FMX commented on PR #3177: URL: https://github.com/apache/celeborn/pull/3177#issuecomment-2771432938
> > Besides, the serialization method used by cpp is actually language-agnostic already as it doesn't introduce any language-dependent component. Other language such as rust, golang, etc could use the cpp's serialization method, including the same header if they want. > > @HolyLow So let's use a different name for the first byte, in case someone might blindly add a new serialization layout when they want to add a new language client in the future. Maybe just V1, V2, or something that won't make the user think it's exclusive to some languages. > > Just out of curiosity, why 0xFF is chosen for the new layout? > > > The new serialization method is mostly based on the protobuf but with some adaptation to java's existing TransportMessage's handling stack. > > I might miss some context, is there a doc that describes the layout? > > > I think reorganizing/refactoring java's message decode/encode framework is the java engine's issue, not the cppClient's issue. > > I agree, but I think we should migrate all implementations to the new serialization layout to simplify the code eventually. Of course, this is out of the scope of CIP-14. Sounds reasonable to me. I think the `LanguageType` can be renamed to get a better semantic meaning. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
