support Junit5 parallel execution #1472#1519
Conversation
|
(eclipse-jdt/eclipse.jdt.ui#378 will need to be resolved before this feature will work correctly) |
|
@fladdimir Thank you for taking look into this and deal with the upstream issue. Will it be fine if I review this PR after the upstream changes are merged? |
@jdneo of course thats fine, thanks for the quick reply |
|
Manual tests (including the locally built upstream fix for the RemoteTestRunner): before - test2 appears not to be run at all, swallowing its failure: fixed - both test failures are shown, comparison failures and traces are placed correctly: |
|
@jdneo the upstream PR has been merged. Unfortunately I am not familiar with the workflow. I assume the upstream artifact ( |
|
Cool! I saw the upstream PR has been merged, great job! The workflow looks like in this way:
I can help update the |
test/test-projects/junit/src/test/java/junit5/ParallelExecutionTest.java
Outdated
Show resolved
Hide resolved
That sounds great, thanks a lot! |
48a5910 to
ff57ce2
Compare
|
Awesome work @fladdimir! When I was looking into this I was noticing the issues regarding the interleaving of messages from the eclipse provided unit test runner - great job finding the cause of the issue and working to resolve it! |
|
I raised a PR to update the dependency in VS Code Java: eclipse-jdtls/eclipse.jdt.ls#2403. since it's about to release a new version this week, that PR should be able to be merged after the release (late this week) |
jdneo
left a comment
There was a problem hiding this comment.
I looked the PR today and overall it looks good! 👍
After the latest pre-release of vscode-java is available, I'll play with it manually and merge this PR if everything works fine.
| private currentStates: Map<TestItem, CurrentItemState> = new Map(); | ||
|
|
||
| // failure info for a test is received consecutively: | ||
| private currentItem: TestItem | undefined; |
There was a problem hiding this comment.
nit: how about give it a more specific name, something like: tracingItem?
There was a problem hiding this comment.
just changed the name 👍🏼
ff57ce2 to
2b6530c
Compare
While debugging I realized that this duplication of diff-messages also occurs during normal test execution on the current main branch, without being related to parallel test execution. Apparently its a problem of the general stack trace processing, occuring when a failing equality assertion stack trace contains multiple entries from the test class(?). (Non-concurrently executed) sample based on the current main branch: There is also a unit-test (part of I'd like to continue investigating, however, would it be ok to open a separate issue for this? |
|
Sorry for the late reply. I'm on vacation these days :) Yes, I can confirm that this bug can also be observed in the marketplace version. And it's totally fine to deal with it in a separate issue. Thank you for the contribution! |
no worries at all - thanks for the review and merge! |


resolves #1472