On Thu, 12 Jul 2001, Craig Sanders wrote: > On Mon, Jun 25, 2001 at 08:43:26PM -0300, Henrique de Moraes Holschuh wrote: > > On Mon, 25 Jun 2001, Ethan Benson wrote: > > > i think someone already wrote something like query-rc.d in order to > > > solve the annoying problem of package upgrades starting daemons that > > > > That's invoke-rc.d... > > wouldn't that be better done in start-stop-daemon? No. start-stop-daemon is in the wrong abstation layer for this. invoke-rc.d is an abstraction layer, on top of fixing the extremely annoying ¨start out of runlevel" design bug we currently have. > have start-stop-daemon call query-rc.d or whatever it's called to find > out if the run-level is appropriate for starting the daemon. No. And what about services that are not daemons? We need to get those under control as well. > if we change start-stop-daemon to support this then *all* init.d scripts > immediately support it without any further changes. Wrong. Many init.d scripts do not use start-stop-daemon. > btw, this implies the need for a --force option to start-stop-daemon > and/or init.d scripts so that the admin can manually force a daemon to > be started no matter what the run-level is. Read the archived discussions, the policy proposal thread, and the already written, tested and debugged code (for both sysv-style links and file-rc) before you reply to this message, please. All these issues have been discussed before. -- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh
Attachment:
pgpUciO3pTv2h.pgp
Description: PGP signature