Re: How to handle incorrectly-encoded public keys

[Redirected back to the list due to some addressing confusion]

Simon Tatham <[email protected]> writes:

>It's especially surprising since it's such an outdated value of e, so you'd
>most likely expect to see it coming from an SSH implementation that was
>outdated in other respects as well, such as not having even noticed rsa-
>sha2-256! The combination of e=3 with an up-to-date RSA variant (even if done
>wrong) is more surprising than either on its own.

It's a really strange mix, alongside the above it disconnects unless you send
the old-format GEX message (so SSH_MSG_KEX_DH_GEX_REQUEST_OLD instead of
SSH_MSG_KEX_DH_GEX_REQUEST) but then it also implements the -etm modes and
other stuff.  I'm waiting to hear back on whether any widely-use SSH
implementation will actually connect to it.

Peter.