Re: Removing packages perhaps too aggressively?

[Abou Al Montacir <abou.almontacir@sfr.fr>]

On Wed, 2018-01-31 at 23:00 +0000, Scott Kitterman wrote:

On January 31, 2018 10:34:28 PM UTC, Michael Biebl <email@michaelbiebl.de> wrote:

Am 31.01.2018 um 22:49 schrieb Don Armstrong:

On Wed, 31 Jan 2018, Abou Al Montacir wrote:

Me too likes to extend the removal notice for few weeks/months.
Especially removal from testing when outside freeze periods.


Packages removed from testing outside of the freeze can be easily
re-added to testing once the underlying RC bugs are fixed. So RMs

should

continue to remove early, and remove often. [When this has happened

with

my packages (see lilypond), it's resulted in more people helping with
the maintenance of them, and brought some issues to a wider

audience.]




I agree. Removals from testing should have no artifical delay. Removals
from the archive (i.e. unstable), a two or four week courtesy delay
seems ok to me, giving the person listed in Maintainers a chance to
reply, seems ok.


So far, every time this comes up, there's no actual volunteer to invest the time to update the removals page to make this reasonable to do in practice.

I think some normal delay is reasonable, but it needs to be integrated into the pending removals page so the FTP team member processing removals gets an indication the request is new.

In general I agree with this as a DD, but when I wear my user hat I don't.

It happened multiple times that I need to install a package that is not widely used (last one I remember was spyder, a Python editor) and it was not in stable, because of FTBFS caused by other packages update! I was obliged to take it from sid and "update" some libs from "unstable" to make my "stable" system install it. This is not user friendly.

Of course here I'm not discussing the pertinence of FTBFS bugs, but saying that sometimes some packages get removed as collateral damage of igration of othe ones especially if the package maintainer does not react because busy.

I think this kind of case need to be discussed and the situation improved, but I don't have a strong mind about what to do with it.

-- 
Cheers,
Abou Al Montacir

Attachment: signature.asc
Description: This is a digitally signed message part