Updated partial auth patch against CVS

[Carson Gaspar]

--On Wednesday, May 02, 2001 11:10 AM -0700 Karl M <karlm30 at hotmail.com> 
wrote:

> Hi Carson...
>
> Please help me understand the issues/constraints here. Regarding
>
> Authorder2 publickey,publickey:password
>
> what part of the syntax do you see as a problem? Is it that an
> authentication sequence is a propper subset of a later authentication
> sequence in the list?

Yes. Although "later" is irrelevant. If publickey succeeds, you're in. 
You'll never be asked for a password. You're suggesting changing that 
behaviour on Cygwin, making it platform-specific. That's bad.

> If I use
>
> Authorder2 publickey:method1,publickey:method2
>
> and the public key succeeds but method1 fails, assuming no agent will the
> user see
>
> publickey
> method1
> method2
>
> or will he see
>
> publickey
> method1
> publickey
> method2
>
> I would hope and vote for the first.

Currently, it should show "method1,method2". This makes the client "do the 
right thing" and let you re-try your flubbed password, without making you 
re-enter your private key passphrase.

> What I would hope to specify is: publickey (when ssh and sshd userids
> match)  and publickey:password (when ssh and sshd userids differ). This
> gets the  password to the sshd so that the userid can be changed in
> CygWin. Do you  have any thoughts on the syntax for this that would be
> compatible with your  syntax?

publickey:cygwinoptpass

and add code to handle the cygwin bizzaro world case.

-- 
Carson Gaspar - carson at taltos.org
Queen trapped in a butch body