On Thu, Jan 05, 2006 at 12:33:33PM -0200, Henrique de Moraes Holschuh wrote:
> > Is "-delete" is no slower than the current "-print0 | xargs rm" code? > > I'll assume so. The -delete action corresponds to a direct call to unlink. Therefore it is faster than the "-print0 | xargs" method. In theory on some systems xargs -P could be faster (i.e. ones with many CPUs and many spindles) but it's rather unlikely. > It might be a bit slower (extra fork, pipe setup), but it won't make much > of a difference, either way. I doubt it is in the easily measurable land. I'd expect the difference to be measurable for deep directory hierarchies, but have not measured it myself. > > How about we make this change but encapsulate the find call in a function > > in order to assist the sarge backporters? The sysvinit packages already > > require modification for use in sarge. > > IMHO It makes no sense to do such a change and place it in a function call, > the function call is probably slower than the original code to begin with :) I agree. The conceptual cost of the additional complexity outweighs the benefit. James. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
