OVC2FZV6RYK2RL27UZOCXF2CFJL4K4PR2FY2TTPWQWIV3CW4N4QQC
C4Z5NJZGREVN3J2DLLISXR7U3FJGG2YES5RF4LNJ6SL3OWZBMBIQC
I believe you might want to use pijul key prove fintohaps@ssh.pijul.com
.
Actually, I’d rather move stuff in the other direction, to try and separate the protocol logic (thrussh) from the key logic (thrussh-keys).
The reason is, there are two different skillsets involved, one is cryptography algorithms, the other one is network/asynchronous things/security, so separating these two would be easier to audit in my opinion.
pijul key prove fintohaps@ssh.pijul.com.
I just tried that and nothing seems to happen, the program just hangs. Perhaps I’ve missed a step? Is there documentation somewhere for this setup?
Actually, I’d rather move stuff in the other direction, to try and separate the protocol logic (thrussh) from the key logic (thrussh-keys).
I think I’ve come to a somewhat similar conclusion, but only by assessing the needs in radicle-keystore#17
. Our need there is to only be generic over the agent::client
. I’m going to create a separate discussion for those changes and close this one :)
I just tried that and nothing seems to happen, the program just hangs. Perhaps I’ve missed a step? Is there documentation somewhere for this setup?
Ah incidentally, the update to my pijul
binary also seems to have fixed this
In the same vein as https://nest.pijul.com/pijul/thrussh/discussions/45, I’d like to move
Named
over tokey.rs
to minimise the appearance ofKeyPair
andPublicKey
inthrussh