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

[jeroen@wolffelaar.nl: Re: Bug#385446: PTS: handling of udebs could be improved]

Hi all,

forwarding from Bug#385446 to the list as I think not everyone
follows qa.debian.org bugs. :-)

Without having thought about it very deeply, I think Jeroen has a
good point. Is it perhaps time to establish an udeb policy and/or
update policy to reflect udeb packaging?

Please CC 385446@bugs.debian.org in replies.


----- Forwarded message from Jeroen van Wolffelaar <jeroen@wolffelaar.nl> -----

From: Jeroen van Wolffelaar <jeroen@wolffelaar.nl>
To: Max Vozeler <max@nusquama.org>, 385446@bugs.debian.org
Subject: Re: Bug#385446: PTS: handling of udebs could be improved
Date: Thu, 31 Aug 2006 13:34:25 +0200
Sender: Jeroen van Wolffelaar <jeroen@A-Eskwadraat.nl>
User-Agent: Mutt/1.5.9i

On Thu, Aug 31, 2006 at 12:24:14PM +0200, Max Vozeler wrote:
> Package: qa.debian.org
> Looking at packages.qa.debian.org/partman-crypto, there are two
> points specific to udebs that could be improved. They would make
> the page more useful (and correct) for udeb-only packages:
> 1. TODO and Problems warn about outdated Standards-Version. The
> current practice for udeb-only packages is to not include a
> Standards-Version header and to add a lintian source override for
> no-standards-version-field.

Two remarks on this:

- Adding a lintian override is wrong -- this isn't a very specific
  exception that can't be fixed in lintian, if it's indeed so decided,
  then lintian can easily not show this warning for udeb-only packages
- I understand it's current practice, however, I think it is an unneeded
  practice. S-V is only in source packages, not in binary packages, so
  space considerations do not apply.  Policy still applies where
  possible to udebs, with the understanding that udebs don't follow
  policy on points where there's good reason to do so. It'd be best if
  also policy is updated to actually make an exception for those areas.
  I'm not involved in d-i, but I do think it'd be best to stop this
  practice. Of course, if it's decided that udeb-only packages will not
  have such header anyway, then lintian and QA pages should be updated
  to understand this exception. Using overrides for this remains
  being a misuse of the override feature in either case.

> 2. Testing Status mentions that "partman-crypto has no binaries 
> on any arch", which is not correct. It actually has one arch: any 
> package and two with arch: all.

This is a "problem" (feature request if you so wish) in britney, the
testing scripts. Britney doesn't yet look at udebs at all, and actually
produces this 'error'. There's ongoing discussion on how to
improve/implement britney's handling of udeb-having packages, but
there is no change yet. The PTS could hide this output in the case of
udeb-only packages, or better, hide the message if it's exactly equal to
the one shown now to cope with improvements in britney. That's probably
best to reduce confusion.


Jeroen van Wolffelaar
Jeroen@wolffelaar.nl (also for Jabber & MSN; ICQ: 33944357)

----- End forwarded message -----

Reply to: