Arne

> Quick comment: the [`@since`](https://docs.oracle.com/en/java/javase/14/docs/api/java.base/java/lang/Deprecated.html#since()) attribute doesn't seem to admit a descriptive comment as you have here, but merely a software version. I think you'd do better to use the...

> It would also be nice to get some more context around the motivation for this change. While "faster and less memory" is noted in the title and description presumably...

> @arne-bdt - looking at the reviews, I think we're nearly done. There was one comment about iterator remove - how many test are affected? > > I think @rvesse's...

@rvesse : Thank you very much for your motivating words. That was quite helpful, when I got frustrated in the last days. I updated the PR description, including the devastating...

> If you point me at `ExtendedIterator.remove` in `org.apache.jena.sparql.*`, I'll check them and rewrite suspect ones. Sorry, I didn't mean to point at anyone. It seems that it's pretty common...

Please comment on the new benchmarks (see PR-description above) for Graph#add, Graph#find, Graph#stream, memory consumption and Graph#contains! My general conclusions are: - GraphMem2 is always faster at Graph#add - GraphMem2...

> @arne-bdt - just checking in to see where we are on this pull request. To sum it up: I estimate this PR to bei ready in August 2022. In...

> Hi @arne-bdt, > > Would it be a good idea to put in this PR in its current state as a replacement for the existing (and difficult to maintain)...

I did not want to close this PR entierly. I just needed to cleanup the mess I did with my fork to get a clean start.

Hello @ktk, thank you for asking. This puts me under pressure to finally get to work and finish it. After my naive and arrogant attempt at a first solution, the...