chaokunyang commented on issue #2609:
URL: https://github.com/apache/fory/issues/2609#issuecomment-3765200727
> For xlang serialization, we do have a concept to distinguish the type
kinds, and allow every kind to use their own id:
>
> @Override
> public void writeClassInfo(MemoryBuffer buffer, ClassInfo classInfo) {
> int xtypeId = classInfo.getXtypeId();
> int internalTypeId = xtypeId & 0xff;
> buffer.writeVarUint32Small7(xtypeId);
> switch (internalTypeId) {
> case Types.NAMED_ENUM:
> case Types.NAMED_STRUCT:
> case Types.NAMED_EXT:
> if (shareMeta) {
> writeSharedClassMeta(buffer, classInfo);
> return;
> }
> assert classInfo.namespaceBytes != null;
> metaStringResolver.writeMetaStringBytes(buffer,
classInfo.namespaceBytes);
> assert classInfo.typeNameBytes != null;
> metaStringResolver.writeMetaStringBytes(buffer,
classInfo.typeNameBytes);
> break;
> case Types.NAMED_COMPATIBLE_STRUCT:
> case Types.COMPATIBLE_STRUCT:
> assert shareMeta : "Meta share must be enabled for compatible
mode";
> writeSharedClassMeta(buffer, classInfo);
> break;
> default:
> break;
> }
> }
> Users can reigster id from 0, and even fory register new interal types, it
won't break users.
>
> We can unify this with java serialization, but I'm not sure whether it
will introduce extra performance overhead, maybe need some benchmark. And it's
a big change, it may take more efforts.
@drse I refactored teh java type system into this in #3153. Now we can
extend out type registration withotu breaking any user-side registeration
anymore
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
