[
https://issues.apache.org/jira/browse/BEAM-474?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15399726#comment-15399726
]
Scott Wegner commented on BEAM-474:
-----------------------------------
These tests verify that the pipeline translation in each runner preserves
display data in translated graph. We don't actually run the pipeline, but the
tests currently use @RunnableOnService and TestPipeline in order to validate
the implementation of each runner.
If these shouldn't be @RunnableOnService or @NeedsRunner-- what are they? Is
there any reason that @RunnableOnService / @NeedsRunner require that the tests
run a pipeline, or can we update the definition?
> Tests which evaluate DisplayData but don't run a pipeline should be marked
> @NeedsRunner not @RunnableOnService
> --------------------------------------------------------------------------------------------------------------
>
> Key: BEAM-474
> URL: https://issues.apache.org/jira/browse/BEAM-474
> Project: Beam
> Issue Type: Bug
> Components: sdk-java-core
> Reporter: Luke Cwik
> Priority: Minor
>
> Tests which evaluate DisplayData but don't run a pipeline should be marked
> @NeedsRunner not @RunnableOnService
> @RunnableOnService represents tests which actually execute a Pipeline while
> @NeedsRunner represents tests that need a runner to construct the pipeline.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
