[APPLIED] Pull request for Raring

[Andy Whitcroft]

On Thu, Nov 08, 2012 at 04:18:05PM -0500, Ben Collins wrote:
> On Nov 8, 2012, at 12:26 PM, Andy Whitcroft <apw at canonical.com> wrote:
> 
> > On Thu, Nov 08, 2012 at 07:41:24AM -0500, Ben Collins wrote:
> >> On Nov 8, 2012, at 5:17 AM, Andy Whitcroft <apw at canonical.com> wrote:
> >> 
> >>> On Wed, Nov 07, 2012 at 10:29:13PM -0500, Ben Collins wrote:
> >>>> Please pull for this commit to master-next:
> >>>> 
> >>>> commit f808fd0abdb75313167cdc145a2fb947a4e625b8
> >>>> Author: Ben Collins <ben.c at servergy.com>
> >>>> Date:   Wed Nov 7 11:49:28 2012 -0500
> >>>> 
> >>>>   UBUNTU: [Config] Update enforce rule for CONFIG_NVRAM to better suit flavours
> >>>> 
> >>>>   Signed-off-by: Ben Collins <ben.c at servergy.com>
> >>>> 
> >>>> From git://github.com/benmcollins/ubuntu-raring-powerpc.git master-next
> >>> 
> >>> I have pulled this in.  Note I have modified the NVRAM check as below to
> >>> match what I believe you intended (based on the assumption the configs
> >>> as generated now are correct) as in its previous form the check was
> >>> effectivly unchecked on non-power:
> >>> 
> >>>   !exists CONFIG_NVRAM | \
> >>>    (arch powerpc &/ value CONFIG_NVRAM y) | \
> >>>    value CONFIG_NVRAM m
> >>> 
> >>> Yes I know the form is somewhat opaque :/.
> >>> 
> >>> -apw
> >> 
> >> 
> >> Thanks
> >> 
> >> Mainly the problem is that the flavour token does not actually allow for more than one flavour to be listed. So the previous check had "(flavour foo bar \& ...)" and it would fail (sometimes) because it seemed to be evaluating bar as it's own predicate. Not sure how this happens, but I didn't feel like digging around in config-check too much.
> > 
> > Ok, I cannot reproduce this behaviour, and the internal test suite seems
> > to pass with similar constructions.  Odd.
> 
> Try two in a row where neither apply to the current flavour. The ADT construct works, but the NVRAM construct after it failed. They had the exact same syntax and the config option was set to the exact same value of =y

I substituted in this valid alternative in using that exact syntax and it seems to
pass for me:

    (flavour powerpc-smp powerpc-e500 powerpc-e500mc &/ value CONFIG_NVRAM y) | \
     value CONFIG_NVRAM m | \
     !exists CONFIG_NVRAM

Confused, perhaps you can send me the failing syntax.

-apw