From a past discussion with hyc on 2.5 requirements: [09:27] <hyc> we also need to audit all of these schema defs [09:27] <hyc> we're supposed to have official, non-experimental OIDs for released schema [09:28] <hyc> accesslog is still using 666, experimental arc [09:29] <hyc> I think this means we should polish up the logschema draft, Informational status, and publish it again as final
How about just giving up this .666-is-experimental nonsense, define .666 assignments as "official" and leave all OID assignments as is?
(In reply to Michael Ströder from comment #1) > How about just giving up this .666-is-experimental nonsense, define .666 > assignments as "official" and leave all OID assignments as is? Something like that. We probably should never have had such a grab-bag "experimental" arc in the first place. A full audit of what uses .666 is still a a necessary step before deciding what else to do though. In the future, we should simply allocate per-feature arcs, which remain unchanged throughout a feature's evolution from experimental to release. Just like the config schema delegates sub-arcs for every backend and overlay, etc. We're probably stuck with the current .666 usages. Again, OIDs aren't supposed to change after they've been used in the wild. But someone should still document what the current .666 uses are.
(In reply to Howard Chu from comment #2) > (In reply to Michael Ströder from comment #1) > > How about just giving up this .666-is-experimental nonsense, define .666 > > assignments as "official" and leave all OID assignments as is? > > Something like that. We probably should never have had such a grab-bag > "experimental" arc in the first place. Yes! > In the future, we should simply allocate per-feature arcs, Full ack! > We're probably stuck with the current .666 usages. Again, OIDs aren't > supposed to change after they've been used in the wild. But someone should > still document what the current .666 uses are. This means e.g. updating this: https://tools.ietf.org/html/draft-chu-ldap-logschema-00 BTW: I can't find this draft in doc/drafts/ though...
Need to move everything to non-experimental OID arc
Keep all OIDs as is Document all OIDs that are not currently documented.
unclear where/how to document