Hi Roger, > CONCURRENCY is now a stable tested feature. Is it realistic > to remove this for wheezy? > > We could > - Just remove the option to set it in rcS, and retain the > existing fallbacks for non-insserv-enabled systems, or > - Remove everything and strictly require insserv. > I guess this depends upon how many systems are still out there > using the fallbacks. Would it be possible to abort install in > the preinst if we encounter such a system for the latter > option? In git, I've gone with the safe former option, but > could do the latter if that's acceptable.
I'd love to see the latter, but fear that there is still a large number of systems where the dep-based-boot conversion has not yet taken place for whatever reason. Petter would probably have a better POV about this. It's definitely the right time for the safe former option. > > > tmpfs > Recently a number of bug reports have surfaced regarding running > out of space on /tmp. I think we do an OK job with the existing > /run and /run/lock defaults, but we can probably do better with > /tmp. Another issue was leaving files hidden on the original > /tmp. I'd like to propose: > - moving mounting of /tmp from kernel to after checkroot but > before mountall. > - merging this with the existing overflowtmp logic, which should > probably be using the same defaults > - cleaning /tmp before mounting the tmpfs (using the existing > bootclean logic). We just move it to before mountall. > Does this sound OK? Yep. Thanks, Kel _______________________________________________ Pkg-sysvinit-devel mailing list [email protected] http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-sysvinit-devel
