On Sun, Nov 20, 2005 at 08:20:59AM +0100, Petter Reinholdtsen wrote: > [Steve Langasek] > > It's perfectly sensible: if the scripts were meant to be run in > > parallel, they shouldn't have the ".sh" extension... > Eh, are you claiming that policy mention sourcing of .sh scripts to > make sure those scripts are not run in paralell? It does not sounds > reasonable to me, as the parallel booting feature have never been > tried before in Debian. Or are you trying to say something else, like > we should change the name of the init.d scripts in /etc/rcS.d/ to > allow parallelization? I imagine that parallelization wasn't a major consideration when this part of policy was written, but *if* parallelization of these scripts is now the most important thing for boot time optimization, then they could be renamed, yes. This would have to be evaluated per-script; the "side-effect" of being able to set env variables for the remaining boot scripts also needs to be taken into consideration. > I suspect this part of policy was written to document the current > implementation, without considering which part of the implementation > should be part of the policy and which part was better left > unspecified for future improvements. Am I wrong? Er, this is a case of policy defining an interface, so that both init script implementors and implementors of runlevel managers both know what to expect. I don't think that's inappropriate for policy. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. vorlon@debian.org http://www.debian.org/
Attachment:
signature.asc
Description: Digital signature