layue13andClaude Opus 5.5
789648c82f
- proxy/：core、plugin-api、luckperms 及其上游子模块 vendor/luckperms - messaging/、backend/ 保持在根目录（messaging 是代理与后端共用的 API） - build-logic/（原 proxy-build-logic） - testing/：smoke/ 与 benchmarks/，不进入发行物 Gradle 项目名不变（settings.gradle.kts 映射目录），任务路径与 Maven 坐标照旧。 脚本的仓库根定位、文档与记录中的链接同步更新。 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
1.6 KiB
Player command interception screening
Run date: 2026-09-26 (Asia/Shanghai). Local Windows loopback, synthetic Minecraft 1.7.10 PLAY echo, four connections, 4,096-byte bodies, 500 warmup and 5,000 measured exchanges per connection, three repetitions, in-flight window 16. Each command started a fresh benchmark JVM. The direct and proxy phases in each command shared the JVM and fake backend. The only proxy configuration difference was whether the frontend command interceptor was installed; no command packets were sent.
.\benchmarks\run-proxy-session.ps1 -Connections 4 -Messages 5000 -Warmup 500 -Payload 4096 -Repeats 3 -Window 16 -CommandInterceptor 1
.\benchmarks\run-proxy-session.ps1 -Connections 4 -Messages 5000 -Warmup 500 -Payload 4096 -Repeats 3 -Window 16 -CommandInterceptor 0
| Interceptor | Proxy roundtrips/s, rounds 1–3 | Proxy p95 latency, ms | Correct echoes |
|---|---|---|---|
| Enabled | 69,444; 65,450; 92,875 | 1.773; 1.951; 0.853 | 60,000 / 60,000 |
| Disabled | 73,820; 84,262; 96,942 | 1.450; 1.241; 0.768 | 60,000 / 60,000 |
The ranges overlap, though all three enabled rounds were slower in this run. Separate JVM runs and visible throughput variation prevent attributing that difference to the interceptor. This short screening run does not resolve a small overhead or establish target-pack capacity. It verifies ordinary framed traffic remained correct with the interceptor installed under these conditions. Raw output: enabled, disabled.