[
https://issues.apache.org/jira/browse/HADOOP-12059?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14574558#comment-14574558
]
Sean Busbey commented on HADOOP-12059:
--------------------------------------
{quote}
bq. it seems to me that wrapLocalFileUri should really be added to
ProviderUtils rather than to the credential provider itself. There is a bit of
a blurred line due to the fact that it creates a localjceks based URI but since
it isn't part of the CredentialProvider interface and is basically needed for
tests - I think it should go in either the ProviderUtils class or just be part
of the test.
I'm happy to place the helper function where ever y'all would like. However,
it's important to note that the wrapper is just implementing the munging
described in the javadocs for LJKSP, it's not the same as the one described in
ProviderUtils.unnest (though that one is a superset). I figure it ought not be
a part of the specific test since it's generally useful for other tests that
need to similarly test against a local jks.
Is something like {{ProviderUtils.nestLocalFileUriInLocalJavaKeyStoreProvider}}
preferable?
{quote}
alternatively, I could mark the method with {{@VisibleForTesting}}, though the
class is already IA.Private.
> S3Credentials should support use of CredentialProvider
> ------------------------------------------------------
>
> Key: HADOOP-12059
> URL: https://issues.apache.org/jira/browse/HADOOP-12059
> Project: Hadoop Common
> Issue Type: Improvement
> Reporter: Sean Busbey
> Assignee: Sean Busbey
> Attachments: HADOOP-12059.1.patch, HADOOP-12059.2.patch
>
>
> Right now S3Credentials only works with cleartext passwords in configs (as a
> secret access key or the URI). The non-URI version should use credential
> providers with a fallback to the clear text option.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
