[
https://issues.apache.org/jira/browse/MCHECKSTYLE-389?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17270889#comment-17270889
]
Hudson commented on MCHECKSTYLE-389:
------------------------------------
Build failed in Jenkins: Maven » Maven TLP » maven-checkstyle-plugin » 31 #2
See
https://ci-builds.apache.org/job/Maven/job/maven-box/job/maven-checkstyle-plugin/job/31/2/
> MCHECKSTYLE-365 introduces regression with 'rules' aggregate count section on
> report
> ------------------------------------------------------------------------------------
>
> Key: MCHECKSTYLE-389
> URL: https://issues.apache.org/jira/browse/MCHECKSTYLE-389
> Project: Maven Checkstyle Plugin
> Issue Type: Dependency upgrade
> Affects Versions: 3.1.0
> Reporter: Jeremy Landis
> Assignee: Enrico Olivelli
> Priority: Major
> Fix For: 3.1.1
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Commit eee0ba1 in CheckstyleReportGenerator.java sets the default 'severity'
> is set as 'error'. The suggestion is that it fixes count issues per adjusted
> comment at that time. While that may or may not be true, what it does do is
> incorrectly shut down the entire section of rules aggregation counts in some
> cases. The severity only takes into account that specific one to then
> aggregate. Selecting just one other than null is going to mess up the logic
> as implemented here. This change needs reverted for that one specific change
> only. The original issue should be reopened as it was not properly fixed.
> The IT tests fails to take into account individual types. It includes both
> 'info' and 'error' thus the only reason it even works.
> In my test case, I'm using 'mybatis-3'. I have configured it to use google
> checks so nothing else setup. It has 0 info, 3151warnings, and 0 errors.
> Switching it to any other value will cause it to only reflect that section.
> So using 'info', nothing shows up. Using 'error' as coded now, nothing shows
> up. Using 'warning', it does show up. Continuing to use null as originally
> the case, it does show up.
> There really was a miscount somewhere, but that was not the right fix. For
> my test, I get a miscount of +5 if using 'warning' or the original null.
> That is entirely better than nothing. And therefore, the best solution for
> now is to revert this change.
> While my test does not have a combination of issues in 'info', 'warning', and
> 'error', presumably most do and thus why this was not reported after almost a
> year since the release of 3.1.0.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)