Thx Dan! Answering inline ..
On 21.02.2018 23:48, Dan Haywood wrote: > Hi Andi > > good catch. > > I think we should change the docs to match the code rather than the other > way around; otherwise we run the risk of changing the behaviour of the > framework which might cause issues. I've updated the doc to conform with current code behavior and fixed the logic of the particular test. > What happens with auditing, is there a discrepancy there too, I wonder? No discrepancy here ... Auditing behaves according to doc: falls back to 'enabled', if invalid or missing configuration Cheers Andi! > > Thx > Dan > > > On Wed, 21 Feb 2018 at 21:08 Andi Huber <[email protected]> wrote: > >> Hi Dan, >> >> I'm in the processes of fixing some tests in module metamodel on the >> 'master' branch. >> >> If I interpret the doc [1] correctly, then we have a discrepancy between >> doc and current implementation: >> >> Given (invalid or missing) configuration property >> isis.services.publish.objects=foobar >> >> according to the doc should be picked up by the framework equivalently to >> isis.services.publish.objects=all >> >> If there is no configuration property in isis.properties then publishing >> is automatically enabled for domain objects. >> >> >> But in code we do the opposite (isis.services.publish.objects=none). >> Which way should it be? >> >> Cheers, Andi >> >> [1] >> https://isis.apache.org/guides/rgant/rgant.html#_rgant-DomainObject_publishing >>
