[oe] [PATCH,V2 0/6] libtool 2.4 upgrade

[Chris Larson]

On Fri, Oct 8, 2010 at 1:48 PM, Frans Meulenbroeks <
fransmeulenbroeks at gmail.com> wrote:

> 2010/10/7 Khem Raj <raj.khem at gmail.com>:
> > On Thu, Oct 7, 2010 at 12:56 AM, Koen Kooi <k.kooi at student.utwente.nl>
> wrote:
> >> -----BEGIN PGP SIGNED MESSAGE-----
> >> Hash: SHA1
> >>
> >> On 07-10-10 03:46, Khem Raj wrote:
> >>> Hi
> >>>
> >>> Following set of patches is v2 of the patches adding support for
> getting libtool 2.4 into
> >>> openemebedded. I have so far built minimal-image for two machines
> successfully using old
> >>> and new libtool.
> >>>
> >>> The libtool sysroot feature knob is added through LIBTOOL_HAS_SYSROOT
> variable. If this is
> >>> set to "yes" then you ought to use libtool > 2.4 by default its set to
> "no" which means
> >>> the current behavior remains.
> >>>
> >>> The big change that libtool 2.4 brings is sysroot support and I have
> added
> >>> support to use this feature. It should make our life easier.
> >>>
> >>> This needs a lot of testing.
> >>>
> >>> Please cherry pick the patch bundle and give it a try in yout
> environment
> >>
> >> Can you apply these to a branch so we can add fixes there if needed?
> >
> > I can do that. However now that backward compatibility is left intact I
> think
> > it would not be that bad to merge it into master and fix things on
> > master. the libtool
> > 2.4 would have DEFAULT_PREFERENCE = "-1" by default it wont be picked by
> anyone.
> > with the number of developers we have it will get better testing before
> we
> > accumulate too many changes on a branch. Secondly there might be changes
> like
> > gnutls one where the patch is only valid for new libtool and should be
> tested
> > in old and new way. If there is any breakage introduced in existing build
> will
> > be caught quickly. Where as on branch resources might be divided and it
> may
> > not progress so well. I am just weighing efforts Vs. risk here and
> > risk seems low
> >
> > Thanks
> > -Khem
> >
> >>
>
> Personally I'm in favour of merging into head. That way it gets
> accepted easier and faster.
>

Agreed, I am as well.  As the new libtool is opt-in, I don't see the harm,
and it'll ensure that any issues which crop up with the compatibility get
fixed asap.
-- 
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics