Re: ssh, /dev/urandom

[neal@cs.uml.edu (Neal H. Walfield)]

>    1) no ssh
>    2) ssh with no security
> 
>    you have advocated (2), right?  It is that statement which I am
>    arguing against.
> 
> No, I have advocated against including a unsecure random translator.
> You are forgetting the third alternative, making ssh use its own
> random pool.  Assuming that it is possible, if it is possible then
> this is the best solution.

There is a fourth alternative: leave everything the way it is.  If you
need some quality entropy copy it from e.g. a GNU/Linux system.