Re: Need for maintainers
Dear Andreas,
> Yes, that's correct. I hope you were facing the according FIXMEs in
> d/changelog as expected. Any better clue how to teach users about this
> would be welcome.
Yes, it was pretty straightforward from there.
> BTW, I will have a Debian packaging workshop at FOSSASIA online meeting
Thanks, I just registered for the conference.
> There is the script /usr/bin/salsa in the devscripts package. However,
> for pure historic reasons I (and several other team members here) use
>
> https://salsa.debian.org/med-team/community/helper-scripts/-/blob/master/inject-into-salsa-git
>
> The tools are a bit separated since invented in different teams. But
> you can download this script and run it inside the root of your new git
> repository. It should do everything correctly - provided you are a
> member of the R-pkg team and have permissions to create a repository.
>
> You also want to file an ITP bug report for your package which you
> can do easily by using
>
> itp_from_debian_dir
>
> (in dh-r package). However, its possibly advisable for pedagogic
> reasons that you do it at least once manually by
>
> reportbug wnpp
Thanks - I just filed an ITP using reportbug for pingr
(https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=982895). I think I
will need to be granted access to the R-pkg team on salsa. I have
created an account (ericeb) and requested access, if that is
appropriate.
Warm regards,
Eric
On Mon, Feb 15, 2021 at 9:49 AM Andreas Tille <andreas@an3as.eu> wrote:
>
> On Mon, Feb 15, 2021 at 08:59:42AM -0500, Eric Brown wrote:
> > Dear all,
> >
> > Thank you for the warm welcome and helpful advice. With your direction
> > I believe I've successfully locally packaged some R packages.
> >
> > > I am not sure if I have missed a reply, just in case that I happen to be
> > > the first: Welcome! Everyone is a bit busy with the upcoming Debian
> > > release, please forgive.
> >
> > Thanks very much Steffen.
> >
> > > Debian always needs hands and brains. I just hope your contribution will
> > > not end with R packages once you are familiar with that process. Is
> > > there a package you would be particularly like to see added to/updated
> > > in the distribution? If not then I would find something for you.
> >
> > Thanks again, as above I was thinking some of the more popular and
> > missing rOpenSci packages deserve a place in Debian. Having built R
> > packages in the Debian repo has saved me in the past. Additionally,
> > since I am most familiar with R packages over other software, I
> > thought it was a good place to start to learn the process of deb
> > packaging. The software I wish were in Debian (e.g. Zotero,
> > shiny-server) are too complex for me to help with currently.
>
> BTW, I will have a Debian packaging workshop at FOSSASIA online meeting
>
> Timing: 19 März, 2021 14:00 (CET) - 19 März, 2021 17:00 (CET)
>
> May be I should advertise this a bit more here. I will have another
> packaging workshop one week earlier but it will be in German language.
>
> > On a fresh Debian unstable VM, I installed dh-r and ran the
> > prepare_missing_cran_package script. I found that I also needed to apt
> > install cme and git-buildpackage for it to work properly (maybe these
> > should be dependencies of dh-r?).
>
> Good catch. I've added these to Recommends in git[1]. However, I'm
> wondering whether this might bloat automatic build environments (or are
> these installed with --no-install-recommends?) Comments are welcome
> what others might think about this Recommends.
>
> > I also needed to configure git and
> > export DEBFULLNAME and DEBEMAIL environment variables.
>
> Yes, that's correct. I hope you were facing the according FIXMEs in
> d/changelog as expected. Any better clue how to teach users about this
> would be welcome.
>
> > My thought is that it would be helpful to have more rOpenSci CRAN R
> > packages in Debian, since these are generally well reviewed,
> > maintained and promoted. So for practice, I ran
> > prepare_missing_cran_package on codemetaR. I saw that it automatically
> > found the CRAN dependencies that were not in Debian and also packaged
> > those. In this case, it packaged pingr. pingr is a utility package
> > which needs C compilation and sees a fair amount of activity on
> > Rstudio's mirrors, so it was good practice and might be a good
> > candidate for inclusion into Debian. prepare_missing_cran appeared to
> > work well, with manual intervention required to replace the FIXME in
> > the Debian copyright file to match the R package copyright file.
>
> Happy to hear so. You need a bit of experience if there is a bigger
> tree of dependencies to package but that should be bearable.
>
> > To test that it compiles/builds properly, I downloaded the original
> > tarball from CRAN and renamed it, then ran debuild. There were no
> > errors and I was able to then install the .deb and run the package in
> > R without error.
>
> Also happy to hear this. :-)
>
> > Is the next step for me to push the git to a new repo on the r-pkg team salsa?
>
> There is the script /usr/bin/salsa in the devscripts package. However,
> for pure historic reasons I (and several other team members here) use
>
> https://salsa.debian.org/med-team/community/helper-scripts/-/blob/master/inject-into-salsa-git
>
> The tools are a bit separated since invented in different teams. But
> you can download this script and run it inside the root of your new git
> repository. It should do everything correctly - provided you are a
> member of the R-pkg team and have permissions to create a repository.
>
> You also want to file an ITP bug report for your package which you
> can do easily by using
>
> itp_from_debian_dir
>
> (in dh-r package). However, its possibly advisable for pedagogic
> reasons that you do it at least once manually by
>
> reportbug wnpp
>
> and follow the menu.
>
> Hope this helps
>
> Andreas.
>
>
> [1] https://salsa.debian.org/r-pkg-team/dh-r/-/commit/c9aa3c2ebee0af84be4db4dd77099e36af4a921f
>
> --
> http://fam-tille.de
>
--
Eric Brown MD MSc FRCPC
For encryption, OpenPGP public key available on request.
Reply to: