Sebastian Schuberth

@oheger-bosch, fixing the test for cyclic dependencies now actually shows that we're running into a [stack overflow](https://github.com/oss-review-toolkit/ort/actions/runs/14040469951/job/39309373901?pr=10083#step:5:1426). From a first glance it seems that calling `breakCycles()` / `removeCycles()` happens too...

> It is unclear to me what you want to achieve with this PR. I'm currently inspecting various "strange" real-world issues that we have when analyzing Gradle or Node projects...

> But do you agree that the test [in its current form](https://github.com/oss-review-toolkit/ort/blob/888c785f7811c9a42aaf6a84cd78e57b4157d6b7/model/src/test/kotlin/utils/DependencyGraphBuilderTest.kt#L199-L213) actually does not check that cycles are properly handled? @oheger-bosch, I'd appreciate some feedback here.

> Maybe rename it to make the difference clearer? I'd appreciate any proposals for titles then, as to me it's totally unclear what this test is supposed to cover, that...

> If you want to remove it, I won't object. Hmm. IMO we should either make the title match the test code, or make the test code match the title....

> So, do you have any idea why the cycle detection does not work here? Please see my [comment above](https://github.com/oss-review-toolkit/ort/pull/10083#issuecomment-2748718443): My assumption was that cycles are not detected / broken...

> Duplicate of #1716 Indeed, please close this issue @mhdabbaghy.

I have a similar use-case where we're using the [ServiceLoader](https://docs.oracle.com/javase/7/docs/api/java/util/ServiceLoader.html) to load classes from our own code base / module. As these classes are not directly referenced, they get removed...

Thanks, any clarification would be appreciated!

> it's good to accept any possible file organization scheme AFAIU this is exactly the objection. The proposal here is to require a fixed file organization scheme, and have a...