On Tue, Aug 30, 2016 at 08:18:07AM +0200, Danny Milosavljevic wrote: > On Fri, 19 Aug 2016 14:19:03 +0200 > Pjotr Prins <pjotr.publi...@thebird.nl> wrote: > > > On Fri, Aug 19, 2016 at 02:15:19PM +0200, Pjotr Prins wrote: > > > ldc is broken since May 2016. > > > > > > http://hydra.gnu.org/job/gnu/master/ldc-0.16.1.x86_64-linux > > > > > > hmmm. > > > > Include file gone missing: > > I've posted a patch to fix this on 08. Aug 2016, subject "Fix ldc". > > >I am not so much interested in the error. I am interested how we can keep > >track of these failures and notify relevant people. > > Good question. Maybe, when hydra fails, git blame the package (the block that > guix edit would have edited) and find the authors that way? Then > automatically e-mail them all? Or is that too extreme? >
I am thinking a bit larger in the sense that we need a service for software discovery, linked, for example to maintainers and documentation (such as wikipedia). The point being that we have valuable information coming from different resources (including, indeed, git history). A graph database (RDF) would be ideal for storing and querying such information. It could track the build history and state of packages and notify maintainers when something goes wrong. If anyone is interested in semantic web technologies and building a (simple) web interface, I (and others) will be happy to mentor and support that effort. I am particularly interested in software packages with their versioning and dependencies that are relevant for Science. We can get a publication out of it, for example in the Journal of Open Source Software. Pj. --
