From: Harald Seiler <hws@denx.de>
To: Jan Kiszka <jan.kiszka@siemens.com>,
"Maxim Yu. Osipov" <mosipov@ilbers.de>,
isar-users@googlegroups.com
Subject: Re: DISTRO is set too late
Date: Tue, 30 Oct 2018 16:56:47 +0100 [thread overview]
Message-ID: <b8b6099002429f0b6c580954f555197f35e70fc0.camel@denx.de> (raw)
In-Reply-To: <30bb5a44-53ee-90fd-7740-27111ff01b83@siemens.com>
On Thu, 2018-10-25 at 14:21 +0100, Jan Kiszka wrote:
> On 23.10.18 13:37, Harald Seiler wrote:
> > On Tue, 2018-10-23 at 14:34 +0300, Maxim Yu. Osipov wrote:
> > > On 10/23/18 1:10 PM, Harald Seiler wrote:
> > > > On Tue, 2018-10-23 at 12:48 +0300, Maxim Yu. Osipov wrote:
> > > > > Hi Harald,
> > > > >
> > > > > The list of supported distros is under isar/meta-isar/conf/distro/*.conf
> > > > > (debian-wheezy is officially not supported but still present for
> > > > > convenience).
> > > >
> > > > We are using `debian-stretch`. If I add
> > > >
> > > > DISTRO = "debian-stretch"
> > > >
> > > > to `local.conf`, I can build. But this should not be necessary as
> > > > $DISTRO is also set from multiconf. The issue is that this happens
> > > > too late, so when $DISTRO_SUITE is expanded, it hasn't happened yet.
> > > >
> > > > > How did you get the value `poky` for DISTRO in your environment?
> > > >
> > > > Hmm, apparently this is kas adding
> > > >
> > > > DISTRO ??= "poky"
> > > >
> > > > to `local.conf` if no distro is supplied in its config. This should,
> > > > again, not be an issue, because the multiconf sets
> > > >
> > > > DISTRO ?= "debian-stretch"
> > > >
> > > > What seems to go wrong is that the multiconf is read too late, so
> > > > that base-apt-helper.bb tries to expand $DISTRO before is has gotten
> > > > its final value. I don't think this is the intended behavior ...
> > > >
> > >
> > > I don't use kas.
> > >
> > > I've ran `bitbake -e multiconfig:qemuarm-stretch:isar-image-base` and
> > > DISTRO/DISTRO_SUITE are expanded properly:
> > >
> > > # $DISTRO [2 operations]
> > > # set /home/myo/work/isar/src/trunk/isar/build/conf/local.conf:30
> > > # [_defaultval] "debian-stretch"
> > > # set?
> > > /home/myo/work/isar/src/trunk/isar/meta-isar/conf/multiconfig/qemuarm-stretch.conf:6
> > > # "debian-stretch"
> > > # pre-expansion value:
> > > # "debian-stretch"
> > > DISTRO="debian-stretch"
> > >
> > > ...
> > >
> > > #
> > > # $DISTRO_SUITE
> > > # set?
> > > /home/myo/work/isar/src/trunk/isar/meta/classes/base-apt-helper.bbclass:5
> > > # "${@ d.getVar('DISTRO', True).split('-')[1]}"
> > > DISTRO_SUITE="stretch"
> > >
> > >
> > >
> > > If I set in local.conf
> > >
> > > DISTRO ??= "debian-jessie"
> > >
> > > DISTRO/DISTRO_SUITE are also expanded properly:
> > >
> > > # $DISTRO [2 operations]
> > > # set /home/myo/work/isar/src/trunk/isar/build/conf/local.conf:30
> > > # [_defaultval] "debian-jessie"
> > > # set?
> > > /home/myo/work/isar/src/trunk/isar/meta-isar/conf/multiconfig/qemuarm-stretch.conf:6
> > > # "debian-stretch"
> > > # pre-expansion value:
> > > # "debian-stretch"
> > > DISTRO="debian-stretch"
> > >
> > > ...
> > >
> > > #
> > > # $DISTRO_SUITE
> > > # set?
> > > /home/myo/work/isar/src/trunk/isar/meta/classes/base-apt-helper.bbclass:5
> > > # "${@ d.getVar('DISTRO', True).split('-')[1]}"
> > > DISTRO_SUITE="stretch"
> >
> > I did a little more testing. Apparently, it works fine if the weak default value
> > contains at least one `-`. The rest of the value can be anything (though I do get
> > some warnings if it is not a valid distro).
> >
> > So this works:
> >
> > DISTRO ??= "invalid-distro"
> >
> > But this doesn't:
> >
> > DISTRO ??= "invalid_distro"
> >
> >
>
> I didn't dig into all details, and maybe this is actually resolving the issue
> better, but I had to fix a similar issue with my kas-driven layer lik this:
> https://github.com/siemens/jailhouse-images/commit/e595c6d62ddc6e8a608b199dc767dab9bc12f216
I am also able to reproduce this behavior with poky/sumo, so this is a bitbake issue and
not directly isar related.
I am unsure if it makes sense to follow up on this; this bug will only ever be triggered
if an invalid default value is provided, which is by itself a "bug" because a default
value that is supposed to be overwritten shouldn't be set in the first place ...
What do you think?
--
Harald
DENX Software Engineering GmbH, Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-62 Fax: +49-8142-66989-80 Email: hws@denx.de
prev parent reply other threads:[~2018-10-30 15:56 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-10-23 9:05 Harald Seiler
2018-10-23 9:48 ` Maxim Yu. Osipov
2018-10-23 10:10 ` Harald Seiler
2018-10-23 11:34 ` Maxim Yu. Osipov
2018-10-23 12:37 ` Harald Seiler
2018-10-25 13:21 ` Jan Kiszka
2018-10-30 15:56 ` Harald Seiler [this message]
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=b8b6099002429f0b6c580954f555197f35e70fc0.camel@denx.de \
--to=hws@denx.de \
--cc=isar-users@googlegroups.com \
--cc=jan.kiszka@siemens.com \
--cc=mosipov@ilbers.de \
/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