[PATCHes] mostly Kbuild fixes for out of tree build, Kconfig support in abuild

[Stefan Reinauer]

See patches

[PATCHes] mostly Kbuild fixes for out of tree build, Kconfig support in abuild

[Myles Watson]

> don't turn on WARNINGS_AS_ERRORS for Qemu per default

I disagree with this change.  I think it should be noticed when changes add
warnings to a target that is free of them.  I think the goal should be to be
able to turn this on for all targets.  There have been several times that
warnings pointed out breakage.  Ignoring them seems counterproductive.

Abuild patch:

Acked-by: Myles Watson

Thanks,

Myles

_____

From: [e-mail] [mailto:[e-mail]]
On Behalf Of Stefan Reinauer
Sent: Friday, January 29, 2010 1:08 PM
To: coreboot
Subject: [coreboot] [PATCHes] mostly Kbuild fixes for out of tree build,
Kconfig support in abuild

See patches

[Stefan Reinauer]

On 1/29/10 9:16 PM, Myles Watson wrote:
> > don't turn on WARNINGS_AS_ERRORS for Qemu per default
>
> I disagree with this change.  I think it should be noticed when
> changes add warnings to a target that is free of them.  I think the
> goal should be to be able to turn this on for all targets.  There have
> been several times that warnings pointed out breakage.  Ignoring them
> seems counterproductive.

Hm... It did not compile without warnings for me when I tried, so that's
why I disabled it at some point.

I generally agree with what you're saying.  Actually I manually deleted
the  "add -Werror-implicit-function-declaration to CFLAGS" part of my
patch because I figured people would feel no love for me if I left it
in. I think it should go in then, too.

Stefan

[PATCHes] mostly Kbuild fixes for out of tree build, Kconfig support in abuild

[Myles Watson]

On Fri, Jan 29, 2010 at 1:22 PM, Stefan Reinauer wrote:
> On 1/29/10 9:16 PM, Myles Watson wrote:
> > don't turn on WARNINGS_AS_ERRORS for Qemu per default
>
> I disagree with this change.  I think it should be noticed when changes add
> warnings to a target that is free of them.  I think the goal should be to be
> able to turn this on for all targets.  There have been several times that
> warnings pointed out breakage.  Ignoring them seems counterproductive.
>
> Hm... It did not compile without warnings for me when I tried, so that's
> why I disabled it at some point.

It happens frequently, but the kbuildall report from 5063 says it built, and
5064 works for me. Hopefully tracking it will make it happen less often.
Each port that gets to the "warnings free" point will cover more and more of
the code base.
> I generally agree with what you're saying.  Actually I manually deleted
> the  "add -Werror-implicit-function-declaration to CFLAGS" part of my patch
> because I figured people would feel no love for me if I left it in. I think
> it should go in then, too.

I guess that depends on how many boards it breaks and how long it would take
us to fix them.

Thanks,
Myles

[PATCHes] mostly Kbuild fixes for out of tree build, Kconfig support in abuild

[Myles Watson]

On Fri, Jan 29, 2010 at 1:34 PM, Myles Watson  wrote:
> On Fri, Jan 29, 2010 at 1:22 PM, Stefan Reinauer wrote:
> > On 1/29/10 9:16 PM, Myles Watson wrote:
> > > don't turn on WARNINGS_AS_ERRORS for Qemu per default
> >
> > I disagree with this change.  I think it should be noticed when changes
> > add warnings to a target that is free of them.  I think the goal should be
> > to be able to turn this on for all targets.  There have been several times
> > that warnings pointed out breakage.  Ignoring them seems counterproductive.
> >
> > Hm... It did not compile without warnings for me when I tried, so that's
> > why I disabled it at some point.
>
> It happens frequently, but the kbuildall report from 5063 says it built,
> and 5064 works for me. Hopefully tracking it will make it happen less
> often.  Each port that gets to the "warnings free" point will cover more and
> more of the code base.
> > I generally agree with what you're saying.  Actually I manually deleted
> > the  "add -Werror-implicit-function-declaration to CFLAGS" part of my patch
> > because I figured people would feel no love for me if I left it in. I think
> > it should go in then, too.
>
> I guess that depends on how many boards it breaks and how long it would
> take us to fix them.

* fix iasl output directory for i945 boards

I'd prefer it if the output went into $(obj)/mainboard/$(MAINBOARDDIR)
instead of just $(obj).  I know it gets moved, but I still think it's nicer.

Minus the qemu change:
Acked-by: Myles Watson

Thanks,
Myles