Skip to content

v3.1.0

Choose a tag to compare

View all tags
@superGekFordJ superGekFordJ released this 30 Jul 18:31
· 124 commits to main since this release
v3.1.0
1ff5f34

在 3.1.0 版本中，我们带来了从底层到 UI 的全面进化。在这个版本里，我们深度同步了上游 Surge 引擎的跨代级事件驱动重构，同时完整保留并优化了 GoAria 独有的自适应调度与界面交互。

在这个版本中，您将体验到：

  • 更精准实时的速率显示：高达 150ms 刷新频率的真实速率视觉反馈。
  • 更聪明的并发过载保护：为同一网站并发下载引入域级排队账本，从源头避免被服务器拉黑。
  • 更具个性的高质感界面：重写了列表 FLIP 动画，并引入全新的 0-100 视觉质量滑块，让您自由调校独属的丝滑渲染体验。

从更省心、更智能的底层线程调度，到更迅捷、更个性化的 UI 交互，3.1.0 致力于为您带来超越以往的纯粹下载体验。

In version 3.1.0, we bring a comprehensive evolution from the core engine to the UI. In this release, we deeply synchronized the generational event-driven refactoring from the upstream Surge engine, while fully retaining and optimizing GoAria's unique adaptive scheduling and interfaces.

In this version, you will experience:

  • More Accurate Real-time Speeds: True speed feedback updated at a blistering 150ms interval.
  • Smarter Overload Protection: A new domain-level capacity ledger for concurrent downloads from the same site, preventing server IP bans at the source.
  • Highly Personalized Premium UI: Rewritten list FLIP animations and a brand-new 0-100 visual quality slider, allowing you to freely tune your unique, silky-smooth rendering experience.

From more effortless and intelligent underlying thread scheduling to swifter and highly personalized UI interactions, 3.1.0 is dedicated to delivering an unprecedented, pure downloading experience.

📖 阅读完整中文更新日志

🎨 渲染引擎革命与性能优化

  • 全新视觉质量滑块：告别了简单的“全开/全关”设置。我们引入了全新的 0-100 视觉质量滑块，并加入了全新的“均衡 (Balanced)”默认档位。
  • 动态性能优化：我们重构了核心渲染管线。由于 Chromium 内核对实时 CSS 辉光呼吸（Glow Breathing）的渲染优化存在历史局限，全效模式下的性能消耗可能偏高。全新的“均衡”档位不仅保留了绝佳的静态玻璃质感与柔和流光，更大幅提升了运行时的流畅度。
  • 无缝无极调节：滑块支持 0-100 平滑过渡。为了将手感打磨到极致，我们引入了分段二次贝塞尔曲线。这项技术让滑块的调节告别了生硬的线性变化，不仅赋予了拖拽时极度直观的顺滑感，更极大地丰富了单个档位内部（从低到高）的视觉跨度。这意味着，即使在同一个“均衡”档位下，您也能调校出从厚重实体到高级通透的多种细腻质感，轻松找到最完美契合您审美的那个点。此外，我们剥离了拖拽时的本地存储写入（Storage I/O），确保实时预览全界面效果时零卡顿！

🖼️ 视觉档位对比

💡 提示：特效不仅仅局限于任务卡片和玻璃框，整个界面的其他元素也会随之变化。推荐在设置页面亲自拖动滑块体验！

✨ 1. 任务边缘辉光

任务卡片的辉光与动态能量条。以下依次展示了在不同主档位下的边缘发光细节与渲染效果：

🍃 精简

⚖️ 均衡 (默认档位)

🌟 全效


🎛️ 2. 无极滑块曲线过渡展示

得益于上述的“分段二次贝塞尔曲线”映射，滑块在每一个大档位区间内拖拽时，面板的玻璃材质都会发生细腻的变化。以下展示了滑块在各区间内从“较低值 (Low)”到“较高值 (High)”的真实演变：

🍃 精简区间 (0 ➔ 30)
快速拉开视觉差距，从厚重实体向毛玻璃过渡

⚖️ 均衡区间 (31 ➔ 70)
手感最为平滑，提供极高宽容度的细腻微调

🌟 全效区间 (71 ➔ 100)
跨度再次放大，向极尽通透与华丽延展

🚀 网络连接与调度优化

  • 1-RTT TLS 会话恢复：现在底层引擎跨传输池共享会话缓存。在空闲连接重建、任务暂停/恢复等场景下，握手延迟被大幅压缩，显著降低了安全连接的建立开销。
  • 轻量级连接预热：在引擎决定增加下载线程时，如果发现没有备用连接，现在会用极低的代价优先将网络连接“捂热”。这个微小的改进能让下载在加速阶段更加顺滑，有效避免了由于冷连接握手导致的短暂停顿或降速，让您更快地达到满速状态。
  • 宏观带宽统计算法重构：彻底重构了底层的可用带宽聚合逻辑。该变更严格隔离了底层调度账本与前端 UI 的平滑展示，修复了由于上游引擎瞬断引发的剩余带宽超发分配问题。
  • 高频实时速率渲染：得益于同步上游引擎的深度重构，现在 UI 界面的实时速度展示更加精准。引擎能够以 150ms 的极高频率向前端持续推送更新，为您带来近乎毫无延迟的真实速率视觉反馈。

🧊 列表动画渲染重构

  • 丝滑的列表 FLIP 动画：得益于底层虚拟滚动组件的升级，我们获得了更大的自定义动画空间，并成功重写、引入了全新的 FLIP 动画引擎。过去在任务状态变动、排序刷新时毫无过渡效果的生硬切换已成为历史。现在，您能直观地看到每一个任务在列表中的顺滑流转与位移！

🚀 同源并发过载保护

  • 域级调度账本：彻底解决了在同一网站（域名）下高速连加多个任务时，因瞬间建立大量连接而触发服务端严苛限速的难题。现在，智能引擎为每个独立域名启用了严格的余量排队账本。后续任务将以极低的连接数平滑起步，并在有带宽余量时再做探索，从源头上避免了“拥堵踩踏”。
  • 同源带宽智能交接：重构了任务完成时的带宽释放与交接逻辑。现在，当同域名的任务下载完成时，其释放的带宽将以“有界倍增”的方式，顺滑、无缝地交接给其他处于等待或慢速状态的任务。

🧠 弱网环境下的自适应优化

  • 精准探底与天花板侦测：针对晚高峰或极差 CDN 节点的环境进行了深度调优。新机制能更敏锐地侦测到“单线程拉起无效果”的速度天花板，防止在死局中盲目增加无用线程，显著降低了系统的无效资源消耗。
  • 探测底线修复：修正了 BBR 自适应引擎在短暂带宽超卖后，由于探测底限计算错误而无法正常缩容的核心缺陷。现在，当网络拥堵时，引擎能毫无阻碍地回调至安全底线并重新结算。

⚙️ 核心机制修复与更新

  • 底层引擎全面换代：全面引入了上游最新的 Surge 引擎事件驱动重构。通过对底层网络调度的跨代级升级，将核心模块进一步解耦，使得引擎在处理海量并发任务时更加稳定、高效。
  • 终端事件总线重构：重写了事件总线（EventBus），精简了引擎空闲时的资源回收机制，进一步降低了待机时的性能消耗。
  • 限速（HTTP 429）退避优化：针对下载遇到限速的情况，优化了底层的智能退避与重试机制，更平滑地对待服务器。
  • 暂停与重试进度指针显示修复：深度同步了上游 Surge 引擎对于特定重试（Backoff）或暂停场景下进度条指针不准确的问题。由于 GoAria 拥有独立的 Verified Progress 机制，实际下载完整性完全不受影响，本次合并主要解决了纯 UI 层面可能出现的指针跳变。
  • 峰值速度与指纹记录修复：修正了长时下载中峰值速度 (PeakSpeed) 采样可能被异常拉高的“撕裂”缺陷，确保网络环境指纹 (PeakEnvKey) 始终与实际会话状态保持严格一致。
  • 限速识别修复：修复了一个会偶尔将“无限速任务”误认为“被服务器限速”的后台缺陷。现在，自适应引擎在面对无限速任务时，能够毫无顾忌地火力全开。
  • 拓展更新与依赖升级：发布官方浏览器拓展包 0.1.1 版本；更新 pnpm 至 11.17.0 并全面同步前端依赖库至最新。

📖 Read Full English Release Notes

🎨 Rendering Engine & Performance

  • Effects Quality Slider: Say goodbye to the simple "on/off" toggle. We've introduced a new 0-100 visual quality slider featuring a brand-new "Balanced" default tier.
  • Dynamic Performance Optimization: We've refactored the core rendering pipeline. Because browser engines (Chromium) have legacy limitations in optimizing real-time CSS glow breathing, the Full tier can be performance-intensive. The new "Balanced" tier retains stunning static glass textures and soft glow while significantly improving runtime smoothness.
  • Seamless Stepless Adjustment: The slider supports a 0-100 smooth transition. To perfect the tactile feel, we introduced Piecewise Quadratic Bezier Curves. This algorithm eliminates rigid linear changes, bringing an intuitively smooth feeling to your adjustments. More importantly, it dramatically expands the visual range from low to high within a single tier. This means that even within the "Balanced" tier, you can fine-tune through a rich variety of textures—from frosted to highly transparent—to find your absolute sweet spot. We also decoupled live previews from storage I/O, ensuring absolutely silky-smooth, stutter-free dragging!

🖼️ Visual Tier Comparison

💡 Tip: Visual effects apply to the entire UI, not just task cards and glass panels. We recommend dragging the slider in settings to experience it yourself!

✨ 1. Task Glow

Glow and dynamic energy bars on task cards. Below shows the edge glow details and rendering effects under the main tiers:

🍃 Reduced

⚖️ Balanced (Default)

🌟 Full


🎛️ 2. Slider Curve Transitions

Thanks to the "Piecewise Quadratic Bezier Curves" mapping, dragging the slider within each tier subtly evolves the glass materials. Below shows the organic evolution from "Low" to "High" values within each range:

🍃 Reduced Range (0 ➔ 30)
Rapid transition from solid to frosted glass

⚖️ Balanced Range (31 ➔ 70)
Smoothest feel, providing high-tolerance fine-tuning

🌟 Full Range (71 ➔ 100)
Expanded span, extending into ultimate transparency and elegance

🚀 Network & Scheduling Optimizations

  • 1-RTT TLS Session Resumption: The engine now shares a global ClientSessionCache across transport pools. During idle connection rebuilds or pause/resume cycles, TLS handshake latency is reduced to 1-RTT, significantly lowering secure connection overhead.
  • Lightweight Prewarm: When the engine decides to scale up download threads, it now intelligently "warms up" new network connections first. This micro-optimization ensures a much smoother acceleration phase, preventing brief stutters or speed drops caused by cold handshakes, allowing your downloads to reach peak speeds even faster.
  • Macro Bandwidth Decoupling: Refactored the available bandwidth aggregation logic. This strict separation between the underlying scheduling ledger and UI smoothing fixes bandwidth overselling issues caused by brief upstream stalls.
  • High-Frequency Real-time Speed: Thanks to deep refactoring synced from the upstream engine, the UI's real-time speed display is now significantly more accurate. The engine continuously pushes updates to the frontend at a blistering 150ms interval, providing near-instantaneous visual feedback on your true download speeds.

🧊 List Animation & Rendering Refactoring

  • Silky List FLIP Animations: Thanks to upgrading vue-virtual-scroller to 3.0+, we unlocked more flexibility for custom animations within the virtual list and successfully introduced a rewritten FLIP animation engine. The previously abrupt and transition-less state changes and sorting updates are now a thing of the past. Now, you can visually enjoy the silky-smooth displacement and flow of every task within the list!

🚀 Domain Overload Protection

  • Domain-Level Ledger: Completely resolved the issue where rapidly adding multiple tasks from the same website (domain) would trigger strict server rate limits due to an instantaneous burst of connections. The smart engine now utilizes a strict domain-level capacity ledger. Subsequent tasks will start smoothly with minimal connections, preventing congestion at the source.
  • Smart Bandwidth Handover: Refactored the bandwidth release and handover logic upon task completion. Now, when a task finishes, its freed bandwidth is smoothly and seamlessly handed over to other waiting or slow tasks on the same domain via a "bounded multiplication" approach.

🧠 Adaptive Optimization for Poor Networks

  • Accurate Ceiling Detection: Deeply tuned for peak hours or extremely poor CDN node environments. The new mechanism keenly detects the speed ceiling where "adding threads yields no effect," preventing blind and useless thread scaling and significantly reducing invalid resource consumption.
  • Probe Floor Recovery: Fixed a core defect in the BBR adaptive engine where it failed to scale down (Probe-Down) properly after temporary bandwidth oversubscription. Now, during network congestion, the engine safely and smoothly pulls back to the baseline for re-evaluation.

⚙️ Core Fixes & Updates

  • Generational Engine Upgrade: Fully introduced the latest upstream Surge engine event-driven refactoring. This generational upgrade to underlying network scheduling further decouples core modules, making the engine more stable and efficient when handling massive concurrent tasks.
  • EventBus Refactoring: Rewrote the terminal EventBus and streamlined resource reclamation mechanisms when the engine is idle, further reducing standby performance overhead.
  • Throttling (HTTP 429) Backoff: Optimized the underlying intelligent backoff and retry mechanism for rate-limited downloads, treating servers more gently.
  • Pause & Retry Pointer Fix: Deeply synchronized fixes from the upstream Surge engine regarding inaccurate progress bar pointers during specific retry (backoff) or pause scenarios. Because GoAria has its own independent Verified Progress mechanism, functional integrity was completely unaffected; this merge resolves purely UI-level pointer glitches.
  • Peak Tearing Fix: Corrected a defect where PeakSpeed sampling could be abnormally inflated during long downloads, ensuring the network fingerprint (PeakEnvKey) accurately matches the actual session state.
  • Throttling Detection Fix: Resolved a background defect that occasionally misidentified "unlimited speed tasks" as being actively rate-limited by the server. Now, the adaptive engine can unleash its full firepower on unlimited tasks without hesitation.
  • Extension & Dependency Updates: Released the official browser extension v0.1.1; bumped pnpm to 11.17.0 and synced all frontend dependencies to their latest versions.

⚠️ 行为变更 / Breaking Changes

  • 同源并发初始单线程分配：当您同时添加多个来自同一网站的任务时，如果第一个任务的测速已接近或达到该网站的限速天花板，那么后续任务在刚开始时将只会分配 1 个连接（线程）。

    • 请不要担心这会拖慢下载！ 后续的单线程任务在下载过程中，智能引擎会持续监测网络。当有其他任务完成，或者引擎检测到当前网速好于预期时，会自动、顺滑地为这些单线程任务动态增加线程数，确保带宽被充分利用。此变更旨在根除以往“一拥而上”导致被服务端彻底拉黑的死局。

  • Single-Thread Initial Allocation for Same-Domain Tasks: When you add multiple tasks from the same website concurrently, if the first task already consumes most of the site's bandwidth limit, subsequent tasks will initially be allocated only 1 connection (thread).

    • Don't worry about this slowing down your downloads! During the download process, the smart engine continuously monitors the network. When other tasks finish or if the engine detects better-than-expected speeds, it will automatically and smoothly scale up the threads for these tasks, ensuring your bandwidth is fully utilized. This change is designed to eradicate the "thundering herd" issue that previously resulted in strict IP bans from servers.

📦 平台支持 / Platform Support

平台 / Platform 格式 / Format
Windows (amd64) .zip (portable exe)
Windows (arm64) .zip (portable exe)
Linux (amd64) .AppImage, .deb, .rpm, .pkg.tar.zst
macOS (apple silicon / intel) .zip (containing .app bundle)

⚠️ 注意 / Notice

  • 状态 / Status: 正式版 (Stable Release). 此版本为无插件纯净版 (generic-no-pack)。
Assets 9
Loading