Resolve PTR names over DNS-over-HTTPS and cache them for the record TTL (closes #238)

`internal/ptr` asks Google, Cloudflare, AdGuard and Quad9 in turn, 2 seconds in all per address. It keeps answers in memory and writes them through to `cache.db`, opened by the new `internal/cachedb`. A reserved address is not sent to the resolvers.

Judgement calls:
- Each resolver gets an equal share of the time left, so a hung resolver does not use up the 2 seconds.
- The negative TTL is the smaller of the SOA minimum and the SOA's own TTL (RFC 2308).
- Names that are not host names are dropped, as `net.Resolver.LookupAddr` does.

Model: opus-5-5

README.md

@@ -498,8 +498,6 @@ How RouteWatch keeps the answers:
   whose SOA record has a TTL or minimum field of 0.
 - A failed lookup is not kept, so the next lookup of the address asks again.
 - Every 10 minutes it deletes expired answers from memory and from `cache.db`.
-- It keeps answers for at most 250,000 addresses, and keeps no answer for a new
-  address while it holds that many.
 
 ## Configuration
 

TODO.md

@@ -56,9 +56,6 @@ problem the run finds.
 - 2026-11-10: the IP lookup finds PTR names over DNS-over-HTTPS and caches them
   for the record's TTL, in memory and in `cache.db` (closes
   https://git.eeqj.de/sneak/routewatch/issues/238).
-- 2026-11-10: the tests no longer fail when a busy host stalls for several
-  seconds. Most such waits are now 15 seconds. The test of the stream's idle
-  timeout runs on a fake clock (closes #240)
 - 2026-11-10: the IP lookup checks the public VPN server lists of Mullvad,
   NordVPN, Private Internet Access and IVPN, and X4BNet's community list. Each
   result has `flags.is_vpn` and `vpn_providers` (closes

cmd/asinfo-gen/main_test.go

@@ -12,7 +12,7 @@ import (
 
 // parseTimeLimit is how long a test waits for parseCSV to return before it
 // fails.
-const parseTimeLimit = 15 * time.Second
+const parseTimeLimit = 2 * time.Second
 
 // TestParseCSVReadsTodaysLayout parses an as.csv with the country code column
 // that the source added after the description. It expects each AS with its

internal/database/database_internal_test.go

@@ -72,34 +72,17 @@ const (
 	// lockTestPrefix is the prefix the batch writes in the checkpoint tests
 	// store.
 	lockTestPrefix = "192.0.2.0/24"
-	// longBusyTimeoutMs is the busy timeout setLongBusyTimeout gives the
-	// writing connection. A checkpoint that waits for an open read takes this
-	// long.
-	longBusyTimeoutMs = 30000
 	// readerCheckpointLimit is how long a checkpoint, and a batch write
-	// waiting for it, may take while a reader is open: well under
-	// longBusyTimeoutMs, with room for a stall on a busy host.
-	readerCheckpointLimit = 15 * time.Second
+	// waiting for it, may take while a reader is open: well under the busy
+	// timeout, which a checkpoint that waits for the reader takes.
+	readerCheckpointLimit = time.Second
 	// truncateCheckpointLimit is how long a checkpoint of a WAL file over
 	// sqliteWALSizeLimitBytes may take to give up while a reader is open: its
-	// wait for the reader plus readerCheckpointLimit.
+	// wait for the reader plus a second, well under the busy timeout.
 	truncateCheckpointLimit = truncateBusyTimeoutMs*time.Millisecond +
-		readerCheckpointLimit
+		time.Second
 )
 
-// setLongBusyTimeout gives the writing connection of db longBusyTimeoutMs, so
-// that a checkpoint that waits for an open read takes far longer than a stall
-// on a busy host.
-func setLongBusyTimeout(ctx context.Context, t *testing.T, db *Database) {
-	t.Helper()
-
-	_, err := db.writer.ExecContext(ctx,
-		fmt.Sprintf("PRAGMA busy_timeout=%d", longBusyTimeoutMs))
-	if err != nil {
-		t.Fatalf("failed to set the busy timeout: %v", err)
-	}
-}
-
 // TestConnectionPoolPragmas holds several pooled connections open at once and
 // checks each one carries the per-connection settings from the DSN, plus the
 // process-wide heap limits, and sees the new file with auto_vacuum
@@ -596,8 +579,6 @@ func TestCheckpointDoesNotWaitForReader(t *testing.T) {
 	ctx := context.Background()
 	ts := time.Now().UTC()
 
-	setLongBusyTimeout(ctx, t, db)
-
 	err = db.UpdatePrefixesBatch(map[string]time.Time{"198.51.100.0/24": ts})
 	if err != nil {
 		t.Fatalf("failed to store the prefix the reader reads: %v", err)
@@ -709,7 +690,6 @@ func TestCheckpointEmptiesWALOnceNoReadIsOpen(t *testing.T) {
 	}
 
 	fillWALPastLimit(ctx, t, db)
-	setLongBusyTimeout(ctx, t, db)
 
 	started := time.Now()
 

internal/iplist/iplist_internal_test.go

@@ -646,13 +646,9 @@ func TestRestartLoadsCopiesFromDisk(t *testing.T) {
 	second.Start(t.Context())
 	second.Stop(t.Context())
 
-	// The download takes its time after writing the file, and the restart
-	// reads the file's time. A busy host can put seconds between the two.
-	const maxUpdatedDifference = 15 * time.Second
-
 	listings := lookup(t, second, listedAddr)
 	if len(listings) != 1 ||
-		listings[0].Updated.Sub(downloaded).Abs() > maxUpdatedDifference {
+		listings[0].Updated.Sub(downloaded).Abs() > time.Second {
 		t.Errorf("after the restart, listings %+v, want one updated at %s",
 			listings, downloaded)
 	}

internal/ptr/cache.go

@@ -1,6 +1,6 @@
 // Package ptr looks up the PTR names of addresses over DNS-over-HTTPS. It
-// keeps answers in memory for as long as DNS allows, and in cache.db so that
-// the answers survive a restart.
+// keeps each answer in memory for as long as DNS allows, and in cache.db so
+// that the answers survive a restart.
 package ptr
 
 import (
@@ -29,10 +29,6 @@ const (
 
 	// nameSeparator separates the names in a row of the hostnames table.
 	nameSeparator = " "
-
-	// maxAddresses limits how many addresses the cache holds, so that a public
-	// lookup cannot grow memory and cache.db without bound.
-	maxAddresses = 250_000
 )
 
 // Params is what New needs.
@@ -56,16 +52,15 @@ type entry struct {
 // Cache answers PTR lookups from memory, and asks the resolvers for an address
 // it holds no current answer for. Start loads it from cache.db.
 type Cache struct {
-	db         *sql.DB
-	resolvers  []string
-	timeLimit  time.Duration
-	maxEntries int
-	logger     *slog.Logger
-	client     *http.Client
-	now        func() time.Time
+	db        *sql.DB
+	resolvers []string
+	timeLimit time.Duration
+	logger    *slog.Logger
+	client    *http.Client
+	now       func() time.Time
 
-	// entries holds the kept answers by address, at most maxEntries of them.
-	// pending holds those that cache.db does not have yet.
+	// entries holds every answer by address. pending holds those that
+	// cache.db does not have yet.
 	mu      sync.RWMutex
 	entries map[string]entry
 	pending map[string]entry
@@ -78,16 +73,15 @@ type Cache struct {
 // New returns an empty Cache. Start loads it.
 func New(params Params) *Cache {
 	return &Cache{
-		db:         params.DB,
-		resolvers:  params.Resolvers,
-		timeLimit:  LookupTimeout,
-		maxEntries: maxAddresses,
-		logger:     params.Logger.With("component", "ptr"),
-		client:     &http.Client{},
-		now:        time.Now,
-		entries:    make(map[string]entry),
-		pending:    make(map[string]entry),
-		done:       make(chan struct{}),
+		db:        params.DB,
+		resolvers: params.Resolvers,
+		timeLimit: LookupTimeout,
+		logger:    params.Logger.With("component", "ptr"),
+		client:    &http.Client{},
+		now:       time.Now,
+		entries:   make(map[string]entry),
+		pending:   make(map[string]entry),
+		done:      make(chan struct{}),
 	}
 }
 
@@ -174,18 +168,11 @@ func (c *Cache) current(address string) (entry, bool) {
 	return found, true
 }
 
-// store keeps an answer in memory and queues it for cache.db. Once entries
-// holds maxEntries addresses it keeps no new address. It still replaces the
-// answer of an address already held.
+// store keeps an answer in memory and queues it for cache.db.
 func (c *Cache) store(address string, answer entry) {
 	c.mu.Lock()
 	defer c.mu.Unlock()
 
-	_, held := c.entries[address]
-	if !held && len(c.entries) >= c.maxEntries {
-		return
-	}
-
 	c.entries[address] = answer
 	c.pending[address] = answer
 }

internal/ptr/cache_internal_test.go

@@ -880,72 +880,3 @@ func TestFlushKeepsAnswersItCouldNotWrite(t *testing.T) {
 		t.Errorf("%d rows after the second write, want 1", got)
 	}
 }
-
-// TestNewAddressIsNotKeptAtTheLimit fills a cache that holds one address. A
-// lookup of a second address still returns its answer, and the cache keeps
-// nothing for it until the sweep has deleted the first address's expired
-// answer.
-func TestNewAddressIsNotKeptAtTheLimit(t *testing.T) {
-	t.Parallel()
-
-	db := openTestDB(t, t.TempDir())
-	resolver := newFakeResolver(t, ptrReply(300, testName))
-	cache, clock := newTestCache(t, db, resolver)
-	cache.maxEntries = 1
-
-	lookupOK(t, cache, testAddress)
-
-	names := lookupOK(t, cache, otherAddress)
-	if !slices.Equal(names, []string{testName}) {
-		t.Errorf("names %v at the limit, want %s", names, testName)
-	}
-
-	_, inEntries := cache.entries[otherAddress]
-	_, inPending := cache.pending[otherAddress]
-
-	if inEntries || inPending {
-		t.Errorf("a new address was kept at the limit: in entries %t, in pending %t",
-			inEntries, inPending)
-	}
-
-	cache.flush(t.Context())
-
-	if got := countRows(t, db); got != 1 {
-		t.Errorf("%d rows at the limit, want 1", got)
-	}
-
-	clock.advance(300 * time.Second)
-	cache.sweep(t.Context())
-	lookupOK(t, cache, otherAddress)
-
-	_, inEntries = cache.entries[otherAddress]
-	if !inEntries {
-		t.Error("a new address was not kept after the sweep made room")
-	}
-}
-
-// TestHeldAddressIsReplacedAtTheLimit lets the answer of the only address a
-// cache holds expire. The new answer replaces it in memory and is queued for
-// cache.db.
-func TestHeldAddressIsReplacedAtTheLimit(t *testing.T) {
-	t.Parallel()
-
-	resolver := newFakeResolver(t, ptrReply(300, testName))
-	cache, clock := newTestCache(t, openTestDB(t, t.TempDir()), resolver)
-	cache.maxEntries = 1
-
-	lookupOK(t, cache, testAddress)
-	clock.advance(300 * time.Second)
-	resolver.setReply(ptrReply(300, otherName))
-	lookupOK(t, cache, testAddress)
-
-	want := []string{otherName}
-
-	if got := cache.entries[testAddress].names; !slices.Equal(got, want) {
-		t.Errorf("entries holds %v, want %v", got, want)
-	}
-
-	if got := cache.pending[testAddress].names; !slices.Equal(got, want) {
-		t.Errorf("pending holds %v, want %v", got, want)
-	}
-}

internal/routewatch/app_integration_internal_test.go

@@ -8,8 +8,6 @@ import (
 	"net/http"
 	"net/http/httptest"
 	"net/netip"
-	"os"
-	"path/filepath"
 	"slices"
 	"strconv"
 	"strings"
@@ -785,74 +783,3 @@ func TestShutdownCheckpointsAndClosesTheDatabaseLast(t *testing.T) {
 		t.Errorf("Shutdown made the calls %q, want %q", store.calls, want)
 	}
 }
-
-// TestPTRCacheAnswersTheIPLookup starts the PTR cache and the HTTP server as
-// Run does, and looks up a reserved address, which asks no resolver. Without
-// the PTR cache the server has no PTR resolver, and every IP lookup answers
-// 500.
-func TestPTRCacheAnswersTheIPLookup(t *testing.T) {
-	t.Parallel()
-
-	store := newMockStore()
-	log := logger.New()
-	port := freePort(t)
-	cfg := &config.Config{Port: port, StateDir: t.TempDir()}
-	s := streamer.New(log, metrics.New())
-
-	// A started server reads its routing table in the background, so it
-	// needs one.
-	rw := New(Dependencies{
-		DB:       store,
-		Streamer: s,
-		Server: server.New(store, routetable.New(routetable.Params{}), s, log,
-			cfg),
-		Logger: log,
-		Config: cfg,
-	})
-
-	err := rw.startPTRCache(t.Context())
-	if err != nil {
-		t.Fatalf("failed to start the PTR cache: %v", err)
-	}
-
-	defer rw.stopPTRCache(t.Context())
-
-	err = rw.server.Start(t.Context())
-	if err != nil {
-		t.Fatalf("failed to start the HTTP server: %v", err)
-	}
-
-	defer func() { _ = rw.server.Stop(t.Context()) }()
-
-	req, err := http.NewRequestWithContext(t.Context(), http.MethodGet,
-		"http://"+net.JoinHostPort("127.0.0.1", strconv.Itoa(port))+
-			"/api/v1/ip/10.0.0.1", nil)
-	if err != nil {
-		t.Fatalf("failed to build the request: %v", err)
-	}
-
-	resp, err := http.DefaultClient.Do(req)
-	if err != nil {
-		t.Fatalf("the IP lookup failed: %v", err)
-	}
-
-	defer func() { _ = resp.Body.Close() }()
-
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		t.Fatalf("failed to read the answer: %v", err)
-	}
-
-	if resp.StatusCode != http.StatusOK {
-		t.Fatalf("status %d, want %d; body %s", resp.StatusCode, http.StatusOK, body)
-	}
-
-	if !strings.Contains(string(body), `"ptr":[]`) {
-		t.Errorf("answer %s, want an empty ptr", body)
-	}
-
-	_, err = os.Stat(filepath.Join(cfg.StateDir, "cache.db"))
-	if err != nil {
-		t.Errorf("cache.db is not in the state directory: %v", err)
-	}
-}

internal/server/handlers_internal_test.go

@@ -17,7 +17,6 @@ import (
 	"strings"
 	"sync/atomic"
 	"testing"
-	"testing/synctest"
 	"time"
 
 	"git.eeqj.de/sneak/routewatch/internal/config"
@@ -175,9 +174,6 @@ type tableStats struct {
 // carried 203.0.113.0/24, the two with kept routes withdrew it, so its origin
 // keeps no route. The oldest and newest routes are those of the other two
 // prefixes.
-//
-// The requests run in a synctest bubble, whose clock is fake, so a stall on
-// the host does not use up the handlers' 4-second limit.
 func TestStatsHandlersAnswerFromMemory(t *testing.T) {
 	t.Parallel()
 
@@ -223,11 +219,7 @@ func TestStatsHandlersAnswerFromMemory(t *testing.T) {
 
 	for name, handler := range handlers {
 		rec := httptest.NewRecorder()
-
-		synctest.Test(t, func(t *testing.T) {
-			handler(rec,
-				httptest.NewRequestWithContext(t.Context(), http.MethodGet, "/", nil))
-		})
+		handler(rec, httptest.NewRequestWithContext(t.Context(), http.MethodGet, "/", nil))
 
 		if rec.Code != http.StatusOK {
 			t.Errorf("%s: status %d, want %d; body %s",
@@ -262,20 +254,15 @@ func TestStatsHandlersAnswerFromMemory(t *testing.T) {
 // TestStatsAnswerCarriesChurnPerMinute checks that the stream object of the
 // /api/v1/stats answer holds route_churn_per_min as a whole number, and no
 // longer route_churn_per_sec, the moving average it replaces
-// (https://git.eeqj.de/sneak/routewatch/issues/160). The request runs in a
-// synctest bubble, whose clock is fake, so a stall on the host does not use up
-// the handler's 4-second limit.
+// (https://git.eeqj.de/sneak/routewatch/issues/160).
 func TestStatsAnswerCarriesChurnPerMinute(t *testing.T) {
 	t.Parallel()
 
 	s := newServerWithTestRoutes(t)
 
 	rec := httptest.NewRecorder()
-
-	synctest.Test(t, func(t *testing.T) {
-		s.handleStats()(rec,
-			httptest.NewRequestWithContext(t.Context(), http.MethodGet, statsPath, nil))
-	})
+	s.handleStats()(rec,
+		httptest.NewRequestWithContext(t.Context(), http.MethodGet, statsPath, nil))
 
 	if rec.Code != http.StatusOK {
 		t.Fatalf("status %d, want %d; body %s", rec.Code, http.StatusOK, rec.Body)
@@ -323,26 +310,18 @@ func (heldBackHandler) HeldBackCount() int                 { return heldBackMess
 
 // TestStatsAnswerGivesHowManyMessagesAHandlerHoldsBack registers a handler
 // that holds back messages. Its figures in /api/v1/stats, which the status
-// page shows, must give how many. The request runs in a synctest bubble, whose
-// clock is fake, so a stall on the host does not use up the handler's 4-second
-// limit.
+// page shows, must give how many.
 func TestStatsAnswerGivesHowManyMessagesAHandlerHoldsBack(t *testing.T) {
 	t.Parallel()
 
 	s := newServerWithTestRoutes(t)
 	s.streamer.RegisterHandler(heldBackHandler{})
 
-	var answer []byte
-
-	synctest.Test(t, func(t *testing.T) {
-		answer = get(t, s, statsPath).Body.Bytes()
-	})
-
 	var body struct {
 		Data StatsResponse `json:"data"`
 	}
 
-	err := json.Unmarshal(answer, &body)
+	err := json.Unmarshal(get(t, s, statsPath).Body.Bytes(), &body)
 	if err != nil {
 		t.Fatalf("decoding the answer: %v", err)
 	}
@@ -386,9 +365,7 @@ type unitStats struct {
 // names. They used to be strings with the unit in them, such as "1h18m34s" or
 // "1.1 GB" (https://git.eeqj.de/sneak/routewatch/issues/199). The stream has
 // never connected, so the uptime is 0. The memory use in /api/v1/stats is the
-// same number of bytes as its gc.heap_alloc_bytes. The requests run in a
-// synctest bubble, whose clock is fake, so a stall on the host does not use up
-// the handlers' 4-second limit.
+// same number of bytes as its gc.heap_alloc_bytes.
 func TestStatsAnswersGiveUnitsInFieldNames(t *testing.T) {
 	t.Parallel()
 
@@ -398,11 +375,7 @@ func TestStatsAnswersGiveUnitsInFieldNames(t *testing.T) {
 	s.SetASNFetcher(fixedIntervalASNFetcher{interval: fetchInterval})
 
 	for _, path := range []string{statusJSONPath, statsPath} {
-		var answer []byte
-
-		synctest.Test(t, func(t *testing.T) {
-			answer = get(t, s, path).Body.Bytes()
-		})
+		answer := get(t, s, path).Body.Bytes()
 
 		var body struct {
 			Data unitStats `json:"data"`
@@ -443,89 +416,82 @@ func TestStatsAnswersGiveUnitsInFieldNames(t *testing.T) {
 // connection_duration_seconds of /api/v1/stats is null while the stream is not
 // connected, and a whole number of seconds once it is. It used to be the
 // string "disconnected", or a Go duration such as "1h18m34s"
-// (https://git.eeqj.de/sneak/routewatch/issues/199). The test runs in a
-// synctest bubble, whose clock is fake, so a stall on the host does not use up
-// the handler's 4-second limit.
+// (https://git.eeqj.de/sneak/routewatch/issues/199).
 func TestStatsAnswerGivesConnectionDurationInSeconds(t *testing.T) {
 	t.Parallel()
 
 	s := newServerWithTestRoutes(t)
 
-	synctest.Test(t, func(t *testing.T) {
-		connectionDuration := func() json.RawMessage {
-			answer := get(t, s, statsPath).Body.Bytes()
+	connectionDuration := func() json.RawMessage {
+		answer := get(t, s, statsPath).Body.Bytes()
 
-			var body struct {
-				Data map[string]json.RawMessage `json:"data"`
-			}
-
-			err := json.Unmarshal(answer, &body)
-			if err != nil {
-				t.Fatalf("decoding the answer: %v; answer %s", err, answer)
-			}
-
-			duration, ok := body.Data["connection_duration_seconds"]
-			if !ok {
-				t.Fatalf("no connection_duration_seconds: %s", answer)
-			}
-
-			return duration
+		var body struct {
+			Data map[string]json.RawMessage `json:"data"`
 		}
 
-		disconnected := connectionDuration()
-		if string(disconnected) != "null" {
-			t.Errorf("disconnected: connection_duration_seconds = %s, want null",
-				disconnected)
+		err := json.Unmarshal(answer, &body)
+		if err != nil {
+			t.Fatalf("decoding the answer: %v; answer %s", err, answer)
 		}
 
-		s.streamer.GetMetricsTracker().SetConnected(true)
-
-		var seconds int64
-
-		connected := connectionDuration()
-
-		err := json.Unmarshal(connected, &seconds)
-		if err != nil || seconds != 0 {
-			t.Errorf("just connected: connection_duration_seconds = %s, want 0",
-				connected)
+		duration, ok := body.Data["connection_duration_seconds"]
+		if !ok {
+			t.Fatalf("no connection_duration_seconds: %s", answer)
 		}
-	})
+
+		return duration
+	}
+
+	disconnected := connectionDuration()
+	if string(disconnected) != "null" {
+		t.Errorf("disconnected: connection_duration_seconds = %s, want null",
+			disconnected)
+	}
+
+	s.streamer.GetMetricsTracker().SetConnected(true)
+
+	var seconds int64
+
+	connected := connectionDuration()
+
+	err := json.Unmarshal(connected, &seconds)
+	if err != nil || seconds != 0 {
+		t.Errorf("just connected: connection_duration_seconds = %s, want 0",
+			connected)
+	}
 }
 
 // TestStreamDurationsCountSeconds checks that the uptime and the connection
 // duration of a stream that connected 90 seconds ago are both 90. A stream
 // that has never connected has an uptime of 0, and one that is not connected
-// has no connection duration. The test runs in a synctest bubble, whose clock
-// is fake, so a stall on the host does not add to the 90 seconds.
+// has no connection duration.
 func TestStreamDurationsCountSeconds(t *testing.T) {
 	t.Parallel()
 
-	synctest.Test(t, func(t *testing.T) {
-		const connectedFor = 90 * time.Second
+	const connectedFor = 90 * time.Second
 
-		connectedSince := time.Now().Add(-connectedFor)
-		want := int64(connectedFor / time.Second)
+	connectedSince := time.Now().Add(-connectedFor)
+	want := int64(connectedFor / time.Second)
 
-		if got := uptimeSeconds(connectedSince); got != want {
-			t.Errorf("uptimeSeconds = %d, want %d", got, want)
-		}
+	if got := uptimeSeconds(connectedSince); got != want {
+		t.Errorf("uptimeSeconds = %d, want %d", got, want)
+	}
 
-		if got := uptimeSeconds(time.Time{}); got != 0 {
-			t.Errorf("uptimeSeconds before any connection = %d, want 0", got)
-		}
+	if got := uptimeSeconds(time.Time{}); got != 0 {
+		t.Errorf("uptimeSeconds before any connection = %d, want 0", got)
+	}
 
-		switch got := connectionDurationSeconds(true, connectedSince); {
-		case got == nil:
-			t.Errorf("connectionDurationSeconds = nil, want %d", want)
-		case *got != want:
-			t.Errorf("connectionDurationSeconds = %d, want %d", *got, want)
-		}
+	switch got := connectionDurationSeconds(true, connectedSince); {
+	case got == nil:
+		t.Errorf("connectionDurationSeconds = nil, want %d", want)
+	case *got != want:
+		t.Errorf("connectionDurationSeconds = %d, want %d", *got, want)
+	}
 
-		if got := connectionDurationSeconds(false, connectedSince); got != nil {
-			t.Errorf("connectionDurationSeconds while not connected = %d, want nil",
-				*got)
-		}
-	})
+	if got := connectionDurationSeconds(false, connectedSince); got != nil {
+		t.Errorf("connectionDurationSeconds while not connected = %d, want nil",
+			*got)
+	}
 }
 
 // newTestTable returns a routing table that holds a route for testPrefixV4,
@@ -1172,9 +1138,7 @@ func (failingDB) GetPrefixFirstSeen(
 // with the driver's error text. It also sends a prefix that does not parse.
 // No answer may carry the driver's or the prefix parser's text. Before the fix
 // the answers sent it to the client
-// (https://git.eeqj.de/sneak/routewatch/issues/108). The requests run in a
-// synctest bubble, whose clock is fake, so a stall on the host does not use up
-// the stats handlers' 4-second limit.
+// (https://git.eeqj.de/sneak/routewatch/issues/108).
 func TestJSONRoutesDoNotSendErrorDetails(t *testing.T) {
 	t.Parallel()
 
@@ -1195,11 +1159,8 @@ func TestJSONRoutesDoNotSendErrorDetails(t *testing.T) {
 
 	for _, tt := range tests {
 		rec := httptest.NewRecorder()
-
-		synctest.Test(t, func(t *testing.T) {
-			s.router.ServeHTTP(rec, httptest.NewRequestWithContext(
-				t.Context(), http.MethodGet, tt.path, nil))
-		})
+		s.router.ServeHTTP(rec, httptest.NewRequestWithContext(
+			t.Context(), http.MethodGet, tt.path, nil))
 
 		if rec.Code != tt.wantCode {
 			t.Errorf("%s: status %d, want %d; body %s",
@@ -1453,11 +1414,8 @@ func TestIPLookupKeepsRouteWhenPTRLookupDoesNotAnswer(t *testing.T) {
 	s := newServerWithTestRoutes(t)
 	useFakeResolver(s, &fakeResolver{ptrNeverAnswers: true})
 
-	// The request's time limit passes well after the PTR lookup's, so that a
-	// stall on a busy host does not use up the difference.
-	const requestTimeLimit = ptr.LookupTimeout + 15*time.Second
-
-	ctx, cancel := context.WithTimeout(t.Context(), requestTimeLimit)
+	// The request's time limit passes after the PTR lookup's.
+	ctx, cancel := context.WithTimeout(t.Context(), 2*ptr.LookupTimeout)
 	defer cancel()
 
 	rec := httptest.NewRecorder()
@@ -2456,13 +2414,9 @@ func settledGoroutineCount() int {
 	return prev
 }
 
-// waitForGoroutines waits up to 15 seconds for the goroutine count to drop to
-// target or below.
+// waitForGoroutines waits until the goroutine count drops to target or below.
 func waitForGoroutines(target int) bool {
-	const maxWait = 15 * time.Second
-
-	deadline := time.Now().Add(maxWait)
-	for time.Now().Before(deadline) {
+	for range 100 {
 		if runtime.NumGoroutine() <= target {
 			return true
 		}

internal/server/home_internal_test.go

@@ -76,9 +76,7 @@ func asLink(asn int) string {
 }
 
 // TestHomePageCardsShowTheTableFigures fills a routing table and the asns
-// table, works out the cards, and checks that each card shows its figure. The
-// cards are worked out in a synctest bubble, whose clock is fake, so a stall on
-// the host does not use up the 4-second limit on their database read.
+// table, works out the cards, and checks that each card shows its figure.
 func TestHomePageCardsShowTheTableFigures(t *testing.T) {
 	t.Parallel()
 
@@ -97,9 +95,7 @@ func TestHomePageCardsShowTheTableFigures(t *testing.T) {
 		t.Fatalf("GetOrCreateASNBatch: %v", err)
 	}
 
-	synctest.Test(t, func(t *testing.T) {
-		s.refreshHomeCards(t.Context())
-	})
+	s.refreshHomeCards(t.Context())
 
 	page := get(t, s, "/").Body.String()
 
@@ -283,7 +279,7 @@ func TestKeepHomeCardsFreshWorksThemOutEveryInterval(t *testing.T) {
 
 	const (
 		interval = 10 * time.Millisecond
-		maxWait  = 15 * time.Second
+		maxWait  = 5 * time.Second
 	)
 
 	table := emptyTable()
@@ -325,7 +321,7 @@ func TestKeepHomeCardsFreshWorksThemOutEveryInterval(t *testing.T) {
 func TestStartedServerWorksOutTheCards(t *testing.T) {
 	t.Parallel()
 
-	const maxWait = 15 * time.Second
+	const maxWait = 5 * time.Second
 
 	s := newServerWithTable(t, newCardsTable())
 

internal/server/middleware_internal_test.go

@@ -14,7 +14,6 @@ import (
 	"regexp"
 	"strings"
 	"testing"
-	"testing/synctest"
 	"time"
 
 	"git.eeqj.de/sneak/routewatch/internal/config"
@@ -164,10 +163,6 @@ func TestJSONAnswersPutNoUnitInAString(t *testing.T) {
 // router and the middleware write. Before the fix a panic in a handler, and a
 // method /status.json or the health check does not take, answered with an
 // empty body (https://git.eeqj.de/sneak/routewatch/issues/201).
-//
-// The requests run in a synctest bubble, whose clock is fake, so a stall on
-// the host does not use up the 2-second limit of the health check or the
-// 4-second limit of the stats handlers.
 func TestJSONAnswersCarryMeta(t *testing.T) {
 	t.Parallel()
 
@@ -186,10 +181,6 @@ func TestJSONAnswersCarryMeta(t *testing.T) {
 	s.router.Get(panicsPath, JSONValidationMiddleware(http.HandlerFunc(
 		func(http.ResponseWriter, *http.Request) { panic(errDiskIO) })).ServeHTTP)
 
-	failing := New(failingDB{}, newTestTable(),
-		streamer.New(logger.New(), metrics.New()), logger.New(), &config.Config{})
-	useFakeResolver(failing, &fakeResolver{})
-
 	tests := []struct {
 		method   string
 		path     string
@@ -225,40 +216,42 @@ func TestJSONAnswersCarryMeta(t *testing.T) {
 		{http.MethodGet, panicsPath, http.StatusInternalServerError},
 	}
 
-	synctest.Test(t, func(t *testing.T) {
-		for _, tt := range tests {
-			checkAnswerCarriesMeta(t, s, httptest.NewRequestWithContext(
-				t.Context(), tt.method, tt.path, nil), tt.wantCode)
-		}
+	for _, tt := range tests {
+		checkAnswerCarriesMeta(t, s, httptest.NewRequestWithContext(
+			t.Context(), tt.method, tt.path, nil), tt.wantCode)
+	}
 
-		// A request whose time limit has already passed answers 504. The
-		// handlers write the answers of the first five routes, and the JSON
-		// check on the route writes the last.
-		pastTimeLimit, cancel := context.WithDeadline(t.Context(), time.Now())
-		defer cancel()
+	// A request whose time limit has already passed answers 504. The handlers
+	// write the answers of the first five routes, and the JSON check on the
+	// route writes the last.
+	pastTimeLimit, cancel := context.WithDeadline(t.Context(), time.Now())
+	defer cancel()
 
-		for _, path := range []string{
-			statusJSONPath, statsPath, "/ip/" + addrInTestPrefixV4,
-			"/api/v1/ip/" + addrInTestPrefixV4, asJSONPath, writesNothingPath,
-		} {
-			checkAnswerCarriesMeta(t, s, httptest.NewRequestWithContext(
-				pastTimeLimit, http.MethodGet, path, nil), http.StatusGatewayTimeout)
-		}
+	for _, path := range []string{
+		statusJSONPath, statsPath, "/ip/" + addrInTestPrefixV4,
+		"/api/v1/ip/" + addrInTestPrefixV4, asJSONPath, writesNothingPath,
+	} {
+		checkAnswerCarriesMeta(t, s, httptest.NewRequestWithContext(
+			pastTimeLimit, http.MethodGet, path, nil), http.StatusGatewayTimeout)
+	}
 
-		// With a store whose reads fail, the health check answers 503, and
-		// the handlers of the other routes 500.
+	// With a store whose reads fail, the health check answers 503, and the
+	// handlers of the other routes 500.
+	failing := New(failingDB{}, newTestTable(),
+		streamer.New(logger.New(), metrics.New()), logger.New(), &config.Config{})
+	useFakeResolver(failing, &fakeResolver{})
+
+	checkAnswerCarriesMeta(t, failing, httptest.NewRequestWithContext(
+		t.Context(), http.MethodGet, healthCheckPath, nil),
+		http.StatusServiceUnavailable)
+
+	for _, path := range []string{
+		statusJSONPath, statsPath, asJSONPath,
+		"/api/v1/ip/" + addrInTestPrefixV4, "/ip/" + addrInTestPrefixV4,
+	} {
 		checkAnswerCarriesMeta(t, failing, httptest.NewRequestWithContext(
-			t.Context(), http.MethodGet, healthCheckPath, nil),
-			http.StatusServiceUnavailable)
-
-		for _, path := range []string{
-			statusJSONPath, statsPath, asJSONPath,
-			"/api/v1/ip/" + addrInTestPrefixV4, "/ip/" + addrInTestPrefixV4,
-		} {
-			checkAnswerCarriesMeta(t, failing, httptest.NewRequestWithContext(
-				t.Context(), http.MethodGet, path, nil), http.StatusInternalServerError)
-		}
-	})
+			t.Context(), http.MethodGet, path, nil), http.StatusInternalServerError)
+	}
 }
 
 // checkAnswerCarriesMeta serves req through the server's router, and checks

internal/server/routes_internal_test.go

@@ -5,7 +5,6 @@ import (
 	"net/http"
 	"net/http/httptest"
 	"testing"
-	"testing/synctest"
 )
 
 // TestRoutesAnswerHEAD sends HEAD for a page and JSON routes through the
@@ -13,23 +12,15 @@ import (
 // lookup of the caller's own address, are in the list because chi's GetHead
 // does not reach them. Before the fix every route answered HEAD with 405
 // (https://git.eeqj.de/sneak/routewatch/issues/111).
-//
-// The requests run in a synctest bubble, whose clock is fake, so a stall on
-// the host does not use up the stats handlers' 4-second limit. The resolver
-// is fake because no time limit would cut off a real DNS lookup in a bubble.
 func TestRoutesAnswerHEAD(t *testing.T) {
 	t.Parallel()
 
 	s := newServerWithTestRoutes(t)
-	s.resolver = &fakeResolver{}
 
 	for _, path := range []string{"/", statusJSONPath, statsPath, "/ip", "/ip/"} {
 		rec := httptest.NewRecorder()
-
-		synctest.Test(t, func(t *testing.T) {
-			s.router.ServeHTTP(rec, httptest.NewRequestWithContext(
-				t.Context(), http.MethodHead, path, nil))
-		})
+		s.router.ServeHTTP(rec, httptest.NewRequestWithContext(
+			t.Context(), http.MethodHead, path, nil))
 
 		if rec.Code != http.StatusOK {
 			t.Errorf("HEAD %s: status %d, want %d; body %s",

internal/server/server_internal_test.go

@@ -22,10 +22,10 @@ func TestStopClosesTheConnectionOfARequestStillRunning(t *testing.T) {
 		stopDeadline = 100 * time.Millisecond
 		// waitTimeout is how long the test waits for the request to reach
 		// the handler, and then for the client to see its connection closed.
-		waitTimeout = 15 * time.Second
+		waitTimeout = 2 * time.Second
 		// readHeaderTimeout bounds the test server's read of the request
 		// header, as every http.Server must.
-		readHeaderTimeout = waitTimeout
+		readHeaderTimeout = time.Second
 	)
 
 	release := make(chan struct{})

internal/streamer/streamer_internal_test.go

@@ -14,7 +14,6 @@ import (
 	"strings"
 	"sync/atomic"
 	"testing"
-	"testing/synctest"
 	"time"
 
 	"git.eeqj.de/sneak/routewatch/internal/logger"
@@ -399,7 +398,7 @@ func TestStopQueuesTheHeldStateMessages(t *testing.T) {
 	const (
 		held    = 2
 		lines   = held + 1
-		maxWait = 15 * time.Second
+		maxWait = 5 * time.Second
 	)
 
 	srv := httptest.NewServer(http.HandlerFunc(
@@ -479,15 +478,15 @@ func TestStopQueuesTheHeldStateMessages(t *testing.T) {
 // TestStopReturnsByItsDeadline stops a streamer whose handler has a full queue
 // and a STATE message held for it, and a worker that stays busy until the test
 // ends, as one does whose handler waits for a database write. Stop must return
-// within its deadline plus a margin instead of waiting for room on the queue
-// and for the worker, and log as not handed over the held message and every
-// message on the queue.
+// within its deadline plus a small margin instead of waiting for room on the
+// queue and for the worker, and log as not handed over the held message and
+// every message on the queue.
 func TestStopReturnsByItsDeadline(t *testing.T) {
 	t.Parallel()
 
 	const (
 		deadline = 100 * time.Millisecond
-		margin   = 15 * time.Second
+		margin   = 2 * time.Second
 	)
 
 	var logs bytes.Buffer
@@ -578,7 +577,7 @@ func TestStopBeforeMessageReachesQueues(t *testing.T) {
 func TestReconnectWhenStreamEnds(t *testing.T) {
 	t.Parallel()
 
-	endResponse := pipeTransport{write: func(context.Context, io.Writer) {}}
+	endResponse := func(http.ResponseWriter, *http.Request) {}
 	checkReconnect(t, endResponse, streamIdleTimeout)
 }
 
@@ -590,8 +589,9 @@ func TestReconnectWhenStreamGoesIdle(t *testing.T) {
 
 	const idleTimeout = 50 * time.Millisecond
 
-	holdResponse := pipeTransport{
-		write: func(ctx context.Context, _ io.Writer) { <-ctx.Done() },
+	holdResponse := func(w http.ResponseWriter, r *http.Request) {
+		flush(w)
+		<-r.Context().Done()
 	}
 	checkReconnect(t, holdResponse, idleTimeout)
 }
@@ -604,51 +604,16 @@ func TestReconnectWhenServerNeverAnswers(t *testing.T) {
 
 	const idleTimeout = 50 * time.Millisecond
 
-	neverAnswer := roundTripFunc(
-		func(req *http.Request) (*http.Response, error) {
-			<-req.Context().Done()
-
-			return nil, req.Context().Err()
-		},
-	)
+	neverAnswer := func(_ http.ResponseWriter, r *http.Request) {
+		<-r.Context().Done()
+	}
 	checkReconnect(t, neverAnswer, idleTimeout)
 }
 
-// pipeTransport answers every request with a response whose body write fills.
-// The body fails when the request's context ends, as a connection's does.
-// write gets that context, so that it can end when the request does.
-type pipeTransport struct {
-	write func(context.Context, io.Writer)
-}
-
-func (p pipeTransport) RoundTrip(req *http.Request) (*http.Response, error) {
-	body, w := io.Pipe()
-
-	context.AfterFunc(req.Context(), func() {
-		_ = w.CloseWithError(req.Context().Err())
-	})
-
-	go func(ctx context.Context) {
-		p.write(ctx, w)
-
-		_ = w.Close()
-	}(req.Context())
-
-	return &http.Response{StatusCode: http.StatusOK, Body: body}, nil
-}
-
-// roundTripFunc makes a function an http.RoundTripper.
-type roundTripFunc func(*http.Request) (*http.Response, error)
-
-func (f roundTripFunc) RoundTrip(req *http.Request) (*http.Response, error) {
-	return f(req)
-}
-
-// TestStreamNotDroppedWhileLinesArrive has a response send lines well within
-// the idle timeout for several times the timeout, then end. Each line must
-// start the idle timeout over, so the stream lasts until the response ends
-// instead of being dropped as idle. It runs in a synctest bubble, whose clock
-// is fake, so a stall on the host does not use up the timeout.
+// TestStreamNotDroppedWhileLinesArrive has the server send lines well within
+// the idle timeout for several times the timeout, then end the response. Each
+// line must start the idle timeout over, so the stream lasts until the server
+// ends it instead of being dropped as idle.
 func TestStreamNotDroppedWhileLinesArrive(t *testing.T) {
 	t.Parallel()
 
@@ -658,35 +623,26 @@ func TestStreamNotDroppedWhileLinesArrive(t *testing.T) {
 		sendFor      = 5 * idleTimeout
 	)
 
-	// The streamer is made outside the bubble. Its metrics start a goroutine
-	// that never returns, and a bubble waits for each of its goroutines to
-	// return.
-	s := New(logger.New(), metrics.New())
-	s.idleTimeout = idleTimeout
-	s.client = &http.Client{Transport: pipeTransport{
-		write: func(ctx context.Context, w io.Writer) {
+	srv := httptest.NewServer(http.HandlerFunc(
+		func(w http.ResponseWriter, _ *http.Request) {
 			deadline := time.Now().Add(sendFor)
 			for time.Now().Before(deadline) {
 				_, _ = io.WriteString(w, updateLine)
 
-				// The wait ends with the stream. A bubble's clock stops when
-				// the test function returns, and a goroutine then still
-				// asleep panics the test binary.
-				select {
-				case <-ctx.Done():
-					return
-				case <-time.After(lineInterval):
-				}
+				flush(w)
+				time.Sleep(lineInterval)
 			}
-		},
-	}}
+		}))
+	defer srv.Close()
 
-	synctest.Test(t, func(t *testing.T) {
-		err := s.stream(context.Background())
-		if !errors.Is(err, errStreamEnded) {
-			t.Fatalf("stream returned %v, want %v", err, errStreamEnded)
-		}
-	})
+	s := New(logger.New(), metrics.New())
+	s.url = srv.URL
+	s.idleTimeout = idleTimeout
+
+	err := s.stream(context.Background())
+	if !errors.Is(err, errStreamEnded) {
+		t.Fatalf("stream returned %v, want %v", err, errStreamEnded)
+	}
 }
 
 // TestStreamReadsLongLine has the server send one UPDATE line of about
@@ -750,7 +706,7 @@ func TestHandlerStatsReadWhileMessagesAreDropped(t *testing.T) {
 	const (
 		lines       = 5
 		wantDropped = lines - 2
-		maxWait     = 15 * time.Second
+		maxWait     = 5 * time.Second
 	)
 
 	srv := httptest.NewServer(http.HandlerFunc(
@@ -864,70 +820,61 @@ func streamLines(t *testing.T, s *Streamer, lines string) {
 // line and then stays open. The streamer must come back for a second
 // response, must not be marked connected when it does, and must handle the
 // UPDATE in it.
-//
-// It runs in a synctest bubble, whose clock is fake, so a stall on the host
-// does not use up the idle timeout. The fake server gets the requests in the
-// order the streamer makes them, because no real connection carries them.
 func checkReconnect(
-	t *testing.T, firstResponse http.RoundTripper, idleTimeout time.Duration,
+	t *testing.T, firstResponse http.HandlerFunc, idleTimeout time.Duration,
 ) {
 	t.Helper()
 
-	const maxWait = 15 * time.Second
+	const maxWait = 5 * time.Second
 
-	// The tracker is made outside the bubble. It starts a goroutine that never
-	// returns, and a bubble waits for each of its goroutines to return.
 	tracker := metrics.New()
+	connectedAtReconnect := make(chan bool, 1)
 
-	synctest.Test(t, func(t *testing.T) {
-		connectedAtReconnect := make(chan bool, 1)
-		laterResponse := pipeTransport{
-			write: func(ctx context.Context, w io.Writer) {
-				_, _ = io.WriteString(w, updateLine)
+	var requests atomic.Int32
 
-				<-ctx.Done()
-			},
-		}
+	srv := httptest.NewServer(http.HandlerFunc(
+		func(w http.ResponseWriter, r *http.Request) {
+			if requests.Add(1) == 1 {
+				firstResponse(w, r)
 
-		var requests atomic.Int32
+				return
+			}
 
-		s := New(logger.New(), tracker)
-		s.client = &http.Client{Transport: roundTripFunc(
-			func(req *http.Request) (*http.Response, error) {
-				if requests.Add(1) == 1 {
-					return firstResponse.RoundTrip(req)
-				}
+			select {
+			case connectedAtReconnect <- tracker.IsConnected():
+			default:
+			}
 
-				select {
-				case connectedAtReconnect <- tracker.IsConnected():
-				default:
-				}
+			_, _ = io.WriteString(w, updateLine)
 
-				return laterResponse.RoundTrip(req)
-			},
-		)}
-		s.idleTimeout = idleTimeout
-		s.minBackoff = time.Millisecond
-		handler := updateHandler{handled: make(chan struct{}, 1)}
-		s.RegisterHandler(handler)
+			flush(w)
+			<-r.Context().Done()
+		}))
+	defer srv.Close()
 
-		err := s.Start(t.Context())
-		if err != nil {
-			t.Fatalf("Start returned error: %v", err)
-		}
-		defer s.Stop(t.Context())
+	s := New(logger.New(), tracker)
+	s.url = srv.URL
+	s.idleTimeout = idleTimeout
+	s.minBackoff = time.Millisecond
+	handler := updateHandler{handled: make(chan struct{}, 1)}
+	s.RegisterHandler(handler)
 
-		select {
-		case <-handler.handled:
-		case <-time.After(maxWait):
-			t.Fatalf("no message from a second response handled within %s", maxWait)
-		}
+	err := s.Start(t.Context())
+	if err != nil {
+		t.Fatalf("Start returned error: %v", err)
+	}
+	defer s.Stop(t.Context())
 
-		if <-connectedAtReconnect {
-			t.Error("streamer was still marked connected when it came back " +
-				"for a second response")
-		}
-	})
+	select {
+	case <-handler.handled:
+	case <-time.After(maxWait):
+		t.Fatalf("no message from a second response handled within %s", maxWait)
+	}
+
+	if <-connectedAtReconnect {
+		t.Error("streamer was still marked connected when it came back " +
+			"for a second response")
+	}
 }
 
 // settledGoroutineCount lets transient goroutines finish, then reports the
@@ -949,13 +896,9 @@ func settledGoroutineCount() int {
 	return prev
 }
 
-// waitForGoroutines waits up to 15 seconds for the goroutine count to drop to
-// target or below.
+// waitForGoroutines waits until the goroutine count drops to target or below.
 func waitForGoroutines(target int) bool {
-	const maxWait = 15 * time.Second
-
-	deadline := time.Now().Add(maxWait)
-	for time.Now().Before(deadline) {
+	for range 100 {
 		if runtime.NumGoroutine() <= target {
 			return true
 		}