Hi, Ben Hutchings <ben@decadent.org.uk> (17/01/2013): > On Thu, 2013-01-17 at 06:58 +0100, Christian PERRIER wrote: > > Quoting Ben Hutchings (ben@decadent.org.uk): > > > > > By the same token, we don't need iproute2 if there is no GUI tool that > > > invokes it. See how silly that argument is? > > > > The point is not saying that iw is useless. My point is that I don't > > think that solving this problem (if there is one) should be done by > > adding iw to a *desktop* task. > > > > > iw is the standard Linux command for configuring wireless interfaces. > > > But you are quite unlikely to have them on a server, otherwise I would > > > suggest it should be priority standard. > > > > > > crda is definitely required for wireless interfaces, but it is normally > > > invoked via a udev hook rather than from a shell or GUI front-end. > > > > > > Unless some other solution exists, isn't it better to use the *laptop* > > task for these? > > As I said before, wireless networking is increasingly used in desktops. > It's a lot easier than retro-fitting Ethernet in a house! yeah, even if the following is probably flawed by my very limited amount of yearly travels, many advertised-by-ISPs-as-if-they-were-magical *DSL boxes make it much more attractive to go wireless than to go wired. While I haven't played extensively with 'iw' myself, it indeed looks like something we want to be shipping along 'ip'; no objection on the size side either: 52K iw_3.4-1_amd64.deb 56K libnl-3-200_3.2.7-4_amd64.deb 20K libnl-genl-3-200_3.2.7-4_amd64.deb and libnl-* are already pulled on CD#1 by network-manager, so iw won't make a huge difference. :-) (crda pulls wireless-regdb and all those are tiny anyway.) I guess I'd be happy to see iw in both the desktop and laptop tasks. Mraw, KiBi.
Attachment:
signature.asc
Description: Digital signature