Official Baidu App Tech Account
WKWebView parses HTML by streaming bytes from the network process to the rendering process, decoding them into characters, tokenizing into HTML tokens, building a DOM tree through node creation and insertion, and finally laying out and painting the document using a doubly‑linked in‑memory structure.
As mobile CPUs plateau, developers increasingly use OpenCL to harness Android GPUs like Qualcomm Adreno and Huawei Mali for heterogeneous computing, leveraging its platform, execution, and memory models to write portable kernels—illustrated by a simple array‑addition example that demonstrates device initialization, kernel creation, buffer management, and parallel execution.
Baidu's T7 browser kernel introduces a ResourceScheduler tuning mechanism that lets front‑end developers set image loading priority via an importance attribute or CSS property, enabling high‑priority above‑fold images to load faster while low‑priority resources are throttled under heavy domain request competition.
This article explains how WKWebView leverages various WebKit cache types—including disk, memory, and page caches—detailing their structures, sizing strategies, eviction policies, and the exact source‑code logic that governs HTTP caching, cleanup, and validation on iOS.
This article analyzes the trade‑offs between software and hardware video decoding on Android, presents performance measurements for different MediaCodec modes, identifies compatibility and latency challenges, and proposes a decoder‑monitoring system and seamless switch architecture to maximize playback efficiency.
San CLI 4.0 introduces a re‑architected command system, decoupled configuration, Webpack 5 support, TypeScript and ESM build options, and a series of performance tweaks that together improve flexibility, reduce bundle size, and speed up both development and production builds.
Paddle.js OCR SDK brings Baidu’s lightweight PaddleOCR models to web browsers, offering init() and recognize() APIs that load the ch_PP-OCRv2 detection (DB) and recognition (CRNN with bidirectional LSTM) models in parallel, achieving 258 ms detection, 60 ms recognition, 0.52 F‑score, and a combined size under 12 MB.
The article details how to create the AI‑driven “Object Hunt Battle” game by processing data, designing and training a PP‑LCNet model with PaddleClas, converting it for Paddle.js, and integrating real‑time WebGL inference on mobile devices, achieving sub‑50 ms latency and encouraging developers to explore further.
The article explains front‑end state management using San’s san‑store, detailing its Flux‑style architecture, action registration, component connection, immutable updates with san‑update, and a time‑travel debugging system that logs state changes, restores snapshots, computes efficient diffs, and discusses its benefits and limitations.
This article examines Baidu's Tekes platform—its evolution from automated component publishing to a full‑stack mobile DevOps service—detailing its architecture, workflow engine, event‑driven pipeline, and how it boosts development efficiency across hundreds of repositories and product lines.
