Extend AnnotationExcluder to catch fully qualified annotations

[cortinico]

Currently AnnotationExcluder is not handling usages of fully qualified annotations (eg. @dagger.Module). This is causing the failure reported in #2360.
I'm extending it to catch both simple and fully qualified annotations.

Fixes #2360

[codecov-io]

Codecov Report

Merging #2361 into master will increase coverage by 0.02%.
The diff coverage is 71.42%.

@@             Coverage Diff              @@
##             master    #2361      +/-   ##
============================================
+ Coverage     82.81%   82.83%   +0.02%     
  Complexity     2162     2162              
============================================
  Files           350      350              
  Lines          6056     6059       +3     
  Branches       1110     1110              
============================================
+ Hits           5015     5019       +4     
  Misses          469      469              
+ Partials        572      571       -1

Impacted Files	Coverage Δ	Complexity Δ
...gitlab/arturbosch/detekt/api/AnnotationExcluder.kt	50% <71.42%> (+27.77%)	5 <4> (ø)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9f64c26...f7d9052. Read the comment docs.

[cortinico]

Don't really know why UnusedPrivateClassSpec > does not report (companion-)object/named-dot references is failing only for Java8 on Travis 🧐

[BraisGabin]

It's failing because the code snippet doesn't compile. We only compile them in Travis Java8 (performance reasons). You can test it with: ./gradlew :detekt-rules:test -Pcompile-test-snippets=true

I think that for really fix this issue we need type solving but this is better that we have for sure.

[cortinico]

It's failing because the code snippet doesn't compile. We only compile them in Travis Java8 (performance reasons). You can test it with: ./gradlew :detekt-rules:test -Pcompile-test-snippets=true

Thanks for the clarification. Still, UnusedPrivateClassSpec is unrelated to my change 🧐

I think that for really fix this issue we need type solving but this is better that we have for sure.

Yeah agree.

[BraisGabin]

Oh, yes, I didn't noticed that. It's strange, master is passing. Can you rebase your branch with the current master? I think that'll solve the problem

[cortinico]

I think that'll solve the problem

Nope is still failing :/

[schalkms]

@cortinico I had the same problem in #2350.
I just haven’t had the time so far to look into it.

On jdk8 we additionally compile the provided code snippets in detekt’s unit tests through Jsr223 in order to validate it, because there were problems with invalid Kotlin code in the past. That’s the reason why this PR only fails on this single build agent.

I suspect this error is because of one of the following reasons:

• Using 8 compile engines for compiling around +1000 test snippets is not enough anymore.
• The error mentions something about coroutines. There might be problems with the previously introduced coroutine ruleset test cases.

[schalkms]

@cortinico btw thanks for your fast fixes. You helped us out multiple times. Thanks for that! 🙂