Bug#959696: debian-rules-uses-as-needed-linker-flag and backports

["Chris Lamb" <lamby@debian.org>]

Hey Mattia,

> […]

You make a number of fair points and many of them have changed the
colour how I feel about this. For example, you are right to highlight
that if a tag exists even for a day then it can have remarkable
impact. I would still insist that it is regrettable that users
hesitate before asking or file bugs but, amongst other things, being
'right' is not the same as being effective.

I would like to add three further bit of nuance from my perspective.
Firstly, one elementary observation that has to be made is that there
is always going to be a severe observation bias towards new tags which
are problematic in some way.  For obvious reasons we never discuss
tags that we add which nobody disagrees with, particularly when they
have few false positives.

Secondly, it is quite demotivating to see Lintian bugs (ie. quite
literally errors in the logic, not tags that are opinionated in any
way) being discussed on debian-devel with the unstated assumption that
the current state of Lintian is how it now will be for all time. Note
that it is not possible to categorise these contributors as "not being
annoyed enough" given that, res ipsa loquitur, they would not have
posted in to a public list in the first instance, unless they were
answering to some obscure psychological need.

Lastly, it is easy to ask that the Lintian maintainers are more
circumspect before adding tags, but there will always be things that
we simply do not know and connections we don't even know we are not
making. We will always implicitly rely on the submitter of the
wishlist bug to have done some of this thinking especially as they are
likely quite knowledgable in their domain. We also rely to a lesser
degree on extremely kindly people (such as yourself, which you have
done many times to great effect) to add comments on outstanding
requests and ask these extremely important "oh, did you think of…?"
questions.

To be explicit, it did not remotely occur to me that backports would
be affected by this tag and it is likely that no deeper analysis or
long walks in the countryside would have made me think of it. Whether
this says something about my suitability to be a maintainer of this
tool is for others to say.

Anyway, this is perhaps not the right venue for this valuable
discussion, so whilst we should continue to think and talk about this
let us now move on to addressing the specifics of this particular tag.

I would like to do a Lintian release today, so given that a) I was led
to believe skipping the tag for backports uploads would be sufficient
and b) that I would not like to leave the resolution of this bug for
another release, I would like to make the concrete proposal that we
remove this tag (or possibly, mark it as pedantic, experimental, and
add a suitable note to the long description). Would you concur?


Regards,

-- 
      ,''`.
     : :'  :     Chris Lamb
     `. `'`      lamby@debian.org 🍥 chris-lamb.co.uk
       `-