Re: [PATCH] bsp-guide: replace meta-intel with meta-xilinx as container layer

Andrea Galbusera

On Fri, May 3, 2019 at 12:56 PM Robert P. J. Day <rpjday@...> wrote:

On Thu, 2 May 2019, Scott Rifenbark wrote:

Great, thanks!

On Thu, May 2, 2019 at 11:21 AM akuster <akuster@...> wrote:

On 5/2/19 10:45 AM, Scott Rifenbark wrote:
The term "Container Layer" was put in the ref-manual by me to describe a
"meta-*" layer that had other "meta-*" layers (see
and the term "Container Layer"). Maybe this was never a good term? I don't
know. Nobody has said anything about that term for many releases.

The problem Robert points out needs to be fixed in the BSP manual as "meta-intel" is
not longer qualifying by my definition. I can fix that.

Is "Container Layer" ok as I have defined it?

in those terms above, yes its fine.
the more i think about it, the more i'm nervous about the phrase
"container layer", as it suggests a "layer" of some kind in the
context of OE. the one distinction i think worth making is that a "BSP
layer" is a layer expressly designed to support identified target
systems (eg., meta-xilinx-bsp), while a non-BSP layer simply packages
functionality (recipes, classes) -- a good example is

in either case, the trivial property of an OE layer is something
that can be specified in a bblayers.conf file. so i'm not sure *how*
you would describe, for example, meta-openembedded.
Maybe just as a "metadata collection repo"? IMHO from a OE
perspective, a "layer" is a metadata set which provides its own
layer.conf file (which, as you say, means you can add it to your
project's bblayers.conf). Bundling more than one layer within the same
git repository (or whatever archive format use to distribute them)
does not qualify this metadata collection as anything otherwise
meaningful for bitbake. That said, meta-openembedded is definitely the
most evident example of such a "concept". I agree the term "container
layer" can be either confused with a OE layer, which technically is
not, or something having to do with containers (as I suspect from
Armin's reply), which is completely unrelated to. Maybe rewording a
little bit the manual here could reduce potential confusion to new
users. Moreover whatever term we choose to explain this concept, we
should make clear that it is orthogonal to any other layer
classification by content (i.e. BSP, distro, ...).

i'm sure i'm overthinking this.



Robert P. J. Day Ottawa, Ontario, CANADA

yocto mailing list

Join to automatically receive all group messages.