Re: NRSS has been deprecated [#696302]

[Adrian Bunk <bunk@stusta.de>]

On Sun, Oct 30, 2016 at 06:28:41AM +0100, Adam Borowski wrote:
>...
> An user interested in future releases is usually a contributor of sorts,
> thus often has "devscripts" installed.

The typical user of Debian stable is running Debian on servers,
and will become interested in a future release after it became stable.

>...
> Once stretch is released, we'd file thousands of ITRs, listen to the
> crickets for a bit, then greatly decruft the archive!
>...

If you want to reduce the amount of cruft in the archive,
you should look at the other end.

When something that was ITP'ed into Debian is declared cruft after it 
has been shipped in stable releases, the question should be why it was 
allowed into Debian in the first place.

Especially when looking at the quality of much of the stuff from GitHub, 
the 10k binary packages that are new in stretch should be a bigger worry
than packages with a history of not causing problems in past releases.

> (although if popcon is 4 or so, that's probably telling enough).

It is telling us that relying on popcon numbers for anything is harmful.

What popcon numbers do you expect for architecture-specific packages 
that are only relevant for a subset of the supported hardware on a port?

Half of the stretch release architectures have a popcon lower than 25.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed