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

Bug#725772: RFS: nfft -- Library for computing Non-uniform Fast Fourier Transforms



On Tue, 2013-10-22 at 08:42 +0200, Andreas Tille wrote:
> Hi Ghislain,
> 
> On Mon, Oct 21, 2013 at 02:54:50PM +0100, Ghislain Vaillant wrote:
> > Just a short update on this RFS,
> > 
> > Following your advice, I have moved the package over to debian-science
> > and followed the debian-science policy. It now hosted on
> > http://anonscm.debian.org/gitweb/?p=debian-science/packages/nfft.git
> 
> Thanks.  In turn (and because I think you are not a member of Debian
> Blends team) I added the -dev package to mathematics-dev, physics-dev
> and engineering-dev:
> 
>    http://anonscm.debian.org/viewvc/blends?view=revision&revision=3887
> 
> Please tell me whether you regard this as sensible (or please ask if you
> have no idea what I'm talking about).  When doing so I realised that
> physics-dev is very badly / not really maintained.  Any volunteer to
> check the package pool for packages that are helpful for developing
> applications for physicists?
> 

I believe nfft should be part of the same groups as fftw.

> > I have removed the version uploaded on mentors.debian.org for now to
> > avoid duplication.
> 
> If you want me to sponsor this as SoB[1] package I personally do not
> require an upload to mentors.d.o because I use the repository anyway.
> 

Good, I'll leave it like this then.

> When looking at the package I have the following hints:
> 
>  d/copyright:
>    The header is not a valid DEP5 header.  You can easily check
>    the file by using
>        cme fix dpkg-copyright
> 

Thanks for the tip.

>  d/*.install:
>    The files are starting by a line "#!/usr/bin/dh-exec"   I admit
>    I have never seen this before even if I suspect this might be
>    somewhere in the docs which you have definitely read in a way more
>    recent version than me.  The line does not harm but to the best of
>    my knowledge you can safely remove it.
> 
I probably took it blindly from somewhere else. Good to know for the
future.

>  d/watch:
>    Pure nitpicking:  The file contains some boilerplate comments like
>     "# Uncomment to examine a Webserver directory"
>    You actually *have* uncommented the line - I'd recommend droping
>    useless comments.
> 
I am happy with nitpicking. That's how I'll learn.

> Regarding building he package:  When using git-buildpackage it comes
> very handy if the pristine-tar information is inside the repository.  I
> think this is even written down in Debian Science policy (did not
> checked; if it is not it should be - it is in several other team
> policies).  Please use
> 
>    git import-orig --pristine-tar <orig.tar.gz>
> 
> to import the original tarball information to get a byte identical
> orig tarball without downloading again from scratch.
> 

I'll have a look at pristine-tar.

> Kind regards
> 
>        Andreas.
> 
> [1] http://wiki.debian.org/DebianPureBlends/SoB
> 
> -- 
> http://fam-tille.de
> 
> 

I'll make an update of the package addressing your comments. Thanks for
your time.

Ghislain


Reply to: