Martin Leitner-Ankerl

Hi, thanks for finding! I'll see what I can do, but I unfortunately currently don't have much time for this project

No problem, it's easy to work around

It would be possible with `typeid` and `abi::__cxa_demangle`, see this: https://stackoverflow.com/a/4541470/48181 Note that this would resolve `typedef`s, so the output is not necessarily the same as the input.

Hi, sorry about being inactive for a while. I blame it on COVID stress... I hope to find more time soon and have a look at the issues that have...

nanobench 4.0.0 is released! With up to date and extensive documentation at https://nanobench.ankerl.com

I think simply checking for a wrap around is not enough. E.g when you insert 3 elements at location 30, 2 elements are at the end and one will wrap...

I also thought about adding an overflow buffer at the end. This might have a slight performance advantage because less need to `& (tablesize-1)`. The DoS attack could also be...

@ktprime I don't see what that benchmarks adds that is not already covered by the other benchmarks?

Any updates on that issue? Is c++98 still a requirement? My PR https://github.com/jgarzik/univalue/pull/79 introduces move semantics which can lead to a very significant speedup of about a factor of 2.

Nice! concept ACK. How about removing the copy constructor and copy assignment from the UniValue class, as done [here](https://github.com/bitcoin-core/univalue-subtree/pull/27/commits/31293dc75a435889a1a22d13fbf7c3ac3963e511#diff-8f61d7162a69fedc1207fb91c3e7eb4c0a10ce28ff9b28bddbb17a24b0ca62b8R43-R54)? Fixing all copies would make the PR much bigger though.