Meow Anti-Xray 1.3.0 更新日志

这是面向 Minecraft 26.2 的兼容与验证版本，重点是升级双 loader 依赖、校准官方版本信息，并补齐 26.2 下的真实压测结果。

兼容性

• 支持 Minecraft 26.2，Java 语言级别保持 25。
• Fabric 侧升级到 Fabric Loader 0.19.3 与 Fabric API 0.153.0+26.2。
• NeoForge 侧升级到 26.2.0.7-beta。
• Fabric metadata 现在限制为 >=26.2 <26.3；NeoForge Minecraft 范围为 [26.2,26.3)。

压测工具

• tools/mc-2612-load-runner.js 新增 --mc-version 26.2 协议 profile，默认使用 Minecraft 26.2 / protocol 776。
• 保留 26.1.2 / protocol 775 profile，便于回归旧版本。
• 修正 26.2 configuration 阶段 custom payload 处理：不再把服务端插件消息原样回发，避免服务端按新的 serverbound payload codec 解码失败。

压测结果

本轮使用默认 async-queue-size=16，8 个网络假玩家通过 RCON spectator 传送持续制造区块发送压力：

Loader	时长	区块包	吞吐	客户端错误	rewrite 平均耗时	sync fallback
Fabric	92s	32,893	357.49 chunks/s	0	2.839ms	42.0%
NeoForge	92s	35,593	386.86 chunks/s	0	2.719ms	39.0%

pressure=saturated 在这种高压跑图下是预期信号，表示默认队列背压正在限制异步任务堆积，优先保护内存稳定。

Paper 对齐

• 已复查 PaperMC/Paper main，远端 HEAD 仍为 76d2ac758cb3abe75aceefa88207443768f585c6，与项目既有 Paper anti-xray parity 基线一致。
• Paper anti-xray 默认配置、hidden/replacement 逻辑与权限节点未发现新的上游变化，本轮不需要额外调整核心反矿透行为。

验证

• node --check tools\mc-2612-load-runner.js 通过。
• Fabric 与 NeoForge 26.2 dev server 均启动到 Done，/antixray status 确认 version=1.3.0、minecraft=26.2。
• `.\gradlew.bat clean test :neoforge:test releaseAllLoaders --c 通过。
• 已检查 build/release/meowantixray-fabric-1.3.0.jar 与 build/release/meowantixray-neoforge-1.3.0.jar 的内嵌 metadata，版本号与 26.2 依赖范围正确。
• 完整压测报告见 reports/1.3.0-minecraft-26.2-pressure-report.md。

---

Meow Anti-Xray 1.3.0 Changelog

This release updates Meow Anti-Xray for Minecraft 26.2, refreshes the loader dependencies from official sources, and adds new 26.2 stress-test results.

Compatibility

• Supports Minecraft 26.2; Java remains 25.
• Fabric now targets Fabric Loader 0.19.3 and Fabric API 0.153.0+26.2.
• NeoForge now targets 26.2.0.7-beta.
• Fabric metadata now requires >=26.2 <26.3; NeoForge's Minecraft range is [26.2,26.3).

Load Runner

• tools/mc-2612-load-runner.js now supports --mc-version 26.2, using Minecraft 26.2 / protocol 776 by default.
• The previous 26.1.2 / protocol 775 profile is still available for regression checks.
• Fixed configuration-stage custom payload handling for 26.2 by no longer echoing server plugin messages back as serverbound payloads.

Stress Test Results

These runs used the default async-queue-size=16 with 8 fake network players driven by RCON spectator teleports:

Loader	Duration	Chunk packets	Throughput	Client errors	Rewrite avg	Sync fallback
Fabric	92s	32,893	357.49 chunks/s	0	2.839ms	42.0%
NeoForge	92s	35,593	386.86 chunks/s	0	2.719ms	39.0%

Seeing pressure=saturated during this synthetic exploration load is expected: the bounded async queue is applying backpressure to protect memory stability.

Paper Parity

• Rechecked PaperMC/Paper main; the remote HEAD is still 76d2ac758cb3abe75aceefa88207443768f585c6, matching this project's existing Paper anti-xray parity baseline.
• No new upstream Paper anti-xray defaults, hidden/replacement behavior, or permission-node changes were found, so no core anti-xray behavior change was needed for Paper parity.

Verification

• node --check tools\mc-2612-load-runner.js passed.
• Both Fabric and NeoForge 26.2 dev servers reached Done; /antixray status confirmed version=1.3.0 and minecraft=26.2.
• `.\gradlew.bat clean test :neoforge:test releaseAllLoaders --c passed.
• Checked embedded metadata in build/release/meowantixray-fabric-1.3.0.jar and build/release/meowantixray-neoforge-1.3.0.jar; both contain the expected version and 26.2 dependency ranges.
• Full stress-test report: reports/1.3.0-minecraft-26.2-pressure-report.md.