On 4 October 2016 at 12:02, Oliver Lietz <apa...@oliverlietz.de> wrote:
> On Tuesday 04 October 2016 11:24:38 Ian Boston wrote: > > Hi, > > I have tried many times to use Pax Exam, and always got stuck in its > > complexity and opacity, this time was no different, hence why I resorted > to > > the crankstart provisioning model. Obviously your mileage varies, but > > thanks for the pointer. > > I tried all of our test modules and setups and finally came to generate Pax > Exam Options from Sling's Features – works like a charm... but YMMV. > > > The feature.xml looks extensive at 557 lines > > compared to the 50 or so far less extensive lines of the provisioning > > model. > > > > I dont really care if the only way to start an OSGi container is to use > > feature.xml as long as I can continue to use a simple provisioning model > to > > generate it, ideally not in XML, and ideally able to be minimalist. > > Well, the features file contains 67 features you can choose from. Comparing > that to 50 lines of provisioning model doesn't make sense. In Sling > Lauchpad > you can choose between Oak Tar and Oak Mongo – that's it. And building a > custom Sling with provisioning model means copying 50 lines around, right? > Sure, agreed. I have a library of models for all sorts of scenarios, mostly in Mesos, Docker, Kubernetes containers, with different configs, different repo-inits etc. Saves me a lot of disk space and a lot of time. Easy to manage and each deployment is 100% clear what it does. Just sharing what has worked for me, because I thought it might help. ymmv. I have nothing against the Karaf XML or Karaf, but I would rather not be forced to use it. I see the proposal on one of these threads was to continue to support provisioning models, so I'll step out. Ian. > > Regards, > O. > > > Best Regards > > Ian > > > > On 4 October 2016 at 10:54, Oliver Lietz <apa...@oliverlietz.de> wrote: > > > On Tuesday 04 October 2016 10:19:06 Ian Boston wrote: > > > > Hi, > > > > > > Hi Ian, > > > > > > > On 3 October 2016 at 11:37, Robert Munteanu <romb...@apache.org> > wrote: > > > > > On Wed, 2016-09-28 at 21:37 +0000, Stefan Seifert wrote: > > > > > > the launchpad should be split up in separates > features/provisioning > > > > > > fragments to make it possible to easily build a "minimal > launchpad" > > > > > > or "minimal launchpad + X" by selection only some but not all > > > > > > features. > > > > > > > > > > +1, this would be very useful. > > > > > > > > The crankstart provisioning model at [1] is a minimal launchpad (24 > > > > bundles). Doesn't even have a repository. > > > > In the case of the integration test is started by [2], but could > equally > > > > > > be > > > > > > > started on the command line. > > > > Model contents can be found at [3], not all are used. > > > > > > > > Is that a starting point for what you were looking for ? > > > > > > the split up is already done for Karaf [1] and generated from that are > > > Options > > > for Pax Exam[2]. Getting a minimal Sling for Pax Exam is just using > > > sling(). > > > As it is in Java, you can assemble different Sling "Launchpads" in the > > > same > > > module (different IT classes) also. > > > > > > Regards, > > > O. > > > > > > [1] https://github.com/apache/sling/blob/trunk/karaf/org. > > > apache.sling.karaf-features/src/main/feature/feature.xml > > > [2] > > > https://github.com/apache/sling/tree/trunk/testing/org. > > > apache.sling.testing.paxexam > > > > > > > Best Regards > > > > Ian > > > > > > > > > > > > 1 > > > > https://github.com/apache/sling/blob/trunk/contrib/ > > > > > > commons/mom/examples/jobs > > > > > > > /it/src/test/java/org/apache/sling/jobs/it/Models.java#L16 2 > > > > https://github.com/apache/sling/blob/trunk/contrib/ > > > > > > commons/mom/examples/jobs > > > > > > > /it/src/test/java/org/apache/sling/jobs/it/TestSuiteLauncher > IT.java#L33 > > > > > > 3 > > > > > > > https://github.com/apache/sling/tree/trunk/contrib/ > > > > > > commons/mom/examples/jobs > > > > > > > /it/src/test/resources > > > > > > > > > Robert > > >
