At the 2019 GMTC Beijing Global Front‑End Conference, Beike engineers shared practical approaches to extreme front‑end performance optimization, Node.js service stability, comprehensive mobile crash and network monitoring, and the challenges and solutions of integrating Flutter into their existing mobile apps.
This article examines Xianyu's adoption of Flutter, detailing its benefits for small and Android‑focused teams, the technical hurdles of integrating Flutter into existing apps, memory‑optimized engine sharing, video rendering improvements, the evolution of the Fish Redux framework, and the automation pipeline that drives efficient, data‑backed development.
This article provides a comprehensive step‑by‑step guide for integrating Flutter into iOS projects, covering required configuration files, script usage, AppDelegate modifications, MethodChannel communication, common pitfalls, debugging tools, performance monitoring, and routing strategies for both static and dynamic navigation.
AspectD is a compile‑time Aspect‑Oriented Programming framework for Dart that traverses Flutter’s dill intermediate files to inject, call, or execute advice code, allowing non‑intrusive instrumentation, logging, and UI event injection without modifying source, and integrates into the build pipeline via a toolchain patch.
To replace the native‑only tracking used by Xianyu after its migration to Flutter, the team built a high‑accuracy solution that mirrors the Flutter navigation stack with an index list, correctly fires enter/leave events on pushes, pops and filtered dialogs, and adds exposure detection based on 50 % visibility for 500 ms, ultimately delivering 100 % tracking accuracy in production.
The UI2CODE Page Framework evolves a unified, MVP‑based architecture—comprising Pages, Cards, and Reactions—that leverages AI‑driven code generation and Redux messaging to enable high‑cohesion, low‑coupling components, dramatically cut development time, and plans further AI integration for fully automatic page creation.
The article describes a custom Flutter photo album component that uses a MethodChannel to bridge native modules for permission, caching, decoding, and OpenGL texture conversion, implements MVC architecture, employs lazy GridView pagination, GPU‑based resource recycling, scaled bitmap preview, and PageView preloading to achieve flexible UI, multi‑selection, and low memory usage, while remaining extensible for future features.
Fish_redux introduces a hierarchical connector that links a child Component to a filtered slice of its parent’s state, using get and set operations to extract and merge sub‑state, enabling efficient, low‑coupling data flow similar to React‑Redux’s connect but within a strict component tree.
Xianyu’s team presents FlutterBoost, a reusable, non‑intrusive hybrid framework that lets large native Android/iOS apps progressively adopt Flutter by sharing a single engine, mapping native containers to Flutter widgets, supporting complex navigation patterns, and optimizing performance for billions of users.
This article explains Xianyu's comprehensive Flutter performance monitoring system—including frame‑rate tracking, page‑load timing, and exception rate metrics—detailing the underlying principles, implementation steps, and real‑world results that help maintain a smooth user experience at massive scale.
This article explains Flutter's multithreading model, comparing isolates and event queues to iOS GCD, detailing async/await, Future, coroutine concepts, and the roles of Platform, UI, GPU, and IO runners, with code examples illustrating isolate communication and asynchronous network requests.
The article examines Flutter's rise as a leading cross‑platform mobile framework, compares it with other solutions like Hybrid, React Native, and Qt, explores its ecosystem, and highlights key breakthroughs such as Code‑Push and Hummingbird that could redefine the mobile development landscape.
Fish‑Redux separates UI from business logic by using a dispatch function that routes actions from a view’s handler to the component’s effect, then to the page store’s reducer, passes through any applied middleware, updates the state, and finally notifies listeners to re‑render the UI.
The article describes how Xianyu’s FXTexImage library (versions 1‑3) transforms product‑detail images in Flutter e‑commerce apps into native‑provided external textures, progressively reducing memory growth, eliminating duplicate downloads, and cutting CPU load via GPU‑thread cleanup and shared‑context texture reuse, boosting page push limits from ten to over thirty, lowering crash rates, and shrinking the app package by more than 900 KB.
Flutter is Google's cross‑platform UI toolkit that uses the Dart language to build native‑compiled applications for iOS and Android, featuring its own rendering engine, widget hierarchy, hot reload, AOT/JIT compilation, networking libraries, state management, and best‑practice development guidelines.
The UI2CODE pipeline converts image‑derived JSON DSL layouts into Flutter code by constructing a hierarchical DSL tree, partitioning it into rows, columns or stacks, applying unsupervised text‑alignment and padding heuristics, generating templated widget code with separate data models, cutting development effort by roughly two‑thirds and now used internally at Xianyu.
To meet Xianyu’s rapid UI change demands and curb its swelling Android package, the team built a Flutter‑specific dynamic template system that compiles Dart widget code into JSON, streams it to the client, reconstructs the widget tree at runtime, and maintains 55‑60 fps performance while planning further widget support and caching.
This article examines the challenges of cross‑platform mobile development, compares WebView and React Native with Flutter, explains Flutter’s layered architecture, JIT/AOT compilation, hot‑reload, integration strategies for iOS and Android, performance measurements, and practical lessons from Mafengwo’s merchant app implementation.
FlutterBoost is a hybrid integration framework that lets large mobile apps progressively embed Flutter using a shared engine and native containers, providing unified page lifecycle, channel‑based messaging, multi‑navigator support, and production‑ready performance, now open‑source after successful deployment in Alibaba’s Xianyu client.
This article introduces Fish Redux, an open‑source Flutter framework from Alibaba that combines Redux‑style state management with a modular component architecture, explains its layered design, performance benefits, and how it simplifies building complex cross‑platform mobile apps.
UI2CODE, an AI‑powered tool from Xianyu’s tech team, automatically transforms UI screenshots into editable Flutter code by extracting visual elements with machine‑vision, classifying them via deep‑learning, generating a domain‑specific language through a recursive neural network, and mapping this DSL onto syntax‑tree templates to achieve pixel‑level precision, full accuracy, and maintainable output.
To meet Xianyu’s rapid‑release targets, the team built a CI‑integrated automation suite that gathers Flutter FPS, CPU, memory and network data via extended Android tools, drives mixed Flutter/Native UI tests with Appium and integration‑test, and generates lane‑by‑lane performance reports that quickly flag regressions across dozens of feature branches.
The article presents a full system for capturing and replaying Flutter UI gestures, detailing how raw pointer data is processed, how gesture recognizers compete, how recordings wrap callbacks and generate synthetic touch packets, and how a timer‑driven replay reproduces user interactions for debugging.
The article examines Flutter’s limited built-in JSON handling, explains why the popular json_serializable package can fall short for complex, generic objects, and introduces fish-serializable—a custom library with a ValueHolder, FishSerializable interface, JsonSerializer, and IntelliJ plugin—already used in production at Alibaba’s Xianyu and soon to be open-sourced.
This article examines the drawbacks of maintaining separate native codebases for iOS and Android, reviews the evolution of cross‑platform solutions such as WebView‑based frameworks, React Native, and Flutter, and shares practical insights, code examples, and future plans for using Flutter in large‑scale mobile projects.
Fish Redux is a lightweight, configuration‑driven Flutter framework that extends Redux principles with a three‑layer View‑Effect‑Reducer architecture, automatic reducer composition, component‑level state decomposition, and high‑performance adapters, enabling centralized, observable data management, easier maintenance, team collaboration, and scalability for medium‑to‑large applications.
This article explains how Flutter renders animations by repeatedly calling setState, describes the roles of Animatable, Animation, Tween, and AnimationController, and provides a complete example of creating a custom BezierTween to achieve a bezier‑curve motion with full source code.
Xianyu’s 2018 technical breakthroughs combine Flutter/Dart for tri‑platform development, a TensorFlow‑driven UI2Code system that converts screenshots into runnable Flutter code, and the SWAK framework that isolates business logic from middleware, boosting productivity, simplifying legacy code, and showcasing playful engineering anecdotes.
The Xianyu tech team’s 2018 Flutter migration cut exception rates from several per‑thousand to under one by targeting two main stack‑trace groups—setState‑related and BuildContext‑related errors—while emphasizing proper lifecycle checks such as verifying mounted before setState and using the correct widget context.
This article introduces JDFlutter, JD's cross-platform mobile development framework built on Flutter, detailing its architecture, integration with existing Android/iOS projects, debugging capabilities, data statistics, exception monitoring, degradation mechanisms, and future outlook as part of JD's dual-engine ARES platform.
This article explains how to replace cumbersome if‑else or manual map‑based routing in Flutter with a lightweight, annotation‑driven solution that automatically generates a page‑to‑URL mapping at compile time, using Dart's source_gen, build, and analyser packages.
The iQIYI Live Broadcast Assistant adopts Flutter to replace native UI layers on Android and iOS, integrating via Flutter fragments or views and AAR packages on Android and through Podfile, Xcode scripts, and FlutterViewController on iOS, achieving near‑native performance and enabling future migration of more screens to a single codebase.
Xianyu tackled Flutter adoption by creating a private CocoaPods CI pipeline, a component‑based fishRedux architecture, and a shared‑GL engine modification that let native middleware run in Flutter, thereby unifying Android, iOS, and Flutter development, improving build speed, and contributing tools back to the community.
This article explains how the Flutter team adapted Flutter's multi‑layer architecture to run on the Web, describing the three prototype approaches, the rendering, layout and painting strategies, CSS Paint API experiments, embedding considerations, portability limits, and the current implementation status.
This article describes a component‑based dynamic rendering framework for Flutter that parses Dart source via the analyzer’s AST, maps widgets to a DataModel, supports atomic to dynamic components, delivers over 50 fps performance, and includes template versioning, download, rollback, with plans for automation, i18n, and personalized content.
This article provides a step‑by‑step tutorial for beginners who want to learn Flutter, covering prerequisites, a quick Dart introduction, installing and running a Hello World app, experimenting with core widgets, UI challenges, personal tips, and a curated list of further learning resources.
At the inaugural Android Green Alliance Conference in Beijing, Xianyu engineer Ji Feng presented a deep‑dive on using Flutter to build efficient cross‑platform applications, covering its unified programming model, performance gains, solutions to memory, font, caching, ArmV7, video, CI, hybrid stack challenges, and introducing a novel Redux‑based Half‑Dumb Component architecture.
JDFlutter is a new cross-platform development framework introduced by the mobile development team at JD.com, designed to integrate with existing Android/iOS apps and provide a high-performance alternative to React Native.
JDFlutter is JD.com’s next‑generation cross‑platform framework that integrates Flutter into existing Android/iOS projects, offering a rich UI component library, native API bridges, debugging tools, data statistics, and a roadmap for future multi‑engine development alongside JDReact.
The article explains how to embed native Android views in Flutter using AndroidView, detailing the three-step setup, size and touch‑event handling, and a Gaode map example, while warning of performance costs and recommending native embedding only when a pure Flutter alternative is unavailable.
This article introduces Flutter as a Google‑created cross‑platform app framework, explains why developers choose it for rapid development, beautiful UI, high performance, native integration and single‑codebase deployment, outlines its widget hierarchy and layout system, and provides step‑by‑step Windows environment setup instructions.
The Xianyu team tackled iOS Flutter app bloat by analyzing App.framework, cutting excessive type‑casting, applying dwarf‑stack‑trace and obfuscation flags, stripping dSYM symbols, and removing duplicate assets, which together shrank the generated framework by over 30 % and improved download efficiency.
At QCon2018 Shanghai, Alibaba’s Xianyu team unveiled a three‑platform development model that uses Flutter for UI and Dart for both client and a lightweight “glue‑layer” server, cutting product‑page implementation time from eight months to 1.5 months while delivering near‑native performance and Java‑comparable backend metrics, and promising unified language, faster iteration, and future AI‑driven UI generation.
Flutter’s sub‑second hot‑reload works by scanning changed Dart files, generating incremental kernel (.dill) files, sending them to the running Dart VM via RPC, triggering the VM’s reload routine, and finally invoking the framework’s reassemble sequence to rebuild the widget tree, all enabled by JIT compilation during development.
The article explains how excessive memory use in a Flutter app, caused by ImageCache and related image handling, was diagnosed with Xcode Instruments and Dart Observatory, then mitigated through refined caching policies, adaptive screenshot resolution, improved FlutterViewController lifecycle management, and targeted garbage‑collection techniques, resulting in lower memory peaks and greater stability.
At Google Developer Day 2018 in Shanghai, the Xianyu team highlighted Flutter’s upcoming 1.0 release, hybrid hot‑reload challenges, native view embedding via platform_views.dart, and performance optimizations such as lighter layout calculations and Skia tweaks, noting rapid adoption by major Chinese firms and plans to open‑source their tools.
The Xianyu team rewrote a product detail page in Flutter, benchmarked it against native and React Native implementations on iPhone 5c, iPhone 6s, Xiaomi 2s and Samsung S8, and found Flutter consistently matched or outperformed native on iOS and surpassed React Native on both platforms in FPS, CPU usage and memory stability, earning it a gold‑medal rating.
Flutter, a cross‑platform framework built on immutable, declarative widgets, lets beginners create UI using basic widgets like Text and Image, layout containers such as Row, Column, Stack, and Flex, manage visibility, and handle state through StatefulWidget lifecycle methods while observing app‑level lifecycle events.
The article explains Flutter’s platform channels—BasicMessageChannel, MethodChannel, and EventChannel—detailing their components (name, messenger, codec), how messages are encoded, decoded, and routed via BinaryMessenger, the various codecs and handlers, and considerations for thread safety, large data transfer, and practical usage.
Flutter reduces CPU/GPU overhead in multi‑video calls by separating drawing from presentation, using a unified LayerTree and external textures that share OpenGL contexts between Flutter and native code, allowing direct Skia rendering of native pixel buffers on iOS and Android with significantly lower latency and memory use.
The article explains how to manage a hybrid Flutter‑Native stack by using a single shared Flutter engine instance linked to native view controllers, unified openUrl routing, native‑driven page transitions, resource‑efficient reuse of native components, and detailed guidance on creating and publishing Flutter plugins and packages.
This article explains Flutter's layered architecture, its high‑performance UI rendering pipeline, Dart language optimizations, typical development patterns, plugin and package management, common pitfalls such as lack of reflection and JSON handling, and provides practical code examples for mobile developers.
Flutter, Google’s open‑source cross‑platform UI framework built on Dart and Skia, provides native‑level performance through a layered widget‑rendering architecture, supports rapid development via hot‑reload (with known limits), offers extensive plugins, can be integrated into Android apps as an AAR, and delivers comparable frame‑rate performance to native code with a modest APK size increase.
The article details how to reconstruct Xianyu’s product detail page using Flutter within existing native iOS/Android apps, covering project structure, environment management, build optimizations for Android and iOS, native‑launch debugging with hot‑reload, and CI integration for reliable hybrid releases.
The article explains how to extract Flutter engine, compiled Dart assets, and custom plugins from a mixed Flutter‑Native project, package them into an AAR (Android) and static libraries (iOS), publish them to an internal Maven repository, and let native code consume these artifacts via CI‑automated versioned dependencies, eliminating the need for a full Flutter SDK.
This article walks through Flutter’s Image.network implementation, explaining its constructors, the ImageState lifecycle, how resolveImage creates an ImageStream, the caching strategy in ImageProvider, the NetworkImage download flow, and multi‑frame handling for animated images.
This article provides a comprehensive analysis of Flutter's Image widget, especially Image.network, detailing its constructors, underlying classes, state management, image loading, caching mechanisms, and source code flow, while illustrating how images are fetched, decoded, and rendered in Flutter applications.
This article explains Flutter’s cross‑platform engine architecture, detailing its four Task Runners, the Dart isolate model, platform‑specific implementations, custom thread configurations, and common pitfalls, helping developers understand and optimise performance for mobile apps.
The article explains how Flutter adopts a React‑style reactive architecture—using immutable widgets, InheritedWidget for communication, Dart Streams and a Redux‑like store with reducers and interrupters—to achieve clear model‑view separation, fine‑grained UI updates, componentized multi‑store widgets, and shares Xianyu’s practical implementation details and code examples.
Xianyu’s team refactored their large iOS and Android codebases by extracting Flutter binaries, plugins, and assets into a remote repository, enabling native projects to compile, build, and debug independently in both Standalone and Flutter modes while still supporting full Flutter features such as hot‑reload and Dart builds.
This article provides a comprehensive overview of Flutter, Google's cross‑platform mobile UI toolkit, covering its purpose, development history, key advantages, sample Dart code for state management and platform channel usage, and resources for quickly getting started with Flutter development.
Xianyu became the first Chinese app to adopt Flutter after its 2018 preview, choosing it for performance and a single codebase, and used it to solve hybrid integration, page‑stack, video and component issues while overcoming Armv7, font, ICU and memory challenges, achieving comparable or better performance, modest APK growth, and a stable production page, though dynamic updates remain limited, prompting continued collaboration with Google.
The article explains how to develop Flutter plugins using platform channels—MethodChannel for invoking native APIs and EventChannel for streaming events—detailing Android and iOS implementation steps, registration, dependency setup, common pitfalls, and advanced considerations such as JSON‑serializable data and texture handling.
The article explains Flutter’s engine thread model—four embedder‑provided task runners for platform, UI, GPU, and I/O—and how Dart isolates run single‑threaded with message‑based communication, highlighting performance benefits over JavaScript bridges, common threading pitfalls, and strategies for managing engine instances across multiple pages.
At Google I/O 2018, Xianyu’s team showcased their large‑scale adoption of Flutter—highlighting a hybrid architecture, AI‑driven UI‑to‑code generation workflow, collaboration with the Flutter team, and insights from the event’s Flutter party and the CEO’s Duplex AI demo.
This article from the Xianyu tech team explains Flutter’s architecture, covering the creation and management of the widget, element, and render object trees, the three‑stage rendering pipeline (build, layout, paint), performance‑optimizing techniques, state lifecycle, data flow, and practical tips for building efficient cross‑platform mobile applications.
This article highlights essential open-source platforms—Arduino, Home Assistant, Node-RED, Zetta, and Flutter—that empower developers to build robust IoT applications across various verticals, offering ease of use, extensive community support, and versatile integration capabilities.
The guide explains Flutter’s three‑layer architecture (Framework, Engine, Embedder), Dart compilation modes, iOS and Android build pipelines, project layout, and how to customize or rebuild the Engine, manage engine versions, and extend the framework for future native integration.
This article explores Flutter's memory management, detailing the Dart runtime's generational GC, comparing Image widget memory usage across Android versions, and evaluating the impact of reusing or not reusing FlutterView on Java and native memory consumption, with practical test results and optimization tips.
The 2018 Qunar Technology Carnival QClient session on May 19 concluded with thanks to judges, hosts, volunteers, and a series of technical talks covering Flutter, ReactNative performance monitoring, instant‑messaging cloud, anti‑crawling platform, YApi API management, Android custom builds, and QRMaps indoor mapping, followed by lively audience interaction and a group photo.
This article examines the evolution of Android architectural patterns—from MVC, MVP, and MVVM to Flutter‑derived frameworks and Android Architecture Components—explaining their layering, communication mechanisms, and how router‑based decoupling can achieve fully modular mobile applications.
This article reviews Android mobile architecture patterns—including MVC, MVP, MVVM, Flutter-derived framework, and Android Architecture Components—explaining their evolution, layer separation, and communication mechanisms such as object holding, interface contracts, and router-based dispatch, while providing illustrative code examples.