[oe] OE recipe tree quality

Richard Purdie rpurdie at rpsys.net
Sat Jul 31 23:12:56 UTC 2010


On Fri, 2010-07-30 at 10:34 +0200, Frans Meulenbroeks wrote:
> 2010/7/30 Koen Kooi <k.kooi at student.utwente.nl>
> > On 30-07-10 09:21, Esben Haabendal wrote:
> > > On Thu, Jul 29, 2010 at 2:07 PM, Philip Balister <philip at balister.org>
> > wrote:
> > >> On 07/29/2010 05:45 AM, Koen Kooi wrote:

> > >>> I would actually advocate removing the 'world' feature from bitbake/OE
> > >>> to stop people from wasting time on looking at bitbake world and have
> > >>> them fix actual problems.
> > >>
> > >> bitbake world seems to be the source of pointless listserv discussions.
> > Does
> > >> it serve any purpose?
> > >
> > > Pointless or not really depends on how you look at quality.
> > >
> > > If you look at it as you, Koen and other OE long-timers, yes, it looks
> > > rather pointless to have bitbake world.
> > > But for those of us who have a different view on what quality is, then
> > > bitbake world serves a purpose.
> >
> > As Thomas points out, as soon as you start blacklisting things (which
> > actually increases quality), bitbake world doesn't work anymore.
> 
> Blacklisting does *NOT* increase quality. It just hides the problem and as
> such it is ostrich behaviour.
> 
> Instead of masking the problem, better fix things. That is what improves
> quality!
> 
> For me a non building recipe is a sign of poor quality.
> 
> BTW: to avoid the blacklisting issue, I've restarted my test with minimal
> distro.
> Results will probably be available after the weekend.

"world" is used by Poky quite successfully. In Poky we expect everything
to build with a known list of things that don't. Our aim is to have that
list consisting of zero items and I intend to see that happens :).

I don't think many people have been watching Poky recently but we've
been quietly having a massive quality control effort on the metadata in
there. This has included:

a) Removal of legacy staging
b) Using BBCLASSEXTEND = native where it makes sense
c) Using nativesdk (no sdk class)
d) Upgrading everything to modern versions of the package concerned
e) Only allowing sane pkgconfig files
f) Enhancing the metadata with licence and other information

Its possible to contemplate this due to Poky's version policy and size
and its still taking significant effort. OE has always been pulled in
many directions and this is both a good and a bad thing. The only ways I
can think of to make a dramatic improvement to OE in this area would be
disliked by some of its users so I don't know what the solution is.

There are other good things happening in Poky as well as the above too
btw :)

Cheers,

Richard








More information about the Openembedded-devel mailing list