[
https://issues.apache.org/jira/browse/HADOOP-15117?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16306632#comment-16306632
]
Chris Douglas commented on HADOOP-15117:
----------------------------------------
bq. In AbstractContractPathHandleTest, do we need the if checks in the catch
after 180? We should've made sure those were true before. Are we covering
anything there?
Yes, that's where the test verifies that if an exception was thrown, it must be
because the rename or the modification was disallowed (because it's the
expression in the {{try}} that throws). If one is allowed, then it must have
been the other that caused the exception. In the case where both are
disallowed, we don't need an assertion because we've already checked both
conditions.
bq. Can we describe the serialize in TestHDFSContractPathHandle? Maybe move it
from Boolean to boolean too.
Sure, I'll add a comment. IIRC the change to {{Boolean}} was in service of an
invocation problem from {{Parameterized}}, but it might have been trying to
work around the {{name}} issue. I'll make sure it's necessary.
> open(PathHandle) contract test should be exhaustive for default options
> -----------------------------------------------------------------------
>
> Key: HADOOP-15117
> URL: https://issues.apache.org/jira/browse/HADOOP-15117
> Project: Hadoop Common
> Issue Type: Test
> Reporter: Chris Douglas
> Assignee: Chris Douglas
> Attachments: HADOOP-15117.000.patch, HADOOP-15117.001.patch,
> HADOOP-15117.002.patch
>
>
> The current {{AbstractContractOpenTest}} covers many, but not all of the
> permutations of the default {{HandleOpt}}. It could also be refactored to be
> clearer as documentation
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
