Re: Bug #32888: The old `base' package.
On Tue, 9 Mar 1999, Branden Robinson wrote:
> On Tue, Mar 09, 1999 at 07:55:03PM +0100, Santiago Vila wrote:
> > I'm trying to understand what kind of harm does the existence of the base
> > package which has to be considered as a "bug".
> > Is there a technical reason why we should make "base" to disappear
> > from old systems?
> > Is there any reason to make "base" to disappear other than aesthetical?
> I'm trying to understand what kind of harm the existence of the old,
> obsolete X font and static library packages cause which necessitates
> their consideration as "bugs".
IMHO, both things (making base to disappear, and making the X upgrade to
be fully automatic, without selecting anything by hand) fall, at least,
into the category of "things that people usually want to happen".
Of course, that people want something to happen is not automatically a bug
by itself, we have to consider whether it is reasonable that it actually
happens or not, and whether we have a solution for it or not. BTW: if
something is a bug, but we don't have a fix, it does not automatically
make it a non-bug, but we can downgrade it to wishlist if it is "very
difficult to fix due to major design considerations", according to the
definition of "wishlist").
IMHO, it is reasonable to expect the X packages to be upgraded
automatically because it is what dselect does by default (i.e. upgrade
everything). Only this way we can be sure that possible bugs in these
packages are actually fixed when the user updates his/her system.
If the package remain "obsolete", this will not happen.
The main difference I see between "base" and the X packages is that base
may not automagically disappear by using current methods, so this is
something that we probably can't fix, while the X packages may indeed be
upgraded automatically by current methods, without using any dirty hack
not allowed by the packaging system, like fiddling with dpkg's database
[ If I'm not mistaken, you said that you would fix this problem "in one
way or another", and I believed you, but apparently you have closed the
bug (#30852) without fixing it at all. Now we can talk about "credibility"
if you like, but I would much prefer to talk exclusively about technical
issues in this list ].
"7be24e4ae60507b0985712004c901c28" (a truly random sig)