Open Stack

On Mon, Nov 11, 2013 at 12:01 AM, Clayton Coleman <ccoleman at redhat.com>wrote:

> The implication I heard from the session is the command infrastructure and
> client code have or will have a well defined interface from the framework,
> although I haven't looked through current state enough to say that's
> totally true today.  If that interface is in place we certainly would be
> better isolated from upstream.
>

OSC has a very simple plugin mechanism that is little more than pre-defined
entry point groups that are scanned for additional commands.  That works
but doesn't properly leverage the desirable common bits, such as auth, that
are available to the built-in clients.  Enhancing this has not been a high
priority before now but needs to be done.

I guess it depends where we want to spend our focus as well - do we write
> something that in the long run we'd end up porting, or benefit from the
> work the common client is doing (doc, common patterns, keystone auth work
> with x509 and potentially Kerberos, bash completion, etc) as well as help
> them out.
>

I'd suggest you focus on putting together a clean and well-defined library
API that implements your REST API.  Consider that work is going on to
refactor the implementation of these layers in existing clients and that
some of the library APIs will be changing.  Specifically, the first of
these efforts is a much-needed cleanup of the Identity Auth interfaces and
the migration of existing clients to use that rather than their own
embedded auth.

Please don't just copy one of the existing clients, down that path lies
madness.

dt

-- 

Dean Troyer
dtroyer at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20131111/110f18a8/attachment.html>