routing/http: fix goroutine/memory leak

I've noticed some memory and goroutine buildup, linked to that function.

It seems like when the context get cancelled, the consumer could be gone, but that function would still attempt to write in the channel and block forever.

routing/http/contentrouter/contentrouter.go

@@ -113,7 +113,7 @@ func (c *contentRouter) Ready() bool {
 
 // readProviderResponses reads peer records (and bitswap records for legacy
 // compatibility) from the iterator into the given channel.
-func readProviderResponses(iter iter.ResultIter[types.Record], ch chan<- peer.AddrInfo) {
+func readProviderResponses(ctx context.Context, iter iter.ResultIter[types.Record], ch chan<- peer.AddrInfo) {
 	defer close(ch)
 	defer iter.Close()
 	for iter.Next() {
@@ -140,10 +140,14 @@ func readProviderResponses(iter iter.ResultIter[types.Record], ch chan<- peer.Ad
 				addrs = append(addrs, a.Multiaddr)
 			}
 
-			ch <- peer.AddrInfo{
+			select {
+			case <-ctx.Done():
+				return
+			case ch <- peer.AddrInfo{
 				ID:    *result.ID,
-				Addrs: addrs,
+				Addrs: addrs}:
 			}
+
 		//lint:ignore SA1019 // ignore staticcheck
 		case types.SchemaBitswap:
 			//lint:ignore SA1019 // ignore staticcheck
@@ -162,9 +166,12 @@ func readProviderResponses(iter iter.ResultIter[types.Record], ch chan<- peer.Ad
 				addrs = append(addrs, a.Multiaddr)
 			}
 
-			ch <- peer.AddrInfo{
+			select {
+			case <-ctx.Done():
+				return
+			case ch <- peer.AddrInfo{
 				ID:    *result.ID,
-				Addrs: addrs,
+				Addrs: addrs}:
 			}
 		}
 	}
@@ -179,7 +186,7 @@ func (c *contentRouter) FindProvidersAsync(ctx context.Context, key cid.Cid, num
 		return ch
 	}
 	ch := make(chan peer.AddrInfo)
-	go readProviderResponses(resultsIter, ch)
+	go readProviderResponses(ctx, resultsIter, ch)
 	return ch
 }