Port to rustls v0.20 #9
Conversation
|
Hmm, nope, that's not it. There are probably a few things I overlooked, let me give this a once-over. |
For this failure, I think you probably need to do something like tokio-rs/tls@c2d1fe6. |
|
Is this PR still alive? |
|
Yes, I was just side-tracked. If it's important to you I can probably get this done pretty quickly. |
|
I'm not really familiar with SSL, so I'm not entirely sure why the tests are failing here. I don't see where an accidental EOF condition could arise. |
|
Not urgent no, thanks for your efforts! I also believe to have found a bug. Without looking into it deeply, my initial guess is that the handshake logic is reading more bytes off the socket than just the handshake. I can create a repo with my noddy server if this helps? This may be related to the EOF error you're seeing above? Let me know if you'd like some help. |
|
Alright, I don't really know what the actual problem is. In this case, I just re-forked from tokio-rustls, and the tests are passing now. |
|
Just to add for the most recent revision, |
|
@smol-rs/admins Any blockers to this being merged? |
| Poll::Ready(result) => result, | ||
| Poll::Pending => Err(io::ErrorKind::WouldBlock.into()), | ||
| } | ||
| self.poll_with(|io, cx| io.poll_flush(cx)) | ||
| } | ||
| } |
There was a problem hiding this comment.
Probably a good fit for the io part of futures_lite
|
Note: this PR should be squash-merged. |
Ports the codebase to the new
rustls0.20 version.cargo testappears to be failing because of expired certificates.Resolves #8