<p dir="ltr">Em 22/10/2013 15:16, "Peter FELECAN" <<a href="mailto:pfelecan@opencsw.org">pfelecan@opencsw.org</a>> escreveu:<br>
> > This is the experimental implementation of "Package rename tracking"<br>
> > as described here:<br>
> > <a href="http://wiki.opencsw.org/projects-available#toc8">http://wiki.opencsw.org/projects-available#toc8</a><br>
> ><br>
> > Basically it is to find out which packages can safely be dropped.<br>
> ><br>
> > We need to see how we can integrate this in our workflow.<br>
><br>
> After you choose a work-flow, please document it, preferably in the<br>
> OpenCSW Manual and let people know that it's documented.</p>
<p dir="ltr">One thing definitely worth documenting, is the procedure of package rename. For example, how you need two catalog releases to complete the rename: in the first one you create the _stub package, and in the second catalog you drop the stub and use the autoclean function in pkgutil.</p>
<p dir="ltr">For now, we'll try to get a nicely readable, aggregated report of what can be done in each catalog. Then we'll group it by maintainer, and we'll be able to either make it a report available on the build farm, or we'll turn it in to a nag email. I don't think there will be any workflow change in terms of how packages are built and uploaded.</p>
<p dir="ltr">One thing we could automate, is the dropping of old _stub packages. Why make a human do it, if an algorithm can. One concern would be that if we allow a script to modify the catalog, we better make sure that the script working correctly.</p>
<p dir="ltr">Maciej</p>