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

Re: Backports vs Pinning Lenny



Apologies for wading in :-p

I can see both sides having important points, possibly based on different ideas on how BPO should be (and actually is) used.

There's probably two main stereotypes for BPO users:

1. The "SysAdmin". The SysAdmin wants to run Debian stable, because it is stable. Inevitably practicalities dictate that some more-modern versions of a few tools are needed because some damned development feature just needs the latest version of java/apache/python/whatever.

The SysAdmin installs < 20 packages from BPO, most are servers or services. SysAdmin probably uses pinning, and begrudges any package being updated that is not exactly necessary. SysAdmin is a little surly ;) SysAdmin is planning a timeline to move to etch some time in 2007. Some time later etch won't be enough for some service and SysAdmin will start adding etch-backports.

2. The "LAG". The "LAG" wants the "Latest-And-Greatest" applications but doesn't want to run testing for some reason (for me it was because I had dial-up and changes to libc that require 0.5Gb download didn't thrill me). LAG is probably a desktop user.

The "LAG" mostly installs applications, and is more likely than the SysAdmin to ever enter the command:
  aptitude -t <release>-backports install gnome
The "LAG" just wanted FF1.5 in sarge. The "LAG" probably updated to etch within a week of it coming out.


sarge-backports:
The LAG is not interested in sarge-backports anymore. The SysAdmin needs sarge-backports, but probably only wants the etch level of packages - ie versions that are in etch, or in security/etch-proposed-updates.

lenny packages in etch-backports:
The LAG is a power-user, but may not be a developer. In the last year before etch was out most packages in testing would not have worked in sarge, so LAG never looked into running packages from testing and may not know how to. BPO is a magical place where new versions of applications come from. LAG expects new versions of applications from lenny to appear in etch-backports.


The least-cost path I would suggest is:

1. "Freeze" sarge-backports such that packages are updated only up to the level of "etch + security", not only does this reduce developer support-load, but makes the transition to etch simpler for SysAdmin.

2. If there is a "simple" way to add a package into etch-backports that _would_ install straight from testing, then a developer can do so ("simple" should mean that when testing updates, moving it back to backports should be less than 5 man-minutes, or possibly scriptable). If there is no "simple" way to do this, tough. The LAG will need to learn how to install testing packages.

Comments?

Greg Whiteley

Micha Lenk wrote:
Hi,
Luis Matos wrote:
Using the porposed scheme for package versions should let users to
upgrade from sarge-bpo to etch-bpo to lenny-bpo easily wich is very
good.

... so what's the point in doing any Debian Release any more?

Supporting the users in weird or at least exceptional environments will
lead to outrageous efforts for supporting their setup. Who will support
these users in favour of keeping the real Debian suites in shape?

i know several software work from lenny to etch and even sarge, but it's
not easy to use. So, is this case what i propose is that bpo should use
the lenny version ( the people who submit, just submits the same version
that is in lenny). In the future, with bpo integrated in debian's
archives, we can have the same file for lenny and bpo.

I'd strongly vote against integrating bpo into Debian as an official
supported suite. This simply makes stuff more complicated than neccessary.

Regarding all the (sublimal) requests for uploads of packages newer than
Etch I would even vote for (technically enfoced) stopping any further
uploads to sarge-backports for packages with higher version than
available in Etch. Why should we bother to supply newer software in
sarge-backports than in Etch? Or again: What's the point of releasing Etch?

I think we should encourage the users to upgrade to our latest stable
release (Etch) instead of providing any new package uploads (other than
security fixes) in sarge-backports.

Just my thoughts...
  Micha


Reply to: