On 05/10/2016 05:10 AM, Daniel P. Berrange wrote: > On Mon, May 09, 2016 at 06:53:01PM -0400, Cole Robinson wrote: >> On 05/09/2016 09:48 AM, Daniel P. Berrange wrote: > > IMHO it is a total failure if we require the application to extend its > parser every time we add a new enum to the domain capabilities. We have > the ability to design something that is data driven - we should not build > something it is forced to be code driven with code changes for every > libvirt addition. This is ignoring the point I made previously that the schema is not already fully introspectable. Most of the current enums cannot be programmatically consumed anyways, because there's no way to map an enum name to its domain XML property. So if that's our goal to be data driven, we should address that issue first. I can't really think of a good way to represent that without nesting deeply or using specially formatted strings. Do you have a suggestion for that? Thanks, Cole -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list