On Thu, Apr 30, 2020 at 03:33:56PM -0600, Theo de Raadt wrote: > Jason McIntyre <[email protected]> wrote: > > > On Thu, Apr 30, 2020 at 10:07:14PM +0100, Stuart Henderson wrote: > > > > > > On 2020/04/30 20:52, Gerhard Roth wrote: > > > > It it too much to expect users to read the ifconfig man page? > > > > > > Printed, it is 28 pages of A4. > > > > > > > ouch. > > > > > Compare with the wifi drivers, you have to look at ifconfig(8) if > > > you want all the details, but EXAMPLES in iwm(4) (and I think all the > > > other > > > drivers) is enough for a quick bare-bones config. That seems a reasonable > > > model. > > > > > > > i think we have lost our way a bit with ifconfig.8. the need to avoid > > repitition drove it, but bridge broke it. > > > > we should start farming out all of the subsections back to their > > respective pages. > > Respective pages?! The respective page for ifconfig commands *is* the > ifconfig page. The drivers impliment various lowlevel and system > behaviours, but ifconfig is documenting the commands. >
there is more than one way to do it. that's all. > > umb.4 is one screenful, but how to use it with > > ifconfig is at the end of ifconfig.8. that can;t be optimal. > > Since the beginning of umb, I've begged for it to stop being > a such special snowflake, and to search for common functionality with > other drivers, to hide the specialness. > > But the previous comment remains. These are ifconfig commands. > They belong in the ifconfig manual page. We don't describe ls options > in filesystem manual pages, we describe them in the ls man page. > > > of course there's an issue, and it's a big one: IEEE 802.11. farming > > that out would inflate a lot of pages, and require care to keep > > consistent. > > It simply makes no sense. Driver options aren't being explained. > All of those drivers have interfaces to low-level network stack. > The network stack is told what to do with the ifconfig command, > using ifconfig options. The docs are in the right place. > since you disagree, it won;t happen. but i don;t agree that the docs are in the right place - there is no "right place". farming it out would solve some of ifconfig's problems, and replace them with some other. it's a trade off. that's all. i think the trade off would be worth it. an analogy: carp documents some sysctls. but you use sysctl to set them. jmc
