On Mon, Jul 18, 2011 at 5:10 AM, Karel Zak <kzak@xxxxxxxxxx> wrote: > I'm not sure if the currently used extra separators (,,) for the > caches is a good idea. Maybe it would be better to force people to > parse the last comment line where is the header for the columns. Speaking as a user of lscpu, I think that forcing people to parse the last comment line is not a particularly good idea - no one is used to doing it, and I'm not sure of any other utility that forces you to parse it's "machine-readable" output. Stability of this format is key. > The ideal solution is to extend the "-p" functionality and allow to > specify expected columns at command line, something like: > > lscpu -p -o cpu,core,book,socket I think that this is the only long-term supportable way to do this. CPU architectures (even in the x86 world that I'm interested in) ARE going to change and evolve. Heck, it's even possible that concepts like the hypervisor scheduling parameters that you mentioned on s390 could eventually make their way down to x86 virtualization, and exposing stuff like that in lscpu would be nice. -- To unsubscribe from this list: send the line "unsubscribe util-linux" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
