Re: [PATCH 2/3] kconfig: remove undocumented type definition alias 'boolean'

[Paul Bolle <pebolle@tiscali.nl>]

On Mon, 2014-12-08 at 20:41 +0100, Paul Bolle wrote:
> On Mon, 2014-12-08 at 20:16 +0100, Daniel Borkmann wrote:
> > On 12/08/2014 08:13 PM, Paul Bolle wrote:
> > > On Mon, 2014-12-08 at 19:51 +0100, Paul Bolle wrote:
> > >> This doesn't apply on next-20141208. What tree did you base this on?
> > >
> > > I got it to apply on next-20141208 by dropping the hunk for
> > > arch/cris/arch-v32/drivers/Kconfig and by making (trivial) context
> > > changes to the hunks for init/Kconfig and net/Kconfig.
> > >
> > > (My experience with actual maintainers is that they won't bother to do
> > > that. Anyhow, Kconfig is basically unmaintained. And I don't know which
> > > tree one should base ones Kconfig patches on.)
> > 
> > Well, I guess if in doubt then this could go into Andrew's tree. :)
> > 
> > > I'll muddle along...
> 
> Checking this on the in tree defconfig files gave me a few hundred, it
> seems, errors like:
>     net/switchdev/Kconfig:7: syntax error
>     net/switchdev/Kconfig:6: unknown option "boolean"
>     drivers/hwmon/pmbus/Kconfig:61: syntax error
>     drivers/hwmon/pmbus/Kconfig:60: unknown option "boolean"
>     drivers/usb/gadget/Kconfig:373: syntax error
>     drivers/usb/gadget/Kconfig:372: unknown option "boolean" 
>     drivers/usb/gadget/Kconfig:386: syntax error
>     drivers/usb/gadget/Kconfig:385: unknown option "boolean"
>     drivers/usb/gadget/Kconfig:404: syntax error
>     drivers/usb/gadget/Kconfig:403: unknown option "boolean"
>     drivers/usb/gadget/Kconfig:418: syntax error
>     drivers/usb/gadget/Kconfig:417: unknown option "boolean"
>     make[1]: *** [oldconfig] Error 1
>     make: *** [oldconfig] Error 2
> 
> I assume these are uses of "boolean" added by trees merged in
> next-20141208 after the tree that you based your patch on.

Changing these last uses of "boolean" did the trick.

> Well, it seems the treewide "boolean" cleanup should be done first.
> Removing support for "boolean" could than be a second, separate step.
> Just to ease review.

This appears to have no effect on the .config files I generated for the
defconfig files in next-20141208. (After porting the patch and changing
those last booleans to bool, that is.) So that's good.

If you'd resend as two patches on top of linux-next, I might add an
Acked-by: or a Tested-by:.

Thanks,


Paul Bolle