[Ovirt-devel] [PATCH 4/5] Find hardware_pool by path; search by parent_id

David Lutterkort dlutter at redhat.com
Wed Aug 13 21:17:38 UTC 2008

On Wed, 2008-08-13 at 13:34 +0100, Daniel P. Berrange wrote:
> On Wed, Aug 06, 2008 at 05:00:18PM -0700, David Lutterkort wrote:
> > Get individual hardware pool with a path, e.g. with a GET to
> >   /ovirt/hardware_pools?path=/default/foo/bar
> > 
> > Get children of a hardware pool from
> >   /ovirt/hardware_pools?parent_id=1
> > 
> > Get specific child with
> >   /ovirt/hardware_pools?parent_id=1&name=foo
> Guess I have similar questions about the URI syntax here as the previous
> mail. Should we be encoding the name of the object as a parameter, vs
> part of the URI path ? I've always thought of the URI path as providing
> lookup for the object, and the parameters as filters/modifiers for the
> data returned / action performed

I assume this is about the path parameter: on one hand it would be
nicer, on the other it would keep us from addressing nested objects,
e.g. the hosts in a hardware pool as /ovirt/hardware_pools/1/hosts - is
that the hosts for the hw pool with id 1 or is it the hw pool 'hosts'
inside the pool '1' ?

In any event, the exact structure of the URL's is an implementation
detail that isn't exposed to somebody scripting against the API. (at
least not in Ruby)


More information about the ovirt-devel mailing list