[
https://issues.apache.org/jira/browse/TINKERPOP3-688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14548467#comment-14548467
]
Bob Briody commented on TINKERPOP3-688:
---------------------------------------
[~pietermartin] I agree that there is significant coupling between these tests
and how the vendor executes the Traversal. That said, I was unable to find a
suitable way of maintaining integration-style test coverage while improving
generality. Additionally, HadoopGraph, Neo4jGraph, TitanGraph, and TinkerGraph
do all pass the tests.
I did add a new test method {code}g_V_out_out_profile_simple{code} which will
provide some fully generic Profile test coverage. I suggest that SQLg graph
OptOut of the other tests as necessary. I'm not fully satisfied with this
approach but I do not have a better alternative at the moment. What do you
think?
> ProfileTest fails on optimized queries
> --------------------------------------
>
> Key: TINKERPOP3-688
> URL: https://issues.apache.org/jira/browse/TINKERPOP3-688
> Project: TinkerPop 3
> Issue Type: Bug
> Components: test-suite
> Reporter: pieter martin
> Assignee: Bob Briody
> Fix For: 3.0.0.GA
>
>
> {code}ProfileTest.g_V_out_out_modern_profile{code} fails because it asserts
> the traversal as executed by TinkerGraph.
> As sqlg optimizes the query the metrics being asserted are incorrect.
> I doubt it will be possible to hardcode metrics in a generic test as
> different backends will have different behaviours.
> eg. Tinkergraph
> {code}pre-strategy:[GraphStep([],vertex), VertexStep(OUT,vertex),
> VertexStep(OUT,vertex), ProfileStep]
> post-strategy:[TinkerGraphStep([],vertex), ProfileStep,
> VertexStep(OUT,vertex), ProfileStep, VertexStep(OUT,vertex), ProfileStep]
> {code}
> eg. Sqlg
> {code}pre-strategy:[GraphStep([],vertex), VertexStep(OUT,vertex),
> VertexStep(OUT,vertex), ProfileStep]
> post-strategy:[SqlgGraphStep([],vertex), ProfileStep, SqlgVertexStepCompiled,
> ProfileStep]
> {code}
> As you can see the number of steps are not the same after the optimization.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
