Hi protobuf team,
I noticed that, all MessageLite serialization APIs perform an explicit or 
implicit flush, except for SerializeWithCachedSizes. Because of this, data 
may remain buffered and not be immediately readable unless an extra flush 
is performed, details at 
https://github.com/protocolbuffers/protobuf/issues/25144.

As a public API, should the behavior of SerializeWithCachedSizes be 
consistent with the other serialization interfaces? If so, is it possible 
to add an flush/trim inside SerializeWithCachedSizes, and provide a 
separate internal-only entry point that preserves the current no-flush 
behavior for internal callers?

Thanks

-- 
You received this message because you are subscribed to the Google Groups 
"Protocol Buffers" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/protobuf/2eec55c0-63a9-4916-a245-46d3325a1e3dn%40googlegroups.com.

Reply via email to