This project hosts the Go implementation for protocol buffers, which is a language-neutral, platform-neutral, extensible mechanism for serializing structured data. The protocol buffer language is a language for specifying the schema for structured data. This schema is compiled into language specific bindings. This project provides both a tool to generate Go code for the protocol buffer language, and also the runtime implementation to handle serialization of messages in Go. See the protocol buffer developer guide for more information about protocol buffers themselves.
This project is comprised of two components:
Code generator: The protoc-gen-go tool is a compiler plugin to protoc, the protocol buffer compiler. It augments the protoc compiler so that it knows how to generate Go specific code for a given .proto file.
Runtime library: The protobuf module contains a set of Go packages that form the runtime implementation of protobufs in Go. This provides the set of interfaces that define what a message is and functionality to serialize message in various formats (e.g., wire, JSON, and text).
See the developer guide for protocol buffers in Go for a general guide for how to get started using protobufs in Go.
This project is the second major revision of the Go protocol buffer API implemented by the google.golang.org/protobuf module. The first major version is implemented by the github.com/golang/protobuf module.
Summary of the packages provided by this module:
proto: Package proto provides functions operating on protobuf messages such as cloning, merging, and checking equality, as well as binary serialization.encoding/protojson: Package protojson serializes protobuf messages as JSON.encoding/prototext: Package prototext serializes protobuf messages as the text format.reflect/protoreflect: Package protoreflect provides interfaces to dynamically manipulate protobuf messages.reflect/protoregistry: Package protoregistry provides data structures to register and lookup protobuf descriptor types.reflect/protodesc: Package protodesc provides functionality for converting descriptorpb.FileDescriptorProto messages to/from the reflective protoreflect.FileDescriptor.testing/protocmp: Package protocmp provides protobuf specific options for the cmp package.testing/prototest: Package prototest exercises the protobuf reflection implementation for concrete message types.types/dynamicpb: Package dynamicpb creates protobuf messages at runtime from protobuf descriptors.types/known/anypb: Package anypb is the generated package for google/protobuf/any.proto.types/known/emptypb: Package emptypb is the generated package for google/protobuf/empty.proto.types/known/timestamppb: Package timestamppb is the generated package for google/protobuf/timestamp.proto.types/known/durationpb: Package durationpb is the generated package for google/protobuf/duration.proto.types/known/wrapperspb: Package wrapperspb is the generated package for google/protobuf/wrappers.proto.types/known/structpb: Package structpb is the generated package for google/protobuf/struct.proto.types/descriptorpb: Package descriptorpb is the generated package for google/protobuf/descriptor.proto.types/pluginpb: Package pluginpb is the generated package for google/protobuf/compiler/plugin.proto.compiler/protogen: Package protogen provides support for writing protoc plugins.cmd/protoc-gen-go: The protoc-gen-go binary is a protoc plugin to generate a Go protocol buffer package.The issue tracker for this project is currently located at golang/protobuf.
Please report any issues there with a sufficient description of the bug or feature request. Bug reports should ideally be accompanied by a minimal reproduction of the issue. Irreproducible bugs are difficult to diagnose and fix (and likely to be closed after some period of time). Bug reports must specify the version of the Go protocol buffer module and also the version of the protocol buffer toolchain being used.
This project is open-source and accepts contributions. See the contribution guide for more information.
This module and the generated code are expected to be stable over time. However, we reserve the right to make breaking changes without notice for the following reasons:
.proto files. This includes declared constants, variables, functions, types, fields in structs, and methods on types. This may break attempts at injecting additional code on top of what is generated by protoc-gen-go. Such practice is not supported by this project.protoc-gen-go/internal_gengo package, the runtime/protoimpl package, and all packages under internal.Any breaking changes outside of these will be announced 6 months in advance to protobuf@googlegroups.com.
Users should use generated code produced by a version of protoc-gen-go that is identical to the runtime version provided by the protobuf module. This project promises that the runtime remains compatible with code produced by a version of the generator that is no older than 1 year from the version of the runtime used, according to the release dates of the minor version. Generated code is expected to use a runtime version that is at least as new as the generator used to produce it. Generated code contains references to protoimpl.EnforceVersion to statically ensure that the generated code and runtime do not drift sufficiently far apart.
This project is the second major revision (released in 2020) of the Go protocol buffer API implemented by the google.golang.org/protobuf module. The first major version (released publicly in 2010) is implemented by the github.com/golang/protobuf module.
The first version predates the release of Go 1 by several years. It has a long history as one of the first core pieces of infrastructure software ever written in Go. As such, the Go protobuf project was one of many pioneers for determining what the Go language should even look like and what would eventually be considered good design patterns and “idiomatic” Go (by simultaneously being both positive and negative examples of it).
Consider the changing signature of the proto.Unmarshal function as an example of Go language and library evolution throughout the life of this project:
// 2007/09/25 - Conception of Go // 2008/11/12 export func UnMarshal(r io.Read, pb_e reflect.Empty) *os.Error // 2008/11/13 export func UnMarshal(buf *[]byte, pb_e reflect.Empty) *os.Error // 2008/11/24 export func UnMarshal(buf *[]byte, pb_e interface{}) *os.Error // 2008/12/18 export func UnMarshal(buf []byte, pb_e interface{}) *os.Error // 2009/01/20 func UnMarshal(buf []byte, pb_e interface{}) *os.Error // 2009/04/17 func UnMarshal(buf []byte, pb_e interface{}) os.Error // 2009/05/22 func Unmarshal(buf []byte, pb_e interface{}) os.Error // 2011/11/03 func Unmarshal(buf []byte, pb_e interface{}) error // 2012/03/28 - Release of Go 1 // 2012/06/12 func Unmarshal(buf []byte, pb Message) error
These changes demonstrate the difficulty of determining what the right API is for any new technology. It takes time multiplied by many users to determine what is best; even then, “best” is often still somewhere over the horizon.
The change on June 6th, 2012 added a degree of type-safety to Go protobufs by declaring a new interface that all protobuf messages were required to implement:
type Message interface {
Reset()
String() string
ProtoMessage()
}
This interface reduced the set of types that can be passed to proto.Unmarshal from the universal set of all possible Go types to those with a special ProtoMessage marker method. The intention of this change is to limit the protobuf API to only operate on protobuf data types (i.e., protobuf messages). For example, there is no sensible operation if a Go channel were passed to the protobuf API as a channel cannot be serialized. The restricted interface would prevent that.
This interface does not behaviorally describe what a protobuf message is, but acts as a marker with an undocumented expectation that protobuf messages must be a Go struct with a specific layout of fields with formatted tags. This expectation is not statically enforced by the Go language, for it is an implementation detail checked dynamically at runtime using Go reflection. Back in 2012, the only types with this marker were those generated by protoc-gen-go. Since protoc-gen-go would always generate messages with the proper layout of fields, this was deemed an acceptable and dramatic improvement over interface{}.
Over the next 10 years, use of Go would skyrocket and use of protobufs in Go would skyrocket as well. With increased popularity also came more diverse usages and requirements for Go protobufs and an increased number of custom proto.Message implementations that were not generated by protoc-gen-go.
The increasingly diverse ecosystem of Go types implementing the proto.Message interface led to incompatibilities, which often occurred when:
Passing custom proto.Message types to the protobuf APIs: A concrete message implementation might work with some top-level functions (e.g., proto.Marshal), but cause others (e.g., proto.Equal) to choke and panic. This occurs because the type only had partial support for being an actual message by only implementing the proto.Marshaler interface or having malformed struct field tags that happened to work with one function, but not another.
Using Go reflection on any proto.Message types: A common desire is to write general-purpose code that operates on any protobuf message. For example, a microservice might want to populate a trace_id field if it is present in a message. To accomplish this, one would use Go reflection to introspect the message type, and assume it were a pointer to a Go struct with a field named TraceId (as would be commonly produced by protoc-gen-go). If the concrete message type did not match this expectation, it either failed to work or even resulted in a panic. Such was the case for concrete message types that might be backed by a Go map instead of a Go struct.
Both of these issues are solved by following the idiom that interfaces should describe behavior, not data. This means that the interface itself should provide sufficient functionality through its methods that users can introspect and interact with all aspects of a protobuf message through a principled API. This feature is called protobuf reflection. Just as how Go reflection provides an API for programmatically interacting with any arbitrary Go value, protobuf reflection provides an API for programmatically interacting with any arbitrary protobuf message.
Since an interface cannot be extended in a backwards compatible way, this suggested the need for a new major version that defines a new proto.Message interface:
type Message interface {
ProtoReflect() protoreflect.Message
}
The new proto.Message interface contains a single ProtoReflect method that returns a protoreflect.Message, which is a reflective view over a protobuf message. In addition to making a breaking change to the proto.Message interface, we took this opportunity to cleanup the supporting functionality that operate on a proto.Message, split up complicated functionality apart into manageable packages, and to hide implementation details away from the public API.
The goal for this major revision is to improve upon all the benefits of, while addressing all the shortcomings of the old API. We hope that it will serve the Go ecosystem well for the next 10 years and beyond.