Errors handling ShardRunnerMessage shut down shards

[DPlayer234]

On next, when handling ShardRunnerMessage variants that send messages to the websocket, encountering an error will shut down the shard without logging any sort of warning or error. The shard is not queued to be restarted in that case either:

serenity/src/gateway/sharding/shard_runner.rs

Lines 267 to 319 in 5209741

	// Handles a received value over the shard runner rx channel.
	//
	// Returns a boolean on whether the shard runner can continue.
	//
	// This always returns true, except in the case that the shard manager asked the runner to
	// shutdown or restart.
	#[cfg_attr(feature = "tracing_instrument", instrument(skip(self)))]
	async fn handle_rx_value(&mut self, msg: ShardRunnerMessage) -> bool {
	match msg {
	ShardRunnerMessage::Restart => {
	self.restart().await;
	false
	},
	ShardRunnerMessage::Shutdown(code) => {
	self.shutdown(code).await;
	false
	},
	ShardRunnerMessage::ChunkGuild {
	guild_id,
	limit,
	presences,
	filter,
	nonce,
	} => self
	.shard
	.chunk_guild(guild_id, limit, presences, filter, nonce.as_deref())
	.await
	.is_ok(),
	ShardRunnerMessage::SetPresence {
	activity,
	status,
	} => {
	if let Some(activity) = activity {
	self.shard.set_activity(activity);
	}
	if let Some(status) = status {
	self.shard.set_status(status);
	}
	self.shard.update_presence().await.is_ok()
	},
	#[cfg(feature = "voice")]
	ShardRunnerMessage::UpdateVoiceState {
	guild_id,
	channel_id,
	self_mute,
	self_deaf,
	} => self
	.shard
	.update_voice_state(guild_id, channel_id, self_mute, self_deaf)
	.await
	.is_ok(),
	}
	}

These cases should unconditionally return true, instead of returning .is_ok(), and log the error.

Alternatively, it could bubble the error back up the original caller, but that might add significant complexity.