Last take on updating netlib (Was: Problems when trying to create Debian packages from latest Netlib)
moving back to the list for some final advise. Markus has spent some
time slices into netlib after upstream has basically said we are on our
own. I'd like to decide now for myself whether we simply should stick
to the version packaged in Debian (which would probably be sufficient
for the purpose it was packaged for).
On Wed, May 11, 2016 at 11:44:53AM +0200, Markus Koschany wrote:
> Am 11.05.2016 um 11:19 schrieb Andreas Tille:
> > Hi Markus,
> > I admit I have no idea how to answer this question (except that I can
> > confirmed that we successfully built f2c).
> > Does Sam's answer provide any help?
> I don't understand the f2c part either because it seems we already ship
> that in Debian.
Yes, it was packaged and the whole effort I'm putting in netlib (and its
predepends as well the packages netlib is a dependency for) is based on
the fact that I was able to get rid of the non-free parts of f2c.
> You probably need to add Maven artifacts to the package
> but that should be doable with maven-repo-helper.
Any commits to the according repositories to make the packages better
are more than welcome.
> Basically you either have to rewrite the build system
That's no task for me, sorry.
> or you go to maven
> central and download the sources for each maven artifact separately.
> Just click of "View pom" and then remove last part of the URL, the file
> name, and you see a directory like
> with a -sources.jar file.
> Depending on your future projects with netlib this might be easier for
> you. Just combine the sources and create a new orig tarball or package
> the artifacts separately.
So what you want to tell me is for instance to package
as say source package netlib-core-1.1.2 point a watch file to it and see
whether even this part might fullfill all requirements I need for my
final targets or whether I also need say:
and package everything using maven-helper. I'm just trying to estimate
my effort compared to simply stick what we have.
> Hope that helps a little
That's definitely a slight step forward but not fully sufficient to let
me draw a final conclusion how to proceed.
Thanks a lot for the time you've spent into this