ssh-keyscan for ssh2 (was Re: openssh-2.9p1)

[mouring at etoh.eviladmin.org]

Better version of conrecycle()...

int
conrecycle(int s)
{
        int ret;
        con *c = &fdcon[s];

        ret = conalloc(c->c_namelist, c->c_output_name, c->c_keytype);
        confree(s);
        return (ret);
}


conalloc() xstrdup() the first two arguments.  So it's not required to do
it again.

- Ben

On Fri, 18 May 2001 mouring at etoh.eviladmin.org wrote:

>
>
> On Fri, 18 May 2001, Wayne Davison wrote:
>
> > On Fri, 18 May 2001 mouring at etoh.eviladmin.org wrote:
> > > > +	int keytype = c->c_keytype;
> > > 	^^^ Is this really needed any more?  It's used once.
> >
> > I though it was needed because I thought confree() destroyed the
> > structure "c" was pointing at, but I was apparently wrong about that.
> > So, your change would simplify things a tad.
> >
>
> Your right.. The the code is current right.
>
> Which means that confree(s) is wrong.  It does not currently does not set
> fdcon[c].keytype to -1.
>
> > > Don't know if we want to move it to -t {rsa1,rsa,dsa} instead of -1 -R
> > > -D.  Would that change -t [timeout] to -T [timeout] ?
> >
> > It would make things a little more uniform at a small incompatibility
> > cost (changing -t).  How would we specify multiple queries?  Multiple
> > -t options, or a comma-separated value?  E.g.
> >
> >    ssh-keyscan -t rsa1,dsa foo.com
> >
> > I think that might be a nice way to go.
> >
>
> I think that is what Markus was refering to.
>
> - Ben
>
>