[BusyBox] Re: [BusyBox-cvs] svn commit: trunk/busybox: include networking

[Rob Landley]

On Thursday 28 July 2005 16:04, Shaun Jackman wrote:
> On 7/28/05, Rob Landley <rob at landley.net> wrote:
> > I'm leaving the CONFIG_ thing with its current semantics because the
> > alternative is an extremely intrusive patch that replaces over 1000
> > existing #ifdefs.
>
> I agree to leave the semantics of CONFIG_FEATURE_xxx as it was. It

Except we don't say CONFIG_FEATURE_XXX, we just say CONFIX_XXX.

> doesn't appear possible to make one name work with all three of
> #ifdef, #if, and if() without taking a binary size hit.
>
> I was suggesting naming the new semantics either ENABLED_FEATURE_xxx,
> FEATURE_xxx, DEFINED_FEATURE_xxx or DEF_FEATURE_xxx.
>
> #if DEF_FEATURE_xxx and if(DEF_FEATURE_xxx) appeal to me the most.

Enable/enabled I don't have a strong feeling for.  "Should we enable this", 
"did we enable this".  If you deeply care one way or the other, it's still 
small enough we can submit a patch. :)

I don't plan to actually _use_ this infrastructure until after I get the mount 
rewrite checked in, and right now I'm off on the tangent of documenting the 
complete history of how loop.c is cursed (which I intend to post here) 
because I've got to untangle the most recent damage as part of the checkin 
and that involves explaining what exactly broke and what's wrong with all the 
obvious alternatives...

> Cheers,
> Shaun

Rob