Date: Sat, 1 Jan 2011 15:23:24 -0800 From: Chip Camden <sterling@camdensoftware.com> To: freebsd-questions@freebsd.org Subject: Re: csup and build question Message-ID: <20110101232324.GF66527@libertas.local.camdensoftware.com> In-Reply-To: <4d1f9d6a.0Y9F5EG3I94PNVKJ%perryh@pluto.rain.com> References: <20110101183822.GB66527@libertas.local.camdensoftware.com> <20110101195330.41ca6a26.freebsd@edvax.de> <20110101192311.GE66527@libertas.local.camdensoftware.com> <20110101204401.62602973.freebsd@edvax.de> <20110101200011.00005459@unknown> <4d1f9d6a.0Y9F5EG3I94PNVKJ%perryh@pluto.rain.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--XaUbO9McV5wPQijU Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Quoth perryh@pluto.rain.com on Saturday, 01 January 2011: > Bruce Cran <bruce@cran.org.uk> wrote: > > On Sat, 1 Jan 2011 20:44:01 +0100 > > Polytropon <freebsd@edvax.de> wrote: > > > If programs or other components of the world do have to > > > be in sync with kernel or system files (sys/*.h), then > > > I conclude that world also has to be compiled. In this > > > regards, it means a change of the kernel, and if the > > > kernel did change, kernel and world are out of sync, > > > which implies that both have to be compiled. > > The ABI of the kernel sometimes changes, in which case despite > > there being no changes in world it still needs to be rebuilt > > against the newer header files. For example, recently: > > > > 20101228: > > The TCP stack has been modified to allow Khelp modules to > > interact with it via helper hook points and store per-connection > > data in the TCP control block. Bump __FreeBSD_version to 900029. > > User space tools that rely on the size of struct tcpcb in tcp_var.h > > (e.g. sockstat) need to be recompiled. >=20 > IIUC, things like that are supposed to happen only within -CURRENT, > not within a -STABLE or security branch. >=20 > I would surmise that, if a csup yielded changes only in /usr/src/sys > and none of those changes touched a .h file, it should not be necessary > to rebuild world. OK, that's about what I had thought. Thanks to all three of you for responding. Obviously, the safe choice is to always rebuild world if anything changes. But if I see that only a couple of .c files in the kernel have been touched, then rebuilding only kernel should be OK in stable -- I think. The only possible gotcha would b= e if something in the kernel gets compiled to a library that world uses, or if something in world includes a .c file from kernel for some reason. --=20 Sterling (Chip) Camden | sterling@camdensoftware.com | 2048D/3A978E4F http://camdensoftware.com | http://chipstips.com | http://chipsquips= .com --XaUbO9McV5wPQijU Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (FreeBSD) iQEcBAEBAgAGBQJNH7dsAAoJEIpckszW26+ReF8H/1KCNB1SG6WcVx83azNDQ3DI VvgBuBnaYsQM8WVFWxEssHNceVOiokrfb5FUFQ3erMK2JElXaJI6oR9eJ+D1AgMA V+Sd3NtWdlZJ68PtdBtWGQC8tCA5KUR6E7jfBJl3X4nWz5YprDrFpJ9+b+AVL0RF Y+7S0tYGZTCN0L9hmdNOcwmQQOjYb/12mnkL4rDZKtNGakJMJU8EB0k+b5HMGnxh vhF7uXCee/YJpArSha1oCD4b/yt6YzDezPcrbiDpjLYXgJVXHE5LntC89GBpE/ih ZO0XV/jXmPTITf74kP2ZI0F1suNY0b4spTrPyy6ufX9Jp5uVZBVxVjzrrjxq5Bo= =HdnW -----END PGP SIGNATURE----- --XaUbO9McV5wPQijU--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20110101232324.GF66527>