ignite-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Oleg Ignatenko (JIRA)" <j...@apache.org>
Subject [jira] [Assigned] (IGNITE-10191) Incorrect comparison of lists in RendezvousAffinityFunctionSimpleBenchmark#testAffinityCompatibility
Date Tue, 20 Nov 2018 13:48:00 GMT

     [ https://issues.apache.org/jira/browse/IGNITE-10191?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Oleg Ignatenko reassigned IGNITE-10191:
---------------------------------------

    Assignee: Oleg Ignatenko

> Incorrect comparison of lists in RendezvousAffinityFunctionSimpleBenchmark#testAffinityCompatibility
> ----------------------------------------------------------------------------------------------------
>
>                 Key: IGNITE-10191
>                 URL: https://issues.apache.org/jira/browse/IGNITE-10191
>             Project: Ignite
>          Issue Type: Bug
>    Affects Versions: 2.6
>            Reporter: Oleg Ignatenko
>            Assignee: Oleg Ignatenko
>            Priority: Major
>              Labels: MakeTeamcityGreenAgain
>
> RendezvousAffinityFunctionSimpleBenchmark#testAffinityCompatibility (in current codebase
muted by renaming to {{_testAffinityCompatibility}}) looks troublesome: apparent bug is incorrect
comparison of lists expecting elements to be always in the same order which doesn't look like
the case for the tested API:
> {code}            List<List<ClusterNode>> assignment0 = assignPartitions(aff0,
nodes, null, backups, 0).get2();
>                   List<List<ClusterNode>> assignment1 = assignPartitions(aff1,
nodes, null, backups, 0).get2();
>                   assertEquals (assignment0, assignment1);
> {code}
> Though test kept failing even after I experimented with replacing comparison to one that
was insensitive to the order of list elements.
> Brief checking of code intended to be tested suggests that maybe it isn't even supposed
to be deterministic - in case if this is correct test should be very thoroughly redesigned.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message