每日新聞摘要

In this tutorial, Sachin Verma demonstrates using Apache POI's HSSFWorkbook within a Spring Boot project to generate Excel (.xls) files in memory, serialize them to byte arrays, and deserialize them back into workbooks. It shows creating a sample workbook with headers, styles and sample employee rows, converting the HSSFWorkbook to bytes via ByteArrayOutputStream, and rebuilding it from a byte[] using ByteArrayInputStream. The article then integrates these utilities into REST endpoints: a GET /download-excel that streams the workbook as application/vnd.ms-excel for download, and a POST /upload-excel that accepts a MultipartFile to reconstruct and inspect the workbook. The examples emphasize in-memory processing to avoid disk I/O, straightforward error handling, and practical use cases such as storing bytes in databases or serving dynamic reports through Spring MVC.

本文演示在 Spring Boot 项目中如何使用 Apache POI 的 HSSFWorkbook 在内存中生成 Excel (.xls)，将工作簿序列化为字节数组并重建回工作簿。示例创建带有标题与样式的员工表格，演示用 ByteArrayOutputStream 将 HSSFWorkbook 转为 byte[]，并用 ByteArrayInputStream 从 byte[] 还原。随后将这些工具整合为两个 REST 端点：GET /download-excel 即时生成并下载文件，POST /upload-excel 接收 MultipartFile 并解析第一个工作表名称。示例强调免磁盘 I/O 的内存处理、基本错误处理以及适用于存储到数据库或生成动态报表的场景。

This tutorial demonstrates how to use Selenium to interact with an autocomplete-enabled search bar (example: eBay). It shows a JUnit-based test that opens Chrome, finds the search input by XPath, and sends the keyword "iphone" character-by-character to ensure the suggestion dropdown appears. The test then uses WebDriverWait with an explicit wait for visibility of the autocomplete list items, verifies at least two suggestions exist, and clicks the second suggestion. The article also lists required Maven dependencies (selenium-java and webdrivermanager) and includes setup and teardown of the WebDriver. The approach emphasizes mimicking human input and using explicit waits to ensure reliable selection of dynamic suggestion items in end-to-end tests.

本文示范如何用 Selenium 与带自动完成功能的搜索栏交互（以 eBay 为例）。通过 JUnit 管理生命周期，开启 Chrome、以 XPath 定位搜索输入框，并将关键词“iphone”逐字符输入以触发建议（一次性输入可能不会弹出下拉）。随后使用 WebDriverWait 等待建议列表出现，确保至少有两个建议后，选择第二项并点击。文章还列出所需的 Maven 依赖（selenium-java 与 webdrivermanager）以及测试的 setup/teardown，实现通过模拟人工输入和显式等待来保证端到端测试中建议项选择的可靠性。

This short blog post by Manish Singh greets Android developers and introduces intermediate concepts of Kotlin Flow. It outlines deeper aspects of reactive streams in Kotlin, explains patterns and considerations for using Flow in Android apps, and pairs technical explanations with real interview questions to help readers both learn and assess their knowledge. The post aims to bridge basic Flow understanding and more advanced practical usage, highlighting common pitfalls and topics frequently raised in interviews. Alongside conceptual discussion, it offers practical tips on applying Flow reliably within app architectures and on preparing concise, interview-ready answers. Designed as a concise yet focused resource, the article helps developers strengthen their conceptual grasp, improve real-world application of Flow, and boost confidence for technical interviews without assuming extensive prior expertise.

Manish Singh 向 Android 开发者问好，本文介绍 Kotlin Flow 的中级概念，并配合真实面试题目。内容阐述 Flow 作为 Kotlin 中响应式流的进阶方面，讨论使用场景、常见陷阱与设计考量，并提供应用建议与面试准备要点。文章定位介于基础与进阶之间，帮助读者巩固概念、提升在实际项目中的运用能力，并增强面试答题信心，适合希望在 Flow 技术上进一步提升的 Android 开发者阅读。

The article stresses that user perception of app performance directly affects retention: when an app is slow, briefly freezes, drains battery, or crashes unexpectedly, users are unlikely to return. It argues that systematic performance analysis and targeted optimizations are essential to prevent these outcomes. Recommended practices include profiling to identify CPU, memory, and I/O bottlenecks; measuring power consumption and startup latency; using instrumentation and real-user monitoring to capture reproducible issues; and prioritizing fixes that improve perceived responsiveness. The piece also emphasizes iterative testing, automated benchmarks, and balancing resource usage against feature functionality. By integrating performance considerations into development cycles—through profiling, testing on representative devices, and monitoring in production—developers can reduce crashes and battery drain, improve responsiveness, and increase user retention.

文章强调应用性能对用户留存有直接影响：当应用感觉慢、短暂卡顿、耗电或无预警崩溃时，用户通常不会再给第二次机会。为避免这些问题，应进行系统化的性能分析并实施有针对性的优化，包括定位 CPU、内存与 I/O 瓶颈、衡量启动时间与能耗、使用探针和真实用户监控以重现问题，并优先改善可感知的响应性。文章还建议采用迭代测试、自动化基准与生产环境监控，并在具代表性的设备上验证改进效果，以减少崩溃与耗电、提升流畅度并提高用户留存率。

In this short piece by Anang Suwasto, the author states that AI is rewriting the rules of software development and challenges Android developers to remain relevant as code-generation tools become more capable. The article highlights the disruptive potential of AI-assisted coding and implies that developers should adapt by focusing on higher-level skills, workflows, and domains that are less likely to be fully automated. While brief, the post serves as a prompt to explore practical strategies for staying valuable—such as learning new tools, emphasizing architecture and design, improving collaboration and problem-solving, and committing to continuous learning—and invites readers to continue to ProAndroidDev for the full discussion.

作者 Anang Suwasto 指出，AI 正在重写软件开发的规则，随着代码生成工具能力增强，Android 开发者面临如何保持相关性的问题。文章简要概述了 AI 对开发角色与流程的冲击，并提示开发者应转向难以被完全自动化的领域与技能，如架构与设计思维、复杂问题解决、团队协作与持续学习等。该短文旨在提醒并促使开发者採取行动來适应变化，並引導讀者到 ProAndroidDev 深入閱讀完整論述。

This article demonstrates how to build a pragmatic image downloader using batch files that downloads a set of images (for example, 100) by running three concurrent downloads at a time while showing real-time progress. It focuses on batching and concurrency control to balance speed and stability, techniques for reporting per-item and overall progress in real time, and practical measures for handling failures, retries, and logging. The goal is to keep the process responsive and observable so users don’t lose track of long-running downloads. The write-up emphasizes simple, robust strategies—grouping work into manageable concurrent units, tracking progress continuously, and applying retry/recording mechanisms—to make automated bulk downloads efficient and reliable without complex tooling.

本文展示如何用批处理文件构建一个实用的图片下载器：将大量图片下载任务分批执行，并以每次三个并发下载来兼顾速度与稳定性，同时在运行过程中实时呈现进度信息。文章讨论了分组与并发控制、逐项与总体进度回报、以及遇到网络或文件错误时的重试与记录策略。目标是提供简明且稳健的实现思路，使自动化批量下载在面对网络波动或错误时仍保持可观测性与可预测性，从而提高运行效率并减少人工干预。

Beyond Monolith outlines how to decouple Android app navigation by combining deep links and dependency injection in Kotlin with the Jetpack Navigation component. Instead of a single monolithic navigation graph, the article advocates treating navigation as a modular responsibility: expose navigation points through deep links, define small module-level navigators, and inject navigation interfaces so feature modules do not depend on app-level routing details. Deep links provide a stable, discoverable contract for navigation across modules and processes, while DI lets the app wire concrete navigator implementations at runtime, keeping modules testable and replaceable. The result is a scalable navigation architecture that supports independent module development, easier testing, and clearer boundaries. Practical examples and patterns help readers migrate from tightly coupled navigation to a modular approach using Kotlin idioms and Jetpack Navigation tooling.

本文介绍如何使用深度链接与依赖注入（DI），结合 Kotlin 与 Jetpack Navigation Component，将单体导航拆解为模块化的可扩展导航架构。作者建议各功能模块通过公开的深度链接作为导航契约，并通过 DI 注入导航接口与实现，避免模块直接依赖全局路由，从而提高可测试性与可替换性。该策略支持模块独立开发、降低耦合并简化测试，文章提供实务模式与迁移方向，帮助开发者将紧耦合的导航迁移到更易维护与扩展的模块化设计。

This article argues that as foldable and tablet devices gain traction, Android developers should build adaptive layouts with Jetpack Compose so apps deliver a good user experience beyond phones. It highlights the mismatch between rising multi-form-factor hardware adoption and the prevailing phone-focused app designs, and recommends applying Compose-driven responsive strategies and best practices to bridge that gap. Rather than prescribing a single API, the piece emphasizes design and implementation considerations — planning for different screen sizes, continuity across states, and thorough testing — so interfaces scale, reflow, and take advantage of larger or split displays. The goal is practical guidance for making Compose UIs that gracefully adapt to foldables and tablets while maintaining usability and visual consistency.

文章指出，随着可折叠设备与平板日益普及，开发者应使用 Jetpack Compose 构建自适应布局，使应用在手机之外的设备上也能提供良好体验。它强调当前的矛盾：硬件形式多样但应用设计仍以手机为中心，因此需要通过 Compose 的灵活性采用响应式策略与实践来弥补差距。建议从布局与交互设计角度出发，考虑不同屏幕尺寸与状态的连续性，并重视测试与迭代，保证界面在更大屏幕或分屏显示时能正确重排并维持视觉和使用一致性。目标是为 Compose UI 在可折叠与平板设备上提供切实可行的适配思路。

Sixtin's article examines the Storage Access Framework (SAF), long the principal mechanism for file access in Android, and frames 2025 as a turning point for storage APIs. With platform evolution and shifting policy and privacy priorities, the piece outlines why developers should reassess how they work with external files, permissions, and user consent. Rather than prescribing a single replacement, the author surveys the changing landscape, highlights implications for app architecture and maintenance, and urges developers to monitor platform announcements, prepare migration paths, and prefer patterns that minimize permissions and respect user data. The article is a call to adapt: maintain compatibility where needed, but be ready to adopt newer APIs and best practices as Android’s storage model continues to evolve.

作者回顾 Storage Access Framework（SAF）在 Android 存储访问上的长期角色，并指出 2025 年成为存储 API 的转折点。文章说明平台演进以及隐私與权限政策的变动，提醒开发者重新评估应用对外部文件的处理、权限申请与用户授权方式。作者并未提供单一替代方案，而是分析生态变化对应用架构、维护和兼容性的影响，建议持续关注 Android 公告、规划迁移路径，并采用能减少权限暴露且尊重用户数据的实现方式。总体主旨是呼吁开发者调整策略，在保持兼容性的同时准备采纳新的 API 与最佳实践。

As agents grow more capable they become harder to debug and to cost accurately: verbose logs are noisy and provider dashboards show only aggregate usage. This article demonstrates adding observability to a Kotlin coding agent using OpenTelemetry and Langfuse. With four lines of integration in Koog, developers get detailed traces showing messages, tool calls, observations, and per-span costs. Traces make it easy to spot failures — the author found a tool bug where requested line ranges exceeded file length — and to measure run-level cost (a 50-task evaluation cost $66). Session attributes let Langfuse aggregate runs for batch evaluations. The article argues that traces are essential not only for debugging but for understanding agent behavior and deciding where to optimize or delegate work.

文章指出，随着代理增加工具和完成定义，传统的调试日志变得冗杂且难以从单次运行中看清成本细节。作者演示只需几行代码将 Koog 与 OpenTelemetry 和 Langfuse 集成，即可获取按步详尽的 trace（包含提示、响应、工具参数与每段成本），進而在调试时快速定位问题——例如发现读取行范围超出文件长度的错误——并能通过为 trace 添加 session ID 在评估运行中聚合时长与总成本。此可观测性不只便于修复错误，也帮助决定后续优化与委派策略。

Based on the article title and available excerpt, this piece focuses on using LazyColumn and LazyRow in Android to display very large lists quickly and efficiently. It highlights the value of lazy/virtualized rendering—only composing and laying out visible items—to minimize memory footprint and CPU work when handling millions of entries. The write-up likely discusses practical considerations such as item reuse, pagination or incremental loading, and performance tuning in Jetpack Compose, and points readers to the full post on Medium for code samples and deeper guidance. Because only a brief excerpt is available here, the summary emphasizes the article’s apparent intent: to show developers how to leverage Compose’s lazy containers to build responsive, scalable list UIs for large datasets.

依据标题与摘录可知，本文聚焦于在 Android（Jetpack Compose）中使用 LazyColumn 与 LazyRow 来高效显示大量数据。核心在于利用延迟/虚拟化渲染，仅构建当前可见项，进而减少内存占用与布局开销，使得处理数百万条或海量数据时界面仍能保持流畅。文章可能涵盖分页或增量加载、项重用与性能调优等实务议题，并以代码示例说明在真实项目中的应用。由于此处仅有简短片段，建议至 Medium 阅读完整内容以获取详细实现与示例。