CertainTLS

The test is able to pass in my debugging session, this makes me think the problem is with deadlock between client-server wait.

Here is a little bit more debugging info. If I step through debugging mode, the test passes, and below is the debugging info ``` Server up and running time="2022-05-31T18:08:03Z" level=info...

> The new code still has problems. You are only accepting the first stream of each connection, so it won't work when the same client dials again. And needless to...

> way I can think of to make this work, is to have an internal goroutine for accepting connections. Then for each accepted connection, start a goroutine for handling the...

I have located the exact line that is causing the flickering in test: https://github.com/HyNetwork/hysteria/blob/master/pkg/core/client.go#L80 If I place a single break point on this line `ctxCancel()`, and manually hit Continue (F5)...

I have some to debug the test a little bit more. The exact breakpoint to make the test succeed isn't L80, but L85 https://github.com/HyNetwork/hysteria/blob/master/pkg/core/client.go#L85 . In other words, if I...

I put breakpoints inside `handleControlStream(qs, stream)`, if I put one on L120 https://github.com/HyNetwork/hysteria/blob/master/pkg/core/client.go#L120, and manually continue, the test can succeed, if I put one on L121, the test would fail.

> Have you tried debugging on the server side instead? Yes, I started the debugging from the server side. > Looks like the server was blocked on `Accept` Right, `s.listener.Accept`...

Just to provide an update: I definitely have not forgot this PR. I switched gear a little bit to improve PT testing (manual and automated) in general in https://github.com/OperatorFoundation/shapeshifter-dispatcher/pull/42 and...

A security audit was conducted by Cure53.de for the PT related work. I copy and paste the relevant parts from the report below: ## Identified Vulnerabilities The following section lists...