Openembedded Core Discussions
 help / color / mirror / Atom feed
From: Richard Purdie <richard.purdie@linuxfoundation.org>
To: "David Nyström" <david.c.nystrom@gmail.com>
Cc: openembedded-core@lists.openembedded.org
Subject: Re: [PATCH] shadow: Create recipe nativesdk-shadow
Date: Wed, 25 Sep 2013 11:03:20 +0100	[thread overview]
Message-ID: <1380103400.18603.318.camel@ted> (raw)
In-Reply-To: <5241A215.4020501@gmail.com>

On Tue, 2013-09-24 at 16:30 +0200, David Nyström wrote:
> I'm have a question about this though.
> 
> diff --git a/meta/recipes-extended/shadow/shadow.inc 
> b/meta/recipes-extended/shadow/shadow.inc
> index 4df5e5e..75b0afc 100644
> --- a/meta/recipes-extended/shadow/shadow.inc
> +++ b/meta/recipes-extended/shadow/shadow.inc
> <snip>
> +pkg_postinst_${PN} () {
> +	if [ "x$D" != "x" ]; then
> +	  rootarg="--root=$D"
> +	else
> +	  rootarg=""
> +	fi
> +
> +	pwconv $rootarg
> +	grpconv $rootarg
> +}
> <snip>
> 
> This will introduce the postinstall hook for both ${BPN}-native and 
> ${BPN} ?. Before the change, the postinstall hook was activated only by 
> ${BPN} afaict. Tried removing it from native* through 
> pkg_postinst_${BPN}, but that does not seem to work.

Basically the postinst is ignored for the -native version. It never gets
packaged so isn't relevant. Adding in special cases for native for each
and every recipe would get ugly very quickly.

> Hmm, slightly off-topic, but when looking at the postinstalls:
> 
> shadow:
> cat /tmp/opkg-extract-574/CONTROL/postinst
> 	if [ "x$D" != "x" ]; then
> 	  rootarg="--root=$D"
> 	else
> 	  rootarg=""
> 	fi
> 
> 	pwconv $rootarg
> 	grpconv $rootarg
> 	update-alternatives --install /usr/bin/passwd passwd 
> /usr/bin/passwd.shadow 200
> 	update-alternatives --install /usr/bin/chfn chfn /usr/bin/chfn.shadow 200
> 	update-alternatives --install /usr/bin/newgrp newgrp 
> /usr/bin/newgrp.shadow 200
> 
> Above is not a problem since OPKG_OFFLINE_ROOT in update-alternatives 
> can be used to redirect symlink creation when creating a from a package 
> repo.
> 
> nativesdk-shadow:
> if [ "x$D" != "x" ]; then
> 	  rootarg="--root=$D"
> 	else
> 	  rootarg=""
> 	fi
> 
> 	pwconv $rootarg
> 	grpconv $rootarg
> 	update-alternatives --install 
> /opt/poky/1.4+snapshot/sysroots/x86_64-pokysdk-linux/usr/bin/passwd 
> passwd 
> /opt/poky/1.4+snapshot/sysroots/x86_64-pokysdk-linux/usr/bin/passwd.shadow 
> 200
> 	update-alternatives --install 
> /opt/poky/1.4+snapshot/sysroots/x86_64-pokysdk-linux/usr/bin/chfn chfn 
> /opt/poky/1.4+snapshot/sysroots/x86_64-pokysdk-linux/usr/bin/chfn.shadow 200
> 	update-alternatives --install 
> /opt/poky/1.4+snapshot/sysroots/x86_64-pokysdk-linux/usr/bin/newgrp 
> newgrp 
> /opt/poky/1.4+snapshot/sysroots/x86_64-pokysdk-linux/usr/bin/newgrp.shadow 
> 200
> 	update-alternatives --install 
> /opt/poky/1.4+snapshot/sysroots/x86_64-pokysdk-linux/usr/bin/chsh chsh 
> /opt/poky/1.4+snapshot/sysroots/x86_64-pokysdk-linux/usr/bin/chsh.shadow 200
> 
> Could we not add a "--root" option to update-alternatives like done to 
> update-rc.d instead of hardcoding paths or using envs.

nativesdk are built to run in that specific path. We're therefore not
installing them to some subdirectory where a root makes sense, the
prefix really was /opt/xxx. We then manipulate things later to make them
more relocatable.

> 
> Also, seems like ${sysconfdir} in the nativesdk-opkg postinstall expands to:
> 
> chmod 0755 $D${sysconfdir}/rcS.d/S${POSTINSTALL_INITPOSITION}run-postinsts
> chmod 0755 
> $D/opt/poky/1.4+snapshot/sysroots/x86_64-pokysdk-linux/etc/rcS.d/S98run-postinsts
> 
> Is this really expected behaviour ?

nativesdk postinstalls are probably badly handled at the moment. I
suspect if things can run at rootfs creation time they work out, if they
can't run then, they never happen at all. We don't have a "first boot"
of the sdk...

Cheers,

Richard



  reply	other threads:[~2013-09-25 10:03 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-09-23 16:33 [PATCHv3] makedevs: add nativesdk to BBCLASSEXTENDS David Nyström
2013-09-23 16:34 ` [PATCH] shadow: Create recipe nativesdk-shadow David Nyström
2013-09-23 18:01   ` Mark Hatle
2013-09-23 19:01     ` David Nystrom
2013-09-23 21:14       ` Richard Purdie
2013-09-24  8:25         ` Richard Purdie
2013-09-24 14:30           ` David Nyström
2013-09-25 10:03             ` Richard Purdie [this message]
2013-09-25 12:24               ` David Nyström
2013-09-23 16:34 ` [PATCHv2] libpam: Avoid host contamination issue w. libprelude David Nyström

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=1380103400.18603.318.camel@ted \
    --to=richard.purdie@linuxfoundation.org \
    --cc=david.c.nystrom@gmail.com \
    --cc=openembedded-core@lists.openembedded.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox