Oleh Dokuka

Just FYI. Having `cancel` racing with `request` is expected behavior as of rule 2.7. None of the existing Reactive libs doing serialization for those calls and that is specification expected....

> the interpretation that 2.7 means that request can not race with request, and cancel can not race with cancel doesn't seem plausible, because [reactive-streams/reactive-streams-jvm#3.5](https://github.com/reactive-streams/reactive-streams-jvm#3.5) says that cancel must be...

@dkhalanskyjb indeed, it sounds like about racing but the intent is different I guess we can clarify it in the spec. Also, I suggest paying more attention to the hint...

@qwwdfsad this PR is not really relevant, so my suggestion is to reject it.

Yeah. Noticed that. I guess I found another fix. Let me try it quickly for the Reactor to see if it works. Apologies for confusing with the first message

Fix -> https://github.com/reactor/reactor-core/pull/2450

@cision-batag we implemented the required mechanism in windowTimeout, thus now you can replace your bufferTimeout with an equal code with windowTimeout ```java import org.junit.jupiter.api.Test; import org.slf4j.Logger; import org.slf4j.LoggerFactory; import reactor.core.publisher.Flux;...

@kfoekao windowTimeout + concatmap(f.collectList, 0) should be a replacement

> @OlegDokuka What does the prefetch value of 0 represent? Can you explain how does this accomplishes the buffer timeout behavior? 0 means concat map does not prefetch extra elements,...

Would love to hear any suggestions/objections/comments before I push the rest of the operators (FYI, they are prepared without tests, found those missing gaps during verification of FluxLimitRate operators, so...