On Mon, Jun 18, 2018 at 9:47 AM Mark Hatle <[email protected]> wrote: > > On 6/18/18 11:30 AM, Burton, Ross wrote: > > On 18 June 2018 at 17:25, Mark Hatle <[email protected]> wrote: > >> On 6/18/18 10:48 AM, Ross Burton wrote: > >>> Both busybox and coreutils provide mktemp, and the only difference > >>> between those > >>> (and standalone mktemp) is that coreutils supports --suffix. > >> > >> I've got no objections to this.. but I do have a question/comment. > >> > >> Is mktemp packaged by itself in coreutils or do you need the larger > >> coreutils to > >> get it. (I'm not actually sure it matters, but in the past there were some > >> cases where you might not want all of coreutils, but wanted mktemp...) > > > > We don't have a small or large coreutils. Maybe that's a WR-ism? > > No, I'm referring to system sizes.. there have been systems developed in the > past that use a number of alternative command line tooling. > > I thought coreutils was broken into various packages already. It may have > been > in the past and is no longer broken up as well -- or maybe the busybox version > was lacking some argument mktemp (coreutils mktemp) supports. > > The point is there was a reason for it in smaller systems, and that reason may > not exist any longer -- but just be aware it might come back up as a problem.
The affected case is where busybox one is not used and coreutils is not used but mktemp is used. I think its in best interest for this usecase to move on to accept a maintained version of utility, be it from coreutils or busybox. -- _______________________________________________ Openembedded-core mailing list [email protected] http://lists.openembedded.org/mailman/listinfo/openembedded-core
