We could allow the User to specify either a --type or a combination of --data-policy and --scope but not all three.
Apparently the introduction of RegionShortcut (and its use for --type) was supposed to simplify things. AttributesFactory was deprecated in favor of using shortcuts. The shortcuts combine Scope, DataPolicy, InterestPolicy, InterestResultPolicy, EvictionAttributes/Algorithm, ExpirationAttributes/Action, etc. Specifying all of the legal combinations with enum Shortcuts was the direction taken to eliminate incompatible combinations. I'm not sure adding --data-policy and --scope would be enough. I think we'd have to add the others as well to give full flexibility to justify having them. It's probably simpler at this point to introduce shortcuts for global and its few variations. Adding anything else probably warrants further discussion. At a minimum, I gather that global regions are not deprecated (just not encouraged due to performance compared with other scopes). I'll file a bug against gfsh because we can't create a global region. -Kirk On Mon, Nov 28, 2016 at 3:40 PM, Udo Kohlmeyer <ukohlme...@pivotal.io> wrote: > I think we need to add --data-policy and --scope to GFSH. > > --Udo > > > > On 11/28/16 15:39, Kirk Lund wrote: > >> Yep, the --type argument in gfsh matches the enum values of >> org.apache.geode.cache.RegionShortcut which doesn't include any shortcuts >> for global. >> >> >> On Mon, Nov 28, 2016 at 3:17 PM, Udo Kohlmeyer <ukohlme...@pivotal.io> >> wrote: >> >> I think it was never added. I think we keep on favoring shortcuts above >>> correct configuration properties. >>> >>> --Udo >>> >>> >>> >>> On 11/28/16 15:15, Kirk Lund wrote: >>> >>> Looks like global regions cannot be created using GFSH. We should either >>>> add support for global scope to GFSH or remove global scope (and the >>>> RegionLock APIs). >>>> >>>> Does anyone know if global scope is officially deprecated? Or why it's >>>> not >>>> supported by GFSH? >>>> >>>> -Kirk >>>> >>>> >>>> >
