Open mloskot opened 5 years ago
Yeah, I would love to see a benchmark BUT full disclosure, I have spent exactly zero time optimizing this code (LOL)
@vinniefalco I wouldn't worry about it. The benchmark compares what is available 'on the market' as it is. AFAIK, there has been just first round of implementation of the Ciere's Spirit-based parser, so it has not been optimised either. Still, it is valuable to have it compared, for both, its developers and its users. I believe, this benchmark is for purely practical purposes and not about building any halls of (sh|f)ame.
The streaming parser is separate from the non-streaming parser now. They have different performance characteristics.
The parser implementation is available from Boost.Beast's demo project here https://github.com/vinniefalco/BeastLounge/tree/develop/include/boost/beast/_experimental/json
UPDATE: Current code lives in https://github.com/vinniefalco/json
/cc @vinniefalco