Jernej Tuljak <jernej.tul...@mg-soft.si> wrote: > > > On 30/09/2021 10:48, Martin Björklund wrote: > > Hi, > > > > Jernej Tuljak <jernej.tul...@mg-soft.si> wrote: > >> Hi, > >> > >> can someone clarify the meaning of the following bullet in Section 11, > >> RFC7950: > >> > >> o A "base" statement may be removed from an "identityref" type, > >> provided there is at least one "base" statement left. > >> > >> This seems to enable the value space of a definition to narrow across > >> revisions of the same module. > > "base" statements are ANDed, so removing one expands the value space. > > > > 9.10.2 says: > > > > Valid values for an identityref are any identities derived from all > > the identityref's base identities. > > > > Oh, my. That is not how I understood "identityref"s with multiple > bases worked at all.
If you want OR-semantics you can use a union. /martin > > Jernej > > > /martin > > > > > >> Most other changes in this section seem > >> to only allow expansion of the value space, relaxing of constraints, > >> etc. What kind of revisions does this bullet cover? > >> > >> Jernej > >> > _______________________________________________ netmod mailing list netmod@ietf.org https://www.ietf.org/mailman/listinfo/netmod
