[Debian-med-packaging] Bug#890783: Autopkgtest for prodigal (#890783)

[Andreas Tille]

Dear Liubov,

On Fri, Mar 16, 2018 at 01:26:18PM +0000, Liubov Chuprikova wrote:
> I have finished and pushed an *autopkgtest *for *prodigal

:-)

> *and also updated
> to the resent standards (by analogy with those changes you've made for
> *coils*).

Very good.  You seem to gain for extra scoring points. ;-)
BTW, when adapting debhelper you need to change two things:

   d/compat (I did so)
   d/control (which you did)

>  Please, have a look at this commit
> <https://salsa.debian.org/med-team/prodigal/commit/10458cc5b6f9e9f9116206f50e291971e2132744>
> .

I'll check soon.
 
> I included a genome sequence from NCBI as test data. Should I indicate the
> source of this data somewhere in the package (e.g., in Readme.tests)?

I think the best place would be debian/copyright since a data file
should come with a license.  I would say something like


Files: debian/tests/test-data
Copyright: yyyy-yyyy Copyright-Owner
License:
Comment:
  This file was obtained by
    wget URL


would be the most precise form we could use.

> Could you please help me with *lintian* messages 1
> <https://lintian.debian.org/tags/hardening-no-fortify-functions.html> and 2

I consider this a false positive - feel free to ignore (but extra
points that you care for the lintian info!)

> <https://lintian.debian.org/tags/hardening-no-bindnow.html>?

Pushed a fix for the hardening options.

> I try not to
> be neglectful about *E *and *W* flags, but there are always some
> information messages. Should I be worried about them?

I usually try to get rid of these to some extend.  Sometimes I'm fixing
even spellings in quilt patches.  I'm doing so if I consider "my time
well spent".  The latter is not always the case.  Two reasons would be:

   1. Upstream is dead any nobody will care (so my patches will not
      become invalid by frequent upstream updates)

   2. Upstream is very active and I have good reasons to assume that
      they will care and take over the patch (in this case I decide
      between quilt patches and using the issue tracker of upstream.

The short answer about lintian issues is: It depends.
 
> With regards,

I'm very happy about your work and your extra care that reaches
beyond the actual task to write tests.

Kind regards

       Andreas.

-- 
http://fam-tille.de