On Tue, Mar 06, 2001 at 01:09:18AM +0100, Daniel Stenberg wrote:
> On Mon, 5 Mar 2001 curl_at_thewrittenword.com wrote:
>
> > I don't have the time to add this to curl at the moment but if there's
> > a TODO list, please add this. You'll need to decide whether or not to
> > remove the current seed method and replace with the method above or
> > simply to augment the current method with what you currently have.
>
> I've made a first attempt now that was just committed to CVS. I intend to add
> a command line option (and libcurl curl_easy_setopt() option) to be available
> in the next release.

Great. I'll check it out.

> > I also think that if there is an autoconf option, --with-egd-socket, it
> > should reject the current method (if not and you wish to keep the current
> > method, let the user know via a warning that the seed is being generated
> > via a weak method).
>
> I intend to have a series of seeding attempts with my weak one at the last
> one in the line, with a logging that it is a weak seed.

Check out stunnel, http://www.stunnel.org/, for how they do it. I
think they go through a few procedures to try and get the seed value.

-- 
albert chin (china_at_thewrittenword.com)