Tony Arcieri

I just realized this logo isn't an SVG. I'll need to rebuild it as one first, although that shouldn't be too difficult.

@kim splitting out the encoding traits (`Decode`/`Encode` and `Reader`/`Writer`) into their own crate (e.g. `ssh-encoding`) is a good idea. I'll see if I can do that before the next release...

#29 extracted an `ssh-encoding` crate

> actually what would be nice is if we could marshal/unmarshal the wire protocol via an ordinary struct with a derive proc macro It wouldn't be terribly hard to add...

~~I've added `ssh-derive` to the list with the goal of supplying custom derive support for `ssh-encoding`, namely the `Decode`/`Encode` traits~~

@japaric there's a `cfg(poly1305_force_soft)` you can set to disable CPU feature autodetection. Unfortunately these `x86_64-unknown-none` targets are something of a longstanding problem which I thought we had fixed at this...

`cpufeatures` does consult `target_feature`, but only to bypass runtime checks. Its main goal is to autodetect CPU features at runtime when they haven't been explicitly enabled. I'm curious how you're...

Is it possible for you to capture this in a debugger and get a backtrace of some sort? I'm curious where the SIGILL is actually occurring and whether or not...

Aha, we've definitely encountered that "LLVM ERROR: Do not know how to split the result of this operator!" before: https://github.com/RustCrypto/AEADs/issues/405 https://github.com/RustCrypto/AEADs/issues/498 https://github.com/RustCrypto/hashes/issues/446 I thought we had successfully worked around it...

Yeah, something very strange is happening