You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-issues@hadoop.apache.org by "Konstantin Boudnik (JIRA)" <ji...@apache.org> on 2010/11/18 23:51:15 UTC
[jira] Commented: (HADOOP-7043) Regressions (tests failing after a
patch applicarion) should be automatically detected and marked as such
[ https://issues.apache.org/jira/browse/HADOOP-7043?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12933591#action_12933591 ]
Konstantin Boudnik commented on HADOOP-7043:
--------------------------------------------
At the beginning this functionality might as simple as producing a JIRA number in a report so a human can take a look if it is relevant or not. More complicated logic might include considerations about JIRA's states. E.g. if JIRA is in 'Fixed' state then test failure is likely to be a regression; if it open then the failure is simply a know issue, and so on.
> Regressions (tests failing after a patch applicarion) should be automatically detected and marked as such
> ---------------------------------------------------------------------------------------------------------
>
> Key: HADOOP-7043
> URL: https://issues.apache.org/jira/browse/HADOOP-7043
> Project: Hadoop Common
> Issue Type: New Feature
> Affects Versions: 0.22.0
> Reporter: Konstantin Boudnik
>
> The idea is pretty simple: whenever a test failure happens test-patch (or Hudson though the meaning of a plugin) needs to check appropriate project JIRAs to see if the test name is mentioned in one of the 'well-known' fields. Perhaps Description or Title are the best candidates for this role.
> The benefits are apparent:
> - saving time on analysis of intermittent failures (Hudson is only able to calculate failure age for a consequent failures i.e. where there's no "gaps" between failure episodes)
> - semi-intelligent check for possible regressions.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.