[OE-core] [oe] waning "it is recommended to set SYSTEMD_PACKAGES as <existing-package>-systemd"

[Andreas Müller]

On Tue, Apr 17, 2012 at 5:45 PM, Koen Kooi <koen at dominion.thruhere.net> wrote:
>
> Op 17 apr. 2012, om 17:39 heeft Martin Jansa het volgende geschreven:
>
>> On Tue, Apr 17, 2012 at 05:22:31PM +0200, Andreas Oberritter wrote:
>>> Hi,
>>>
>>> the following warnings keep appearing on every fresh invocation of
>>> bitbake since a few months, I think:
>>>
>>> WARNING: .../meta-oe/recipes-devtools/cloud9/cloud9_0.6.bb: it is recommended to set SYSTEMD_PACKAGES as <existing-package>-systemd
>>> WARNING: .../meta-oe/recipes-connectivity/gateone/gateone_git.bb: it is recommended to set SYSTEMD_PACKAGES as <existing-package>-systemd
>
> The above 2 only have systemd support, so the WARNING is completely bogus and should be silenced.
The warning is created in case the SYSTEMD_PACKAGES = "${PN}" instead
of SYSTEMD_PACKAGES = "${PN}-systemd". Intention was to force the
separation to keep systemd stuff out of images not supporting systemd.
Problem with this: Images need to be changed e.g need to add
gateone-systemd instead of gateone.

In the RFC for systemd.bbclass I suggested to make systemd a disto
feature together with an automatic RDEPENDS for *-systemd packages.
This was commented as no go [1]. The alternate suggestion making
systemd as image feature ended up in nothing.

Another suggestion was made to separate out systemd into an own layer
[2]. Guess this is not in due to release hysteria.

I faced similar situation with gtk-iconcache runonce [3]: In case a
solution is not perfect it simply ends up in nothing although the
current situation is not acceptable.

I lost interest making gurus happy so do what you want with it

Andreas
[1] http://lists.linuxtogo.org/pipermail/openembedded-devel/2012-February/037788.html
[2] http://lists.linuxtogo.org/pipermail/openembedded-devel/2012-April/038932.html
[3] http://lists.linuxtogo.org/pipermail/openembedded-core/2012-March/019760.html