On Wed, 2011-04-06 at 22:17 +0200, Johannes Berg wrote:
> On Wed, 2011-04-06 at 22:15 +0200, Johannes Berg wrote:
> > On Wed, 2011-04-06 at 22:10 +0200, Paul Bolle wrote:
> > > On Wed, 2011-04-06 at 14:38 -0400, John W. Linville wrote:
> > > > The syntax may seem strange,
> > >
> > > It does!
> > >
> > > > but basically it just says "don't let me by y if RFKILL is m"
> > >
> > > ... but, besides that, I can be any value. So in effect it's shorthand
> > > for
> > > depends on RFKILL=y || RFKILL=m && m || RFKILL=n
> > >
> > > (which actually looks equally strange). Is that correct?
> >
> > I don't think it is, I believe that an expression like "RFKILL=y" has a
> > bool type, and a tristate type value that depends on a bool type can
> > still take the value m.
>
> Err, which is of course perfectly fine since if RFKILL is built in this
> can be any value, and in the RFKILL=m case you force it to m by making
> it depend on m directly. So yes, you're right.
Whoops ... sorry about the talking to self ...
I still think the original is easier to understand. After all, just
depends on RFKILL
is trivial to understand even with tristates. And knowing that RFKILL
will provide no-op inlines when it is unconfigured, you add
depends on !RFKILL
for that case.
johannes
