You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Sebastian Piu (JIRATEST)" <ji...@apache.org> on 2018/01/21 18:12:03 UTC

[jira] [Commented] (SPARK-8294) Break down large methods in YARN code

    [ https://issues-test.apache.org/jira/browse/SPARK-8294?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16265593#comment-16265593 ] 

Sebastian Piu commented on SPARK-8294:
--------------------------------------

I had a quick look and all those methods seem to have been refactored now, is this issue still relevant?

> Break down large methods in YARN code
> -------------------------------------
>
>                 Key: SPARK-8294
>                 URL: https://issues-test.apache.org/jira/browse/SPARK-8294
>             Project: Spark
>          Issue Type: Sub-task
>          Components: YARN
>    Affects Versions: 1.4.0
>            Reporter: Andrew Or
>            Priority: Major
>
> What large methods am I talking about?
> Client#prepareLocalResources ~ 170 lines
> ExecutorRunnable#prepareCommand ~ 100 lines
> Client#setupLaunchEnv ~ 100 lines
> ... many others that hover around 80 - 90 lines
> There are several things wrong with this. First, it's difficult to follow / review the code. Second, it's difficult to test it at a fine-granularity. In the past we as a community has been reluctant to add new regression tests for YARN changes. This stems from the fact that it is difficult to write tests, and the cost is that we can't really ensure the correctness of the code easily.



--
This message was sent by Atlassian JIRA
(v7.6.0#76001)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org