bug 58002

[Rocco Stanzione]

On Friday 08 September 2006 6:14 pm, Matt Zimmerman wrote:
> On Tue, Aug 29, 2006 at 04:45:44PM -0500, Rocco Stanzione wrote:
[snip]
> > My recommendation (hadn't thought of it by the time I decided to wash my
> > hands of the launchpad bug) is to put an empty /etc/skel/.viminfo into
> > the vim package, which would fix this and have no other consequences that
> > I can think of.  Modifying sudoers to always set home was also suggested,
> > but I think that would have some unintended consequences.  I figure the
> > answer to that will determine which package, if any, this bug should be
> > filed against.
>
> This is a workaround, not a fix, and one which I'm not sure is worth the
> tradeoff.  The underlying issue needs to be addressed with a globally
> consistent policy.  I've adjusted the description of the bug accordingly,
> and encourage the QA team to mark similar bugs as duplicates of that one,
> since it needs a general solution.

What is traded off?  I don't see a cost or risk involved with putting an 
empty /etc/skel/.viminfo in the vim package.

> My gut feeling is that we probably want to adjust some of the default
> environment variable settings and sudo's list of preserved variables, to
> preserve the convenience provided by the current setup, and then switch to
> always_set_home. I think this needs serious thought and discussion before
> we agree on a solution, however.

I'd be very hesitant to change the behavior of sudo - it's behaved that way 
for a long time, and I for one have scripts and habits that depend on $HOME 
being preserved when I run sudo without -H.

As far as any far-reaching policy, maybe vim and nano should be handled 
differently than other similar situations.  They are both included in the 
default install, so the first user would benefit from empty /etc/skel 
entries, which (correct me if I'm wrong) would solve the problem with no 
other consequences.  That wouldn't work for a package that has to be 
installed separately.

Rocco