[
https://issues.apache.org/jira/browse/SPARK-6746?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14483561#comment-14483561
]
Ilya Ganelin commented on SPARK-6746:
-------------------------------------
SPARK-5945 requires updating the logic for handling fetch failures. This will
introduce merge conflicts with this patch so this one should be merged first
since it cleans up the DAG Scheduler code to help understand what's going on.
> Refactor large functions in DAGScheduler to improve readibility
> ---------------------------------------------------------------
>
> Key: SPARK-6746
> URL: https://issues.apache.org/jira/browse/SPARK-6746
> Project: Spark
> Issue Type: Sub-task
> Components: Scheduler
> Reporter: Ilya Ganelin
>
> The DAGScheduler class contains two huge functions that make it
> very hard to understand what's going on in the code. These are:
> 1) The monolithic handleTaskCompletion
> 2) The cleanupStateForJobAndIndependentStages function
> These can be simply modularized to eliminate some awkward type casting and
> improve code readability.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
