Re: [PATCH v1 10/11] backports: prefix c-file / h-file auto backport with BPAUTO

[Luis R. Rodriguez]

On Tue, Nov 04, 2014 at 10:50:50PM +0100, Johannes Berg wrote:
> On Tue, 2014-11-04 at 22:50 +0100, Luis R. Rodriguez wrote:
> > On Tue, Nov 04, 2014 at 11:14:52AM +0100, Johannes Berg wrote:
> > > On Tue, 2014-11-04 at 00:43 -0800, Luis R. Rodriguez wrote:
> > > 
> > > > -            mf.write('compat-$(%s%s) += %s\n' % (bp_prefix, sym, ' '.join(o_files)))
> > > > +            if bp_prefix not in sym:
> > > > +                mf.write('compat-$(%s%s) += %s\n' % (bp_prefix, sym, ' '.join(o_files)))
> > > > +            else:
> > > > +                mf.write('compat-$(%s) += %s\n' % (sym, ' '.join(o_files)))
> > > 
> > > I don't think these changes are correct, the bp_prefix is needed anyway,
> > > and there's no way it could already be. Looks like this comes from a
> > > previous attempt where you misunderstood and thought
> > > CONFIG_BACKPORT_BACKPORT_... was actually wrong, when in fact it is
> > > correct (and would now be CONFIG_BACKPORT_BPAUTO_...)
> > 
> > I did at one point have BACKPORT_BPAUTO_* stuff on the compat/Kconfig for
> > the auto stuff but figured that was superfluous. I'll respin with it.
> 
> Not sure what you mean? It seems to me you should just drop the changes
> like the one I quoted above.

If we keep BACKPORT_BPAUTO as prefix on compat/Kconfig for auto backport
stuff we'll end up with BACKPORT_BACKPORT_BPAUTO, while technically correct
as you have pointed out, I find it personally superfluous. If we however
only use BPAUTO_ prefix on the compat/Kconfig we'll end up with BACKPORT_BPAUTO.
Its subjective then, but I was opting in to prefer to just keep BPAUTO_ prefix
with the resulting CPTCFG_BPAUTO for packaging and CONFIG_BACKPORT_BPAUTO for
integration for these, if you however feel its best to double the BACKPORT
prefix that's fine too, it just seemed odd (although I realize correct).

To do what I just described I think you're right will try removing that,
testing that now. The other case that I ran into issues was the backport
versioning variables but I have solved that through other non intrusive means
to packaging (just use the C define with CPTCFG_ prefix, keeping the original
Makefile variables) which will be part of my next respin.

We still do need to decide if we want double BACKPORT_ prefix for BPAUTO
though, that remains subjective and I don't feel strongly about any way
we go.

  Luis
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/