Re: SSH Tasks

[Dale Anson]

Yes, the ssh task that I wrote does not follow the telnet task 
convention. It is a fairly simple task and does not do anything with the 
response other than log it. I can think of situations where this would 
be useful, and similarity with other tasks is always desirable.  
Getting, parsing, and responding to the server response is doable, but 
not implemented.

Also I would imagine that the scp task should model the copy task, which 
is does not, at this point. None of the fileset/filterset/mapper stuff 
has been done. The posts on this list from "Joe Consumer" look like the 
scp task he's done is further along in implementing these features, and 
I believe he has also based his task on jsch. He also mentioned 
something about key stores being implemented, which Robert Anderson has 
said is a necessary feature.

Let me know how I can help move forward with this.
Dale Anson

Antoine Levy-Lambert wrote:
The <read> <write> interactivity could be used to be able to respond to
different situations like
   ls -l foo
   ls: foo: No such file or directory
   ...
  or
  ls -l foo
  -rwxr-xr-x    1 Administ Kein         1538 Feb 20 19:03 foo
It is true that ssh does not need to be interactive.
The type of "dialog" I am referring to is better handled by a shell script
or an ant build file at the remote end.
It looks like the telnet task needs <read><write> for the login; this is not
the case for ssh.
Antoine
----- Original Message -----
From: "Anderson, Rob H - VSCM" <[EMAIL PROTECTED]>
To: "'Ant Developers List'" <[EMAIL PROTECTED]>
Sent: Thursday, February 20, 2003 6:42 PM
Subject: RE: SSH Tasks
 

My comments (RA>) below. I encourage others to chime in.
-----Original Message-----
From: Stefan Bodewig [mailto:[EMAIL PROTECTED]
Sent: Thursday, February 20, 2003 9:12 AM
To: [EMAIL PROTECTED]
Subject: Re: SSH Tasks
On Thu, 20 Feb 2003, Rob H. Anderson <[EMAIL PROTECTED]>
wrote:
   

The ssh task is different from the telnet task in that is not meant
to handle an interactive shell session, beyond the password
authentication.
     

Why not?
RA> Because there is no need to interact with the shell.
   

It is meant to provide the ability to execute commands remotely as
if you were running ssh as follows:
ssh [EMAIL PROTECTED] 'command arg'
     

Which could be seen as a degenerate case with just a single <write>.
RA> Sure. And the this feature could have been left out of ssh, forcing
   

you
 

to interact with
RA> the shell. Don't you think it is unnecessary to force interactivity
where none is needed?
RA> Other than "consistency with the telnet command", what other value
   

would
 

<read>/<write> add?
   

This sort of thing would be much more difficult to program (and
maintain) as well.
     

I must admit that I'm not familiar with the jsch code at all, so I
can't comment on this.
Stefan
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
   


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]