[gentoo-dev] [PATCH] acct-user.eclass: die explicitly if HOME is missing in preinst

[gentoo-dev] [PATCH] acct-user.eclass: die explicitly if HOME is missing in preinst

[Mike Gilbert]

Bug: https://bugs.gentoo.org/691478
---
 eclass/acct-user.eclass | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/eclass/acct-user.eclass b/eclass/acct-user.eclass
index 60009643c144..008dc6bb7d68 100644
--- a/eclass/acct-user.eclass
+++ b/eclass/acct-user.eclass
@@ -334,6 +334,14 @@ acct-user_pkg_preinst() {
 		if [[ -z ${ACCT_USER_HOME_OWNER} ]]; then
 			ACCT_USER_HOME_OWNER=${ACCT_USER_NAME}:${ACCT_USER_GROUPS[0]}
 		fi
+		# Path might be missing due to INSTALL_MASK, etc.
+		# https://bugs.gentoo.org/691478
+		if [[ ! -e "${ED}/${ACCT_USER_HOME#/}" ]]; then
+			eerror "Home directory is missing from the installation image:"
+			eerror "  ${ACCT_USER_HOME}"
+			eerror "Check INSTALL_MASK for entries that would cause this."
+			die "${ACCT_USER_HOME} does not exist"
+		fi
 		fowners "${ACCT_USER_HOME_OWNER}" "${ACCT_USER_HOME}"
 		fperms "${ACCT_USER_HOME_PERMS}" "${ACCT_USER_HOME}"
 	fi
-- 
2.23.0.rc2

Re: [gentoo-dev] [PATCH] acct-user.eclass: die explicitly if HOME is missing in preinst

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 1122 bytes --]

On Thu, 2019-08-15 at 11:43 -0400, Mike Gilbert wrote:
> Bug: https://bugs.gentoo.org/691478
> ---
>  eclass/acct-user.eclass | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/eclass/acct-user.eclass b/eclass/acct-user.eclass
> index 60009643c144..008dc6bb7d68 100644
> --- a/eclass/acct-user.eclass
> +++ b/eclass/acct-user.eclass
> @@ -334,6 +334,14 @@ acct-user_pkg_preinst() {
>  		if [[ -z ${ACCT_USER_HOME_OWNER} ]]; then
>  			ACCT_USER_HOME_OWNER=${ACCT_USER_NAME}:${ACCT_USER_GROUPS[0]}
>  		fi
> +		# Path might be missing due to INSTALL_MASK, etc.
> +		# https://bugs.gentoo.org/691478
> +		if [[ ! -e "${ED}/${ACCT_USER_HOME#/}" ]]; then
> +			eerror "Home directory is missing from the installation image:"
> +			eerror "  ${ACCT_USER_HOME}"
> +			eerror "Check INSTALL_MASK for entries that would cause this."
> +			die "${ACCT_USER_HOME} does not exist"
> +		fi
>  		fowners "${ACCT_USER_HOME_OWNER}" "${ACCT_USER_HOME}"
>  		fperms "${ACCT_USER_HOME_PERMS}" "${ACCT_USER_HOME}"
>  	fi

LGTM.  Thanks for taking care of it.

-- 
Best regards,
Michał Górny


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 618 bytes --]

Re: [gentoo-dev] [PATCH] acct-user.eclass: die explicitly if HOME is missing in preinst

[Michael Orlitzky]

On 8/15/19 11:43 AM, Mike Gilbert wrote:
> 
> +		# Path might be missing due to INSTALL_MASK, etc.
> +		# https://bugs.gentoo.org/691478
> +		if [[ ! -e "${ED}/${ACCT_USER_HOME#/}" ]]; then
> +			eerror "Home directory is missing from the installation image:"
> +			eerror "  ${ACCT_USER_HOME}"
> +			eerror "Check INSTALL_MASK for entries that would cause this."
> +			die "${ACCT_USER_HOME} does not exist"
> +		fi

This suggests that the problem is with the user's INSTALL_MASK, and not
with the package that crashed. I would rather they file a bug than
change their mask entry. But, I guess it's pretty easy these days to
grep the whole acct-user category for issues like this that I missed on
the mailing list.

Re: [gentoo-dev] [PATCH] acct-user.eclass: die explicitly if HOME is missing in preinst

[Mike Gilbert]

On Thu, Aug 15, 2019 at 3:01 PM Michael Orlitzky <mjo@gentoo.org> wrote:
>
> On 8/15/19 11:43 AM, Mike Gilbert wrote:
> >
> > +             # Path might be missing due to INSTALL_MASK, etc.
> > +             # https://bugs.gentoo.org/691478
> > +             if [[ ! -e "${ED}/${ACCT_USER_HOME#/}" ]]; then
> > +                     eerror "Home directory is missing from the installation image:"
> > +                     eerror "  ${ACCT_USER_HOME}"
> > +                     eerror "Check INSTALL_MASK for entries that would cause this."
> > +                     die "${ACCT_USER_HOME} does not exist"
> > +             fi
>
> This suggests that the problem is with the user's INSTALL_MASK, and not
> with the package that crashed. I would rather they file a bug than
> change their mask entry. But, I guess it's pretty easy these days to
> grep the whole acct-user category for issues like this that I missed on
> the mailing list.
>

The purpose of the message is to give the user a hint on how they may
resolve the issue themselves.