Thomas Skjølberg

@arminhaaf looks very interesting. I am about to merge a big change into the master branch, so perhaps hold your horses a little bit.

@arminhaaf very interesting with the bag logic. The recent v2.x release is a major improvement, do you think you'd be able to migrate the code?

First add a default module name, then later module info (a bigger job).

Bump

@lhazlewood can't seem to find any module descriptor or automatic module name entry in the manifest? Ref https://repo1.maven.org/maven2/io/jsonwebtoken/jjwt-impl/0.11.0/

So disabling the module system is possible, and then all will run 'as before'. But that is hardly proper support. You might be interested in this: https://dzone.com/articles/automatic-module-name-calling-all-java-library-maintainers I have best...

The whole JWT header (as a String) should be mapped to a cache entry containing - Header - SignatureAlgorithm - Key - CompressionCodec Also, this means that the SigningKeyResolver only...

So something [like this](https://github.com/skjolber/java-jwt-benchmark/blob/cacheHeaderExperiment/frameworks/jjwt-bench/src/main/java/com/github/skjolber/bench/jjwt/DefaultJwtParser.java#L233).

Run the benchmark using `./gradlew --stop && ./gradlew clean build jmhClasses jmh --info -x :frameworks:jjwt-bench:spotbugsMain --stacktrace` with java 8 for the following output: | Benchmark | Mode | Cnt |...

@bdemers Remote keys must be downloaded before evaulating tokens, so the set of possible headers will be known in time (at runtime), but lets save the 'permutation' approach for later...