On 06/26/2014 07:36 AM, Daniel P. Berrange wrote: >>> Both of those return a list of lease structures and the only difference >>> is the presence of the @mac argument. >>> >>> We could unify those two APIs into one with the following signature: >>> >>> int >>> virNetworkGetDHCPLeases(virNetworkPtr network, >>> const char *mac, >>> virNetworkDHCPLeasePtr **leases, >>> unsigned int flags) >>> >>> And tweak the semantics of @mac where when the user passes NULL we'd >>> return the complete unfiltered list. I'm in favor of this simplification as well. > Of course what I completely missed was that even in the ForMAC > case, we have to return a dynamic list of leases, because you > can have both IPv4 and IPv6 leases for the same MAC. This > basically kills the main compelling reason to have 2 separate > APIs. I remember the earlier debate, and think I kind of missed that point at the time, as well. > > So in retrospect I was wrong, and I agree with Peter that we > should kill the ForMAC API and just add an (optional) macaddr > parameter to the first API. Of course we can only decided to > do this now before we release. Yes, now is the time to make the fix, before RC-2 is spun (so we still have some test time), and before it is baked into the 1.2.6 release. -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org
Attachment:
signature.asc
Description: OpenPGP digital signature
-- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list