[
https://issues.apache.org/jira/browse/YARN-9052?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16984454#comment-16984454
]
Szilard Nemeth commented on YARN-9052:
--------------------------------------
Hi [~sunilg]!
Uploaded patch009.
1. To answer your question about MockRM: This class is still used as it was.
I only extracted the code from MockRM that submits applications, so this is why
I have MockRMAppSubmitter.
As MockRM is a mocked out version of RM (extended from ResourceManager) that is
used extensively in almost all the tests, it's pretty hard to
replace it with anything more lightweight. Doing that is certainly worth its
own jira and this is out of the scope of this jira.
2. Fixed javadoc checkstyle issues. The rest of the checkstyle issues seems a
lot but they are either:
* HiddenField errors (in MockRMAppSubmissionData.Builder) that we can ignore, I
think.
* Coming from code that I just moved, like "MethodLength" or
"LocalVariableName" issues.
What do you think?
> Replace all MockRM submit method definitions with a builder
> -----------------------------------------------------------
>
> Key: YARN-9052
> URL: https://issues.apache.org/jira/browse/YARN-9052
> Project: Hadoop YARN
> Issue Type: Improvement
> Reporter: Szilard Nemeth
> Assignee: Szilard Nemeth
> Priority: Minor
> Attachments:
> YARN-9052-004withlogs-patch-unit-hadoop-yarn-project_hadoop-yarn_hadoop-yarn-server_hadoop-yarn-server-resourcemanager.txt,
> YARN-9052-testlogs003-justfailed.txt,
> YARN-9052-testlogs003-patch-unit-hadoop-yarn-project_hadoop-yarn_hadoop-yarn-server_hadoop-yarn-server-resourcemanager.txt,
> YARN-9052-testlogs004-justfailed.txt, YARN-9052.001.patch,
> YARN-9052.002.patch, YARN-9052.003.patch, YARN-9052.004.patch,
> YARN-9052.004.withlogs.patch, YARN-9052.005.patch, YARN-9052.006.patch,
> YARN-9052.007.patch, YARN-9052.008.patch, YARN-9052.009.patch,
> YARN-9052.testlogs.002.patch, YARN-9052.testlogs.002.patch,
> YARN-9052.testlogs.003.patch, YARN-9052.testlogs.patch
>
>
> MockRM has 31 definitions of submitApp, most of them having more than
> acceptable number of parameters, ranging from 2 to even 22 parameters, which
> makes the code completely unreadable.
> On top of unreadability, it's very hard to follow what RmApp will be produced
> for tests as they often pass a lot of empty / null values as parameters.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
