[Debian-med-packaging] bedtools 2.26.0: new upstream release, problem with non-free material.

[Charles Plessy]

> On Thu, Jul 07, 2016 at 02:05:35PM +0900, Charles Plessy wrote:
> > >  
> > > +  Problem: docs/content/images/tool-glyphs/barski_binding_site.png is from
> > > +  Jothi et al, 2008, doi:10.1093/nar/gkn488, licensed under no-DFSG CC-BY-NC 2.0.

Le Thu, Jul 07, 2016 at 07:34:07AM +0200, Andreas Tille a écrit :
> 
> I'd suggest to just drop it from the doc via the following patch:
> 
> diff --git a/docs/content/tools/genomecov.rst b/docs/content/tools/genomecov.rst
> index 2cad65e..1c33e2e 100755
> --- a/docs/content/tools/genomecov.rst
> +++ b/docs/content/tools/genomecov.rst
> @@ -346,12 +346,6 @@ overlaps to be performed.
>  Coverage by fragment
>  ==============================================================================
>  
> -|
> -
> -.. image:: ../images/tool-glyphs/barski_binding_site.png
> -
> -|
> -
>  In ChiP-Seq the binding site is usually not at the coordinate where reads map,
>  but in the middle of the fragment. For this reason we often try to estimate average fragment size
>  for single-read experiment and extend the reads in the 5’-3’ direction up to the estimated fragment length.
> 
> 
> and remove the image from the source tarball.  Unfortunately I have no
> idea how to do this with the non-default repository structure.  What
> do you think about the solution in general and do you have any advise
> how to do this technically (except of turning the repository into a
> default one :-P )

Hi Andreas,

To update the Debian branch (currently debian/unstable) using gbp import-orig
(which I guess is what you want to do in combination with uscan), you need
to create an Upstream branch that tracks the tarball contents and
name it in debian/gbp.conf instead of the master branch (that tracks the
GitHub repository).

Another possibility would be to first try to see if Upstream feels like
removing this file by themselves.

Have a nice day,

-- 
Charles