Date: Wed, 29 Dec 2004 23:16:57 +0200 From: Maxim Sobolev <sobomax@portaone.com> To: Colin Percival <colin.percival@wadham.ox.ac.uk> Cc: "current@freebsd.org" <current@freebsd.org> Subject: Re: Building 4.x releases on 5.x and 6.x Message-ID: <41D31EC9.5050909@portaone.com> In-Reply-To: <41D31CEE.5040803@wadham.ox.ac.uk> References: <41D31B8E.7030305@portaone.com> <41D31CEE.5040803@wadham.ox.ac.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
Colin Percival wrote: > Maxim Sobolev wrote: > >> Does anybody knows if subject is possible? I guess that it is not due >> to usage of vn(4) in the release building process, has anybody any >> insights about actual state of things and/or another potential >> problems in mind. If the only problem is vn(4) vs. md(4), I think it >> can be solved quite easily, while providing a good way to avoid having >> separate test machines for oldest branches. > > > I'm fairly certain that this is possible; certainly what I do on my FreeBSD > Update buildbox (building 4.x worlds and kernels inside jails while running > a 5.x kernel) is pretty much equivalent. Making buildworld & buildkernel is not the same as building release, since make release involves creating floppies, even for the CD-based install, which in turn requires vn(4) or md(4). > One problem you may encounter involves /dev -- you'll need to mount a devfs > inside the 4.x chroot rather than trying to MAKEDEV everything. I'm not > sure if the `make release` code handles this. Checked release building scripts and found that there is some conditional code already which uses vn(4) on older system and md(4) on newer ones. Will try now to see if I can get it working. Thanks for turning me in the right direction. -Maxim
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?41D31EC9.5050909>