Re: Hurdish applications for persistence

[ness]

address@hidden wrote:
>    We can. If we write new ones. But some existing program might rely
>    on a working chroot call.
>
> And we can make chroot() just return ENOSUPPORT, and be happy.

Really cool. And every program that depends on chroot will fail with an
error message. (Before you ask, I dunno what programs that are, but I
guess not that much). If you could give reasons for not having chroot
that don't have sth. to with that we can't _implement_ it easily, I
maybye agreed with you. But shouldn't it be possible to decrase the cost
of starting a sub hurd for chroot? With some optimisations? This was
more clear, IMHO.
--
-ness-