[
https://issues.apache.org/jira/browse/TIKA-4470?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18015358#comment-18015358
]
Tilman Hausherr commented on TIKA-4470:
---------------------------------------
I was wondering why the test jar is needed at all. I asked ChatGPT:
bq. You create a test JAR when you want to share test code/resources between
modules or projects, not because you want to ship your actual tests as part of
your main library.
I then ran a build without the test-jar and it still worked. (But maybe the
build retrieved the test-jar from the repository)
I also searched out pom.xml files for test-jar and there are a lot. Do we
really need all these?
About the actual problem: It happens in FSBatchTestBase, there is an
explanation why in the class comment (so that test files are deleted in the
next clean). This is since TIKA-1330 in 2015.
I have excluded the directory. Alternatively, how about just using
"target/test-output"?
> tika-batch-*-tests.jar contain test output XML files
> ----------------------------------------------------
>
> Key: TIKA-4470
> URL: https://issues.apache.org/jira/browse/TIKA-4470
> Project: Tika
> Issue Type: Bug
> Components: batch
> Affects Versions: 3.0.0, 3.1.0, 3.2.0, 3.2.1, 3.2.2
> Reporter: Hervé Boutemy
> Priority: Major
>
> discovered while rebuilding to check Reproducible Builds
> https://github.com/jvm-repo-rebuild/reproducible-central/blob/master/content/org/apache/tika/README.md
> see
> https://github.com/jvm-repo-rebuild/reproducible-central/blob/master/content/org/apache/tika/tika-3.2.0.diffoscope
> for example, that shows content in {{test-output/tika-batch-output-root-*}}
> in the jar, published to Maven Central
> I suppose either this jar should not be published, or it should not contain
> the output
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
