Re: [RFC] add "enable" to the kconfig language

[Steven Rostedt <rostedt@goodmis.org>]

On Fri, 2012-09-14 at 10:14 +1000, Stephen Rothwell wrote:
> Hi all,
> 
> I have noticed that we use the following paradigm quite a bit theses days:
> 
> config <something
> 	select HAVE_<config1>
> 
> config <config1>
> 	depends on HAVE_<config1>
> 
> or similar
> 
> I was wondering if it would make sense to replace this with:
> 
> config <something
> 	enable <config1>
> 
> config <config1>
> 	depends on enabled
> 
> The advantage of this is that we would not have all the HAVE_ config
> names in our .config files and the generated include files.  Of course,
> if config1 does not depend on "enabled", then the "enable <config1>"
> would have no effect (we may want to warn about this).
> 
> Comments?

I always hated those HAVE_* configs, especially because they were always
a source of confusion when asking someone if something is enabled...

Them: "The function tracer doesn't work, I don't see any debugfs files"

me: "Do you have FUNCTION_TRACER enabled?"

Them: "Yes, I checked"

me: "You have *FUNCTION_TRACER* not *HAVE_FUNCTION_TRACER*?"

Them: "Oh, no I just have HAVE_FUNCTION_TRACER"

BAH!

If you can make 'enable' work, I would definitely Ack converting over to
it.

-- Steve