Peter Gafert

@tfij sorry that I was a little slow on this! I looked into the PR, thanks a lot! I think it's a good compromise after our discussions 🙂 Please let...

Sorry, but I think in case of `long` there is no helper method yet. If this was a non-primitive you could maybe have used `com.tngtech.archunit.core.domain.JavaClass#getTransitiveDependenciesFromSelf`, but primitives like `long` are...

I'm not sure if this is possible (but I'm happy to be proven wrong :wink:). Because in the end ArchUnit reads those modifiers from the bytecode. And the bytecode has...

Sounds like a good extension :slightly_smiling_face: If you're willing to contribute a PR I'm all open to support where necessary! Right now I think we're just filtering `module-info.class` (at least...

I'm not completely sure I'm getting the difference. Why do you think the general arguments against field injection would not apply to `@Value`? It's also a dependency injection, right? No...

You can use the `LocationProvider` API for that. E.g. create a custom location provider ``` class PathLocationProvider implements LocationProvider { @Override public Set get(Class testClass) { return Collections.singleton(Location.of(Paths.get("/some/path"))); } }...

Yes, ArchUnit works on bytecode base, so even if you do `importPackages("foo.bar")` it will internally ask the `ClassLoader` "please give me the resource path of the compiled classes" and import...

Hey, sorry for the late response :see_no_evil: I don't think right now there is a good option for this :thinking: The problem for me is that a) The line numbers...

Yes, I tried to be as little invasive as possible, because removing the whole class is a bigger impact. For now you can just adjust the rule to ``` classes()...

The problem is that I'm not sure if there is a general rule to deal with synthetic code. And once excluded there will be no way anymore for users to...