protocol: Do not panic on trivially empty duration

Any connected peer can send a `Subscribe` with `since` after `until`.
The service passes the range to `Store::filtered`, which asserts
`from <= to`. The panic stops the service thread.

Remove the assertion. The query already returns no rows when
`from > to`, so the peer gets no history.

crates/radicle-node/src/tests.rs

@@ -601,6 +601,27 @@ fn announcement_rebroadcast_timestamp_filtered() {
     assert_eq!(relayed, second);
 }
 
+#[test]
+fn announcement_rebroadcast_inverted_range() {
+    let mut amy = Peer::amy();
+    let bob = Peer::bob();
+    let cid = Peer::cid();
+
+    amy.connect_to(&bob);
+    amy.connect_from(&cid);
+    amy.outbox().for_each(drop);
+
+    amy.receive(
+        *cid.nid(),
+        Message::Subscribe(Subscribe {
+            filter: Filter::default(),
+            since: Timestamp::MAX,
+            until: Timestamp::MIN,
+        }),
+    );
+    assert_eq!(amy.relayed(*cid.nid()).count(), 0);
+}
+
 #[test]
 fn announcement_relay() {
     let mut amy = Peer::amy();

crates/radicle-protocol/src/service/gossip/store.rs

@@ -54,11 +54,6 @@ pub trait Store {
 
     /// Get all the latest gossip messages of all nodes, filtered by inventory filter and
     /// announcement timestamps.
-    ///
-    /// # Panics
-    ///
-    /// Panics if `from` > `to`.
-    ///
     fn filtered<'a>(
         &'a self,
         filter: &'a Filter,
@@ -188,8 +183,6 @@ impl Store for Database {
              WHERE timestamp >= ?1 and timestamp < ?2
              ORDER BY timestamp, node, type",
         )?;
-        assert!(*from <= *to);
-
         stmt.bind((1, &from))?;
         stmt.bind((2, &to))?;