fix: encode topic len as LE in compute_message_id

According to the spec, the message ID should encode the topic len in little-endian.
Encoding it as big-endian doesn't make a practical difference, but it's better to follow the spec.

Author: MegaRedHand

crates/net/p2p/src/lib.rs

@@ -506,7 +506,7 @@ fn compute_message_id(message: &libp2p::gossipsub::Message) -> libp2p::gossipsub
         Err(_) => (MESSAGE_DOMAIN_INVALID_SNAPPY, &message.data),
     };
     let topic = message.topic.as_str().as_bytes();
-    let topic_len = (topic.len() as u64).to_be_bytes();
+    let topic_len = (topic.len() as u64).to_le_bytes();
     hasher.update(domain);
     hasher.update(topic_len);
     hasher.update(topic);