XXX_Marshal returns wrong encoded byte slice

[zqkgo]

What version of protobuf and what language are you using?
protobuf version:

$ protoc --version
libprotoc 3.7.1

language: Go

What did you do?
Step 1
define test.proto file

syntax = "proto3";
package proto_test;

message Education {
    string college = 1;
}

message Person {
    int32 age = 1;
    string name = 2;
    Education edu = 3;
}

Step 2

generate Go code

protoc --go_out=. test.proto

Step 3

check encoded message

p := proto_test.Person{
    Age:  666,
    Name: "Tom",
    Edu: &proto_test.Education{
        College: "SOMEWHERE",
    },
}
var b []byte
out, err := p.XXX_Marshal(b, true)
if err != nil {
    log.Fatalln("fail to marshal with error: ", err)
}
fmt.Printf("hexadecimal format:% x \n", out)

What did you expect to see?

it should output,

hexadecimal format:08 9a 05 12 03 54 6f 6d 1a 0b 0a 09 53 4f 4d 45 57 48 45 52 45 

What did you see instead?

but it outputs,

hexadecimal format:08 9a 05 12 03 54 6f 6d 1a dd 97 d1 08 0a 09 53 4f 4d 45 57 48 45 52 45 

the dd 97 d1 08 part should be 0b.

I debug the process of encoding Person struct, here, when Education struct is marshaled, u.cachedsize() tries to check if the sizecache field is valid, when value of u.sizecache is zeroField it assumes valid, then it returns the wrong int which starts at offset 0. In this case, u.sizecache should be 40, and p.Ed.XXX_sizecache is 11 which encoded as 0b.

[dsnet]

Hi, thanks for the report.

The XXX methods and fields are not to be interacted with directly. As you have correctly noted, XXX_Marshal makes assumptions about the validity of XXX_sizecache. This is working as intended. Instead, the proto.Marshal function should be used.

I should note that the compatibility section reserves the right to add/remove XXX methods. In fact, they will be going away in the future. See #276.