status

[Francis Dupont]

> 4) My cvs skillz are rusty and what I'd wanted to come up with was a
> patch for dhcp from this branch that applied against the mainline...

=> if you have access to the ISC wiki there are some hints about CVS
(like cvs rdiff -kk -u -r <branch1> -r <branch2> <product>).
BTW you have similar hints about git in the bind10 trac/wiki,
and I don't know for svn (but does svn need them? :-).

> Some more comments below.
> 
> On Mon, Mar 12, 2012 at 3:30 PM, Francis Dupont <fdupont at isc.org> wrote:
> > AFTR: tested the pseudo SD-AFTR code, including the ICMP (which requires
> > a default private clause to raise the expected error in natin()),
> > saved the config files. Copied and updated the echo match extension
> > for netfilter from the old netfilter-based AFTR branch, including
> > the ping which comes with it, so one can route on the ICMP echo ID.
> 
> This is what is in one of the repos I know about?

=> the SD-AFTR is in rt28100 but I merged it into rt28354 so you can
forget rt28100 (if you don't want to review it). All these branches
are in the aftr/usermode[.git] on repo.isc.org

> > PCP: nothing new (but just received an announce about a new upnp ietf I-D)
> 
> I see pcpd and various other binaries here, not a lot of
> documentation. Which of these binaries
> is required on the server and which on the client.

=> the doc is in the doc directory and as it is not yet a product
(nor the PCP base spec a RFC) there is no generic thing.
To summary client library is libpcp et client server is pcpd.

> > CPE: tested the sdctld stuff, move the build to autoreconf --install
> > (i.e., same than bind10): should be easier to package for OpenWRT?

> preferred packaging method is from a release tarball + patches,
> using autotools.

=> so you prefer the autotools to broken Makefiles, don't you?
I left only the configure.ac and the Makefile.am files so
autoreconf --install runs the 4 autotools and configure + make
are enough to build the whole stuff.

> > Demo: still have no idea (at the exception it is CPE, not PCP demo)
> > about what we'll be supposed to do...
> 
> I thought we were going with server -> cpe -> one or more clients

=> yes but which hardware, virtual machines on which hardware, etc...

Regards

Francis Dupont <fdupont at isc.org>