Date: Fri, 17 Feb 2006 15:41:18 -0800 From: Brooks Davis <brooks@one-eyed-alien.net> To: Doug Barton <dougb@FreeBSD.org> Cc: freebsd-current@FreeBSD.org, Xin LI <delphij@delphij.net> Subject: Re: New RCorder: abi loaded too late Message-ID: <20060217234118.GA22643@odin.ac.hmc.edu> In-Reply-To: <43F65A70.7080608@FreeBSD.org> References: <1140187193.731.47.camel@spirit> <20060217181842.GA21033@odin.ac.hmc.edu> <43F65A70.7080608@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--y0ulUmNC+osPPQO6 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Fri, Feb 17, 2006 at 03:21:20PM -0800, Doug Barton wrote: > Brooks Davis wrote: > > On Fri, Feb 17, 2006 at 10:39:53PM +0800, Xin LI wrote: > >> Hi, > >> > >> With the recent rcorder change, we were able to > >> start /usr/local/etc/rc.d in the same rcorder pass. However, because > >> abi was started late in the startup process, we have potentially broke > >> the services written for older FreeBSD releases, or other operating > >> systems, which were written to use rc.d scripts. > >> > >> Shall we move abi to be started before SERVERS? > >=20 > > Seems fine, I'd recommend: > >=20 > > # REQUIRE: mountcritremote > > # BEFORE: SERVERS > >=20 > > The current requirement of LOGIN makes no sense. >=20 > Agreed on the last, but can we modify SERVERS to REQUIRE abi instead? I > would prefer not to add new instances of BEFORE in /etc/rc.d without good > reason. I'd also vote for REQUIRE syslogd in abi personally, but I won't > quibble on that point. In the case of pseudo targets, I'm of the opinion that the BEFORE statements are more consistent. I'm not all that thrilled with them in general, but IMO inconsistent of use BEFORE and REQUIRE is the biggest problem. I don't see any logical reason for syslogd to be required. kldload doesn't call syslog, I seriously double linux ldconfig does, and the abi script will always return success. > One quick meta-issue if I can hijack Xin's thread for a minute, now that = we > are including the local scripts in the overall rcorder, and as more ports > are converted to take advantage of this, we will see more of these instan= ces > where the stuff in the base that always "just worked" before will need to= be > twiddled. This is expected, and I thank Xin for bringing this one to our > attention. Agreed. -- Brooks --=20 Any statement of the form "X is the one, true Y" is FALSE. PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4 --y0ulUmNC+osPPQO6 Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQFD9lz5XY6L6fI4GtQRAltfAJ9g4S6SeOEJv7P8mIXqydGFq2mpWQCeN4aH K/+CbM/5wb631M3lGEKN+oU= =PnCs -----END PGP SIGNATURE----- --y0ulUmNC+osPPQO6--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060217234118.GA22643>