Subject: Re: sysinst and LFS
To: None <current-users@NetBSD.org>
From: Jochen Kunz <email@example.com>
Date: 11/30/2006 23:18:34
On Thu, 30 Nov 2006 20:48:44 +0000
David Laight <firstname.lastname@example.org> wrote:
> I've inverted the sense of that test...
Well. The test is redundant. You don't need fsck_lfs(8) (most likely),
so it can be removed from the instalation system and thus from the test.
If mount_lfs(8) or newfs_lfs(8) are missing on the instalation system
you can't install on LFS anyway.
> It might be needed to stop sysinst trying to mount the partitions
> - but I've not really looked at that part of sysinst for a while.
Not mounting LFS would be OK if the _target_ system doesn't contain the
needed executables. But in that case the test should test the _target_
system, not the instalation system. The test, if performed at all,
should test for lfs_cleanerd(8) in addition and it must be ensured that
the instaled kernel has "file-system LFS" enabled. "file-system LFS" is
in GENERIC and all LFS user land tools are part of base.tgz. So they are
allways there. Therefore I don't see any rationale for a test like this.