Change
Using JEP 394: Pattern Matching for instanceof for #equals in Node and Triple implementations
Background:
While working with benchmarks Jena 6 seemed slower than Jena 5.6 although the measured code was almost identical. The profiler showed more time spent in those #equals methods, so the OpenJDK 21 optimizer seemed to have skipped those. Maybe I was just hunting ghosts but nevertheless, the code is more concise now and the benchmarks 'magically' now show the expected comparable results for both jena releases.
Are you interested in contributing a pull request for this task?
Yes
Change
Using JEP 394: Pattern Matching for instanceof for
#equalsinNodeandTripleimplementationsBackground:
While working with benchmarks Jena 6 seemed slower than Jena 5.6 although the measured code was almost identical. The profiler showed more time spent in those
#equalsmethods, so the OpenJDK 21 optimizer seemed to have skipped those. Maybe I was just hunting ghosts but nevertheless, the code is more concise now and the benchmarks 'magically' now show the expected comparable results for both jena releases.Are you interested in contributing a pull request for this task?
Yes