On Thu, Dec 15, 2022 at 05:19:14AM +0000, Klemens Nanni wrote: > On Wed, Dec 14, 2022 at 04:27:37PM -0800, Andrew Hewus Fresh wrote: > > On Mon, Dec 12, 2022 at 06:48:36AM +0000, Jason McIntyre wrote: > > > On Fri, Dec 09, 2022 at 09:32:13PM +0000, Jason McIntyre wrote: > > > > hi. > > > > > > > > two points about the recent ability to use lladdr: > > > > > > > > - the example of "bridge0" made sense when bridge was regarded as a > > > > separate entity and not integrated with ifconfig. plus a list of one > > > > example looks rubbish. now that we have a second example (lladdr) and > > > > bridge is not flagged as a special case, i think we can simply the > > > > text and reduce it to two examples > > > > I thought it was an example showing that it works for "dynamic" interfaces, > > but that's probably obvious enough without. > > > > > > > > - i'm not sure about using "lladdr". although we use this term in > > > > ifconfig(8), we explain it. and people may miss it if they are > > > > thinking > > > > of mac address. i've attempted to both write the term fully as "link > > > > layer local address" and add a "(MAC)". i suppose you could argue that > > > > people who think of the term as "lladdr" might miss that (!) but i > > > > don;t think that is a real worry. > > > > I think that since `ifconfig` prints out "lladdr" that even folks who > > don't know what it means should be able to figure out how to use it. > > I am a fan of defining terms though. > > > > > > > > so here's my cut at tweaking... > > > > jmc > > > > > > > > > > on the back of afresh's prioroty flip diff, here's a revision of my diff > > > after some input from deraadt. it will need adjusting again if we flip > > > priority... > > > > This looks OK to me. I'm waiting to see if kn@ or others have feedback > > on the swap and other installer bits before committing, but it seems the > > idea of preferring lladdr has been given the go-ahead, so it's just a > > matter of time. > > Yes, I agree with Theo here that lladdr is more specific and should win > present. > > > > > I tend think of the interface name as its driver + number, so I did find > > the distinction that the name is just the part without the number to be > > surprising. ifconfig(8) does say it is "name unit" with the "name" not > > including the number, so this is correct. > > jmc's might as well say "by name and unit" and "name/unit" later, no? > Then the difference between ifconfig(8) and hostname.if(8) is smaller. > > I'm not super happy with how we currently explain this > name[unit]/group/driver dance, but another diff can deal with that. > > The "Priority" sentence is subject to change, of course, but the rest is > OK kn. > > Doesn't matter much if this goes in first and the prio swap diff changes > it or we wait for the prio diff to land first. >
thanks all for feedback. i comitted it, using the "unit" suggestion. when/if priority flips, please adjust that piece of the text. jmc