Hi Michael, Michael Biebl <[email protected]> writes: > Am 26.07.2014 21:12, schrieb Michael Biebl: >> I completely forgot that we defer the actual start/stop/restart to >> invoke-rc.d in most cases (i.e. if there is a matching .service and SysV >> init script). >> >> So, yeah, we'd actually need to special case .socket units here or we >> change invoke-rc.d to actually stop .service *and* corresponding .socket >> files. > > For completeness sake: The "service" command already does stop both > .service and corresponding .socket unit(s) [1]. > > So far we deemed invoke-rc.d a lower level tool, similar to systemctl, > meaning it doesn't do too much magic. > > > Feedback/comments from other team members would be welcome on this matter. I think we should implement some logic in dh-systemd for the case where there is no SysV init script, and then port the logic that we have in “service” to “invoke-rc.d” for the legacy use-case.
Just to make it clear: my comments from #29 still stand. If one of you could tackle this, it’d be good. -- Best regards, Michael _______________________________________________ Pkg-systemd-maintainers mailing list [email protected] http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-systemd-maintainers
