On Tuesday 20 September 2016 16:04:07 Carsten Ziegeler wrote:
> > On Tuesday 20 September 2016 14:57:40 Robert Munteanu wrote:
> >> On Tue, 2016-09-20 at 13:49 +0200, Carsten Ziegeler wrote:
> >>> I think we created a monster when it comes to configuration
> >>> requirements
> >>> with the Oak server bundle.
> >>> Have a look at  where I had to add dozens of configurations just
> >>> to
> >>> get a JCR repository running. And this requires not only OSGi
> >>> configurations but also framework properties.
> >>> Let's start with that one: It seems the tar implementation is fine
> >>> without a system property and creates a directory in the current dir.
> >>> However, the indexing fails and requires "repository.home" to be set.
> >>> I
> >>> think this should default to something valuable as well. There is no
> >>> point in failing.
> >>> Then we have all these OSGi configurations - I really think we should
> >>> not need them and simply deploying the right bundles should be
> >>> enough.
> >>> It's a nightmare right now.
> >> I think most of the configurations are required by Oak itself, as
> >> opposed to the oak-server bundle.
> > Right, this is an Oak issue. See  for our Oak Server bundle.
> >> Do you see something that we can fix in Sling?
> > Lucene Indexing should be optional, SLING-5796.
> I have no special indexing bundle added, but nevertheless I got an
> exception with the above mentioned message.
oak-lucene – when it's optional in our Oak Server we can use an in-memory node
store in our tests and get rid of any file paths configurations (and disk
> > And use Testing PaxExam which comes with all required configurations.
> I want to keep my dependencies low
One bundle (3 classes) vs. lots of boilerplate code... YMMV.
But Oak really should provide sensible defaults and document what's
> > Regards,
> > O.
> > 
> > https://svn.apache.org/repos/asf/sling/trunk/bundles/jcr/oak-server/src/m
> > ain/java/org/apache/sling/jcr/oak/server/internal/OakSlingRepositoryManage
> > rConfiguration.java>
> >> Robert