为何Firebase FCM通知示例采用Legacy API？能否用非Legacy API实现？

Great question! Let's break this down like we're troubleshooting a dev problem together over a quick coffee.

Why the sample uses Legacy API #

Looking at line 74 of that functions sample, the author uses sendToDevice() which relies on the Legacy FCM API. Here are the key reasons:

• Sample age: This example was built back when FCM's HTTP v1 API (the non-legacy option) wasn't the official go-to standard. At the time, the Legacy API was the simplest way to demonstrate push notification functionality.
• Simplicity: The Legacy API (and its sendToDevice() wrapper) requires minimal setup—you don't need to construct a full, structured Message object, and it works with straightforward payloads. For a sample meant to show core functionality, this keeps the code clean and easy for new developers to follow.
• Backward compatibility: When this sample was published, many existing projects were already using the Legacy API. Sticking with it made it easier for developers to adapt the code to their existing setups without major rewrites.

Can we use the non-legacy FCM API to achieve the same effect? #

Absolutely! In fact, Google now recommends using the FCM HTTP v1 API for all new projects—it’s more secure (uses OAuth2 authentication instead of static server keys) and supports more advanced features like batch message handling, detailed delivery statuses, and finer control over message behavior.

Here’s how you’d rewrite that line using the non-legacy approach with the Firebase Admin SDK:

// Replace the sendToDevice call with this non-legacy implementation
const message = {
  tokens: [token], // Supports multiple tokens for batch pushes if needed
  notification: payload.notification,
  data: payload.data,
  android: options.android,
  apns: options.apns
};

return admin.messaging().sendMulticast(message);

This sendMulticast() method uses the HTTP v1 API under the hood. It also returns more detailed results—you can check exactly which tokens succeeded or failed, making error handling and retries much more straightforward compared to the older sendToDevice() method.

If you’re only sending to a single device, you could also use send() with a single token field instead of tokens, but sendMulticast() gives you flexibility to scale to multiple devices later without changing your core logic.

内容的提问来源于stack exchange，提问作者Jek