Clément MATHIEU

Thanks for the hint @agentgt . State of nullness checking was so depressing that I put myself into hibernation a few years ago waiting for some sanity to appear.

An intern reported this issue to me last month. Being confused, I quickly re-read the papers and noticed that `e` has been used in the initial paper but `2` is...

What I said was wrong. This one breaks source compatibility since it removes a checked exception from method signature. I did a hprof session on one of our batch today...

I will have another API change in ICardinality to submit in few days (regarding the merge method). Let's see if you agree with the other one, and then decide what...

I agree with Matt. I don't see the point to introduce this kind of complexity since estimator can be efficiently merged. I just ran my benchmark on your branch (single...

Before doing that it would be good to benchmark the proposed implementation against plain old locks for a realistic multi-threaded use case. Wrapping with locks provides greater flexibility (you can...

Here are the preliminary results. Please remember these are micro-benchmarks and it is very easy to get them wrong or meaningless. I tried to check the overhead in single thread...

> Regarding the cost of merge, I don't see a huge cost since the merge cost will be amortized across many, many additions. It depends of your use case. HLL(++)...

[Here](https://github.com/cykl/stream-lib-bench) Keep in mind that the paint is wet. Comment, contribution & criticism are welcome. Let me known if you need help.

@lmolkova Would you mind describing how _offset lag for received/**checkpointed** events_ would be computed? A few months ago, I filled https://github.com/Azure/azure-sdk-for-java/issues/19391 and ended up implementing some code to track the...