Re: [PATCH 1/1] shadow: upgrade from 4.1.4.3 to 4.2.1

[Richard Purdie <richard.purdie@linuxfoundation.org>]

On Mon, 2014-06-30 at 14:48 +0800, Chen Qi wrote:
> Upgrade shadow from 4.1.4.3 to 4.2.1.
> 
> Changes during this upgrade are as following.
> 
> 1. Remove the "merged" patches. These patches are either merged or
>    the same functionality has been implemented upstream.
> 
>    add_root_cmd_groupmems.patch
>    add_root_cmd_options.patch
>    fix-etc-gshadow-reading.patch
>    shadow-4.1.4.2-env-reset-keep-locale.patch
>    shadow-4.1.4.2-groupmod-pam-check.patch
>    shadow-4.1.4.2-su_no_sanitize_env.patch
>    shadow.automake-1.11.patch
>    shadow_fix_for_automake-1.12.patch
>    useradd.patch
> 
> 2. Remove the unneeded patch.
>    The following patch has been removed because the logic in the related
>    codes of the new version has been changed. In specific, the codes now
>    can handle the 'NULL' return value. So there's no need for the following
>    patch.
> 
>    slackware_fix_for_glib-2.17_crypt.patch
> 
> 3. Teak the current patch to match the new version.
> 
>    allow-for-setting-password-in-clear-text.patch
> 
> 4. Add a patch to fix compilation failure.
> 
>    usermod-fix-compilation-failure-with-subids-disabled.patch
> 
> 5. Add a patch to fix the installation failure.
> 
>    fix-installation-failure-with-subids-disabled.patch
> 
> 5. Add a patch to fix the failure at rootfs time if extrausers is inherited.
> 
>    commonio.c-fix-unexpected-open-failure-in-chroot-env.patch
> 
> 6. Fix the bad section in the recipe.
> 
> 7. Disable the new subids feature in the new version as it doesn't support
>    cross compilation for now.
> 
> 8. Modify the pkg_postinst to `exit 1' if the `pwconv' or `grpconv' fails.
>    Also, fix the arguments to use '--root $D' instead of '--root=$D'.

I think this introduces the following issue:

https://autobuilder.yoctoproject.org/main/builders/nightly-qa-logrotate/builds/154/steps/BuildImages/logs/stdio

since the shutdown group is probably now installed in the core?

Cheers,

Richard