[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: An old idea, brought back to life

On Sun, Dec 20, 1998 at 02:18:44PM -0500, Raul Miller wrote:
> The following would be very good as documentation (orientation,
> suggestions, whatever) for new users.  However, a few nits:  this is
> about systems, not about people.  Also it's a spectrum -- someone might
> have a system which runs unstable till they get burned, then back off
> for a while till they need some wholesale change from unstable again.
> And some who need stability might even be reluctant to upgrade to a more
> recent "stable" -- especially if they have significant locally installed
> software which isn't properly registered in dpkg's database.

Face it Raul, you cannot be classified.  ;>

Obviously those who need "mostly stable" will use stable with select
packages from elsewhere and that sort of thing.  These aren't hard
classifications but are rather a good indication of who the dists in this
system would be catering to.

> Personally, I've got systems in most of these categories (except for
> the rabid side of bleeding edge).

It does vary by the machine and by the person.  FWIW, I have to admit it
makes some changes (major libc type for example) a little harder because
we have to plan the changes more because of the enforced compatibility
with stable.  But I think this is probably a good thing considering we
don't do this sometimes now.

> > Formal testing is not something Debian can do reliably.  Unless you're
> > volunteering to be the one to test a good portion of those 2700 packages
> > I don't see how we can test any better than we do now.
> Eh?  There's a lot that can be done with `make test`.  This should happen
> before the .deb is even generated.  If we had a few test machines we
> could do some basic configuration testing as well (like: install it
> on on a minimal system and check for basic functionality, install it
> on a maxed out system and check for basic functionality).  We'd need a
> debian script to verify functionality on a per-package basis, and we'd
> initially have to punt on a lot of things, but this is doable.

Perhaps, but you can't give the program a full workover and be sure the
thing is going to work as advertised.  I mean I've had a few epic bugs
that I've found before anyone else because I use the program that much. 
Usually they're my fault, but not always.  =>

> > This should become easier under the new system, but there are always
> > going to be problems in anything but stable.
> Agreed.  But there's still lots of room for improvement.

This is just one suggestion for improvement.  It may or may not be a good
one.  If it is considered to be a good one by others, we should see if it
can be implemented (archive managers willing of course..)

If this doesn't fix the problems or makes more than it fixes we should of
course find a better solution and hold off doing anything until we have

NO ONE expects the Spanish Inquisition!

Reply to: