Create the image's app user without the system-user flag (closes #248)

`useradd -r` with UID 1000 made the image build warn, since 1000 is above the system range. The user keeps UID 1000, so the data directory's owner does not change.

Model: opus-5-5

README.md

@@ -143,9 +143,9 @@ and origin AS can have two kept routes, one or none.
   AS path, a next hop and a time. The next hop is the router the peer sends the
   prefix's traffic to. RouteWatch records the peer itself as the next hop. The
   page also lists the prefix's origin changes, newest first. An origin change is
-  recorded when the first peer starts to carry the prefix with an origin AS, or
+  an AS starting to announce the prefix, when its first peer carries it, or
-  when the last peer stops. The page shows at most the newest 5,000. Each row
+  stopping, when its last peer no longer does. The page shows at most the newest
-  shows:
+  5,000. Each row shows:
     - the time
     - announced or withdrawn
     - the origin AS
@@ -197,9 +197,8 @@ also carries an `@meta` object:
 - `api_version`: `1`
 - `execution_time_ms`: the number of milliseconds the request took
 
-Every value with a unit is a number. The unit ends the field name, as in `_ms`,
+Every value with a unit is a number, and the unit ends the field name. Examples
-`_seconds`, `_bytes`, `_per_sec` and `_per_min`. `latitude` and `longitude` are
+are `_ms`, `_seconds`, `_bytes`, `_per_sec` and `_per_min`.
-in degrees, and their names do not say so.
 
 The paths:
 
@@ -503,13 +502,11 @@ How RouteWatch keeps the answers:
   there.
 - It writes new answers every 5 seconds to `cache.db` in the state directory
   (see Configuration), and loads them from there at startup.
-- Each DNS record carries a TTL, the time for which it may be kept. An answer
+- An answer expires after the TTL of its records, the shortest one when there
-  expires after the shortest TTL of its records. RouteWatch keeps it for at
+  are several. RouteWatch keeps it for at least 60 seconds and at most 24 hours.
-  least 60 seconds and at most 24 hours.
+- When an address has no PTR names, that answer expires after the SOA record's
-- When an address has no PTR names, the answer's SOA record says how long that
+  minimum field, or the record's own TTL when that is shorter. RouteWatch keeps
-  answer may be kept. The answer expires after the SOA record's minimum field,
+  it for at most 1 hour.
-  or the record's own TTL when that is shorter. RouteWatch keeps it for at most
-  1 hour.
 - An answer with no PTR names and no SOA record is not kept. Neither is one
   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.
@@ -523,8 +520,7 @@ An IP lookup finds where each address is, on a best-effort basis. It uses two
 sources:
 
 - [geojs](https://www.geojs.io/) is asked first. One request carries every
-  address of the lookup that the cache does not hold. geojs has 1 second to
+  address of the lookup, and it has 1 second to answer.
-  answer.
 - [DB-IP City Lite](https://db-ip.com/db/download/ip-to-city-lite) answers when
   geojs has no answer for the address. It is a database file that RouteWatch
   reads itself.
@@ -706,10 +702,8 @@ What happens at each limit:
   allocating more.
 - The SQLite hard heap limit: a database statement fails with an out-of-memory
   error, which the handler logs. The prefix handler writes the routing table's
-  changes and its origin changes again in a later write. Everything else in a
+  changes again in a later write. Everything else in the failed write is
-  handler's failed write is dropped. The PTR cache and the geolocation part
+  dropped. The process keeps running.
-  write their answers to `cache.db` again in a later write. The process keeps
-  running.
 - The handler queues: the handlers are the four parts that take the feed's
   messages (see Design). The prefix handler's queue holds at most 100,000
   messages, about 100 MiB. Each of the other three holds at most 20,000. When a
@@ -1076,8 +1070,8 @@ calls one of them. The scripts:
 - `script/projectname`: print the project name, `routewatch`, which
   `script/docker` uses as the image tag.
 - `script/test`: build the `test` phase of the `Dockerfile` alone. It runs
-  `go test -short -count=1 -timeout 90s -race -cover ./...`. On failure it runs
+  `go test -short -count=1 -timeout 90s -race -cover ./...`, and reruns it with
-  the tests again with `-v` and without `-cover`.
+  `-v` on failure.
 - `script/lint`: build the `lint` phase of the `Dockerfile` alone. It runs
   `golangci-lint` v2.14.0 with `.golangci.yml`.
 - `script/fmt`: format the Go code with `go fmt` and the markdown with

TODO.md

@@ -22,54 +22,37 @@ formatting of the Go code and the markdown. Then it builds the image.
 The image sets memory ceilings for a 5 GiB container (README "Memory"). README
 "Running under upaas" says how to run it.
 
-Five runs of `next` on the live feed have ended since 2026-10-27, from three
+A five-day run of `1c65656` on the live feed, from 2026-10-27 to 2026-11-01,
-hours to five days long. The runs found the handlers dropping updates for five
+found three problems:
-causes, and the WHOIS fetcher asking a registry that had refused it. Each is
-fixed on `next`, and its Completed Steps entry below names its issue. A 21-hour
-run of `742e759`, which has every fix but one, dropped no update except after
-the feed stalled. That cause, https://git.eeqj.de/sneak/routewatch/issues/229,
-is fixed on `next`.
 
-A run of `e58c663` on the live feed began at 09:19 UTC on 2026-11-10
+- The prefix handler dropped updates while a peer's session went down
-(https://git.eeqj.de/sneak/routewatch/issues/232). It is to confirm that fix.
+  (https://git.eeqj.de/sneak/routewatch/issues/216).
-Next Step lists what its outcome reports. The seven IP lookup additions of
+- The AS handler dropped updates when the feed burst
-https://git.eeqj.de/sneak/routewatch/issues/203 landed after it began, so it
+  (https://git.eeqj.de/sneak/routewatch/issues/217).
-does not exercise them.
+- The WHOIS fetcher kept asking a registry that had refused it
+  (https://git.eeqj.de/sneak/routewatch/issues/218).
+
+All three are fixed on `next`. So is
+https://git.eeqj.de/sneak/routewatch/issues/222, which the third fix brought up.
+A run of `26d939a` on the live feed showed the prefix handler still dropping
+updates while the home page's cards were worked out.
+https://git.eeqj.de/sneak/routewatch/issues/226 fixed that on `next`.
+https://git.eeqj.de/sneak/routewatch/issues/225, also on `next`, takes the peers
+whose sessions go down together out of the table in one walk. A 24-hour run of
+`9bd0e91` on the live feed is checking whether any handler still drops updates.
 
 The order of the project's remaining work is kept on
 https://git.eeqj.de/sneak/routewatch/issues/86.
 
 # Next Step
 
-Post the outcome of the run of `e58c663` on
+Post the outcome of the 24-hour run of `9bd0e91` on
-https://git.eeqj.de/sneak/routewatch/issues/232. The run lasts at least 24
+https://git.eeqj.de/sneak/routewatch/issues/226: whether any handler still drops
-hours. If the feed has not stalled or reconnected by then, the run goes on until
+updates, and the process's peak memory against 5 GiB. File an issue for each
-it does, up to 72 hours in all. The outcome says:
+problem the run finds.
-
-- whether the feed stalled or reconnected
-- whether any handler dropped an update, and the most messages the prefix
-  handler's queue held at once
-- the most messages the prefix handler held back at once
-- the size of the `origin_changes` table on disk
-- the process's peak memory against 5 GiB
-
-File an issue for each problem the run finds.
 
 # Completed Steps
 
-- 2026-11-11: two more tests no longer fail on a busy host. The test of the home
-  page's card refresh runs on a fake clock. The geo test in which geojs never
-  answers counts each request as it is sent, not when it reaches the server
-  (closes https://git.eeqj.de/sneak/routewatch/issues/256).
-- 2026-11-10: `github.com/mattn/go-sqlite3` is at v1.14.53, so building on the
-  host no longer prints its two C compiler warnings (closes
-  https://git.eeqj.de/sneak/routewatch/issues/245).
-- 2026-11-10: `README.md` and this file are read whole against the rules of
-  https://git.eeqj.de/sneak/routewatch/issues/157, and every statement except
-  measured figures and what outside services do is checked against the code on
-  `next`. Four README sentences failed the prose rules. Five disagreed with the
-  code. Each is rewritten. Status and Next Step describe the live-feed run in
-  progress (closes https://git.eeqj.de/sneak/routewatch/issues/251).
 - 2026-11-10: the image creates the `routewatch` user as a regular user with UID
   1000, so the build no longer warns about the system UID range (closes
   https://git.eeqj.de/sneak/routewatch/issues/248).
@@ -80,18 +63,16 @@ File an issue for each problem the run finds.
 - 2026-11-10: the IP lookup checks the `icloud_private_relay` list of Apple's
   iCloud Private Relay networks, downloaded once a day. Each result has
   `flags.is_relay` (closes https://git.eeqj.de/sneak/routewatch/issues/237).
-- 2026-11-10: the IP lookup checks two ProtonVPN lists from a public mirror of
+- 2026-11-10: the IP lookup checks the `protonvpn` list of ProtonVPN's /24
-  Proton's server list. `protonvpn` holds the /24 networks of ProtonVPN's
+  networks and the `protonvpn_entry` list of its servers' IPv4 and IPv6
-  servers, and `protonvpn_entry` holds the servers' IPv4 and IPv6 addresses
+  addresses, from a public mirror of Proton's server list (closes
-  (closes https://git.eeqj.de/sneak/routewatch/issues/236).
+  https://git.eeqj.de/sneak/routewatch/issues/236).
 - 2026-11-10: the IP lookup finds PTR names over DNS-over-HTTPS and caches them
-  in memory and in `cache.db` for the record's TTL. A TTL is the time for which
+  for the record's TTL, in memory and in `cache.db` (closes
-  a DNS record may be kept (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 waits for something that should happen are now 15 seconds. Tests
+  seconds. Most such waits are now 15 seconds. The test of the stream's idle
-  that depend on a short idle timeout or a request limit run on a fake clock
+  timeout runs on a fake clock (closes #240)
-  (closes https://git.eeqj.de/sneak/routewatch/issues/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

go.mod

@@ -9,7 +9,7 @@ require (
 	github.com/google/btree v1.1.3
 	github.com/google/uuid v1.6.0
 	github.com/hashicorp/golang-lru/v2 v2.0.7
-	github.com/mattn/go-sqlite3 v1.14.53
+	github.com/mattn/go-sqlite3 v1.14.29
 	github.com/maxmind/mmdbwriter v1.2.0
 	github.com/miekg/dns v1.1.73
 	github.com/oschwald/maxminddb-golang/v2 v2.9.0

go.sum

@@ -10,8 +10,8 @@ github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
 github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
 github.com/hashicorp/golang-lru/v2 v2.0.7 h1:a+bsQ5rvGLjzHuww6tVxozPZFVghXaHOwFs4luLUK2k=
 github.com/hashicorp/golang-lru/v2 v2.0.7/go.mod h1:QeFd9opnmA6QUJc5vARoKUSoFhyfM2/ZepoAG6RGpeM=
-github.com/mattn/go-sqlite3 v1.14.53 h1:mnTM0owzUMrKuLo+0w51NUiB0ir2Fo7kFXlPt/l3BGs=
+github.com/mattn/go-sqlite3 v1.14.29 h1:1O6nRLJKvsi1H2Sj0Hzdfojwt8GiGKm+LOfLaBFaouQ=
-github.com/mattn/go-sqlite3 v1.14.53/go.mod h1:6JTjA44L93a0QCyJef5YvlPoKXntQPjzWv5gtm9sB6w=
+github.com/mattn/go-sqlite3 v1.14.29/go.mod h1:Uh1q+B4BYcTPb+yiD3kU8Ct7aC0hY9fxUwlHK0RXw+Y=
 github.com/maxmind/mmdbwriter v1.2.0 h1:hyvDopImmgvle3aR8AaddxXnT0iQH2KWJX3vNfkwzYM=
 github.com/maxmind/mmdbwriter v1.2.0/go.mod h1:EQmKHhk2y9DRVvyNxwCLKC5FrkXZLx4snc5OlLY5XLE=
 github.com/miekg/dns v1.1.73 h1:uhT8nJxmTrPJYClxVxTCX+CVn6qnzSiybRk72Z6DgrE=

internal/geo/locator_internal_test.go

@@ -90,12 +90,14 @@ func parseAddresses(t *testing.T, addresses ...string) []netip.Addr {
 	return addrs
 }
 
-// fakeGeoJS is a geojs server. It sends body, or status when that is set. It
+// fakeGeoJS is a geojs server. It sends body, or status when that is set, or
-// keeps each request's address list.
+// waits for the request to end when hang is true. It keeps each request's
+// address list.
 type fakeGeoJS struct {
 	server *httptest.Server
 	body   string
 	status int
+	hang   bool
 
 	mu       sync.Mutex
 	requests []string
@@ -117,6 +119,12 @@ func (f *fakeGeoJS) serve(w http.ResponseWriter, r *http.Request) {
 	f.requests = append(f.requests, r.URL.Query().Get("ip"))
 	f.mu.Unlock()
 
+	if f.hang {
+		<-r.Context().Done()
+
+		return
+	}
+
 	if f.status != 0 {
 		w.WriteHeader(f.status)
 
@@ -134,20 +142,6 @@ func (f *fakeGeoJS) requestsSeen() []string {
 	return append([]string{}, f.requests...)
 }
 
-// hangingTransport never answers: each request waits until its context ends.
-// It counts the requests as they are sent, so the count does not depend on
-// whether one reaches a server before its time limit runs out.
-type hangingTransport struct {
-	requests atomic.Int32
-}
-
-func (h *hangingTransport) RoundTrip(r *http.Request) (*http.Response, error) {
-	h.requests.Add(1)
-	<-r.Context().Done()
-
-	return nil, r.Context().Err()
-}
-
 // openCacheDB opens a new cache.db in a temporary directory.
 func openCacheDB(t *testing.T, dir string) *sql.DB {
 	t.Helper()
@@ -391,15 +385,13 @@ func TestLookupFallsBackToTheDatabase(t *testing.T) {
 
 			fake := newFakeGeoJS(t, tt.body)
 			fake.status = tt.status
+			fake.hang = tt.hang
 			dir := t.TempDir()
 			locator := newTestLocator(t, fake, openCacheDB(t, dir), dir)
-			requestCount := func() int { return len(fake.requestsSeen()) }
+			locator.timeLimit = shortTimeLimit
 
-			if tt.hang {
+			if !tt.hang {
-				hanging := &hangingTransport{}
+				locator.timeLimit = testTimeLimit
-				locator.client = &http.Client{Transport: hanging}
-				locator.timeLimit = shortTimeLimit
-				requestCount = func() int { return int(hanging.requests.Load()) }
 			}
 
 			giveDatabase(t, locator)
@@ -421,8 +413,8 @@ func TestLookupFallsBackToTheDatabase(t *testing.T) {
 			// asks geojs again.
 			lookupOne(t, locator, testAddress)
 
-			if requests := requestCount(); requests != 2 {
+			if requests := fake.requestsSeen(); len(requests) != 2 {
-				t.Errorf("%d requests, want 2", requests)
+				t.Errorf("%d requests, want 2", len(requests))
 			}
 		})
 	}

internal/server/home_internal_test.go

@@ -275,66 +275,49 @@ func TestPagesServeTheCopyOfTheCards(t *testing.T) {
 	}
 }
 
-// TestKeepHomeCardsFreshWorksThemOutEveryInterval runs keepHomeCardsFresh. It
+// TestKeepHomeCardsFreshWorksThemOutEveryInterval runs keepHomeCardsFresh with
-// works the cards out at once, again an interval after the table changes, and
+// a short interval. It works the cards out at once, again after the table
-// stops when its context ends. It runs in a synctest bubble, whose clock is
+// changes, and stops when its context ends.
-// fake, so a stall on the host cannot fail it.
 func TestKeepHomeCardsFreshWorksThemOutEveryInterval(t *testing.T) {
 	t.Parallel()
 
-	const interval = time.Minute
+	const (
+		interval = 10 * time.Millisecond
+		maxWait  = 15 * time.Second
+	)
 
 	table := emptyTable()
 	announce(table, "1.0.0.0/24", twoPrefixASN, testPeer(1), testRouteTime(0))
 
 	s := newServerWithTable(t, table)
 
-	// newestRouteOrigin returns the origin AS of the newest route on the
+	ctx, cancel := context.WithCancel(t.Context())
-	// cards, or 0 when there is none.
+	stopped := make(chan struct{})
-	newestRouteOrigin := func() int {
-		cards := s.cards.Load()
-		if cards == nil || cards.NewestRoute == nil {
-			return 0
-		}
 
-		return cards.NewestRoute.OriginASN
+	go func() {
+		s.keepHomeCardsFresh(ctx, interval)
+		close(stopped)
+	}()
+
+	newestRouteASNIs := func(asn int) bool {
+		cards := s.cards.Load()
+
+		return cards != nil && cards.NewestRoute != nil &&
+			cards.NewestRoute.OriginASN == asn
 	}
 
-	synctest.Test(t, func(t *testing.T) {
+	waitUntil(t, maxWait, func() bool { return newestRouteASNIs(twoPrefixASN) })
-		ctx, cancel := context.WithCancel(t.Context())
-		stopped := make(chan struct{})
 
-		go func() {
+	announce(table, "8.8.8.0/24", newestRouteASN, testPeer(2), testRouteTime(1))
-			s.keepHomeCardsFresh(ctx, interval)
+	waitUntil(t, maxWait, func() bool { return newestRouteASNIs(newestRouteASN) })
-			close(stopped)
-		}()
 
-		synctest.Wait()
+	cancel()
 
-		if got := newestRouteOrigin(); got != twoPrefixASN {
+	select {
-			t.Fatalf("at the start, the newest route's AS is %d, want %d",
+	case <-stopped:
-				got, twoPrefixASN)
+	case <-time.After(maxWait):
-		}
+		t.Fatal("keepHomeCardsFresh did not stop when its context ended")
-
+	}
-		announce(table, "8.8.8.0/24", newestRouteASN, testPeer(2),
-			testRouteTime(1))
-		time.Sleep(interval)
-		synctest.Wait()
-
-		if got := newestRouteOrigin(); got != newestRouteASN {
-			t.Fatalf("an interval after the table changed, the newest "+
-				"route's AS is %d, want %d", got, newestRouteASN)
-		}
-
-		cancel()
-		synctest.Wait()
-
-		select {
-		case <-stopped:
-		default:
-			t.Fatal("keepHomeCardsFresh did not stop when its context ended")
-		}
-	})
 }
 
 // TestStartedServerWorksOutTheCards starts a server and waits until its home