client: fix FailOnNonTempDialError and add a test for it (#2276)

clientconn.go

@@ -294,6 +294,13 @@ func DialContext(ctx context.Context, target string, opts ...DialOption) (conn *
 			s := cc.GetState()
 			if s == connectivity.Ready {
 				break
+			} else if cc.dopts.copts.FailOnNonTempDialError && s == connectivity.TransientFailure {
+				if err = cc.blockingpicker.connectionError(); err != nil {
+					terr, ok := err.(interface{ Temporary() bool })
+					if ok && !terr.Temporary() {
+						return nil, err
+					}
+				}
 			}
 			if !cc.WaitForStateChange(ctx, s) {
 				// ctx got timeout or canceled.

clientconn_test.go

@@ -31,6 +31,7 @@ import (
 	"google.golang.org/grpc/credentials"
 	"google.golang.org/grpc/internal/backoff"
 	"google.golang.org/grpc/internal/leakcheck"
+	"google.golang.org/grpc/internal/transport"
 	"google.golang.org/grpc/keepalive"
 	"google.golang.org/grpc/naming"
 	"google.golang.org/grpc/resolver"
@@ -443,6 +444,26 @@ func TestDialContextCancel(t *testing.T) {
 	}
 }
 
+type failFastError struct{}
+
+func (failFastError) Error() string   { return "failfast" }
+func (failFastError) Temporary() bool { return false }
+
+func TestDialContextFailFast(t *testing.T) {
+	defer leakcheck.Check(t)
+	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
+	defer cancel()
+	failErr := failFastError{}
+	dialer := func(string, time.Duration) (net.Conn, error) {
+		return nil, failErr
+	}
+
+	_, err := DialContext(ctx, "Non-Existent.Server:80", WithBlock(), WithInsecure(), WithDialer(dialer), FailOnNonTempDialError(true))
+	if terr, ok := err.(transport.ConnectionError); !ok || terr.Origin() != failErr {
+		t.Fatalf("DialContext() = _, %v, want _, %v", err, failErr)
+	}
+}
+
 // blockingBalancer mimics the behavior of balancers whose initialization takes a long time.
 // In this test, reading from blockingBalancer.Notify() blocks forever.
 type blockingBalancer struct {

dialoptions.go

@@ -349,6 +349,9 @@ func WithStatsHandler(h stats.Handler) DialOption {
 // error, gRPC will fail the connection to the network address and won't try to
 // reconnect. The default value of FailOnNonTempDialError is false.
 //
+// FailOnNonTempDialError only affects the initial dial, and does not do
+// anything useful unless you are also using WithBlock().
+//
 // This is an EXPERIMENTAL API.
 func FailOnNonTempDialError(f bool) DialOption {
 	return newFuncDialOption(func(o *dialOptions) {