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

Botched mass bug filing (Re: Packaging _still_ wasteful for many large packages)

On Wed, Feb 18, 2004 at 03:46:13PM -0500, Joey Hess wrote:

> It's interesting that this entire thread neglects the issue of Packages
> file bloat. If we do eventually get a reorganised archive that allows
> per-architecture partial mirroring, then I think that limiting sizes of
> Packages files and avoiding unnecessary complexity (such as -doc or
> -common packages) will perhaps be more important than saving the minimal
> amount of archive space some of the packages on the posted lists take
> up. Where's the dividing line?

In the beginning of the thread, the proposal was to only consider larger
binary packages (.deb >= 10MB), of which there are only about 150 total in
the archive (and surely not all of these meet the other criteria).  He even
posted a list:


which only has 14 source packages in it, for an upper bound of 14 wishlist
bugs.  This seemed like a perfectly reasonable proposition to me, and it
received some support from other members of the mailing list as well.  I
think that the archive size savings in that scenario were well worth the
potential Packages file bloat (surely less than 1k in Packages.gz).

Then, DURING THE PROCESS of filing these bugs, and without ANY further
discussion, Steve changed his mind:


completely changing the criteria, resulting in hundreds of bugs being filed,
against packages like the ~1MB moon-lander.  The bugs were also incorrectly
filed with Severity: normal, and sent to submit@ rather than maintonly@.

 - mdz

Reply to: