[
https://issues.apache.org/jira/browse/TINKERPOP-2622?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Stephen Mallette closed TINKERPOP-2622.
---------------------------------------
Fix Version/s: 3.7.0
3.6.2
Assignee: Stephen Mallette
Resolution: Fixed
Fixed via CTR:
https://github.com/apache/tinkerpop/commit/4d347ce69e1d14bc7c539e6ba55c7a4a32c3c6f7
> Enforce ordering semantics in feature tests
> -------------------------------------------
>
> Key: TINKERPOP-2622
> URL: https://issues.apache.org/jira/browse/TINKERPOP-2622
> Project: TinkerPop
> Issue Type: Improvement
> Components: test-suite
> Affects Versions: 3.6.0
> Reporter: Stephen Mallette
> Assignee: Stephen Mallette
> Priority: Major
> Fix For: 3.7.0, 3.6.2
>
>
> The semantics for ordering aren't always clear and aren't particularly
> enforced especially when it comes to side-effects. We tend to take the
> position that Gremlin processes results as they arrive and it is that order
> that is maintained, but the test suite can be at odds with that as many tests
> are written without a specific processing order in mind and some
> implementations may not internally process traversers in the same order. For
> example consider {{g_V_group_byXageX}} which currently fails for Spark as
> {{V()}} returns in a different order there than how the assertion is written:
> {code}
> gremlin> g.V().group().by("age")
> ==>[null:[v[3],v[5]],32:[v[4]],35:[v[6]],27:[v[2]],29:[v[1]]]
> gremlin> g.V().order().by('name').group().by("age")
> ==>[32:[v[4]],null:[v[3],v[5]],35:[v[6]],27:[v[2]],29:[v[1]]]
> {code}
> We could try to enforce order as shown above but a more flexible assertion
> system for gherkin seems like a better choice for these cases. As of right
> now, tests that fail in this fashion are ignored.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
