdns and it
seems to work.  I need to do some more testing and ship it as a
feature that can be enabled at compile time.  However, there is a
distinct lack of folks asking me for it.  That's never stopped me
before, I didn't hear from any confirmed users until 0.07, and as of
0.08 the count is still at 1.  I've had a lot of distractions over the
last month or so, though.  I'll get back to it eventually on my own,
but a request from an interested end-user would make it happen a lot
sooner.

I suppose, though, that DNSCurve support for dnscache isn't all that
interesting as a feature until there are DNSCurve enabled
authoritative servers deployed to bang on.  Your test server is the
only one I know of.  A DNSCurve patch for tinydns would maybe liven
things up a bit, but a cleaned up version of your DNSCurver forwarder
would let anybody join the party.

Also, at some point, we're going to need a version of NaCl that will
actually install libnacl.  I'm sure Dan will eventually roll his own
solution, but no telling how long that will take.  I've thought about
doing an autotool conversion, but the pile 'o shell scripts that he's
using stopped me in my tracks.  I suppose I could just let his stuff
handle compilation and have Libtool take care of installation once
libnacl.a is built, but I really don't want to deal with having to
support that.  When his stuff works, it works, but when it doesn't,
you have to go digging in the logfile and reading through a bunch of
shell scripts to figure out what went wrong.  No thanks.