resolver: a reply marked recursion available from a root server's address stops the lookup (closes #300)

dnswatcher never asks for recursion and a root server never offers it,
so a reply marked recursion available from a root server's address
came from a resolver on the network answering in its place. The first
such reply now stops the lookup at once with the existing "this network
intercepts DNS queries", naming the address, and no other root server
is asked, whether or not that resolver had the name cached. Before, a
cached name was checked against the resolver's second-hand answer.

On such a network dnswatcher and its tests fail with that error, by
design; nothing works around the interception. A live test passes
public resolvers as the root servers and expects the error.

Model: opus-5-5

README.md

@@ -564,13 +564,14 @@ This approach ensures:
 A root server never offers recursion, so a reply marked "recursion available"
 from a root server's address came from a recursive resolver on the network
 answering in its place, as on a network where a router rule, a VPN or a DNS
-filter sends all DNS to one resolver. Such a reply is passed over for the next
+filter sends all DNS to one resolver. The first such reply from a root server's
-root server. When every root server refuses a query or gives such a reply, the
+address stops the lookup at once with the error "this network intercepts DNS
-lookup fails with the error "this network intercepts DNS queries". Every lookup
+queries", and no other root server is asked. A lookup that every root server
-starts at the root servers, so on such a network every lookup fails with that
+refuses fails with the same error. Every lookup starts at the root servers, so
-error, whether or not that resolver has the name cached, and dnswatcher's tests
+on such a network every lookup fails with that error, whether or not that
-fail there too. That is by design: nothing asks again another way or otherwise
+resolver has the name cached, and dnswatcher's tests fail there too. That is by
-works around the interception.
+design: nothing asks again another way or otherwise works around the
+interception.
 
 A watched name's records are stored as its nameservers return them, CNAME
 included. When they return a CNAME and no address, the chain of every CNAME

TODO.md

@@ -19,10 +19,10 @@ milestone: https://git.eeqj.de/sneak/dnswatcher/milestone/7
 
 # Completed Steps
 
-- 2026-10-16: a reply marked "recursion available" from a root server's address
+- 2026-10-16: the first reply marked "recursion available" from a root server's
-  is not used; when every root server refuses or gives one, the lookup fails
+  address stops the lookup with "this network intercepts DNS queries", so on a
-  with "this network intercepts DNS queries", so on such a network dnswatcher
+  network that sends DNS to a resolver dnswatcher and its tests fail, by design
-  and its tests fail, by design (closes #300).
+  (closes #300).
 - 2026-10-15: a query its caller's context ends before it is sent records no
   reason about the nameserver, and a try the context's deadline cuts short is a
   timeout, so a nameserver whose only failures were timeouts is `timeout`, not a

internal/resolver/errors.go

@@ -53,16 +53,9 @@ var (
 		"reply truncated and its retry over TCP failed",
 	)
 
-	// ErrRecursionAvailable is the reason a reply from a root server's
+	// ErrIntercepted is returned when a root server's address gave a
-	// address marked recursion available is not used. Root servers offer
+	// reply marked recursion available, which ends the lookup at once,
-	// no recursion, so the reply came from something on the network
+	// or when every root server refused a query. Root servers do
-	// answering in their place.
-	ErrRecursionAvailable = errors.New(
-		"reply marked recursion available, so not from a root server",
-	)
-
-	// ErrIntercepted is returned when every root server refused a
-	// query or gave a reply marked recursion available. Root servers do
 	// neither, so those replies came from something on the network
 	// answering in their place. On such a network dnswatcher fails with
 	// it, by design: nothing works around the interception.

internal/resolver/iterative.go

@@ -284,11 +284,14 @@ func shuffled(
 }
 
 // queryServers asks servers, the servers of zone, about name in a random
-// order until one gives a reply it uses (see replyError). A server that
+// order until one gives a usable reply. A server that times out, refuses
-// times out, refuses or gives a reply it does not use is passed over for
+// or gives a reply that is not usable is passed over for the next. When
-// the next. When every server refused, the error says so. When they are
+// every server refused, the error says so, and when they are the root
-// the root servers and each one refused or gave a reply marked recursion
+// servers it is ErrIntercepted. When they are the root servers, the
-// available, the error is ErrIntercepted.
+// first reply marked recursion available also ends the lookup with
+// ErrIntercepted, naming the address it came from, and no other root
+// server is asked: root servers offer no recursion, so that reply came
+// from something on the network answering in their place.
 func (r *Resolver) queryServers(
 	ctx context.Context,
 	servers []string,
@@ -299,7 +302,6 @@ func (r *Resolver) queryServers(
 	var lastErr error
 
 	refused := 0
-	recursionAvailable := 0
 
 	for _, ip := range shuffled(servers, rand.Shuffle) {
 		if checkCtx(ctx) != nil {
@@ -307,8 +309,18 @@ func (r *Resolver) queryServers(
 		}
 
 		resp, err := r.queryDNS(ctx, ip, name, qtype)
-		if err == nil {
+		if err == nil && zone == "." && resp.RecursionAvailable {
-			err = replyError(resp, ip, zone, name)
+			return nil, fmt.Errorf(
+				"query %s @%s: reply marked recursion available, "+
+					"so not from a root server: %w",
+				name, ip, ErrIntercepted,
+			)
+		}
+
+		if err == nil && !usableReply(resp, zone, name) {
+			err = fmt.Errorf(
+				"query %s @%s: %w", name, ip, ErrUnusableReply,
+			)
 		}
 
 		if err == nil {
@@ -319,10 +331,6 @@ func (r *Resolver) queryServers(
 			refused++
 		}
 
-		if errors.Is(err, ErrRecursionAvailable) {
-			recursionAvailable++
-		}
-
 		lastErr = err
 	}
 
@@ -333,14 +341,6 @@ func (r *Resolver) queryServers(
 		)
 	}
 
-	if refused+recursionAvailable == len(servers) && zone == "." {
-		return nil, fmt.Errorf(
-			"every root server refused a query for %s or gave a reply "+
-				"marked recursion available: %w",
-			name, ErrIntercepted,
-		)
-	}
-
 	if refused == len(servers) {
 		return nil, fmt.Errorf(
 			"every server of %s refused a query for %s: %w",
@@ -351,27 +351,6 @@ func (r *Resolver) queryServers(
 	return nil, fmt.Errorf("all servers failed: %w", lastErr)
 }
 
-// replyError returns why resp, the reply of the server at ip, one of the
-// servers of zone, to a query about name, is not used, or nil when it
-// is. A reply that is not usable is not used. Nor is a reply from a root
-// server marked recursion available: root servers offer no recursion,
-// so it came from something on the network answering in their place.
-func replyError(resp *dns.Msg, ip string, zone string, name string) error {
-	if zone == "." && resp.RecursionAvailable {
-		return fmt.Errorf(
-			"query %s @%s: %w", name, ip, ErrRecursionAvailable,
-		)
-	}
-
-	if !usableReply(resp, zone, name) {
-		return fmt.Errorf(
-			"query %s @%s: %w", name, ip, ErrUnusableReply,
-		)
-	}
-
-	return nil
-}
-
 // isErrorReply reports whether msg is an error reply: one with any code
 // but NOERROR and NXDOMAIN, such as SERVFAIL, NOTIMP or FORMERR. An error
 // reply says nothing about the name's records.

internal/resolver/resolver_test.go

@@ -853,12 +853,12 @@ func TestQueryServers_NotEveryRootServerRefused(t *testing.T) {
 // TestQueryServers_RecursiveResolversAsRootServers passes public
 // recursive resolvers of four operators to QueryServers as the servers
 // of the root zone, as a network that sends every DNS query to a
-// resolver would. Each refuses the query or gives a reply marked
+// resolver would. Their replies are marked recursion available, which
-// recursion available, which no root server does: SERVFAIL, or an
+// no root server's is: SERVFAIL, or an answer from a resolver's cache,
-// answer from its cache, as 1.1.1.1 gives for com., which would be used
+// as 1.1.1.1 gives for com., which would be used if a root server gave
-// if a root server gave it. Either way the error is ErrIntercepted, and
+// it. The first such reply ends the lookup with ErrIntercepted, naming
-// since 1.1.1.1 and 8.8.8.8 refuse no such query, it says a reply was
+// the resolver that gave it. A resolver that refuses or gives no reply
-// marked recursion available.
+// is passed over for the next, so one of them answering is enough.
 func TestQueryServers_RecursiveResolversAsRootServers(t *testing.T) {
 	t.Parallel()
 
@@ -877,9 +877,11 @@ func TestQueryServers_RecursiveResolversAsRootServers(t *testing.T) {
 			func(ctx context.Context) error {
 				_, err = r.QueryServers(ctx, resolvers, ".", name, dns.TypeNS)
 
-				// When the error is another, one resolver may have given
+				// When no resolver gave a reply marked recursion
-				// no reply at all, so the attempt is tried again.
+				// available, each refused or gave no reply at all, so
-				if err != nil && !errors.Is(err, resolver.ErrIntercepted) {
+				// the attempt is tried again.
+				if err != nil &&
+					!strings.Contains(err.Error(), "marked recursion available") {
 					return fmt.Errorf(
 						"%w: %w", livednstest.ErrNoAnswer, err,
 					)
@@ -890,12 +892,18 @@ func TestQueryServers_RecursiveResolversAsRootServers(t *testing.T) {
 		)
 
 		require.ErrorIs(t, err, resolver.ErrIntercepted)
-		require.EqualError(
+
-			t, err,
+		// The error names the resolver whose reply ended the lookup.
-			"every root server refused a query for "+name+" or gave a "+
+		messages := make([]string, 0, len(resolvers))
-				"reply marked recursion available: "+
+		for _, ip := range resolvers {
-				"this network intercepts DNS queries",
+			messages = append(messages,
-		)
+				"query "+name+" @"+ip+": reply marked recursion "+
+					"available, so not from a root server: "+
+					"this network intercepts DNS queries",
+			)
+		}
+
+		require.Contains(t, messages, err.Error())
 	}
 }