From: Gunnar Andersson <gandersson@genivi.org>
To: yocto@yoctoproject.org
Cc: "automotive-discussions@lists.linuxfoundation.org"
<automotive-discussions@lists.linuxfoundation.org>
Subject: Idea for bbappend / layer organization and naming convention
Date: Mon, 14 Aug 2017 21:36:15 +0200 [thread overview]
Message-ID: <1502739375.26251.126.camel@genivi.org> (raw)
Yocto community,
Triggered by the previous email request to yocto@ about best practices for
layer organization I'm finally firing off this email for (hopefully) some
feedback on an idea we first kicked around, discussed for rough consensus
and then decided to implement in the Yocto based GDP project [1]. You can
see it as a trial, anything can be changed, but so far so good...
We adopted a somewhat novel (but actually not really unique, see inside)
naming convention [2] for all modifications to components that belong to
other layers.
This convention shows what layers are being modified by the .bbappends by
actually naming the layer it is (intending to**) modify. The initiative
also aims to highlight and separate .bbappends (modifications) from uniquely
added components in our project (new .bb files) in GDP.
**Note that we are well aware this does NOT control bitbake behavior (as
neither does the recipe-xxx/ directory convention), and is therefore just a
guide. If the project is misconfigured, bitbake could do something
different than the naming convention implies and it could even be
misleading, but at least the *programmer intent* is shown clearly.
I think it also makes .bbappends more explicitly visible and might trigger
the idea: Do we really need this? Are we really modifying *that* adopted
layer, and if so, why?
To Russel who wrote the previous request for guidance - feel free to
consider, but I won't recommend this directly since it's by no means an
agreed best practice in all of the community. I'm rather asking for other
experienced OE developers to consider and comment on the idea.
For oldtimers it's probably easy to find this odd and just be negative
against it since well, "standard practice", inertia, NIH and all that, but
I'll stick my neck out anyway because it seems to solve some issues of
organization and understanding at least for us (and I'm a sucker for some
flamebait ;)
If the intro is too formal and long, just look at the final examples. I
think you'll get it. Feel free to ask.
- Gunnar
--
Gunnar Andersson <gandersson@genivi.org>
Development Lead
GENIVI Alliance
[1] https://github.com/GENIVI/genivi-dev-platform
[2] Naming strategy for bitbake extension layers:
https://at.projects.genivi.org/wiki/x/w4Xk
next reply other threads:[~2017-08-14 19:37 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-08-14 19:36 Gunnar Andersson [this message]
2017-08-15 3:02 ` Idea for bbappend / layer organization and naming convention Bruce Ashfield
2017-08-18 14:46 ` Gunnar Andersson
2017-08-18 18:47 ` Bruce Ashfield
2017-08-18 19:35 ` Mark Hatle
2017-08-20 19:02 ` Stephane Desneux
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=1502739375.26251.126.camel@genivi.org \
--to=gandersson@genivi.org \
--cc=automotive-discussions@lists.linuxfoundation.org \
--cc=yocto@yoctoproject.org \
/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.