Rob Hague

Closing per https://github.com/sshnet/SSH.NET/issues/562#issuecomment-2168778374 (sorry)

The SSH.NET-NETCONF package is a fork of this one, with (only) the following changes: https://github.com/sshnet/SSH.NET/compare/develop...declspec:SSH.NET:develop That could explain why it works.

I am also unfamiliar with netconf @peter2145 you may try the above change with the latest code here (which has ecdsa) and see if it works

@peter2145 you should download the code, make the changes to NetConfSession.cs and build it. Then you can add a project reference and see if it works 

Thanks for checking. There is actually a similar change in PR #946 which seems to also fix the issue.

> * Consecutive calls to Encrypt/Decrypt **must not** preserve state. Two consecutive calls with the same data input should return the same output. Doesn't that contradict https://github.com/zybexXL/SSH.NET/blob/ada6ccbcc20e1e26463fbe22004556c9f884da57/src/Renci.SshNet/Security/Cryptography/Ciphers/AesCipher.cs#L225-L232 ? It is...

> This is mostly to pass some test cases which are not expecting correct padding functionality. Calling `TransformFinalBlock` instead of `TransformBlock` in that code would give the same output with...

> I believe an SSH server using AES-CBC will add correct PKCS7 padding Where have you read this? I don't think it is true.

Sure, thanks very much. You can use an email of mine in the commit history

Seems sensible