Google uses AI technology to translate content into your preferred language. AI translations can contain errors.

功能與 API

Android 17 為開發人員推出了強大的新功能和 API。以下各節會簡要說明這些功能，協助您開始使用相關 API。

如需新增、修改及移除 API 的詳細清單，請參閱 API 差異比較表。如要進一步瞭解新的 API，請參閱 Android API 參考資料 - 新的 API 會醒目顯示，以利於查看。

此外，也請查看平台變更可能對應用程式造成的影響。詳情請參閱下列頁面：

核心功能

Android 17 新增了下列與 Android 核心功能相關的功能。

新的 ProfilingManager 觸發條件

Android 17 adds several new system triggers to ProfilingManager to help you collect in-depth data to debug performance issues.

The new triggers are:

TRIGGER_TYPE_COLD_START: Trigger occurs during app cold start. It provides both a call stack sample and a system trace in the response.
TRIGGER_TYPE_OOM: Trigger occurs when an app throws an OutOfMemoryError and provides a Java Heap Dump in response.
TRIGGER_TYPE_KILL_EXCESSIVE_CPU_USAGE: Trigger occurs when an app is killed due to abnormal and excessive CPU usage and provides a call stack sample in response.
TRIGGER_TYPE_ANOMALY: Detect system performance anomalies such as excessive binder calls and excessive memory usage.

To understand how to set up the system trigger, see the documentation on trigger-based profiling and how to retrieve and analyze profiling data documentation.

Profiling trigger for app anomalies

Android 17 introduces an on-device anomaly detection service that monitors for resource-intensive behaviors and potential compatibility regressions. Integrated with ProfilingManager, this service allows your app to receive profiling artifacts triggered by specific system-detected events.

Use the TRIGGER_TYPE_ANOMALY trigger to detect system performance issues such as excessive binder calls and excessive memory usage. When an app breaches OS-defined memory limits, the anomaly trigger allows developers to receive app-specific heap dumps to help identify and fix memory issues. Additionally, for excessive binder spam, the anomaly trigger provides a stack sampling profile on binder transactions.

This API callback occurs prior to any system imposed enforcements. For example, it can help developers collect debug data before the app is terminated by the system for exceeding memory limits.

val profilingManager =
    applicationContext.getSystemService(ProfilingManager::class.java)
val triggers = ArrayList<ProfilingTrigger>()
triggers.add(ProfilingTrigger.Builder(ProfilingTrigger.TRIGGER_TYPE_ANOMALY))
val mainExecutor: Executor = Executors.newSingleThreadExecutor()
val resultCallback = Consumer<ProfilingResult> { profilingResult ->
    if (profilingResult.errorCode != ProfilingResult.ERROR_NONE) {
        // upload profile result to server for further analysis
        setupProfileUploadWorker(profilingResult.resultFilePath)
    }
    profilingManager.registerForAllProfilingResults(mainExecutor,
                                                    resultCallback)
    profilingManager.addProfilingTriggers(triggers)
}

JobDebugInfo API

Android 17 introduces new JobDebugInfo APIs to help developers debug their JobScheduler jobs--why they aren't running, how long they ran for, and other aggregated information.

The first method of the expanded JobDebugInfo APIs is getPendingJobReasonStats(), which returns a map of reasons why the job was in a pending execution state and their respective cumulative pending durations. This method joins the getPendingJobReasonsHistory() and getPendingJobReasons() methods to give you insight into why a scheduled job is not running as expected, but simplifies information retrieval by making both duration and job reason available in a single method.

For example, for a specified jobId, the method might return PENDING_JOB_REASON_CONSTRAINT_CHARGING and a duration of 60000 ms, indicating the job was pending for 60000ms due to the charging constraint not being satisfied.

支援允許閒置時的鬧鐘，減少喚醒鎖定

Android 17 introduces a new variant of AlarmManager.setExactAndAllowWhileIdle that accepts an OnAlarmListener instead of a PendingIntent. This new callback-based mechanism is ideal for apps that currently rely on continuous wakelocks to perform periodic tasks, such as messaging apps maintaining socket connections.

隱私權

Android 17 包含下列新功能，可提升使用者隱私。

支援 Encrypted Client Hello (ECH) 的平台

Android 17 introduces platform support for Encrypted Client Hello (ECH), a significant privacy enhancement for network communications. ECH is a TLS 1.3 extension that encrypts the Server Name Indication (SNI) during the initial TLS handshake. This encryption helps protect user privacy by making it more difficult for network intermediaries to identify the specific domain an app is connecting to.

The platform now includes the necessary APIs for networking libraries to implement ECH. This includes new capabilities in DnsResolver to query for HTTPS DNS records containing ECH configurations, and new methods in Conscrypt's SSLEngines and SSLSockets to enable ECH by passing in these configurations when connecting to a domain. Developers can configure ECH preferences, such as enabling it opportunistically or mandating its use, through the new <domainEncryption> element within the Network Security Configuration file, applicable globally or on a per-domain basis.

Popular networking libraries such as HttpEngine, WebView, and OkHttp are expected to integrate these platform APIs in future updates, making it easier for apps to adopt ECH and enhance user privacy.

For more information, see the Encrypted Client Hello documentation.

Android 聯絡人選擇工具

Android 聯絡人選擇工具是標準化的可瀏覽介面，使用者可透過這個工具與應用程式分享聯絡人。這項工具適用於搭載 Android 17 (API 級別 37) 以上版本的裝置，可做為廣泛 READ_CONTACTS 權限的替代方案，提供隱私權保護。應用程式不會要求存取使用者的完整通訊錄，而是指定需要的資料欄位 (例如電話號碼或電子郵件地址)，並由使用者選取要分享的特定聯絡人。這項功能只會授予應用程式所選資料的讀取權限，確保您能精細控管資料，同時提供一致的使用者體驗，包括內建搜尋、切換設定檔和多選功能，不必建構或維護使用者介面。

詳情請參閱聯絡人挑選器說明文件。

安全性

Android 17 新增了下列功能，可提升裝置和應用程式安全性。

Android 進階保護模式 (AAPM)

Android 進階保護模式為 Android 使用者提供一系列強大的全新安全防護功能，在保護使用者 (尤其是高風險使用者) 免於遭受複雜攻擊方面，邁出重要的一步。AAPM 是一項可選擇啟用的功能，只要設定一次即可啟用，使用者隨時都能開啟這項功能，套用一組預設的安全防護措施。

這些核心設定包括禁止從不明來源安裝應用程式 (側載)、限制 USB 資料訊號，以及強制執行 Google Play 安全防護掃描，大幅縮減裝置的攻擊面。開發人員可以透過 AdvancedProtectionManager API 整合這項功能，偵測模式狀態，讓應用程式在使用者啟用時自動採用強化安全措施，或限制高風險功能。

PQC APK 簽署

Android now supports a hybrid APK signature scheme to future-proof your app's signing identity against the potential threat of attacks that make use of quantum computing. This feature introduces a new APK Signature Scheme, which lets you pair a classical signing key (such as RSA or EC) with a new post-quantum cryptography (PQC) algorithm (ML-DSA).

This hybrid approach ensures your app remains secure against future quantum attacks while maintaining full backward compatibility with older Android versions and devices that rely on classical signature verification.

Impact on developers

Apps using Play App Signing: If you use Play App Signing, you can wait for Google Play to give you the option to upgrade a hybrid signature using a PQC key generated by Google Play, ensuring your app is protected without requiring manual key management.
Apps using self-managed keys: Developers who manage their own signing keys can utilize updated Android build tools (like apksigner) to rotate to a hybrid identity, combining a PQC key with a new classical key. (You must create a new classical key, you cannot reuse the older one.)

連線能力

Android 17 新增下列功能，可提升裝置和應用程式的連線能力。

受限的衛星網路

實作最佳化功能，讓應用程式在低頻寬的衛星網路上也能有效運作。

使用者體驗和系統 UI

Android 17 包含下列異動項目，可提升使用者體驗。

專屬的 Google 助理音量串流

Android 17 introduces a dedicated Assistant volume stream for Assistant apps, for playback with USAGE_ASSISTANT. This change decouples Assistant audio from the standard media stream, providing users with isolated control over both volumes. This enables scenarios such as muting media playback while maintaining audibility for Assistant responses, and the other way around.

Assistant apps with access to the new MODE_ASSISTANT_CONVERSATION audio mode can further improve the volume control consistency. Assistant apps can use this mode to provide a hint to the system about an active Assistant session, ensuring the Assistant stream can be controlled outside of the active USAGE_ASSISTANT playback or with connected Bluetooth peripherals.

Handoff

Handoff is a new feature and API coming to Android 17 that app developers can integrate with to provide cross-device continuity for their users. It allows the user to start an app activity on one Android device and transition it to another Android device. Handoff runs in the background of a user's device and surfaces available activities from the user's other nearby devices through various entry points, like the launcher and taskbar, on the receiving device.

Apps can designate Handoff to launch the same native Android app, if it is installed and available on the receiving device. In this app-to-app flow, the user is deep-linked to the designated activity. Alternatively, app-to-web Handoff can be offered as a fallback option or directly implemented with URL Handoff.

Handoff support is implemented on a per-activity basis. To enable Handoff, call the setHandoffEnabled() method for the activity. Additional data may need to be passed along with the handoff so the recreated activity on the receiving device can restore appropriate state. Implement the onHandoffActivityDataRequested() callback to return a HandoffActivityData object which contains details that specify how Handoff should handle and recreate the activity on the receiving device.

即時更新 - 語意色彩 API

With Android 17, Live Update launches the Semantic Coloring APIs to support colors with universal meaning.

The following classes support semantic coloring:

Coloring

Green: Associated with safety. This color should be used for the case where it lets people know you are in the safe situation.
Orange: For designating caution and marking physical hazards. This color should be used in the situation where users need to pay attention to set better protection setting.
Red: Generally indicates danger, stop. It should be presented for the case where need people's attention urgently.
Blue: Neutral color for content that is informational and should stand out from other content.

The following example shows how to apply semantic styles to text in a notification:

  val ssb = SpannableStringBuilder()
        .append("Colors: ")
        .append("NONE", Notification.createSemanticStyleAnnotation(SEMANTIC_STYLE_UNSPECIFIED), 0)
        .append(", ")
        .append("INFO", Notification.createSemanticStyleAnnotation(SEMANTIC_STYLE_INFO), 0)
        .append(", ")
        .append("SAFE", Notification.createSemanticStyleAnnotation(SEMANTIC_STYLE_SAFE), 0)
        .append(", ")
        .append("CAUTION", Notification.createSemanticStyleAnnotation(SEMANTIC_STYLE_CAUTION), 0)
        .append(", ")
        .append("DANGER", Notification.createSemanticStyleAnnotation(SEMANTIC_STYLE_DANGER), 0)

    Notification.Builder(context, channelId)
          .setSmallIcon(R.drawable.ic_icon)
          .setContentTitle("Hello World!")
          .setContentText(ssb)
          .setOngoing(true)
              .setRequestPromotedOngoing(true)

Android 17 適用的 UWB 下行鏈路 TDoA API

Downlink Time Difference of Arrival (DL-TDoA) ranging lets a device determine its position relative to multiple anchors by measuring the relative arrival times of signals.

The following snippet demonstrates how to initialize the Ranging Manager, verify device capabilities, and start a DL-TDoA session:

Kotlin

class RangingApp {

    fun initDlTdoa(context: Context) {
        // Initialize the Ranging Manager
        val rangingManager = context.getSystemService(RangingManager::class.java)

        // Register for device capabilities
        val capabilitiesCallback = object : RangingManager.RangingCapabilitiesCallback {
            override fun onRangingCapabilities(capabilities: RangingCapabilities) {
                // Make sure Dl-TDoA is supported before starting the session
                if (capabilities.uwbCapabilities != null && capabilities.uwbCapabilities!!.isDlTdoaSupported) {
                    startDlTDoASession(context)
                }
            }
        }
        rangingManager.registerCapabilitiesCallback(Executors.newSingleThreadExecutor(), capabilitiesCallback)
    }

    fun startDlTDoASession(context: Context) {

        // Initialize the Ranging Manager
        val rangingManager = context.getSystemService(RangingManager::class.java)

        // Create session and configure parameters
        val executor = Executors.newSingleThreadExecutor()
        val rangingSession = rangingManager.createRangingSession(executor, RangingSessionCallback())
        val rangingRoundIndexes = byteArrayOf(0)
        val config: ByteArray = byteArrayOf() // OOB config data
        val params = DlTdoaRangingParams.createFromFiraConfigPacket(config, rangingRoundIndexes)

        val rangingDevice = RangingDevice.Builder().build()
        val rawTagDevice = RawRangingDevice.Builder()
            .setRangingDevice(rangingDevice)
            .setDlTdoaRangingParams(params)
            .build()

        val dtTagConfig = RawDtTagRangingConfig.Builder(rawTagDevice).build()

        val preference = RangingPreference.Builder(DEVICE_ROLE_DT_TAG, dtTagConfig)
            .setSessionConfig(SessionConfig.Builder().build())
            .build()

        // Start the ranging session
        rangingSession.start(preference)
    }
}

private class RangingSessionCallback : RangingSession.Callback {
    override fun onDlTdoaResults(peer: RangingDevice, measurement: DlTdoaMeasurement) {
        // Process measurement results here
    }
}

Java

public class RangingApp {

    public void initDlTdoa(Context context) {

        // Initialize the Ranging Manager
        RangingManager rangingManager = context.getSystemService(RangingManager.class);

        // Register for device capabilities
        RangingManager.CapabilitiesCallback capabilitiesCallback = new RangingManager.RangingCapabilitiesCallback() {
            @Override
            public void onRangingCapabilities(RangingCapabilities capabilities) {
                // Make sure Dl-TDoA is supported before starting the session
                if (capabilities.getUwbCapabilities() != null && capabilities.getUwbCapabilities().isDlTdoaSupported()) {
                    startDlTDoASession(context);
                }
            }
        };
        rangingManager.registerCapabilitiesCallback(Executors.newSingleThreadExecutor(), capabilitiesCallback);
    }

    public void startDlTDoASession(Context context) {
        RangingManager rangingManager = context.getSystemService(RangingManager.class);

        // Create session and configure parameters
        Executor executor = Executors.newSingleThreadExecutor();
        RangingSession rangingSession = rangingManager.createRangingSession(executor, new RangingSessionCallback());
        byte[] rangingRoundIndexes = new byte[] {0};
        byte[] config = new byte[0]; // OOB config data
        DlTdoaRangingParams params = DlTdoaRangingParams.createFromFiraConfigPacket(config, rangingRoundIndexes);

        RangingDevice rangingDevice = new RangingDevice.Builder().build();
        RawRangingDevice rawTagDevice = new RawRangingDevice.Builder()
                .setRangingDevice(rangingDevice)
                .setDlTdoaRangingParams(params)
                .build();

        RawDtTagRangingConfig dtTagConfig = new RawDtTagRangingConfig.Builder(rawTagDevice).build();

        RangingPreference preference = new RangingPreference.Builder(DEVICE_ROLE_DT_TAG, dtTagConfig)
                .setSessionConfig(new SessionConfig.Builder().build())
                .build();

        // Start the ranging session
        rangingSession.start(preference);
    }

    private static class RangingSessionCallback implements RangingSession.Callback {

        @Override
        public void onDlTdoaResults(RangingDevice peer, DlTdoaMeasurement measurement) {
            // Process measurement results here
        }
    }
}

Out-of-Band (OOB) Configurations

The following snippet provides an example of DL-TDoA OOB configuration data for Wi-Fi and BLE:

Java

// Wifi Configuration
byte[] wifiConfig = {
    (byte) 0xDD, (byte) 0x2D, (byte) 0x5A, (byte) 0x18, (byte) 0xFF, // Header
    (byte) 0x5F, (byte) 0x19, // FiRa Sub-Element
    (byte) 0x02, (byte) 0x00, // Profile ID
    (byte) 0x06, (byte) 0x02, (byte) 0x20, (byte) 0x08, // MAC Address
    (byte) 0x14, (byte) 0x01, (byte) 0x0C, // Preamble Index
    (byte) 0x27, (byte) 0x02, (byte) 0x08, (byte) 0x07, // Vendor ID
    (byte) 0x28, (byte) 0x06, (byte) 0xCA, (byte) 0xC8, (byte) 0xA6, (byte) 0xF7, (byte) 0x6F, (byte) 0x08, // Static STS IV
    (byte) 0x08, (byte) 0x02, (byte) 0x60, (byte) 0x09, // Slot Duration
    (byte) 0x1B, (byte) 0x01, (byte) 0x0A, // Slots per RR
    (byte) 0x09, (byte) 0x04, (byte) 0xE8, (byte) 0x03, (byte) 0x00, (byte) 0x00, // Duration
    (byte) 0x9F, (byte) 0x04, (byte) 0x67, (byte) 0x45, (byte) 0x23, (byte) 0x01  // Session ID
};

// BLE Configuration
byte[] bleConfig = {
    (byte) 0x2D, (byte) 0x16, (byte) 0xF4, (byte) 0xFF, // Header
    (byte) 0x5F, (byte) 0x19, // FiRa Sub-Element
    (byte) 0x02, (byte) 0x00, // Profile ID
    (byte) 0x06, (byte) 0x02, (byte) 0x20, (byte) 0x08, // MAC Address
    (byte) 0x14, (byte) 0x01, (byte) 0x0C, // Preamble Index
    (byte) 0x27, (byte) 0x02, (byte) 0x08, (byte) 0x07, // Vendor ID
    (byte) 0x28, (byte) 0x06, (byte) 0xCA, (byte) 0xC8, (byte) 0xA6, (byte) 0xF7, (byte) 0x6F, (byte) 0x08, // Static STS IV
    (byte) 0x08, (byte) 0x02, (byte) 0x60, (byte) 0x09, // Slot Duration
    (byte) 0x1B, (byte) 0x01, (byte) 0x0A, // Slots per RR
    (byte) 0x09, (byte) 0x04, (byte) 0xE8, (byte) 0x03, (byte) 0x00, (byte) 0x00, // Duration
    (byte) 0x9F, (byte) 0x04, (byte) 0x67, (byte) 0x45, (byte) 0x23, (byte) 0x01  // Session ID
};

If you can't use an OOB configuration because it is missing, or if you need to change default values that aren't in the OOB config, you can build parameters with DlTdoaRangingParams.Builder as shown in the following snippet. You can use these parameters in place of DlTdoaRangingParams.createFromFiraConfigPacket():

Kotlin

val dlTdoaParams = DlTdoaRangingParams.Builder(1)
    .setComplexChannel(UwbComplexChannel.Builder()
            .setChannel(9).setPreambleIndex(10).build())
    .setDeviceAddress(deviceAddress)
    .setSessionKeyInfo(byteArrayOf(0x01, 0x02, 0x03, 0x04))
    .setRangingIntervalMillis(240)
    .setSlotDuration(UwbRangingParams.DURATION_2_MS)
    .setSlotsPerRangingRound(20)
    .setRangingRoundIndexes(byteArrayOf(0x01, 0x05))
    .build()

Java

DlTdoaRangingParams dlTdoaParams = new DlTdoaRangingParams.Builder(1)
    .setComplexChannel(new UwbComplexChannel.Builder()
            .setChannel(9).setPreambleIndex(10).build())
    .setDeviceAddress(deviceAddress)
    .setSessionKeyInfo(new byte[]{0x01, 0x02, 0x03, 0x04})
    .setRangingIntervalMillis(240)
    .setSlotDuration(UwbRangingParams.DURATION_2_MS)
    .setSlotsPerRangingRound(20)
    .setRangingRoundIndexes(new byte[]{0x01, 0x05})
    .build();

功能與 API 透過集合功能整理內容 你可以依據偏好儲存及分類內容。

核心功能

新的 ProfilingManager 觸發條件

Profiling trigger for app anomalies

JobDebugInfo API

支援允許閒置時的鬧鐘，減少喚醒鎖定

隱私權

支援 Encrypted Client Hello (ECH) 的平台

Android 聯絡人選擇工具

安全性

Android 進階保護模式 (AAPM)

PQC APK 簽署

Impact on developers

連線能力

受限的衛星網路

使用者體驗和系統 UI

專屬的 Google 助理音量串流

Handoff

即時更新 - 語意色彩 API

Coloring

Android 17 適用的 UWB 下行鏈路 TDoA API

Kotlin

Java

Out-of-Band (OOB) Configurations

Java

Kotlin

Java

功能與 API