[Openembedded-architecture] Switch sstate and other pieces of the project to xz?

Christopher Larson clarson at kergoth.com
Sat Jan 9 17:38:59 UTC 2016


On Sat, Jan 9, 2016 at 10:03 AM, Khem Raj <raj.khem at gmail.com> wrote:

> On Sat, Jan 9, 2016 at 8:48 AM, Richard Purdie
> <richard.purdie at linuxfoundation.org> wrote:
> > Times change and xz is becoming the defacto compression format and
> > utility. It has better compression ratios than gz and similar
> > compression/decompression times. Its better than bzip2 decompression in
> > particular.
> >
> > Given the bzip2 lack of performance, switching to xz for the SDK
> > appears to be an easy winning change.
> >
> > Switching sstate over to xz may be a little more controversial and
> > invasive, not least that we'd need to update the docs to mention the xz
> > dependency.
> >
> > Equally, I think I'm coming down strongly in favour of switching. DOes
> > anyone have any concerns about that?
> >
> > There are no compatibility of sstate feed issues, objects just change
> > name at some point and the old and new can coexist side by side if
> > needed.
> >
> > [NB: I know that some have proposed making sstate compression format
> > configurable. I still have mixed feelings about that, not least that it
> > encourages differences in a part of the system I'd prefer to try and
> > keep standardised and figure out the best selection. In this case, xz
> > does look like a clear winner over gzip]
>
> I think its fine idea.May be we can top it up with pxz[1] to harness
> multicores
> and also use xz as default for source archive generation as well.


Seems like a good idea to me. We could think about it for mirror tarballs
as well eventually.
-- 
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openembedded.org/pipermail/openembedded-architecture/attachments/20160109/9cb44be0/attachment-0002.html>


More information about the Openembedded-architecture mailing list