Guillaume Toison

> @gtoison yes, that's true, once Java 22 is available, the current behavior will be correct. But until then it isn't. 😄 ... and once everybody has switched to Java...

Thank you for the samples! I tried to reproduce the problem but strangely got an other error which ultimately boils down to [BCEL-370](https://issues.apache.org/jira/browse/BCEL-370) so I submitted an fix attempt here:...

Hum, I thought this was fixed by #2828 ... The test cases in https://github.com/gtoison/spotbugs/blob/b5893ac7a825413dd6085135318ff7a5a7c49e49/spotbugsTestCases/src/java21/Issue2782.java seem to be handled correctly so maybe your code is a bit different? What compiler (and...

So this should be fixed in the upcoming release now that the problem in BCEL is solved and #2807 is merged

Thanks for the samples, I have added them in #2937 and they are flagged indeed I'd have to investigate why they are not when running the test on my computer

You might run into false positives with 4.8.5 because I had not realized that the bytecode instruction for a switch is either `LOOKUPSWITCH` or `TABLESWITCH`. Your compiler might use `TABLESWITCH`...

Hello @romani, this was fixed by #2807 (as well as #2819), right?

Just to clarify what you're saying: do you mean that you started using JEP 456's "unnamed variables" with Java 22 and you're not getting false positives (because TABLESWITCH switches weren't...

Javac 22 seems to be producing bytecode with a load, even if the variable was unnamed. A further improvement could be for SpotBugs to warn that a named variable _could_...

This is currently not supported unfortunately, there are clevers way to get this working but I'm not sure if we want to go that route: https://community.sonarsource.com/t/how-to-use-sonarcloud-with-a-forked-repository-on-github/7363/34