Bug#846002: blends-tasks must not be priority:important (was Re: Bug#846002: Lowering severity)

[Steve McIntyre]

On Sat, Dec 24, 2016 at 02:25:48AM +0100, Philip Hands wrote:
>Raphael Hertzog  writes:
>...
>> So I agree with Cyril and the d-i team, we should be cautious here.
>>
>> Let's focus everybody's energy on getting Phil's patch merged instead
>> of continuing this discussion.
>
>The latest incarnation of which I think is close to ready:
>
>  https://anonscm.debian.org/cgit/d-i/pkgsel.git/log/?h=pu/simple_tasksel
>
>I've squashed the commits together, so we now have the first (aae3196)
>which implements the feature, and would probably be fine as it is (once
>comments to the translators have been added as appropriate).
>
>The second commit (1bb1feb) adds a level of indirection in the
>template, with code to populate it from some new debconf settings,
>which allows one to then customise the menu via preseeding.  This is not
>in any way essential to the task in hand, but might well be useful for
>others.

I'll be honest - that code scares me right now. If this was simple,
obvious stuff then I'd be pushing to try and get this in. But it's
not. Comments like

+   # there is no need to do  this twice, and it breaks [back] behaviour if 
you do

don't help, and I honestly don't understand what

+   db_subst pkgsel/simplified-tasksel $(echo $i | tr "a-z" 
"A-Z") "$subst"

is doing when I read the code at 2am. Can you explain this better
please?

To make this kind of change for stretch, we'll also need updates to
translations directly in the installer and in the installation
guide. I'm worried that we're doing this too late in the cycle - as
KiBi says.

-- 
Steve McIntyre, Cambridge, UK.st...@einval.com
"... the premise [is] that privacy is about hiding a wrong. It's not.
 Privacy is an inherent human right, and a requirement for maintaining
 the human condition with dignity and respect."
  -- Bruce Schneier

Bug#846002: blends-tasks must not be priority:important (was Re: Bug#846002: Lowering severity)

[Philip Hands]

Raphael Hertzog  writes:
...
> So I agree with Cyril and the d-i team, we should be cautious here.
>
> Let's focus everybody's energy on getting Phil's patch merged instead
> of continuing this discussion.

The latest incarnation of which I think is close to ready:

  https://anonscm.debian.org/cgit/d-i/pkgsel.git/log/?h=pu/simple_tasksel

I've squashed the commits together, so we now have the first (aae3196)
which implements the feature, and would probably be fine as it is (once
comments to the translators have been added as appropriate).

The second commit (1bb1feb) adds a level of indirection in the
template, with code to populate it from some new debconf settings,
which allows one to then customise the menu via preseeding.  This is not
in any way essential to the task in hand, but might well be useful for
others.

I have tested this, with these preseed settings, and it does what one
would hope (adding "Minimal system..." as a third option):

d-i pkgsel/simplified-tasksel/choices string standard ("${DESKTOP}") desktop, 
standard server [text-only console & 'ssh' remote access], Minimal system (adds 
no more packages), other use cases
d-i pkgsel/simplified-tasksel/choices-c string ${DESKTOP}-desktop;standard, 
ssh-server;standard, ;;, ;
d-i pkgsel/simplified-tasksel/longdesc string You can now choose between 
installing a standard desktop, a standard server, a minimal system, or 
alternatively to use the task selection menu to have finer grained control over 
installing tasks and blends.

The use of ; and ;; in the choices-c needs documenting -- ; is being
used as a separator for the tasks to be selected.  A lone ';' is being
used as a marker for the "continue to tasksel" case.  ';;' does not
match that, so converts to selecting no tasks -- I suspect leaving it
blank would work just as well, but have not tried that yet.

If anyone knows how to set choices-c via preseeding, then we might not
need (all of) the second commit.

Cheers, Phil.
-- 
|)|  Philip Hands  [+44 (0)20 8530 9560]  HANDS.COM Ltd.
|-|  http://www.hands.com/http://ftp.uk.debian.org/
|(|  Hugo-Klemm-Strasse 34,   21075 Hamburg,GERMANY


signature.asc
Description: PGP signature

Bug#846002: blends-tasks must not be priority:important (was Re: Bug#846002: Lowering severity)

[Cyril Brulebois]

Hi,

Philip Hands  (2016-12-20):
> Just as a reminder for the upcoming alpha that I was trying to do
> something about this by adding an extra simplified tasksel promt:
> 
> Philip Hands  writes:
> ...
> > The menu is now:
> >
> > --> standard ("${DESKTOP}") desktop   <--
> > standard server [text-only console & 'ssh' remote access]
> > other use cases
> 
> I just applied that as a patch to pkgsel and pushed it as pu/simple_tasksel
> 
> It just occurred to me that if we made the list of tasks for each choice
> be in the template, then it avoids hard-wiring the tasks in the code,
> and it should be easier to preseed, so might serve as a customisable menu
> for the likes of debian-edu -- I've added that as a subsequent commit,
> but that's yet to be tested.  I should have time to test that this evening.
> 
> The template needs proper anotation for translators.
> 
> I think it might well be better to replace 'standard' with 'default'.
> 
> BTW as it stands, the server option doesn't bother installing CUPS,
> despite that being in the default tasksel set -- That was based on
> Colin's comment that task-cups needs a rethink, and is currently a bit
> useless.

So I've just looked at the proposed changes, and adding a prompt at this
point is not an option: we're changing logic during the freeze, and
adding translatable material (not the kind of hidden stuff that might
happen with obscure preseeding values, but something that is going to
hit all users) is not a good thing either.

This can be (re)considered during the next release cycle (long before
the freeze if at all possible, this time).


KiBi.


signature.asc
Description: Digital signature

Bug#846002: blends-tasks must not be priority:important (was Re: Bug#846002: Lowering severity)

[Raphael Hertzog]

On Wed, 21 Dec 2016, Ole Streicher wrote:
> I am quoting popcon here since they give a lower estimate of the number
> of users who actually did the test. Nothing more. Nothing about importance.

It gives an estimate of users who ran debootstrap and got the package
installed. It does not give an estimate of how many people ran
debian-installer and saw this menu.

> "Confusion" is not just something mythical that we can't see empirical.
> We will see it in help requests, blog posts, also bug reports, and other
> complaints. If the raise of confusion is "inacceptable" as stated here,
> I would really ask for some empirical evidence for this. At least, I did
> some homework by checking that no complaints popped up somewhere in the
> net (except the one I already mentioned).

Except that the persons who are installing a weekly build of testing
are advanced users that are less likely to be confused than the stable
users.

So while I like to be able to refer to real complains and real problem,
in this specific case it's hard to do so except when it's too late.

So I agree with Cyril and the d-i team, we should be cautious here.

Let's focus everybody's energy on getting Phil's patch merged instead
of continuing this discussion.

Cheers,
-- 
Raphaël Hertzog ◈ Debian Developer

Support Debian LTS: http://www.freexian.com/services/debian-lts.html
Learn to master Debian: http://debian-handbook.info/get/