[oe] legacy staging: call for action

[Frans Meulenbroeks]

2010/8/1 Richard Purdie <rpurdie at rpsys.net>

> On Sun, 2010-08-01 at 13:49 +0200, Frans Meulenbroeks wrote:
> > In the attached file (which I hope will pass the mailing list) you find
> the
> > results.
> > I've also uploaded the file to http://filebin.ca/xotgad/do_stage so if
> it is
> > not attached you can find it there.
>
> I'm a little dubious of doing it this way. It also includes some rather
> simple cases where do_stage does:
>
> do_stage () {
>        autotools_stage_all
> }
>

Yeah, I'm aware of these.
There are some 400 of them.
I think I can cook up a sed or awk script to filter them out and then
visually inspect the resulting diffs for correctness, but I am not sure if
that is a desired way forward.
(I don't think I can test all those changes, also I seem to recall that
someone already did something similar; but that needs committing, would be
kind of a massive commit anyway (with 400 or so files affected).

What i proposed is a starter to divide the work. Of course I will not keep
anyone from fixing the other recipes.
And of course there are some which are definitely orphaned (e.g. the
nslu2-linux owned ones, some 70 I think).
Then again I guess these are all pretty old.

But if you see a better way, please share.


>
> which are null operations and can be removed in pretty much all cases. I
> was surprised to only find myself listed for about 5% of that list :).
>
> In the spirit of helping:
>
>
> http://git.openembedded.net/cgit.cgi/openembedded/log/?h=rpurdie/for-testing
>
> but I've not tested this so I'm going to need some assistance. I've made
> the changes by a combination of inspection and knowledge of what Poky
> does. I concentrated on the cross components since people are perhaps
> more wary of them.
>

Thanks alot.
Guess Khem should have a peek at the gcc ones.

Wrt the binutils ones:
I see in your patch
binutils_2.14.90.0.6.bb
binutils_2.14.90.0.7.bb
binutils_2.16.91.0.6.bb
binutils_2.16.91.0.7.bb

Given that binutils is at 2.20 now and I expect the diff between the .6 and
.7 subversions to be very small, I would prefer to remove those .6 variants.

Frans