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

Re: Question to all candidates about stable point releases



Anthony Towns <aj@azure.humbug.org.au> writes:
> On Sat, Mar 04, 2006 at 01:02:20PM +0100, Marc 'HE' Brockschmidt wrote:
>> Though Martin 'Joey' Schulze as stable release manager presents lists of
>> packages that are accepted into the next stable point release on a
>> regular basis, they normally are not released "roughly two months after
>> the last update" (which is the official plan).
>> 
>> Do you know why this doesn't work as planned? What would you do to 
>> make regular point releases possible?
> I think the first thing to note is that irregular point releases aren't
> a big deal -- since they are almost solely security updates that are
> already available via security.debian.org, and TTBOMK there haven't
> been any installer updates to either woody or sarge.

They're not really important as technical upgrade, but they give the
impression that Debian releases something - after our last painful
release cycle, many people have decided not to use Debian because of our
(seemingly) slow development. Stable point releases are a nice touch to
get a bit of trust back.

> That means point releases get a fairly low priority when other things
> are going on, and if those are the only concerns, that's what it
> *should* mean.

I have to admit that I don't know what other stuff was going on in the
last 2 years, the ftp-team is not really transparent. But if the
workload is too high for its current members, it might be a good idea to
add other people to the team.

> Anyway, I already partially blogged about what I think will improvement,
>
>    http://azure.humbug.org.au/~aj/blog/2005/11/26#2005-11-26-niv2
>
> which is to change the queue structure so that uploads don't enter
> proposed-updates until approved by the SRM. That allows rejections to
> happen immediately, rather than only when a point release happens,
> and reduces the work that has to happen at a point release pretty
> significantly too.

After following the thread on here on -vote, I have the impression that
this fixes something that's not a problem - as it doesn't reduce the
work needed to be done by the ftp-team, which seems to be the current
bottleneck.

Marc
-- 
Fachbegriffe der Informatik - Einfach erklärt
25: Multithreaded
       Wir mußten ein Flußdiagramm malen, um es zu debuggen. (Kristian Köhntopp)

Attachment: pgpxpfvAdlv1_.pgp
Description: PGP signature


Reply to: