Stefan Bodewig wrote:
> On Thu, 09 Aug 2001, Richard Emberson <[EMAIL PROTECTED]> wrote:
>
> > Well, we are secure ... we all have RSA passphrases ... so still
> > need to enter text into a running ant task.
>
> You could use ssh-agent before starting Ant.
>
> Stefan
Thanks for the suggestion.
I ultimately would like to use Ant to push builds to our production
site. Now
I don't want anyone in Engineering, Test, or QA to have a window open
running ssh-agent that would enable anyone to 'get at' the production
site.
So, using your suggestion, I am going to place the call to start the
ssh-agent
and the call to ssh-add within the build.sh file. If the build.sh file
detects that
the user is passing the "push to production" Ant target to Ant, then it
requires the user to enter their ssh passphrase. If that target is not
requested,
then there is no ssh-x started or passphrase requested. Now after Ant
runs, whether or not it success in push to production, the ssh-agent
process
is killed.
I think that ought to work, providing the desired capability while
retaining security.
RME
