All of lore.kernel.org
 help / color / mirror / Atom feed
From: Richard Purdie <richard.purdie@linuxfoundation.org>
To: Otavio Salvador <otavio@ossystems.com.br>
Cc: Patches and discussions about the oe-core layer
	<openembedded-core@lists.openembedded.org>
Subject: Re: [RFC] Utilizing LAYERVERSION_ to identify breaking changes to a layer
Date: Fri, 21 Jun 2013 09:11:56 +0100	[thread overview]
Message-ID: <1371802316.20823.286.camel@ted> (raw)
In-Reply-To: <CAP9ODKoKYJ+nzTjPoNCdFA9OxEDpTzmeNiEUb7HChgvYEv0irQ@mail.gmail.com>

On Thu, 2013-06-20 at 21:33 -0300, Otavio Salvador wrote:
> On Thu, Jun 20, 2013 at 7:56 PM, Flanagan, Elizabeth
> <elizabeth.flanagan@intel.com> wrote:
> > I know I've run into this issue and I'm going to bet that other folks
> > have as well.
> >
> > When running some sort of continuous integration system (in my case,
> > yocto-autobuilder), I need to build to at least two releases back. The
> > problem is, sometimes, local.conf and bblayers.conf requirements
> > change, for example, a layer needs a BBMASK for the prior release but
> > can't have it for the current one or a layer gets split so that I need
> > to add two directories to BBLAYERS as opposed to just the one that I
> > used a few commits back. I currently have no great way of figuring
> > that out and it has causes immense amounts of pain.
> 
> I am sorry but I didn't follow what it will help. Can you please
> provide an example?
> 
> I always thought the right way to handle it is to have a branch...

In the autobuilder there is information about what it builds, lets say
that it should build "core image-sato". Sometimes we do
add/remove/rename or otherwise change things (e.g. we might drop
meta-toolchain now the -c populate_sdk target works for images).

What is being proposed here is that we should start using the layer
version information to allow the autobuilder to have a better idea of
what to run when.

LAYERVERSION_core already exists for other reasons and this would seem a
logical extension to it. If we're worried about the major version
changes being a problem, we could use a MAJOR.minor type scheme which
would still give the autobuilder the information it needs.

The aitobuilder can then look at the version and run "bitbake
meta-toolchain" in one case but populate_sdk for later versions, or
whatever.

Cheers,

Richard





  reply	other threads:[~2013-06-21  8:12 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-06-20 22:56 [RFC] Utilizing LAYERVERSION_ to identify breaking changes to a layer Flanagan, Elizabeth
2013-06-21  0:33 ` Otavio Salvador
2013-06-21  8:11   ` Richard Purdie [this message]
2013-06-21  8:13 ` Richard Purdie

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1371802316.20823.286.camel@ted \
    --to=richard.purdie@linuxfoundation.org \
    --cc=openembedded-core@lists.openembedded.org \
    --cc=otavio@ossystems.com.br \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.