Bug 393073 - NPE in the hasNext method

[berezovskyi]

Current status is that the empty nextPage property:

<oslc:nextPage rdf:resource="" />

Causes Jena to deserialise a full triple statement with a non-empty
Object:

subject = "http://example.com/provider/query?oslc.where=ex%3Aproduct%3D%22Product%20A%2"
predicate = "http://open-services.net/ns/core#nextPage"
object = "http://example.com/provider/query"

[eclipsewebmaster]

Issue tracker reference:
https://bugs.eclipse.org/bugs/show_bug.cgi?id=393073

[berezovskyi]

@jamsden I will let you handle this client PR

[berezovskyi]

If you need my help, I can try updating the PR.

[jamsden]

As far as I can see, this looks ok. Have the test been run and demonstrated correct results?

[berezovskyi]

Yes, here is the test: https://github.com/eclipse/lyo.client/pull/5/files#diff-0865771cb59c619976d66fa48ba77727R75

[berezovskyi]

There is a hiccup with re-running the tests. Seems like newer Jena versions have become more strict regarding the URIs, the test fails because of an unencoded space character. Will notify you once it's fixed.

…

-- /Andrew 15 feb. 2019 kl. 22:10 skrev Jim Amsden <notifications@github.com<mailto:notifications@github.com>>: @jamsden commented on this pull request. As far as I can see, this looks ok. Have the test been run and demonstrated correct results? — You are receiving this because you were assigned. Reply to this email directly, view it on GitHub<#5 (review)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AAD83nN4QKjArOxPpV_PbGB_cKMIxBkKks5vNyJKgaJpZM4UFN8X>.

[berezovskyi]

Strange! Now the test that used to pass a year ago does not pass...

<oslc:nextPage rdf:resource="" />

This gets recognised as a triple ?s oslc:nextPage ?s (object does not have the query param though)!

Hmm... according to https://www.w3.org/TR/rdf-syntax-grammar/#section-baseURIs

An empty same document reference "" resolves against the URI part of the base URI

[berezovskyi]

@jamsden I think d2026d9#diff-5776f1d67363c3a733d3c944bd2f24f3R146 fix is even better to protect against NPEs. Ready for review/merge.

[berezovskyi]

need to check for possible regression