[
https://issues.apache.org/jira/browse/SUREFIRE-2015?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Tibor Digana closed SUREFIRE-2015.
----------------------------------
Resolution: Fixed
https://gitbox.apache.org/repos/asf?p=maven-surefire.git;a=commit;h=307dee3dfc1b392c8caba3b1a261fcc7c1d72833
> Implement testRunId and RunMode in the SimpleReportEntry
> --------------------------------------------------------
>
> Key: SUREFIRE-2015
> URL: https://issues.apache.org/jira/browse/SUREFIRE-2015
> Project: Maven Surefire
> Issue Type: Improvement
> Reporter: Tibor Digana
> Assignee: Tibor Digana
> Priority: Major
> Fix For: 3.0.0-M6
>
>
> This ticketis the most important in SUREFIRE-1860 because it implements
> RunMode and testRunId in the ReportEntry, TestSetReportEntry,
> SimpleReportEntry, CategorizedReportEntry and WrappedReportEntry in
> appropriate implementation form.
> Additionally, we should distinguish between two events (OutputReportEntry,
> TestOutputReportEntry) and therefore the interface TestReportListener handles
> them via generics dependeing on the situation.
> The first situation happens when provider's listeners handles the event
> OutputReportEntry from System.out and System.err via the
> ConsoleOutputCapture. The ConsoleOutputCapture does not have any notion about
> runMode and testRunId, and therefore the only provider's listener would add
> runMode and testRunId to a recreated entry which would be finally propagated
> to the ForkingRunListener and TestSetRunListener. The runMode and testRunId
> are determined opon the events testStarted and any test end and Thread local.
> The second situation happens when ForkingRunListener and TestSetRunListener
> handles TestOutputReportEntry which contains runMode and testRunId.
> Some classes are deprecated in the module surefire-junit47 and can be freely
> removed after the XML Reporter would be fully capable of parallel tests (must
> support events chaos in one thread). That's the reason why parallel tests are
> badly handled in JUnit5 reports. The provider surefire-junit47 also supports
> parallel tests and we do not see the problem with undeterministic XML report,
> but the reason is that the provider has a cache, eats the memory including
> the std/out logs and all the messages and test events are replayed in a
> sequence as if the tests were executed in one thread.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
