Protobuf

Last updated on 2025-08-08 (history).

# Distribute protobuf files

A general idea is to distribute .proto files via a git repo and managing releases from this central place.

What this does is:

Put all release related tooling at single place.
Version control .proto files.
Provide a review process on contracts.

References are:

# Useful tools

Online Protobuf Decoder - https://protobuf-decoder.netlify.app/

# Tips and tricks

## Decode protobuf bytes locally

### With unknown type

1
2
3
echo '0a 0a 08 c0 f7 be c3 06 10 d0 86 03 10 01 1a 06 68 65 6c 6c 6f 21 22 03 66 6f 6f 22 03 62 61 72' \
  | xxd -r -p \
  | protoc --decode_raw

Its output will look like this:

1
2
3
4
5
6
7
8
1 {
  1: 1752153024
  2: 50000
}
2: 1
3: "hello!"
4: "foo"
4: "bar"

### With known type

Save message.proto with the following content:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
syntax = "proto3";

import "google/protobuf/timestamp.proto";

message Message {
  enum State {
    UNKNOWN = 0;
    PENDING = 1;
    PROCESSED = 2;
    FAILED = 3;
  }

  google.protobuf.Timestamp receiveTime = 1;
  State state = 2;
  string content = 3;
  repeated string tags = 4;
}

1
2
3
echo '0a 0a 08 c0 f7 be c3 06 10 d0 86 03 10 01 1a 06 68 65 6c 6c 6f 21 22 03 66 6f 6f 22 03 62 61 72' \
  | xxd -r -p \
  | protoc --decode Message ./message.proto

Its output will look like this:

1
2
3
4
5
6
7
8
receiveTime {
  seconds: 1752153024
  nanos: 50000
}
state: PENDING
content: "hello!"
tags: "foo"
tags: "bar"

## Encode protobuf bytes locally

With the same message.proto mentioned above, we can decode a message like this:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
echo '
receiveTime {
  seconds: 1752153024
  nanos: 50000
}
state: 1
content: "hello!"
tags: "foo"
tags: "bar"
' | protoc --encode Message ./message.proto | base64

Note: base64 is used just to avoid binary junk being printed on terminal.

### Caveats

How will an enum value be decoded when encoded enum number is not present in schema that is used for decoding?

When encoding, we are passing a number for enum, but when decoding, we usually get a string for enum values. With example:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
echo '
receiveTime {
  seconds: 1752153024
  nanos: 50000
}
state: 1
content: "hello!"
tags: "foo"
tags: "bar"
' | protoc --encode Message ./message.proto | protoc --decode Message ./message.proto

This will print the following output:

1
2
3
4
5
6
7
8
receiveTime {
  seconds: 1752153024
  nanos: 50000
}
state: PENDING
content: "hello!"
tags: "foo"
tags: "bar"

Valid numbers for enum are 0-3 as per our definition from message.proto, but what if we pass an unknown value like 10?

Handling of this depends on implementation. protoc prints it as number. For example,

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
echo '
receiveTime {
  seconds: 1752153024
  nanos: 50000
}
state: 10
content: "hello!"
tags: "foo"
tags: "bar"
' | protoc --encode Message ./message.proto | protoc --decode Message ./message.proto