Re: [gentoo-dev] [PATCH] acct-user.eclass: Support var overrides for user properties

[Patrick McLean]

On Tue, 05 Jan 2021 00:54:58 +0100
Michał Górny <mgorny@g.o> wrote:

> On Mon, 2021-01-04 at 15:50 -0800, Patrick McLean wrote:
> > On Tue, 05 Jan 2021 00:16:49 +0100
> > Michał Górny <mgorny@g.o> wrote:
> > > On Mon, 2021-01-04 at 14:58 -0800, Patrick McLean wrote:
> > > > On Mon,  4 Jan 2021 18:08:02 +0100
> > > > Michał Górny <mgorny@g.o> wrote:
> > > > > Introduce a few variables to allow easy overrides of common user account
> > > > > proprerties, that is:
> > > > >
> > > > > - ACCT_USER_<username>_SHELL
> > > > > - ACCT_USER_<username>_HOME
> > > > > - ACCT_USER_<username>_HOME_OWNER
> > > > > - ACCT_USER_<username>_HOME_PERMS
> > > > > - ACCT_USER_<username>_GROUPS
> > > > > - ACCT_USER_<username>_GROUPS_ADD
> > > >
> > > > Please also add a way to override the UID/GID for the user/group.
> > >
> > > Damn it, and I thought I'd avoid that ;-).  But do we really need it?
> > > The eclass doesn't enforce UID/GID by default if the user exists
> > > already, so it's a bit tangential to the original problem.
> > >
> >
> > The user needs to already exist for that to be helpful. When one using
> > automation to build/deploy large numbers of Gentoo systems, it's quite
> > useful to have control over that sort of things. At the moment, the
> > only way is to fork the ebuilds, which of course means they need to be
> > kept in sync.
>
> Ok, I'll keep that mind.  However, I suppose you won't mind me
> addressing that separately?  Unlike the patch sent, ID-related logic
> needs to be done twice (due to pkg_pretend).  Ideally, could you report
> a feature request on Bugzilla?

Sure, I don't mind it being addressed separately. I created a feature
request on Bugzilla: https://bugs.gentoo.org/763615