+1, seems right, although I did’t do the benchmark as Huxing said, I found some others’ result like [1], it shows that Optional’s performance is not bad, or even better.
And more importantly, it will improve readability much. [1] https://stackoverflow.com/questions/34696884/performance-of-java-optional <https://stackoverflow.com/questions/34696884/performance-of-java-optional> > 在 2019年1月21日，上午10:50，LiZhenNet <[email protected]> 写道： > > +1 , > Now ,Dubbo‘s source code language level is old ,We should use some new > features of java to make the source code more concise and understandable. > > > Ian Luo <[email protected]> 于2019年1月21日周一 上午10:45写道： > >> The potential performance penalty should be trivial, and we should >> encourage use Optional in Dubbbo. I believe the static optimization from >> the compiler and the dynamic optimization from the JIT compiler will take >> care of it. >> >> Thanks, >> -Ian. >> >> >> On Fri, Jan 18, 2019 at 8:05 PM Kun Song <[email protected]> wrote: >> >>> Hi community, >>> >>> We will support Java 8 since Dubbo 2.7 according to Huxing’s comments >>> here [1], one big improvement of Java 8 is `Optional`, which is >> introduced >>> primarily to avoid `null` and `NullPointerException`. >>> >>> `Optional` will make programs more readable and less bug prone, and >>> some projects such as Apache Pulsar do use `Optional` a lot. However >> heavy >>> use of `Optional` may produce many small objects and hence introduce some >>> performance costs, my question is should we encourage `Optional` in >> Dubbo? >>> >>> Looking forward to hearing from you. >>> >>> [1] >>> >> https://github.com/apache/incubator-dubbo/issues/3186#issuecomment-454236757 >>> < >>> >> https://github.com/apache/incubator-dubbo/issues/3186#issuecomment-454236757 >>>> >>
