[yocto] OpenEmbedded sole base meta layer and poky base meta layers in sync, or not (should be The Same meta base layer, used in different distros)???

Zoran Stojsavljevic zoran.stojsavljevic at gmail.com
Mon Oct 23 04:39:43 PDT 2017


> poky is generated using combo-layer, which effectively cherry-picks
commits from the parent repositories (in this case bitbake,
> oe-core, meta-poky, meta-yocto-bsp).  This explains why the SHAs don't
match, as the history doesn't match.

This is exactly my original point. We should have here original repos'
SHAs. Not "modified" ones. The "modified" note should be included at the
end of this report (to understand the differences explained by you). For
the better understanding from where/which repos' commits/clones are coming,
and why.

Thank you,
Zoran

On Mon, Oct 23, 2017 at 11:25 AM, Burton, Ross <ross.burton at intel.com>
wrote:

> On 23 October 2017 at 10:08, Zoran Stojsavljevic <
> zoran.stojsavljevic at gmail.com> wrote:
>
>> > Or better yet use
>> > http://www.yoctoproject.org/docs/2.3.2/mega-manual/mega-ma
>> nual.html#ref-classes-image-buildinfo
>>
>> Since I am also (relatively) new to YOCTO, reading these emails makes me
>> think about dependencies between YOCTO distros and layers within distros.
>>
>> Looking what is written here, I decided to do couple of the tests on my
>> own to try to understand better how OpenEmbedded (oe-core) base layer and
>> Poky distro (with two layers on the top of oe-core) are correlated with
>> each other. As my best understanding, poky master tree should do nothing
>> more that follow dependencies (with two additional layers) on the
>> OpenEmbedded master tree.
>>
>> OpenEmbedded layer is YOCTO QEMU project on its own. And one can build
>> well known QEMU targets on single oe-core.
>>
>> In order to verify this, i did git clone (at the same time) on two
>> distros: oe-core master tree, and on poky master tree.
>>
>> Now, after building oe-core, I have the following in qemux86-64 target
>> /etc/build:
>>
>> -----------------------
>> Build Configuration:  |
>> -----------------------
>> DISTRO = nodistro
>> DISTRO_VERSION = nodistro.0
>> -----------------------
>> Layer Revisions:      |
>> -----------------------
>>
>> *meta              = master:3b413a80578caacd9a7f405f3c51a3921d78a60d*
>>
>> At the same time, for poky, I got the following (qemux86-64 target
>> /etc/build):
>>
>> -----------------------
>> Build Configuration:  |
>> -----------------------
>> DISTRO = poky
>> DISTRO_VERSION = 2.4
>> -----------------------
>> Layer Revisions:      |
>> -----------------------
>> *meta                  = rocko:65d23bd7986615fdfb0f1717b615534a2a14ab80
>> -- modified*
>> meta-poky         = rocko:65d23bd7986615fdfb0f1717b615534a2a14ab80 --
>> modified
>> meta-yocto-bsp = rocko:65d23bd7986615fdfb0f1717b615534a2a14ab80 --
>> modified
>> _______
>>
>> Now, the question: should meta - rocko and meta - master have (in *BOLD*)
>> the SAME master HEAD #???
>>
>
> No, they're different repositories so have different SHAs.  However:
>
> $ git show 65d23bd7986615fdfb0f1717b615534a2a14ab80
> commit 65d23bd7986615fdfb0f1717b615534a2a14ab80
> Author: Randy MacLeod <Randy.MacLeod at windriver.com>
> Date:   Mon Oct 16 11:00:44 2017 -0400
>
>     README.qemu: qemuppc64 is not supported
>
>     (From OE-Core rev: 3b413a80578caacd9a7f405f3c51a3921d78a60d)
>
> They're the same commit.
>
> poky is generated using combo-layer, which effectively cherry-picks
> commits from the parent repositories (in this case bitbake, oe-core,
> meta-poky, meta-yocto-bsp).  This explains why the SHAs don't match, as the
> history doesn't match.
>
> Ross
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.yoctoproject.org/pipermail/yocto/attachments/20171023/9f14655f/attachment.html>


More information about the yocto mailing list