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

Re: "Automatic" backports

On Thu, 29 Sep 2011 09:45:42 +0200, Mike Hommey wrote:

Hi Mike.

I'm one of those people believing that packages that do grant a backport
shouldn't require modification to do so. As such, all the packages I
backport have no modifications in their source compared to the version
in testing/unstable, besides the obvious debian/changelog change.

Sadly for you, I'm not of one those people.

This means such packages could just automatically flow into backports. I
must say it would make my life easier if it were the case ; I happen
to regularly forget to push security updates to lenny-backports.

For Security updates, I completely understand this is a problem, but when somebody performs a backport, he has to maintain the backport and work with the maintainers.

In some cases, this is not done, but I don't think it's a reason for generalizing "automated backports". I think putting notifications to the backport maintainer could be better (based on the diffstats page), but there is no actual magical solution for this.

I think it's better to warn and educate people about backporting is not only "edit debian/changelog; debuild...; send a mail to backport ML, and it's done", like Alexander Wirt did a week ago with Benjamin Sonntag.

What do the backports people think about that?

Also, I don't agree with the idea of automatic backports because some packages may not need a backport. And others packages may cause stability problems to Debian Stable.

If people want to use all packages from Debian Testing, they can upgrade to Testing, instead of asking for backports.



PS: Please Cc me, I'm not subscribed.

By the way, I CC'ed you, but you can subscribe to this mailing list, you will have a better view on what exactly is the backport process.

Cyril "Davromaniak" Lavier

Reply to: