Java, Objective-C, and Other Languages

Relevant source files

This page documents the packaging and distribution strategies for Protocol Buffer language implementations not covered in previous sections. It focuses on how Protocol Buffers integrates with language-specific package managers and build systems, including Maven (Java), CocoaPods (Objective-C), PyPI (Python), and npm (JavaScript).

Java Implementation

The Java implementation is a multi-module project managed by Maven. It provides separate runtimes for full features and resource-constrained environments (Lite).

Maven Module Structure

The Java implementation provides five distinct artifacts coordinated through protobuf-parent and protobuf-bom:

Java Module Hierarchy

Module	Artifact ID	Purpose	Target Platform
Parent	`protobuf-parent`	Maven parent POM managing all modules	Build coordination
BOM	`protobuf-bom`	Bill of Materials for version management	Dependency management
Core	`protobuf-java`	Full-featured runtime with reflection and descriptors	Server-side Java
Lite	`protobuf-java-lite`	Stripped-down runtime without reflection	Android, Embedded
Util	`protobuf-java-util`	JSON conversion and utilities	Web/JSON Interop

Sources: java/pom.xml1-8 java/pom.xml284-291 java/core/src/main/java/com/google/protobuf/GeneratedMessageV3.java1-100 java/core/src/main/java/com/google/protobuf/GeneratedMessageLite.java1-100 java/util/src/main/java/com/google/protobuf/util/JsonFormat.java1-100

Build and Compatibility

The Java runtime maintains compatibility with Java 8 and Android API level 21.

Core Runtime: Uses GeneratedMessageV3 java/core/src/main/java/com/google/protobuf/GeneratedMessageV3.java1-50 as the base for generated code.
Lite Runtime: Uses GeneratedMessageLite java/core/src/main/java/com/google/protobuf/GeneratedMessageLite.java1-50 to minimize binary size and method count.
IO Layer: Handled by CodedInputStream java/core/src/main/java/com/google/protobuf/CodedInputStream.java1-100 and CodedOutputStream java/core/src/main/java/com/google/protobuf/CodedOutputStream.java1-100

Objective-C Implementation

The Objective-C implementation distributes through CocoaPods as the Protobuf pod. It uses a custom runtime optimized for Apple platforms.

Runtime Architecture

The Objective-C runtime is centered around GPBMessage and uses a descriptor-based approach for message metadata.

Objective-C Code Entity Mapping

Message Base: All generated messages inherit from GPBMessage objectivec/GPBMessage.m1-100
Descriptors: Metadata is stored in GPBDescriptor objectivec/GPBDescriptor.m1-100
Well-Known Types: Provided as pre-generated files like GPBAny.pbobjc.m objectivec/GPBAny.pbobjc.m1-50 and GPBTimestamp.pbobjc.m objectivec/GPBTimestamp.pbobjc.m1-50
Field Storage: Uses a C-struct storage pattern (e.g., GPBAny__storage_) within the Objective-C class to manage memory efficiently objectivec/GPBAny.pbobjc.m57-61

Sources: objectivec/GPBMessage.m1-100 objectivec/GPBDescriptor.m1-100 objectivec/GPBAny.pbobjc.m50-61 objectivec/GPBTimestamp.pbobjc.m50-61

Python Implementation

Python provides multiple backends (Pure Python, C++ extension, and UPB) distributed via PyPI.

Multi-Backend Architecture

The Python implementation dynamically selects the best available backend based on the environment.

Python Backend Interaction

C++ Extension: Implemented in python/google/protobuf/pyext/ python/google/protobuf/pyext/message.cc1-100 It wraps the core C++ runtime for maximum performance.
UPB Extension: Implemented in python/message.c python/message.c1-100 and python/descriptor.c python/descriptor.c1-100 This is a lightweight alternative using the UPB runtime.
Message Handling: The PyUpb_Message struct manages the mapping between Python objects and underlying C messages python/message.c169-180
Conversion: Handled by python/convert.c python/convert.c1-50 which manages type translation between Python types and wire format.

Sources: python/message.c1-61 python/google/protobuf/pyext/message.cc1-100 python/descriptor.c1-100 python/google/protobuf/internal/python_message.py1-100

JavaScript and Packaging

The JavaScript implementation is distributed via npm. The build system uses Bazel to coordinate the generation of these language-specific artifacts.

Bazel Build Rules

The protobuf.bzl file defines the rules for generating code across these languages:

Objective-C: _ObjcOuts generates .pbobjc.h and .pbobjc.m files protobuf.bzl53-58
Python: _PyOuts generates _pb2.py files protobuf.bzl60-64
Ruby/C#: Defines standard naming conventions for generated output protobuf.bzl66-76

Sources: protobuf.bzl41-76 python/dist/BUILD.bazel1-50

Summary Table

Language	Distribution	Key Classes/Files	Base Runtime
Java	Maven	`GeneratedMessageV3`, `CodedInputStream`	Java JRE
Objective-C	CocoaPods	`GPBMessage`, `GPBDescriptor`	Foundation
Python	PyPI	`Message`, `Descriptor`, `pyext/`	CPython/UPB
JavaScript	npm	`jspb.Message`	Node.js/Browser

Sources: java/core/src/main/java/com/google/protobuf/GeneratedMessageV3.java1-50 objectivec/GPBMessage.m1-50 python/message.c1-50 protobuf.bzl1-100