Clean up the integration tests of stuff that is never used - #12887
Conversation
|
@elharo do you want to review also these ones? |
elharo
left a comment
There was a problem hiding this comment.
"I run the fix through integration tests in embedded and in forked mode, with mimir and without mimir and every time I run them twice, just to be sure there is nothing bombing out on already existing files. The first commit is the actual removal the second commit is the renaming of the tests. So that if one removes test 2 and leaves test 3, there is no gap in the numbering. The diff might be somehow confusing in the renumbering commit, because some tests were bit-to-bit copies, just with different settings.xml."
I appreciate the effort, but I think that's making this harder to follow, and I don't think we should rename tests or move files just to keep the numbering intact. That mucks up the history. Can you revert the second commit so this is just about removing files that aren't used and nothing else is touched? Thanks
gnodet
left a comment
There was a problem hiding this comment.
Clean, well-scoped deletion of 51 unreferenced integration test resource files. Verified each category:
- profiles.xml files (MNG-2309, MNG-2605, MNG-3259, MNG-3701, MNG-4106): Maven 2 artifacts rendered dead by MNG-4060 (profiles.xml support removed in Maven 3). Tests use settings.xml and POM profiles only.
- Orphaned test variant directories (mng-3948/test-1, mng-3970/test-1, mng-3983/test-2, mng-4036/legacy): Corresponding test methods were removed previously but resource directories were left behind.
- Unreferenced settings.xml (MNG-3259, MNG-8572): Tests never pass
--settingsfor these files.
CI is fully green across all platforms and modes.
This review was generated by an AI agent and may contain inaccuracies. Please verify all suggestions before applying.
Claude Code on behalf of Guillaume Nodet
|
@elharo Please assign appropriate label to PR according to the type of change. |
|
I kept the d112029 around if you people change opinion. So that the test numbering is tidy. |
When working on #12871 I realized that there are several unused configuration files and tests some dating from the beginning of this century. So, I made this follow-up commit removing all stuff that is not referenced anywhere. It included even a hardcoded user path on his Mac and several snippets that can mislead someone that is grepping for best practices.
I run the fix through integration tests in embedded and in forked mode, with mimir and without mimir and every time I run them twice, just to be sure there is nothing bombing out on already existing files. The first commit is the actual removal the second commit is the renaming of the tests. So that if one removes test 2 and leaves test 3, there is no gap in the numbering. The diff might be somehow confusing in the renumbering commit, because some tests were bit-to-bit copies, just with different settings.xml.
Following this checklist to help us incorporate your
contribution quickly and easily:
Note that commits might be squashed by a maintainer on merge.
This may not always be possible but is a best-practice.
mvn verifyto make sure basic checks pass.A more thorough check will be performed on your pull request automatically.
If your pull request is about ~20 lines of code you don't need to sign an
Individual Contributor License Agreement if you are unsure
please ask on the developers list.
To make clear that you license your contribution under
the Apache License Version 2.0, January 2004
you have to acknowledge this by using the following check-box.