iOS: stabilize QR onboarding pairing

This commit is contained in:
Mariano Belinky
2026-02-15 16:05:49 +00:00
committed by Mariano Belinky
parent 7cbaae15e0
commit 4825bb52c4
8 changed files with 237 additions and 31 deletions

View File

@@ -6,10 +6,10 @@ struct GatewayTrustPromptAlert: ViewModifier {
private var promptBinding: Binding<GatewayConnectionController.TrustPrompt?> {
Binding(
get: { self.gatewayController.pendingTrustPrompt },
set: { newValue in
if newValue == nil {
self.gatewayController.clearPendingTrustPrompt()
}
set: { _ in
// Keep pending trust state until explicit user action.
// `alert(item:)` may set the binding to nil during dismissal, which can race with
// the button handler and cause accept to no-op.
})
}
@@ -39,4 +39,3 @@ extension View {
self.modifier(GatewayTrustPromptAlert())
}
}

View File

@@ -57,6 +57,11 @@ final class NodeAppModel {
var gatewayRemoteAddress: String?
var connectedGatewayID: String?
var gatewayAutoReconnectEnabled: Bool = true
// When the gateway requires pairing approval, we pause reconnect churn and show a stable UX.
// Reconnect loops (both our own and the underlying WebSocket watchdog) can otherwise generate
// multiple pending requests and cause the onboarding UI to "flip-flop".
var gatewayPairingPaused: Bool = false
var gatewayPairingRequestId: String?
var seamColorHex: String?
private var mainSessionBaseKey: String = "main"
var selectedAgentId: String?
@@ -1547,6 +1552,8 @@ extension NodeAppModel {
func disconnectGateway() {
self.gatewayAutoReconnectEnabled = false
self.gatewayPairingPaused = false
self.gatewayPairingRequestId = nil
self.nodeGatewayTask?.cancel()
self.nodeGatewayTask = nil
self.operatorGatewayTask?.cancel()
@@ -1576,6 +1583,8 @@ extension NodeAppModel {
private extension NodeAppModel {
func prepareForGatewayConnect(url: URL, stableID: String) {
self.gatewayAutoReconnectEnabled = true
self.gatewayPairingPaused = false
self.gatewayPairingRequestId = nil
self.nodeGatewayTask?.cancel()
self.operatorGatewayTask?.cancel()
self.gatewayHealthMonitor.stop()
@@ -1605,6 +1614,10 @@ private extension NodeAppModel {
guard let self else { return }
var attempt = 0
while !Task.isCancelled {
if self.gatewayPairingPaused {
try? await Task.sleep(nanoseconds: 1_000_000_000)
continue
}
if await self.isOperatorConnected() {
try? await Task.sleep(nanoseconds: 1_000_000_000)
continue
@@ -1680,8 +1693,13 @@ private extension NodeAppModel {
var attempt = 0
var currentOptions = nodeOptions
var didFallbackClientId = false
var pausedForPairingApproval = false
while !Task.isCancelled {
if self.gatewayPairingPaused {
try? await Task.sleep(nanoseconds: 1_000_000_000)
continue
}
if await self.isGatewayConnected() {
try? await Task.sleep(nanoseconds: 1_000_000_000)
continue
@@ -1768,11 +1786,52 @@ private extension NodeAppModel {
self.showLocalCanvasOnDisconnect()
}
GatewayDiagnostics.log("gateway connect error: \(error.localizedDescription)")
// If pairing is required, stop reconnect churn. The user must approve the request
// on the gateway before another connect attempt will succeed, and retry loops can
// generate multiple pending requests.
let lower = error.localizedDescription.lowercased()
if lower.contains("not_paired") || lower.contains("pairing required") {
let requestId: String? = {
// GatewayResponseError for connect decorates the message with `(requestId: ...)`.
// Keep this resilient since other layers may wrap the text.
let text = error.localizedDescription
guard let start = text.range(of: "(requestId: ")?.upperBound else { return nil }
guard let end = text[start...].firstIndex(of: ")") else { return nil }
let raw = String(text[start..<end]).trimmingCharacters(in: .whitespacesAndNewlines)
return raw.isEmpty ? nil : raw
}()
await MainActor.run {
self.gatewayAutoReconnectEnabled = false
self.gatewayPairingPaused = true
self.gatewayPairingRequestId = requestId
if let requestId, !requestId.isEmpty {
self.gatewayStatusText =
"Pairing required (requestId: \(requestId)). Approve on gateway, then tap Resume."
} else {
self.gatewayStatusText = "Pairing required. Approve on gateway, then tap Resume."
}
}
// Hard stop the underlying WebSocket watchdog reconnects so the UI stays stable and
// we don't generate multiple pending requests while waiting for approval.
pausedForPairingApproval = true
self.operatorGatewayTask?.cancel()
self.operatorGatewayTask = nil
await self.operatorGateway.disconnect()
await self.nodeGateway.disconnect()
break
}
let sleepSeconds = min(8.0, 0.5 * pow(1.7, Double(attempt)))
try? await Task.sleep(nanoseconds: UInt64(sleepSeconds * 1_000_000_000))
}
}
if pausedForPairingApproval {
// Leave the status text + request id intact so onboarding can guide the user.
return
}
await MainActor.run {
self.gatewayStatusText = "Offline"
self.gatewayServerName = nil

View File

@@ -51,13 +51,17 @@ struct OnboardingWizardView: View {
@State private var selectedMode: OnboardingConnectionMode?
@State private var manualHost: String = ""
@State private var manualPort: Int = 18789
@State private var manualPortText: String = "18789"
@State private var manualTLS: Bool = true
@State private var gatewayToken: String = ""
@State private var gatewayPassword: String = ""
@State private var connectMessage: String?
@State private var statusLine: String = "Scan the QR code from your gateway to connect."
@State private var connectingGatewayID: String?
@State private var issue: GatewayConnectionIssue = .none
@State private var didMarkCompleted = false
@State private var didAutoPresentQR = false
@State private var pairingRequestId: String?
@State private var discoveryRestartTask: Task<Void, Never>?
@State private var showQRScanner: Bool = false
@State private var scannerError: String?
@@ -131,6 +135,7 @@ struct OnboardingWizardView: View {
}
}
}
.gatewayTrustPromptAlert()
.alert("QR Scanner Unavailable", isPresented: Binding(
get: { self.scannerError != nil },
set: { if !$0 { self.scannerError = nil } }
@@ -147,6 +152,7 @@ struct OnboardingWizardView: View {
},
onError: { error in
self.showQRScanner = false
self.statusLine = "Scanner error: \(error)"
self.scannerError = error
},
onDismiss: {
@@ -203,6 +209,24 @@ struct OnboardingWizardView: View {
.onChange(of: self.discoveryDomain) { _, _ in
self.scheduleDiscoveryRestart()
}
.onChange(of: self.manualPortText) { _, newValue in
let digits = newValue.filter(\.isNumber)
if digits != newValue {
self.manualPortText = digits
return
}
guard let parsed = Int(digits), parsed > 0 else {
self.manualPort = 0
return
}
self.manualPort = min(parsed, 65535)
}
.onChange(of: self.manualPort) { _, newValue in
let normalized = newValue > 0 ? String(newValue) : ""
if self.manualPortText != normalized {
self.manualPortText = normalized
}
}
.onChange(of: self.gatewayToken) { _, newValue in
self.saveGatewayCredentials(token: newValue, password: self.gatewayPassword)
}
@@ -211,21 +235,36 @@ struct OnboardingWizardView: View {
}
.onChange(of: self.appModel.gatewayStatusText) { _, newValue in
let next = GatewayConnectionIssue.detect(from: newValue)
self.issue = next
if self.step == .connect && (next.needsAuthToken || next.needsPairing) {
// Avoid "flip-flopping" the UI by clearing pairing info when the underlying connection
// transitions through intermediate statuses (e.g. Offline after we pause reconnects).
if self.issue.needsPairing, next.needsPairing {
// Keep the requestId sticky even if the status line omits it after we pause.
let mergedRequestId = next.requestId ?? self.issue.requestId ?? self.pairingRequestId
self.issue = .pairingRequired(requestId: mergedRequestId)
} else if self.issue.needsPairing, !next.needsPairing {
// Ignore non-pairing statuses until the user explicitly retries/scans again, or we connect.
} else {
self.issue = next
}
if let requestId = next.requestId, !requestId.isEmpty {
self.pairingRequestId = requestId
}
if next.needsAuthToken || next.needsPairing {
self.step = .auth
}
if !newValue.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty {
self.connectMessage = newValue
self.statusLine = newValue
}
}
.onChange(of: self.appModel.gatewayServerName) { _, newValue in
guard newValue != nil else { return }
self.step = .success
self.statusLine = "Connected."
if !self.didMarkCompleted, let selectedMode {
OnboardingStateStore.markCompleted(mode: selectedMode)
self.didMarkCompleted = true
}
self.onClose()
}
}
@@ -253,6 +292,7 @@ struct OnboardingWizardView: View {
VStack(spacing: 12) {
Button {
self.statusLine = "Opening QR scanner…"
self.showQRScanner = true
} label: {
Label("Scan QR Code", systemImage: "qrcode")
@@ -270,6 +310,13 @@ struct OnboardingWizardView: View {
.buttonStyle(.bordered)
.controlSize(.large)
}
.padding(.bottom, 12)
Text(self.statusLine)
.font(.footnote)
.foregroundStyle(.secondary)
.multilineTextAlignment(.center)
.padding(.horizontal, 24)
.padding(.horizontal, 24)
.padding(.bottom, 48)
}
@@ -331,6 +378,7 @@ struct OnboardingWizardView: View {
LabeledContent("Mode", value: selectedMode.title)
LabeledContent("Discovery", value: self.gatewayController.discoveryStatusText)
LabeledContent("Status", value: self.appModel.gatewayStatusText)
LabeledContent("Progress", value: self.statusLine)
} header: {
Text("Status")
} footer: {
@@ -413,7 +461,7 @@ struct OnboardingWizardView: View {
TextField("Host", text: self.$manualHost)
.textInputAutocapitalization(.never)
.autocorrectionDisabled()
TextField("Port", value: self.$manualPort, format: .number)
TextField("Port", text: self.$manualPortText)
.keyboardType(.numberPad)
Toggle("Use TLS", isOn: self.$manualTLS)
@@ -441,13 +489,19 @@ struct OnboardingWizardView: View {
private var authStep: some View {
Group {
Section("Authentication") {
TextField("Gateway Auth Token", text: self.$gatewayToken)
.textInputAutocapitalization(.never)
.autocorrectionDisabled()
SecureField("Gateway Password", text: self.$gatewayPassword)
if self.issue.needsAuthToken {
TextField("Gateway Auth Token", text: self.$gatewayToken)
.textInputAutocapitalization(.never)
.autocorrectionDisabled()
SecureField("Gateway Password", text: self.$gatewayPassword)
Text("Gateway rejected credentials. Scan a fresh QR code or update token/password.")
.font(.footnote)
.foregroundStyle(.secondary)
} else {
Text("Auth token looks valid.")
.font(.footnote)
.foregroundStyle(.secondary)
}
}
@@ -469,7 +523,7 @@ struct OnboardingWizardView: View {
} header: {
Text("Pairing Approval")
} footer: {
Text("Run these commands on your gateway host to approve this device.")
Text("Approve this device on the gateway, then tap \"Resume After Approval\" below.")
}
}
@@ -485,6 +539,20 @@ struct OnboardingWizardView: View {
}
}
.disabled(self.connectingGatewayID != nil)
Button {
self.resumeAfterPairingApproval()
} label: {
Label("Resume After Approval", systemImage: "arrow.clockwise")
}
.disabled(self.connectingGatewayID != nil || !self.issue.needsPairing)
Button {
self.openQRScannerFromOnboarding()
} label: {
Label("Scan QR Code Again", systemImage: "qrcode.viewfinder")
}
.disabled(self.connectingGatewayID != nil)
}
}
}
@@ -535,7 +603,7 @@ struct OnboardingWizardView: View {
TextField("Host", text: self.$manualHost)
.textInputAutocapitalization(.never)
.autocorrectionDisabled()
TextField("Port", value: self.$manualPort, format: .number)
TextField("Port", text: self.$manualPortText)
.keyboardType(.numberPad)
Toggle("Use TLS", isOn: self.$manualTLS)
TextField("Discovery Domain (optional)", text: self.$discoveryDomain)
@@ -569,15 +637,35 @@ struct OnboardingWizardView: View {
if let password = link.password {
self.gatewayPassword = password
}
self.saveGatewayCredentials(token: self.gatewayToken, password: self.gatewayPassword)
self.showQRScanner = false
self.connectMessage = "Connecting via QR code…"
self.step = .connect
self.statusLine = "QR loaded. Connecting to \(link.host):\(link.port)"
if self.selectedMode == nil {
self.selectedMode = link.tls ? .remoteDomain : .homeNetwork
}
Task { await self.connectManual() }
}
private func openQRScannerFromOnboarding() {
// Stop active reconnect loops before scanning new credentials.
self.appModel.disconnectGateway()
self.connectingGatewayID = nil
self.connectMessage = nil
self.issue = .none
self.pairingRequestId = nil
self.statusLine = "Opening QR scanner…"
self.showQRScanner = true
}
private func resumeAfterPairingApproval() {
// We intentionally stop reconnect churn while unpaired to avoid generating multiple pending requests.
self.appModel.gatewayAutoReconnectEnabled = true
self.connectMessage = "Retrying after approval…"
self.statusLine = "Retrying after approval…"
Task { await self.retryLastAttempt() }
}
private func detectQRCode(from data: Data) -> String? {
guard let ciImage = CIImage(data: data) else { return nil }
let detector = CIDetector(
@@ -622,6 +710,7 @@ struct OnboardingWizardView: View {
self.manualTLS = true
}
}
self.manualPortText = self.manualPort > 0 ? String(self.manualPort) : ""
if self.selectedMode == nil {
self.selectedMode = OnboardingStateStore.lastMode()
}
@@ -635,6 +724,15 @@ struct OnboardingWizardView: View {
self.gatewayToken = GatewaySettingsStore.loadGatewayToken(instanceId: trimmedInstanceId) ?? ""
self.gatewayPassword = GatewaySettingsStore.loadGatewayPassword(instanceId: trimmedInstanceId) ?? ""
}
let hasSavedGateway = GatewaySettingsStore.loadLastGatewayConnection() != nil
let hasToken = !self.gatewayToken.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty
let hasPassword = !self.gatewayPassword.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty
if !self.didAutoPresentQR, !hasSavedGateway, !hasToken, !hasPassword {
self.didAutoPresentQR = true
self.statusLine = "No saved pairing found. Scan QR code to connect."
self.showQRScanner = true
}
}
private func scheduleDiscoveryRestart() {
@@ -658,6 +756,7 @@ struct OnboardingWizardView: View {
private func connectDiscoveredGateway(_ gateway: GatewayDiscoveryModel.DiscoveredGateway) async {
self.connectingGatewayID = gateway.id
self.connectMessage = "Connecting to \(gateway.name)"
self.statusLine = "Connecting to \(gateway.name)"
defer { self.connectingGatewayID = nil }
await self.gatewayController.connect(gateway)
}
@@ -699,6 +798,7 @@ struct OnboardingWizardView: View {
guard !host.isEmpty, self.manualPort > 0, self.manualPort <= 65535 else { return }
self.connectingGatewayID = "manual"
self.connectMessage = "Connecting to \(host)"
self.statusLine = "Connecting to \(host):\(self.manualPort)"
defer { self.connectingGatewayID = nil }
await self.gatewayController.connectManual(host: host, port: self.manualPort, useTLS: self.manualTLS)
}
@@ -706,6 +806,7 @@ struct OnboardingWizardView: View {
private func retryLastAttempt() async {
self.connectingGatewayID = "retry"
self.connectMessage = "Retrying…"
self.statusLine = "Retrying last connection…"
defer { self.connectingGatewayID = nil }
await self.gatewayController.connectLastKnown()
}

View File

@@ -86,10 +86,10 @@ struct RootCanvas: View {
.environment(self.gatewayController)
}
.onAppear { self.updateIdleTimer() }
.onAppear { self.evaluateOnboardingPresentation(force: false) }
.onAppear { self.maybeAutoOpenSettings() }
.onChange(of: self.preventSleep) { _, _ in self.updateIdleTimer() }
.onChange(of: self.scenePhase) { _, _ in self.updateIdleTimer() }
.onAppear { self.evaluateOnboardingPresentation(force: false) }
.onAppear { self.maybeShowQuickSetup() }
.onChange(of: self.gatewayController.gateways.count) { _, _ in self.maybeShowQuickSetup() }
.onAppear { self.updateCanvasDebugStatus() }
@@ -196,6 +196,7 @@ struct RootCanvas: View {
private func maybeAutoOpenSettings() {
guard !self.didAutoOpenSettings else { return }
guard !self.showOnboarding else { return }
guard self.shouldAutoOpenSettings() else { return }
self.didAutoOpenSettings = true
self.presentedSheet = .settings

View File

@@ -106,4 +106,22 @@ import Testing
@Test func parseGatewaySetupCodeRejectsInvalidInput() {
#expect(GatewayConnectDeepLink.fromSetupCode("not-a-valid-setup-code") == nil)
}
@Test func parseGatewaySetupCodeDefaultsTo443ForWssWithoutPort() {
let payload = #"{"url":"wss://gateway.example.com","token":"tok"}"#
let encoded = Data(payload.utf8)
.base64EncodedString()
.replacingOccurrences(of: "+", with: "-")
.replacingOccurrences(of: "/", with: "_")
.replacingOccurrences(of: "=", with: "")
let link = GatewayConnectDeepLink.fromSetupCode(encoded)
#expect(link == .init(
host: "gateway.example.com",
port: 443,
tls: true,
token: "tok",
password: nil))
}
}

View File

@@ -101,16 +101,9 @@ private func withUserDefaults<T>(_ updates: [String: Any?], _ body: () throws ->
host: "gateway.example.com",
port: 443,
useTLS: true,
stableID: "manual:gateway.example.com:443")
stableID: "manual|gateway.example.com|443")
let loaded = GatewaySettingsStore.loadLastGatewayConnection()
guard case let .manual(host, port, useTLS, stableID) = loaded else {
Issue.record("Expected manual last-gateway connection")
return
}
#expect(host == "gateway.example.com")
#expect(port == 443)
#expect(useTLS == true)
#expect(stableID == "manual:gateway.example.com:443")
#expect(loaded == .manual(host: "gateway.example.com", port: 443, useTLS: true, stableID: "manual|gateway.example.com|443"))
}
}
@@ -120,7 +113,7 @@ private func withUserDefaults<T>(_ updates: [String: Any?], _ body: () throws ->
"gateway.last.host": "",
"gateway.last.port": 0,
"gateway.last.tls": false,
"gateway.last.stableID": "manual:bad:0",
"gateway.last.stableID": "manual|invalid|0",
]) {
let loaded = GatewaySettingsStore.loadLastGatewayConnection()
#expect(loaded == nil)

View File

@@ -34,9 +34,9 @@ public struct GatewayConnectDeepLink: Codable, Sendable, Equatable {
let hostname = parsed.host, !hostname.isEmpty
else { return nil }
let scheme = parsed.scheme ?? "ws"
let scheme = (parsed.scheme ?? "ws").lowercased()
let tls = scheme == "wss"
let port = parsed.port ?? 18789
let port = parsed.port ?? (tls ? 443 : 18789)
let token = json["token"] as? String
let password = json["password"] as? String
return GatewayConnectDeepLink(host: hostname, port: port, tls: tls, token: token, password: password)

View File

@@ -133,10 +133,16 @@ public actor GatewayChannelActor {
private var lastAuthSource: GatewayAuthSource = .none
private let decoder = JSONDecoder()
private let encoder = JSONEncoder()
private let connectTimeoutSeconds: Double = 6
private let connectChallengeTimeoutSeconds: Double = 3.0
// Remote gateways (tailscale/wan) can take a bit longer to deliver the connect.challenge event,
// and we must include the nonce once the gateway requires v2 signing.
private let connectTimeoutSeconds: Double = 12
private let connectChallengeTimeoutSeconds: Double = 6.0
// Some networks will silently drop idle TCP/TLS flows around ~30s. The gateway tick is server->client,
// but NATs/proxies often require outbound traffic to keep the connection alive.
private let keepaliveIntervalSeconds: Double = 15.0
private var watchdogTask: Task<Void, Never>?
private var tickTask: Task<Void, Never>?
private var keepaliveTask: Task<Void, Never>?
private let defaultRequestTimeoutMs: Double = 15000
private let pushHandler: (@Sendable (GatewayPush) async -> Void)?
private let connectOptions: GatewayConnectOptions?
@@ -175,6 +181,9 @@ public actor GatewayChannelActor {
self.tickTask?.cancel()
self.tickTask = nil
self.keepaliveTask?.cancel()
self.keepaliveTask = nil
self.task?.cancel(with: .goingAway, reason: nil)
self.task = nil
@@ -257,6 +266,7 @@ public actor GatewayChannelActor {
self.connected = true
self.backoffMs = 500
self.lastSeq = nil
self.startKeepalive()
let waiters = self.connectWaiters
self.connectWaiters.removeAll()
@@ -265,6 +275,29 @@ public actor GatewayChannelActor {
}
}
private func startKeepalive() {
self.keepaliveTask?.cancel()
self.keepaliveTask = Task { [weak self] in
guard let self else { return }
await self.keepaliveLoop()
}
}
private func keepaliveLoop() async {
while self.shouldReconnect {
try? await Task.sleep(nanoseconds: UInt64(self.keepaliveIntervalSeconds * 1_000_000_000))
guard self.shouldReconnect else { return }
guard self.connected else { continue }
// Best-effort outbound message to keep intermediate NAT/proxy state alive.
// We intentionally ignore the response.
do {
try await self.send(method: "health", params: nil)
} catch {
// Avoid spamming logs; the reconnect paths will surface meaningful errors.
}
}
}
private func sendConnect() async throws {
let platform = InstanceIdentity.platformString
let primaryLocale = Locale.preferredLanguages.first ?? Locale.current.identifier
@@ -458,6 +491,8 @@ public actor GatewayChannelActor {
let wrapped = self.wrap(err, context: "gateway receive")
self.logger.error("gateway ws receive failed \(wrapped.localizedDescription, privacy: .public)")
self.connected = false
self.keepaliveTask?.cancel()
self.keepaliveTask = nil
await self.disconnectHandler?("receive failed: \(wrapped.localizedDescription)")
await self.failPending(wrapped)
await self.scheduleReconnect()