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

Re: [MoM] Packaging mindthegap (Was: [MoM] Packaging mindthemap)



Hi Shayan,

On Wed, Jul 10, 2019 at 04:59:03PM +0100, Shayan Doust wrote:
> Nicely constructed description within your last couple of emails, thank you!

:-)
 
> Just to clarify before I push: within my current workspace, two deb
> packages are generated: the origin mindthegap and the new
> mindthegap-examples. The main mindthegap package now recommends
> mindthegap-examples. I have excluded file repetitions within both
> packages so we do not overlap, so I have simply swapped some files in
> favour of the example package. This does not affect autopkgtest. Within
> the example package, I have *cherrypicked* to include the scripts
> directory within /usr/share/mindthegap/test as this contains our python
> dataset generation scripts that generally should not have any influence
> to anything within mindthegap as it is not used within the project scope
> and is there for manual user invocation. I have also thrown in the
> jenkins folder as by looking over the script files, they seem
> interesting enough to be classed as an example. Should these files be
> located within /usr/share/doc/mindthegap or
> /usr/share/doc/mindthegap-examples?

I consider /usr/share/doc/mindthegap better since this is where users
might look for it first.  All other things you are describing sound
pretty sensible.
 
> > Sometimes I use the name PACKAGE-data but in this case this
> > might be misleading since there are other data which most probably are
> > needed to run mindthegap and are not just optional.
> 
> I agree. The data folder is occupied by sample data generated by the
> python files by upstream which is used within the test scripts, which
> are invoked via autopkgtest.
> 
> Now, a quick question. As these are basically static files that are good
> as references, should the mindthegap-examples package do anything else
> rather than simply act as an optional addition/split from mindthegap
> itself. It shouldn't need any form of compilation. At least, not within
> python as that is interpreter-based and python3 as a dep should be good
> enough?

Yes.  That means it is "Architecture: all".
 
> Additionally, I have also eliminated more lintian outputs so things are
> beginning to seem a lot cleaner. How should
> "new-package-should-close-itp-bug" be handled / possibly rectified?

You can file a so called ITP bug using reportbug manually.  I guess the
new developer documentation will explain something about it.  Or you
take the "lazy approach" and call a script I have written for the
r-pkg-team since I've automated the packaging quite a lot.  You can
download

    https://salsa.debian.org/r-pkg-team/dh-r/blob/master/scripts/itp_from_debian_dir

may be try to understand it or just go to your packaging dir and call

    itp_from_debian_dir

It assembles all needed information for the ITP bug and files it (you
have one option to add additional receiver e-mails for this bug but
you can also just press enter.  The only thing which is needed to
make this work is that your ~/.reportbugrc contains the line

   smtphost reportbug.debian.org

I think this will be created if you call reportbug the first time - but
please double check that line.  It ensures that there is a valid
smtphost to send the mail created by reportbug.

Background information:  Its highly recommendet to file an ITP bug
BEFORE you start packaging since this ensures nobody else will start a
competing packaging effort.  However, we *intentionally* ignore this
recommendation in the Debian Med team since chances that some random
developer will grab our usually low popcon packages and do the packaging
without talking here on the list about it are extremely low.  The
advantage of the approach to do the main packaging first is that we
will assemble all needed information in a structured way inside the
packaging and the ITP bug filing can be done by a script.
 
> I will review over the recent emails transacted by you so I can see if I
> have missed out anything - hopefully I haven't as I have the slight
> tendancy of skimming over things :-). I will again push everything as
> soon as possible.

Fine.
 
> Thanks for your assistance,

You are welcome

    Andreas.

-- 
http://fam-tille.de


Reply to: