Struct MessagingClient 

pub struct MessagingClient { /* private fields */ }

Firebase Cloud Messaging client.

Unlike crate::auth::AuthClient, there is no unauthenticated emulator mode: every FCM v1 and Instance ID call requires a live OAuth2 bearer token. Build one with MessagingClientBuilder.

Implementations

impl MessagingClient

pub fn builder(project_id: impl Into<String>) -> MessagingClientBuilder

Starts building a new client for the given Firebase project.

pub async fn send( &self, message: &Message, dry_run: bool, ) -> Result<String, MessagingError>

Sends a single message, returning FCM’s assigned message id.

When dry_run is true, the message is validated but not actually delivered (FCM v1’s validate_only).

pub async fn send_each( &self, messages: &[Message], dry_run: bool, ) -> Result<BatchResponse, MessagingError>

Sends up to MAX_BATCH_SIZE messages, returning a per-message success/failure result. Each message is sent as its own HTTP request, concurrently — mirroring the official Admin SDKs’ sendEach/sendEachForMulticast, which dispatch every request before awaiting any of them (Promise.allSettled) rather than serializing round-trips. A failure sending one message does not prevent the others from being sent.

pub async fn send_each_for_multicast( &self, message_template: &Message, tokens: &[String], dry_run: bool, ) -> Result<BatchResponse, MessagingError>

Sends one message to up to MAX_BATCH_SIZE device registration tokens, individually. Equivalent to building one Message per token from a shared template and calling Self::send_each.

pub async fn subscribe_to_topic( &self, tokens: &[String], topic: &str, ) -> Result<TopicManagementResponse, MessagingError>

Subscribes device registration tokens to a topic.

A single call can partially fail: see TopicManagementResponse.

pub async fn unsubscribe_from_topic( &self, tokens: &[String], topic: &str, ) -> Result<TopicManagementResponse, MessagingError>

Unsubscribes device registration tokens from a topic.

A single call can partially fail: see TopicManagementResponse.

pub fn legacy_http_transport_enabled(&self) -> bool

Whether Self::send_each/Self::send_each_for_multicast were configured (via MessagingClientBuilder::enable_legacy_http_transport) to send each message over its own HTTP/1.1 request instead of multiplexing over HTTP/2.

This crate’s HTTP client (reqwest over hyper) negotiates HTTP/2 automatically when the server supports it and otherwise falls back to HTTP/1.1, so this flag is a no-op today; it exists so callers migrating from the official Admin SDKs (where this setting works around a legacy Node.js HTTP/2 bug) have an equivalent method to call without their code failing to compile.