Re: [PATCH 1/2] kconfig: Print full defined and depends for multiply-defined symbols

[Paul Bolle <pebolle@tiscali.nl>]

On Sat, 2015-04-11 at 23:46 +0200, Stefan Hengelein wrote:
> 2015-04-11 22:23 GMT+02:00 Paul Bolle <pebolle@tiscali.nl>:
> > That's a NAK to this patch, isn't it?
> 
> That's not for me to decide. Maybe I missed something!
> But I wouldn't merge it in the current state.

Thanks. That's all I needed to hear.

> > I'd really prefer things to be simpler: how is anyone reading the
> > Kconfig entries I quoted going to realize all that?
> 
> No one has to, most of the things i explained to you come from a few
> years of experience with Kconfig. FRAME_POINTER is a complicated
> example, it is selected although it has dependencies or a prompt AND
> it is redefined in many architectures.
> AFAIUI, the "depends on" or "selected by" output should give hints
> what you have to enable to get a prompt for that option or simply
> enable it.
> 
> Wouldn't mentioning a symbol two times (because there are two
> declarations) also confuse users if they search for FRAME_POINTER? But
> at least it would give hints were both declarations are defined
> without mixing them up.

I think we can forget about this patch.

Let's focus, for example, on m32r and FRAME_POINTER. The m32r entry for
that symbol reads:
        config FRAME_POINTER
                bool "Compile the kernel with frame pointers"
                help
                  If you say Y here [...]

0) If one is building for m32r is that all there's to it? If so, "make
menuconfig"'s search facility is serving the people building for m32r a
load of crap.

1) If it's actually more complicated than that I think that anyone
reading arch/m32r/Kconfig.debug is being duped. Things look simple but
actually they are quite complicated. I think that's just wrong.

What am I missing here?


Paul Bolle