CODEWITHSUNDEEP
javac++protocolsprotocol-buffers
Mona Wade
Mona Wade

Reputation: 192

How are Google Protocol repeated values structured? What are their limitations and advantages?

I'm trying to create an array that is going to transfer data from C++ application to a Java app. Because of the two languages, how do you format the a Google Protocol message?

Also, what are their limitations and advantages?

So far, I have the following: 

message Config {
    repeated string types = 1;
}

I think I'm missing somethings!

Upvotes: 1

Views: 38

Answers (1)

Marc Gravell
Marc Gravell

Reputation: 1062590

If you mean on the wire - then firstly: it is the serializer's job to worry about that, but in this case it is implemented simply as (per repeated element):

0x0a = "field 1, length prefixed" (strings are always length prefixed)
{number of bytes as UTF8, varint-encoded}
{that many bytes of UTF8}

So if you have 3 items repeated you'll have 3 lots of 0x0a, varint, payload slammed one after the other. In the case of some primitive types - integers etc, there is a "packed" encoding that removes the need for a header per element, by prepending the byte-length of the entire array first, then just the array contents as raw non-tokenized data.

If you mean from the caller's perspective, then: it is down to the implementation, but essentially you should expect some familiar API - common to that framework. This might mean that repeated data is exposed as an array / vector, or it might mean a "collection" class with an Add etc API. It will depend entirely on what is sensible for the target framework and the implementing author's whimsy.

Upvotes: 1

Related Questions