From: Henning Schild <henning.schild@siemens.com>
To: "Dalamagkidis,
Konstantinos (MO RS EN CCIT TEC AR)"
<konstantinos.dalamagkidis@siemens.com>
Cc: "Kiszka, Jan (CT RDA IOT SES-DE)" <jan.kiszka@siemens.com>,
"isar-users@googlegroups.com" <isar-users@googlegroups.com>
Subject: Re: [PATCH] Add support for supplying more types of debian package relationships
Date: Mon, 14 Oct 2019 19:01:00 +0200 [thread overview]
Message-ID: <20191014190100.329e37e5@md1za8fc.ad001.siemens.net> (raw)
In-Reply-To: <AM0PR10MB3652D02A1279556378533986E3B10@AM0PR10MB3652.EURPRD10.PROD.OUTLOOK.COM>
Am Wed, 11 Sep 2019 11:32:08 +0200
schrieb "Dalamagkidis, Konstantinos (MO RS EN CCIT TEC AR)"
<konstantinos.dalamagkidis@siemens.com>:
> > On 11.09.19 10:41, [ext] Dalamagkidis, Konstantinos wrote:
> > > We need to be able to specify other package relationships, such as
> > > Pre-Depends or Provides/Replaces/Conflicts.
> > >
> > > This change supports supplying DEBIAN_PREDEPENDS, DEBIAN_REPLACES,
> > > DEBIAN_CONFLICTS and DEBIAN_PROVIDES, as well as bitbake-style
> > > RDEPENDS or RDEPENDS_{PN} variables. The DEBIAN_* take priority.
> > >
> >
> > This, as well as your DPKG_ARCH patch, is enhancing and further
> > establishing debianize.bbclase as an official interface. I already
> > urged Henning to do that - which includes writing documentation for
> > this. Your change here should extend that (not yet existing)
> > documentation as well.
> >
> > One of the thing to consider while designing & enhancing the recipe
> > API for debianize: We should not create something that is as
> > complex as writing Debian control files directly - or even more
> > complex than that.
> >
> > BTW, what kind of application packages are you generating for
> > dpkg-raw?
>
> We have ca 20 packages that use dpkg-raw, split more or less
> equally between configuration, scripts and some pre-built
> binaries. The conflicts/replaces is used in a couple of
> configuration packages. We are also porting recipes from a
> previous version that was using "plain yocto", so it is easier if
> we can just use RDEPENDS, RPROVIDES, etc. Personally I would
> prefer to use dpkg over dpkg-raw everywhere, but other colleagues
> find dpkg-raw easier.
My suggestion here is to keep that patch in your layer. You having that
problem 20 times does not make it common, 20 independent users would ;).
So you could put that magic into a dpkg-my-layer-yocto-glue class. The
deps in Isar are a pain, because we are dealing with two worlds. I
personally do not think that "pretending this away" is a good idea.
If you actually have configure + compile tasks in your dpkg-raw you are
on the wrong track and should stop pretending its yocto! Or at least
make sure all that gets done in buildchroot. If you want please explain
what you are doing, like Jan asked before.
Henning
> Rgds
> Konstantinos
next prev parent reply other threads:[~2019-10-14 17:01 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-09-11 8:41 Dalamagkidis, Konstantinos
2019-09-11 8:57 ` Jan Kiszka
2019-09-11 9:02 ` Jan Kiszka
2019-09-11 9:32 ` Dalamagkidis, Konstantinos
2019-10-14 17:01 ` Henning Schild [this message]
2019-10-15 10:38 ` Baurzhan Ismagulov
2019-09-13 8:32 ` Baurzhan Ismagulov
2019-10-14 16:49 ` Henning Schild
2019-09-17 9:07 ` Henning Schild
2019-10-08 11:01 ` Baurzhan Ismagulov
2019-10-08 11:26 ` Jan Kiszka
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=20191014190100.329e37e5@md1za8fc.ad001.siemens.net \
--to=henning.schild@siemens.com \
--cc=isar-users@googlegroups.com \
--cc=jan.kiszka@siemens.com \
--cc=konstantinos.dalamagkidis@siemens.com \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox