[
https://issues.apache.org/jira/browse/TINKERPOP-1258?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15241347#comment-15241347
]
pieter martin commented on TINKERPOP-1258:
------------------------------------------
I added a similar same optimization to Sqlg.
However the test still fails, this time on the {{assertNotEquals(traversala1,
traversalb1);}}
The reason for this is because the default {{GraphStep.equals}} ultimately
relies on {{GraphStep.ids}} and Sqlg's optimization/compilation rewrites the
{{ids}} as a {{HasStep}}.
So for both traversals the ids are empty and thus equal.
This means I need to have a better custom {{equals}} implementation.
So I am ok with excepting that this failure is due to Sqlg's incorrect
{{equals}}.
However, this test is testing a particular strategy which is where implementors
will differ so perhaps the test itself should not be in the
{{ProcessStandardSuite}} but rather a {{TinkerGraph}} specific test?
At least for the part that asserts the traversal equality.
> HasTest.g_V_hasId_compilationEquality makes GraphStep assumptions
> -----------------------------------------------------------------
>
> Key: TINKERPOP-1258
> URL: https://issues.apache.org/jira/browse/TINKERPOP-1258
> Project: TinkerPop
> Issue Type: Bug
> Components: test-suite
> Affects Versions: 3.2.1
> Reporter: pieter martin
> Priority: Minor
>
> {{HasTest.g_V_hasId_compilationEquality}} assumes that implementations that
> extends {{GraphStep}} has optimized {{hasId}}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
