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

Re: best policies for third party Debian packaging and get-orig-source target




Hi Andreas,

Thank you for the friendly message.

I've now put the project in question on Bitbucket at the following
URL: https://bitbucket.org/faheem/corrmodel. This is still unfinished
work, but the package should at least build.

Any comments, and particularly those on packaging, are appreciated. I
still need to figure out a suitable make-orig-source for creating an
orig.tar.gz.

On Wed, 31 Jul 2013, Andreas Tille wrote:

Hi Faheem,

On Wed, Jul 31, 2013 at 06:41:59PM +0000, Faheem Mitha wrote:

Additionally, I wrote on June 7th asking if the project wanted any
help with the ggplot2 package, which was out of date, and asking to
be added to the Project Member List. (I'm faheem-guest on alioth.)
I did not get a reply to this message.

That's a bit sad and definitely not the usual way we are dealing
with requests.  The savest way is to use the alioth interface to ask
for membership but at least now you are member of Debian Med

Thanks.

I'm a little puzzled why ggplot2 is being packaged by Debian
Med. It is a general purpose R graphing package.

If nobody else is packaging preconditions we are simply doing it
in our own Vcs.  And this is just a precondition:

I see.

$ apt-cache rdepends r-cran-ggplot2
r-cran-ggplot2
Reverse Depends:
 r-bioc-cummerbund

And you are right - it is not uploaded yet which is most probably
due to the fact that Charles is quite occupied since some time.
What I currently can see from the status in Git the preparation for
the latest upstream is done but the Build-Depends r-cran-scales is
just missing.  It is waiting in

  http://ftp-master.debian.org/new.html

since one month when Charles has uploaded this package.  So we can hope
that it will be accepted soonish so r-cran-ggplot2 will follow.

r-cran-scales is now in Debian, but r-cran-ggplot2 has not yet been updated. I see that r-cran-scales is not installable, because it depends on r-cran-munsell, which is not in Debian, at least not currently on my mirror for amd64.

I'm in the process of reworking my repos so I can make it public. When
I have done so I'll post the repos location to Debian Med, and then
you (the Debian Med developers) can decide if you have any interest in
the software.

We can also work on preconditions for your package - you mentioned that
there are some R packages missing.

Sure thing. Here is a list of R packages I am using. If a package is
in Debian, the Debian package is in a second column. As you can see,
five of these are not in Debian. I certainly think it would be a good
idea to get these into Debian. If so, how should one proceed?

yaml
ggplot2          r-cran-ggplot2 (currently out of date in Debian)
gridExtra
gtable           r-cran-gtable
reshape2         r-cran-reshape2
RPostgreSQL
tikzDevice
Hmisc            r-cran-hmisc
RJSONIO

In the meantime, I'd still appreciate feedback on the general issues I
originally posted about in this thread.

I think you mean the way you should create the private package?  I'd
recommend applying the very same rules as for any other Debian package.
You are welcome to do the development in Debian Med Git / SVN at your
preference (as I said you now have commit permissions).  The rules
are explained in our group policy[1].

I use mercurial. Would you allow mercurial repositories in your
system? If not, it would be difficult to sync between two version
control systems, I think. At least, I have no experience in doing so.

The software uses SQL (PostgreSQL), Python, R and C++, so it is a bit
of a hotchpotch, and while a small package, it has a large number of
dependencies. Unfortunately the R packages I use are not all in
Debian, last I checked.

So why not changing at least the availablity of the preconditions?

I'm not sure what you mean.

You wrote: "Unfortunately the R packages I use are not all in Debian" -
so lets try to get those packages you need in.  As you see we are taking
also quite general packages if they are preconditions.  The rationale is
simply that there is no such thing like an R packaging team (which is a
shame but we need to cope with this).

Ok.

BTW, please do us a favour and if you somehow feel ignored (for
instance by failing to accept your alioth membership request) please
ping again.  That's usually not the way we deal with newcomers and
it should not happen again - but pinging somehow helps specifically
it seems that it were in a quite busy time with release
preparations, Debian Med sprint etc.  So sorry if something did not
went as smooth as you would have expected.

It's no problem, but thank you for your concern.

                                Regards, Faheem


Reply to: