0Pricing
Spring Boot 4 Complete Guide · 课时

AOT 处理与原生构建流水线

了解 Spring 的提前编译引擎，并使用 GraalVM 工具链构建原生镜像。

AOT 处理与原生构建流水线 是 CoddyKit 上的免费 Spring Boot 4 Complete Guide 课时。 这是第 1 节课，共 4 节。 你可以在下方免费阅读本课时的完整内容 — 然后在浏览器中使用内置代码编辑器和全天候 AI 导师进行实践。 这是 Spring Boot 4 Complete Guide 学习路径的一部分，你的进度在网页和 CoddyKit 应用中同步。 Spring Boot 4 Complete Guide 课程共包含 4 节课。

本课时的部分内容尚未翻译，以英文显示。

Why Native Images?

A traditional Spring Boot app runs on the JVM: bytecode is loaded, classes are verified, and the JIT compiler warms up over time. That gives great peak throughput but pays a cost at startup and in memory footprint.

GraalVM native images flip the model. Instead of shipping bytecode plus a JVM, you produce a single, self-contained executable where almost all the work normally done at runtime is moved to build time.

  • Startup drops from seconds to tens of milliseconds.
  • Memory footprint shrinks dramatically (no JIT, no class metadata bloat).
  • Trade-off: longer, heavier builds and a closed-world assumption.

This makes native images ideal for serverless, CLIs, and high-density containers.

The Closed-World Assumption

GraalVM's native compiler (native-image) performs static analysis of your entire program and only includes code it can prove is reachable. Everything must be known at build time — this is the closed-world assumption.

The features that make Spring flexible at runtime are exactly the ones that break under closed-world analysis:

  • Reflection — calling methods/fields discovered by name at runtime.
  • Dynamic proxies — Spring AOP, @Transactional, repository interfaces.
  • Resource loading — files looked up by path at runtime.
  • Serialization and runtime class generation.

Spring's AOT engine exists to bridge this gap: it analyzes your application ahead of time and emits the metadata and code GraalVM needs.

What Spring AOT Actually Does

When you build for native (or just enable AOT), Spring runs an ahead-of-time processing phase that transforms your dynamic application context into static, pre-computed form.

Concretely, Spring AOT generates:

  • Bean definition code — instead of scanning and parsing at runtime, Spring emits Java source (*__BeanDefinitions.java) that registers beans programmatically.
  • An ApplicationContextInitializer that wires the context without classpath scanning.
  • GraalVM reachability metadata — JSON hints for reflection, resources, proxies, and serialization.

The runtime context becomes effectively frozen: the bean set is fixed at build time. You cannot add beans dynamically after AOT processing.

Enabling AOT in the Build

Spring AOT is driven by build plugins. With Maven, the spring-boot-maven-plugin exposes a process-aot goal; the native profile wires it together with GraalVM's native-maven-plugin.

The key dependency for native builds is the GraalVM toolchain plugin. Here is a typical Maven setup that activates AOT and native compilation.

<plugin>
  <groupId>org.graalvm.buildtools</groupId>
  <artifactId>native-maven-plugin</artifactId>
</plugin>
<plugin>
  <groupId>org.springframework.boot</groupId>
  <artifactId>spring-boot-maven-plugin</artifactId>
  <executions>
    <execution>
      <id>process-aot</id>
      <goals>
        <goal>process-aot</goal>
      </goals>
    </execution>
  </executions>
</plugin>

The Native Build Pipeline, Step by Step

Building a native image is a multi-stage pipeline. Understanding the order helps you locate failures.

  • 1. Compile — normal javac compilation of your sources.
  • 2. AOT processing — Spring runs process-aot, generating bean-definition sources and reachability metadata under target/spring-aot.
  • 3. AOT compile — the generated sources are compiled alongside your code.
  • 4. native-image — GraalVM performs static analysis (the closed-world step) and emits a native executable.

Trigger it with Maven via the native profile:

# Produce the native executable in target/
./mvnw -Pnative native:compile

# Or build a native container image with buildpacks
./mvnw -Pnative spring-boot:build-image

Runtime Hints: The Programmatic API

Spring AOT detects most reflection automatically, but for your own dynamic code (e.g., a class you load reflectively), you must declare hints. The idiomatic way is a RuntimeHintsRegistrar.

You register the registrar with @ImportRuntimeHints on a configuration or component. At AOT time Spring invokes it and folds your hints into the GraalVM metadata.

import org.springframework.aot.hint.MemberCategory;
import org.springframework.aot.hint.RuntimeHints;
import org.springframework.aot.hint.RuntimeHintsRegistrar;

public class MyRuntimeHints implements RuntimeHintsRegistrar {

    @Override
    public void registerHints(RuntimeHints hints, ClassLoader classLoader) {
        hints.reflection().registerType(
            com.example.PaymentProcessor.class,
            MemberCategory.INVOKE_DECLARED_CONSTRUCTORS,
            MemberCategory.INVOKE_PUBLIC_METHODS);

        hints.resources().registerPattern("config/*.properties");
    }
}

Wiring Hints into the Context

A RuntimeHintsRegistrar does nothing until Spring knows about it. Attach it with @ImportRuntimeHints so it participates in AOT processing.

This keeps native-specific knowledge close to the code that needs it, rather than in a separate JSON file you must hand-maintain.

import org.springframework.context.annotation.Configuration;
import org.springframework.context.annotation.ImportRuntimeHints;

@Configuration
@ImportRuntimeHints(MyRuntimeHints.class)
public class NativeConfig {
    // Beans defined here are processed with the hints above
}

The @RegisterReflectionForBinding Shortcut

The most common reason to need hints is serialization/deserialization of DTOs — Jackson reflects over your classes. Writing a full registrar for every DTO is tedious.

Spring offers @RegisterReflectionForBinding, which automatically registers the reflection metadata needed to bind (serialize/deserialize) the listed types.

import org.springframework.aot.hint.annotation.RegisterReflectionForBinding;
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RestController;

@RestController
@RegisterReflectionForBinding({ OrderResponse.class, OrderItem.class })
class OrderController {

    @GetMapping("/orders/latest")
    OrderResponse latest() {
        return new OrderResponse("A-100", List.of(new OrderItem("sku-1", 2)));
    }
}

The GraalVM Tracing Agent

For third-party libraries that use reflection but ship no metadata, the AOT engine can't always infer the hints. The fallback is the GraalVM tracing agent.

You run your app on the JVM with the agent attached and exercise its code paths (tests, a smoke run). The agent records every reflective access, resource load, and proxy and writes them as metadata JSON.

  • Output lands under META-INF/native-image/.
  • It only captures paths you actually execute — incomplete test coverage means missing hints.
  • Treat it as a last resort; prefer reachability metadata repos and Spring's automatic detection first.
# Attach the tracing agent while running tests on the JVM
java -agentlib:native-image-agent=config-output-dir=src/main/resources/META-INF/native-image \
     -jar target/app.jar

Build-Time vs Runtime Initialization

A subtle but critical concept: native-image can run static initializers at build time and bake the resulting state into the image. Spring Boot's defaults push most app initialization to build time for speed.

This causes a classic bug: anything that captures environment-specific or time-sensitive state in a static initializer gets frozen at build time.

  • A static final SecureRandom seed computed at build time would be identical on every machine.
  • Reading an environment variable in a static block captures the build server's value, not production's.

The fix: defer such work to runtime (e.g., a @Bean method or lazy init) so it executes when the executable actually starts.

@Configuration
class CryptoConfig {

    // Created when the native executable starts, NOT at build time
    @Bean
    SecureRandom secureRandom() {
        return new SecureRandom();
    }
}

Verifying and Debugging the Native Build

Native builds fail differently from JVM apps. Two habits save hours:

  • Test the AOT path on the JVM first. Run with the springAot mode or the generated context before doing the slow native-image step — most bean-wiring issues surface here in seconds.
  • Read the static-analysis errors. A ClassNotFoundException or No instances of X are allowed in the image heap at run time almost always means a missing reflection/resource hint or an accidental build-time initialization.

You can run the AOT-processed app on the JVM to validate the frozen context quickly:

# Run the AOT-optimized context on a regular JVM (fast feedback loop)
./mvnw spring-boot:run -Dspring-boot.run.profiles=default \
  -Dspring.aot.enabled=true

Quick Check: AOT and the Native Pipeline

Test your understanding of how Spring's AOT engine cooperates with GraalVM.

Recap: AOT Processing and the Native Pipeline

You now understand how Spring Boot 4 turns a dynamic application into a native executable:

  • Why native: millisecond startup and low memory, at the cost of slow builds and a closed-world assumption.
  • Spring AOT runs at build time, emitting bean-definition code, an ApplicationContextInitializer, and GraalVM reachability metadata; the bean set becomes frozen.
  • The pipeline: compile → process-aot → compile generated sources → native-image static analysis.
  • Hints: use RuntimeHintsRegistrar + @ImportRuntimeHints for custom needs, @RegisterReflectionForBinding for DTOs, and the tracing agent as a last resort for opaque libraries.
  • Watch out for build-time initialization freezing state; defer environment- and time-sensitive work to runtime beans.
  • Debug fast by running the AOT context on the JVM before the slow native build.

常见问题解答

「AOT 处理与原生构建流水线」课时是免费的吗？

是的 — 「AOT 处理与原生构建流水线」的完整文本可在网页上免费阅读。要进行交互式练习（内置代码编辑器和全天候 AI 导师）并解锁 Spring Boot 4 Complete Guide 课程的其余内容，请升级到 CoddyKit PRO。 Spring Boot 4 Complete Guide 课程共包含 4 节课。

「AOT 处理与原生构建流水线」这节课中我会学到什么？

了解 Spring 的提前编译引擎，并使用 GraalVM 工具链构建原生镜像。 你通过在浏览器中直接运行的动手代码来练习 Spring Boot 4 Complete Guide，全天候 AI 导师会在你学习这节课的过程中回答你的问题。

学习 Spring Boot 4 Complete Guide 需要有经验吗？

无需任何先前经验。CoddyKit 上的 Spring Boot 4 Complete Guide 课程适合初学者到高级学习者，你可以从这里开始或从头开始，按照自己的节奏学习。 这是第 1 节课，共 4 节。

「AOT 处理与原生构建流水线」课时需要多长时间？

大多数 CoddyKit 课程大约需要 5–10 分钟。每节课都很精短且互动，所以你能稳步进步，并在网页和应用中从离开的地方继续。

我能在这节 Spring Boot 4 Complete Guide 课中编写并运行代码吗？

能。每节 Spring Boot 4 Complete Guide 课都包含内置代码编辑器，你可以在浏览器中直接编写并运行真实代码，并获得即时 AI 反馈 — 无需本地设置。

此课程中的所有课时

  1. AOT 处理与原生构建流水线
  2. 反射与资源的运行时提示
  3. 类数据共享与 JVM 启动调优
  4. 诊断并修复原生兼容性问题
← 返回 Spring Boot 4 Complete Guide