google protobuf - PB_BYTES_ARRAY_T(n) - what is the use of .size field?-CodePudding

My protobuf file is

message Msg{
    // User Authentication data as bytes. 
    bytes MsgData = 1 [(nanopb).max_size = 2048];
}

When I generate the C API, the relevant parts are:

#define PB_BYTES_ARRAY_T(n) struct { pb_size_t size; pb_byte_t bytes[n]; }

/* Struct definitions */
typedef PB_BYTES_ARRAY_T(2048) Msg_AuthenticationData_t;

/* Maximum encoded size of messages (where known) */
#define Msg_size  2051

QUESTION

Who sets the field pb_size_t size? And why isn't it set in the macro somehow?

Since the size of the payload is known - 2048 in this case - shouldn't pb_size_t size always by 2048?

Note: I understand that 2051 is for the encoded size and 2048 is for the size of payload.

CodePudding user response：

Since the size of the payload is known - 2048 in this case - shouldn't pb_size_t size always by 2048?

When you set (nanopb).max_size = 2048, it is the maximum size. The actual size of the data can be anything from 0 to 2048 bytes.

The .size field should be set by your code prior to encoding, and it will be set by nanopb when decoding a protobuf message.

If the length of the data is always the same, you can additionally set (nanopb).fixed_length = true. Then the .size field is not generated. Any incoming messages with length different than max_size will cause pb_decode() to return an error.