John Sonnenschein wrote:
On 31-Jan-07, at 11:42 AM, [EMAIL PROTECTED] wrote:
It is "super easy" (IMO) for people to get Solaris, and the
OpenSolaris code. The hack on it and c
ontribute part is hard because of closed_bins and the integration
process respectively.
What's difficult about the closed bins apart from not being able
to port to a different architecture or chance the bits in closed_bins?
Nobody likes the closed_bins; but it's not under our control....
Rubbish.
It can be reimplemented. Are we seriously to believe that sun doesn't
have the enigneering muscle to reimplement 150 small utility
functions?The good chunk of closed bins can be taken from gnu/bsd,
there's only a couple libs (ipsec is one, and the critical one that you
can't build ON at all, even in a degraded state, is libc_i18n.a )
There is the issue of tainting, one has to be very careful which
engineers can work on this. To be safe you probably shouldn't use any
of the engineers that have ever worked on the existing closed code when
doing a clean room implementation, even if that is "just porting" from
another os.
BTW for libc_i18n.a I highly suspect that using a GPL implementation
could be problematic!
If you really want to help with the IKE/IPsec case (all of IPsec is open
source already it is just the userland IKE daemon/utils that is not)
then please come over the security community and help us port something
like Racoon.
--
Darren J Moffat
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
