[bind10-dev] config experiment

[JINMEI Tatuya / 神明達哉]

At Wed, 16 Sep 2009 11:43:17 +0200,
Jelte Jansen <jelte at isc.org> wrote:

> > BTW: are these notifications supposed to be delivered via
> > "communication channels"?  If so, the interface may also need to be
> > generalized accordingly (I don't have a specific image about how it
> > looks like, though).

> i've been thinking a bit on another approach, based on the
> specification-language stuff i sent in
> https://lists.isc.org/pipermail/bind10-dev/2009-September/000113.html

[snip]

> For the module part, the generator could build something with a predictable API
> that simply needs to be included in the module (as in if the model contains a
> field 'data' it would make a getData()), in the language of choice of the module
> writer.

It may depend on further details but I think the proposed architecture
looks reasonable.  To me, the important point is to make the API for
the "module"s as general as possible so that the underlying "channel"
will be hidden from the module developers (it may be a local function
call or some communication channel to a remote manager; the channel
could also be implemented several ways).

I'd not worry about how the "skeleton" is loaded at the manager
(although I may not understand the point).  If it's almost empty I
guess we could simply hardcode it in the manager implementation like
BIND9 hardcodes configuration defaults.  Even if it's more complicated
and/or expected to be changed relatively frequently, I guess the
loading overhead is marginal since it will be a rare event.

---
JINMEI, Tatuya