Re: When SSH standards noncompliance is a "feature"

Mouse <[email protected]> Sun, 13 Jun 2021 09:03:30 -0400 (EDT)
Newsgroups gmane.ietf.secsh
Message-ID <[email protected]>
> What it's sending isn't just a truncated SSH ID but the truncated ID
> followed by a string of space characters, which aren't immediately
> visible unless you're looking at a hex dump.

Does it have the second dash?  As in, SSH-2.0-      ?  If so, I think
it's technically conformant; it's just a zero-length softwareversion
string and a comment containing a bunch of spaces.  If not, of course,
it's broken.

> I don't want to publicly name and shame but if someone wants the
> server FQDN for testing against, let me know.  In brief, if your
> client can connect to this server then it's not implementing SSH
> correctly.

Is there a requirement that a client refuse to accept a nonconformant
banner from the server?  I don't recall any.  I would treat that as an
error condition whose recovery is unspecified, meaning that pretty much
any behaviour on the part of the client is allowed.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		[email protected]
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B