when a message is serialized its known fields should be written
sequentially by field number, as in the provided C++, Java, and Python
Like the encoded wire format, is the above guaranteed for the C++
Serialization API (assuming no unknown fields)?
I'm designing a file format based on protobuf. I plan to have the
first field as a fixed size file type magic value and second field as
a fixed size checksum value, followed by other fields. While opening
such a file, I'd like to match the magic value and verify the
checksum, before parsing the whole file.
You received this message because you are subscribed to the Google Groups
"Protocol Buffers" group.
To post to this group, send email to proto...@googlegroups.com.
To unsubscribe from this group, send email to
For more options, visit this group at