fix: stop the device picker hanging on an offer

Two causes. The connect step had no timeout, so an unreachable device was
retried indefinitely instead of falling through to hold-for-later; it now
gives up after 15s and holds the offer as designed.

The picker also waited on the whole exchange, which includes a person on the
other device deciding — up to two minutes. It now closes on tap and reports
the outcome as a message, and a decline or an unanswered offer is shown as
information rather than an error, since the offer did arrive.

crates/vnidrop/src/runtime/contacts.rs

@@ -4,7 +4,7 @@
 //! [`crate::pairing`]; this is where those meet the endpoint and the UniFFI
 //! surface.
 
-use std::sync::Arc;
+use std::{sync::Arc, time::Duration};
 
 use anyhow::{Context, Result};
 use iroh::{EndpointAddr, EndpointId};
@@ -27,6 +27,12 @@ use crate::{
     util::now_ms,
 };
 
+/// How long to wait for a device to answer before treating it as not running.
+///
+/// Without this an offline peer never fails, it just keeps being retried, and
+/// the offer is never handed to the hold-for-later path.
+const OFFER_CONNECT_TIMEOUT: Duration = Duration::from_secs(15);
+
 /// Whether a device may be polled again yet.
 ///
 /// Split out because the surrounding call needs two live nodes to exercise,
@@ -419,9 +425,9 @@ impl CoreInner {
     ) -> Result<OfferResponse> {
         let addr = self.contact_addr(endpoint_id).await?;
         let client = OfferService::client(self.endpoint.clone(), addr);
-        let challenge = client
-            .request_challenge()
+        let challenge = tokio::time::timeout(OFFER_CONNECT_TIMEOUT, client.request_challenge())
             .await
+            .map_err(|_| VnidropError::transfer(anyhow::anyhow!("device did not answer in time")))?
             .context("device is not reachable")
             .map_err(VnidropError::transfer)?;