Re: su in debian.rules

["brian (b.c.) white" <bcwhite@bnr.ca>]

>>> The issue is that *I* consider, for *myself*, the need to become
>>> root to build a package to be a *serious* impediment.
>> I have one problem with this.  Since you need to be root in order to
>> install packages, why does it matter that you have to give root
>> access to anybody developing packages?  If you are not actually
>> installing the packages you create (even just as a test), then why
>> would anybody else even consider installing them?  I'd like to
>> believe that any package I install has at least been tested by the
>> person who created it.
>
>You are jumping to some far-fetched conclusions.  They aren't
>suggesting that packages not be tested.  I think the idea is that it
>is good to minimize the number of tasks that must be performed as
>root.

Yes, I realize that it has to be tested.  My thought was that since the
first thing that I do after "debian.rules binary" is a "dpkg --install",
then it makes little difference whether the binary is built as root
or not since it required being root immediately thereafter, anyway.

>Accidents always happen but they can be much less destructive if you
>are not root at the time.  For instance, a stupid mistake in one of my
>debian.rules files once caused the deletion of / instead of
>./debian-tmp.  :-)

Oops!  Okay.  I can see your reasoning now.

<leaves quietly remembering a "rm -rf .*" from /tmp>

                                        Brian
                                 ( bcwhite@bnr.ca )

-------------------------------------------------------------------------------
    In theory, theory and practice are the same.  In practice, they're not.