Please implement VncAuth authentication

[julian-seward1]

Connecting tigervnc's viewer to the server I want to connect to produces this
CConnection: Using RFB protocol version 3.8
CConnection: Choosing security type VncAuth(2)

Is "VncAuth(2)" different from "VncAuth" ? I have no idea. But anyway, that's
what the server appears to request.

[whitequark]

Unfortunately I can't do this because the only Rust implementation of DES is currently broken because of crossbeam-rs/crossbeam#75.

[whitequark]

@julian-seward1 OK, I've split out the binaries from the vnc crate, implemented the plumbing for authentication, and inlined the DES code from the des-rs crate so that it doesn't pull in broken crossbeam. Then I've discovered that something in the challenge-response handshake doesn't work correctly, because Xvnc refuses to let me in.

I've spent several hours digging through libvnc and des-rs sources and I have no idea what's wrong with it. I am sorry but by this point I will not spend any more of my time figuring out a completely braindead authentication scheme using a broken algorithm. If you have any idea how to fix it, the des branch has code that's supposed to work according to the RFB protocol description.

[whitequark]

@julian-seward1 So I've randomly revisited this and, amusingly enough, I am far from the only one having this problem: http://www.catonmat.net/blog/curious-case-of-des-algorithm/. To quote:

Turns out that the VNC authentication reverses the order of bits in every byte of the password.

I am astounded by the depth of depravity, but I'll implement it now for you.

[whitequark]

@julian-seward1 That works! Let me know if you would like any minimal examples to go with your use case, or anything else.

[whitequark]

Also published 0.4.0 on crates.io, with the client and proxy split off.