Updates for migrating from standalone Fuseki to fuseki.war. Address F…

[ruebot]

…CREPO-1467.

https://jira.duraspace.org/browse/FCREPO-1467

[ajs6f]

I have nothing against this PR, but Fuseki 1 is getting a bit longer in the tooth and Fuseki 2 is clearly where the Jena folks are putting more attention. See here for an example of how to use Fuseki 2 to support integration tests with the Maven plugin for Cargo.

[ruebot]

@ajs6f this is for moving to Fuseki 2 in fcrepo-exts/fcrepo-vagrant#18. That said, are you saying that we should update the integration tests to account for that? I think that is what you're saying, but I just want to make sure 😄

[ajs6f]

I don't care at all about fcrepo4-vagrant in this context (not saying it's unimportant, just saying it doesn't matter to me when looking at ITs here). Generally, I'm saying we should move to Fuseki 2 in all places. Including these ITs.

[ajs6f]

Because EmbeddedFusekiServer is a Fuseki 1 class.

[acoburn]

The current integration test uses fuseki 1.1.1, and I am also in favor of moving that to 2.x, but this PR is completely unrelated to that. That is, this PR is about updating the default properties so that they assume a war-deployed fuseki rather than a standalone fuseki.

[ruebot]

Shall I create a JIRA ticket?

[acoburn]

@ruebot that would be great, thanks

[ajs6f]

If they are assuming a war-deployed Fuseki, why is there any use of EmbeddedFusekiServer?

[ruebot]

Also, whoever wrote those integration tests, they were extremely helpful for getting this updated. So, thank you very much 😄

[acoburn]

@ajs6f the EmbeddedFusekiServer class is for doing integration tests. The vagrant project is for actually running the fuseki server.

[ajs6f]

Urg, that's my point. If this PR is about relying on a deployed WAR, then using EmbeddedFusekiServer is a fail, because it is exactly relying on something other than a WAR.

[acoburn]

@ajs6f The point of the integration test in question is to test how the camel route interacts with an HTTP endpoint backed by some triplestore. In this case it happens to be fuseki. My point is that it shouldn't matter what version or what deployment strategy is used for that HTTP endpoint: the endpoint should be the same. The camel route is completely indifferent about what triplestore is used and what deployment strategy is used for that triplestore.

[ajs6f]

Sure. This isn't about the validity of the test. It's about the dependency footprint (which is always an issue) and whether the tests are useful in explaining to people how to set up their own integrations, which is one of the purposes of writing integration tests. As long as it gets fixed, I don't really care where it gets fixed, in this PR or another, but using EmbeddedFusekiServer is outdated practice and we shouldn't be doing it.

[ruebot]

Let me know if this captures what y'all are looking for.

[acoburn]

@ajs6f That's a fair point, and I agree about updating fuseki. Do you have a preference between cargo-maven2-plugin and jetty-maven-plugin? (We're currently using jetty for other tests, but I'm not particularly attached to it)

[ajs6f]

I'm inclined towards Cargo because it generalizes (e.g. you can quickly install profiles to test in other containers, which is a really good habit that we rarely evidence).

[awoods]

It sounds like we have landed on moving towards a cargo-based, war deployment of Fuseki2 for the integration tests in this project... per the ticket created by @ruebot:
https://jira.duraspace.org/browse/FCREPO-1708

Otherwise, this current PR looks good. Are there objections to moving the PR into master? ...pending more testing.

[acoburn]

I'm 👍 on merging this PR with master, provided it works well with the vagrant setup

[ruebot]

From the JIRA ticket:

Updated fcrepo-camel-toolbox, built, and deployed artifact on vagrant VM. I added a new object and was successful.

DEBUG 11:25:24.899 (InternalAuditor) Event detected: <anonymous> /audit/4f/b0/a3/4f/4fb0a34f-eda7-45b9-9186-8e0716d9fa85
DEBUG 11:25:24.899 (AuditUtils) Constructed event type URIs: http://fedora.info/definitions/v4/repository#NODE_ADDED,http://fedora.info/definitions/v4/repository#PROPERTY_ADDED
INFO 11:25:25.281 (FedoraLdp) HEAD for: 4e/20/c6/f3/4e20c6f3-2605-4107-a14b-a6f491a73d64
INFO 11:25:25.304 (FedoraLdp) HEAD for: 4e/20/c6/f3/4e20c6f3-2605-4107-a14b-a6f491a73d64
WARN 11:25:25.347 (RDFLanguages) java-jsonld classes not on the classpath - JSON-LD input-output not available.
WARN 11:25:25.367 (RDFLanguages) Minimum jarfiles are jsonld-java, jackson-core, jackson-annotations
WARN 11:25:25.368 (RDFLanguages) If using a Jena distribution, put all jars in the lib/ directory on the classpath
INFO 11:25:25.441 (FedoraLdp) GET resource '4e/20/c6/f3/4e20c6f3-2605-4107-a14b-a6f491a73d64'
INFO 11:25:25.484 (FedoraLdp) GET resource '4e/20/c6/f3/4e20c6f3-2605-4107-a14b-a6f491a73d64'
INFO 11:25:26.353 (SolrRouter) Indexing Solr Object /4e/20/c6/f3/4e20c6f3-2605-4107-a14b-a6f491a73d64
INFO 11:25:26.509 (FedoraTransform) GET transform, 'default', for '4e/20/c6/f3/4e20c6f3-2605-4107-a14b-a6f491a73d64'
INFO 11:25:26.967 (audit) Audit Event: /4e/20/c6/f3/4e20c6f3-2605-4107-a14b-a6f491a73d64 :: <http://example.com/event/4f/b0/a3/4f/4fb0a34f-eda7-45b9-9186-8e0716d9fa85>
[2015-08-25 11:25:27] Fuseki INFO [11] POST http://localhost:8080/fuseki/test/update
[2015-08-25 11:25:27] Fuseki INFO [11] POST /test :: 'update' :: [application/x-www-form-urlencoded] ?
INFO 11:25:27.665 (TriplestoreRouter) Indexing Triplestore Object /4e/20/c6/f3/4e20c6f3-2605-4107-a14b-a6f491a73d64
INFO 11:25:27.716 (FedoraTransform) GET transform, 'default', for '4e/20/c6/f3/4e20c6f3-2605-4107-a14b-a6f491a73d64'
[2015-08-25 11:25:27] Fuseki INFO [12] POST http://localhost:8080/fuseki/test/update
[2015-08-25 11:25:27] Fuseki INFO [12] POST /test :: 'update' :: [application/x-www-form-urlencoded] ?
[2015-08-25 11:25:27] Fuseki INFO [11] 200 OK (732 ms)
[2015-08-25 11:25:28] Fuseki INFO [12] 200 OK (293 ms)